1274849 九、發明說明： 【發明所屬之技術領域】 本發明係關於一種形貌測量方法（Surface Profile Measuring Method)及其測量裝置，尤其是一種透過使用 寬頻光源，以非接觸之方式進行之形貌測量方法及其測量 裝置。 【先前技術】 按，利用白光干涉（white light interferometry) 之特性，以非接觸方式對物體表面形貌進行測量之形貌測 量裝置，係廣泛應用於對精密度有高度要求之產品，諸如·· 半導體晶圓、液晶顯示器之玻璃面板等。 請參照第一圖所示，係一典型非接觸式形貌測量裝置 之示意圖。如圖中所示，此形貌測量裝置具有一寬頻光源 1〇、一準直透鏡（collimation lens) 20、一 45 度分光鏡 (Splitter) 30、一成像目鏡40、一影像感測裝置5〇、一 干涉顯微物鏡組60、一掃描平台70與一電腦系統8〇。寬 頻光源10所提供之光線，透過準直透鏡2〇形成平行光， 投射至45度分光鏡30。此平行光係受到45度分光鏡3〇 之反射，照射至干涉顯微物鏡組。 干涉顯微物鏡組60係位於掃描平台7〇之上方，並且， 對準掃描平台70所承載之待測物體9〇的表面。此干涉顯 微物鏡組60具有一顯微物鏡62、一反射鏡64與一分光鏡 (beam splitter) 66。其中，顯微物鏡62係位於反射鏡 64之上方，而反射鏡64係位於分光鏡66之上方。來自45 1274849 度分光鏡30之光線’透過顯微物鏡62，照射至分光鏡66， 而被分光鏡66分成兩道光線。其中一道光線係照射至反射 鏡64，而另一道光線則是照射至待測物體9〇表面。 此二道光線分別經由反射鏡64與待測物體9〇表面反 射後’投射回分光鏡66再度重合。值得注意的是，由於此 二道光線行經距離之差異（即此二道光線之光程差 (Optical Path Difference，OPD))，將在重合後之光線 内產生干涉效果。而此光線再向上投射，經過顯微物鏡62 與45度分光鏡30，最後，透過成像目鏡4〇聚焦於影像感 測裝置50 〇 刚述一道光線之光程差的大小’係受到干涉顯微物鏡 組60與知描平台70之距離的影響。因此，藉由改變干涉 顯微物鏡組60或是掃描平台70之垂直位置，即可改變光 程差之大小，而使影像感測裝置5〇接收到一系列具有不同 光程差之干涉影像。透過電腦系統80分析，計算此一系列 干涉影像中，各個畫素位置在不同之干涉影像中的光強 度’即可形成如第二圖之干涉圖譜。此干涉圖譜係一理想 之干涉圖譜，透過計算此干涉圖譜中之波包的峰值，即可 推導出零光程差之位置，以確認此晝素位置所對應之高 度。以同樣的方式，計算其他晝素位置所對應之高度，即 可得到此待測物體90之表面形貌。 基本上’計算干涉圖譜之波形峰值的方法大致可分為 兩類·一係相移法（phase shifting interferometry, psi)’ 一係垂直掃描干涉法（ver_^cai scani^呢 interferometry, VSI)〇 7 1274849 係利Γ干涉圖譜之規律的頻率相位變化，透過 触錄料㈣，轉干涉圖譜 1日士置，以叶异出物體之表面形貌。值得注 物體表面有較大之兩度差（所形成之 !:=大於光波長之•-)，相移= 付物體之表面形貌。 、垂直掃描干涉法係利用干涉圖譜對稱於零光程差處， 步在零光程差處之光強度最大的特性，分析干 ^ Ί找出零光程差之位置，並藉以計算物體之表面形 :值得，意的是，垂直掃描干涉法雖然並不受到待測物 -表面之高度差的聞，但是，在測量之精確度上，卻不 及相移法。以下係列舉—些用以估算零光程差位置之技術。 美國專利案第5633715號係使用質心法(Centroid Approach)，以干涉圖譜之質心位置為零光程差之位置，藉 以計算出待測表面之垂直高度。此方法雖然具有速度快之 優·、沾仁疋，右疋干涉圖譜在波包處之干涉條紋不對稱， 則以貝心法计算之質心位置，必然與零光程差之位置有明 顯之誤差產生。此外，不可避免的，在以質心法計算質心 位置之過程中，亦將干涉圖譜内之雜訊一併計算進去。因 此，在系統雜訊比較大或是干涉圖譜之垂直掃描範圍較大 (通常係使用於表面高度之範圍較大的待測物體）的情況 下’此方法之量測誤差將更形明顯。 美國專利案第5133601號之方法在形成干涉圖譜之過 程中’係以相位差為9〇度之掃描步幅進行垂直掃描，以獲 8 1274849 得干涉圖譜。隨後，再卿干涉圖譜在零絲差處，其干 涉條紋之光強度對比最大的特性，以干涉圖譜上，連續三 點或五點之光強度數據，求取其光強度對比，以尋找干涉 條紋光強度對比最大處（亦即尋找零階干涉條紋 fringe))。接下來，再以相位補償之方式，在此尋得之零 階干涉條紋上，精確求取零光程差之位置。 " 基本上，此方法具有下列缺點：一、條紋對比之計算 量龐大，需要耗費相當多的時間；二、此方法對於雜訊之 抵抗能力不佳，一旦在零階干涉條紋附近有明顯之雜訊， 所求得光強度對比最大之干涉條紋偏離零階干涉條紋，則 後續計算所獲得之零光程差位置，將與實際上之零光程差 位置有360度相位差之整數倍的差異。 美國專利案第5398113號之方法係利用傅利葉轉換 (Fourier analysis)等頻域轉換方式，將垂直掃描所獲 得之干涉資訊轉移至頻率域（frequency domain)做一^ 列處理’再搭配最小平方逼近法 推算零光程差位置。值得注意的是，此方法雖然可以精確 計算出待測物體之表面形貌，但是，頻域轉換與最小平方 逼近法所產生之資料量卻非常龐大，而需要耗費相當多之 時間。 如前所述，習知方法在估算零光程差之過程，或是無 法有效排除雜訊之影響，而容易產生誤差，或是在計算過 程中產生龐大之資料量，而耗費大量之時間，而無法同時 兼顧量測之精確度與運算之速度。 9 1274849 【發明内容】 本發明之一目的’係提供一種形貌量測方法，可以有 效降低干涉圖譜内的雜訊之影響。 本發明之另一目的，係提供一種形貌量測方法，在顧 及量測精確度之前提，同時顧及運算速度之需求。 本發明所提供之形貌測量方法（surface pr〇file measuring method)，係用以準確且快速地在干涉圖譜中， 確，零光程差之位置。首先，掃描此干涉圖譜，找出對應 於最大光強度之第-干涉條紋。然後，於第一干涉條紋及 其附近之干涉條財，找出—第二干涉條紋，使干涉圖譜 相對於此第二干涉條紋具有最佳對稱性。隨後，在此第二 干涉條紋上，利用相位補償法，找出第二干涉條紋之波峰 所對應之高度值。 β依據前述之雜制方法，本發明—併提供—種形貌 測量裝置。此形貌測量裝置具有—寬縣源、—分光鏡、 -位移模組、-影像感測模組、—取樣模組、—圖譜掃描 模組、一對稱性判斷模組與一計算模組。其中，寬頻光源 係產生-寬縣。分光賴將此寬頻光分束，分別照射^ 待測物體表面與-參考面。位移模組係可以—定之步幅， 改變待測健表面齡光鏡之_轉。影佩測模組係 用以收集寬縣經制物财面與參考φ反射*形成 涉影像。並且，經此影像_模_獲取之干涉影像 各個晝素，係分卿應於铜物體絲之各個特定位置、 取樣模組個以獲取前述各個畫素的強度值。並且，Ρ 位移模域齡_體表_分錢之間_離，取^ 1274849 位置，分別形成一高度值 組係對待測物體表面之各個特定 對應於影像強度之干涉圖譜。 靜掃描干涉圖譜，以找出對應於最大光 ==>邊、文的第-資料點。對稱性判斷模组，係於此 第-貝枓點及其週邊-定範咖之資料點中，找出可以使 :涉=具有最佳對稱性之第二資料點。計算模組係依據 第-貝枓點及其鄰近之#料點，找出第二龍點所座落之 干涉條紋的辨於干涉目譜上賴應的冑度值。1274849 IX. Description of the Invention: [Technical Field] The present invention relates to a Surface Profile Measuring Method and a measuring device thereof, and more particularly to a non-contact manner by using a broadband source Measuring method and measuring device thereof. [Previous technique] A topography measuring device that measures the surface topography of a surface in a non-contact manner by utilizing the characteristics of white light interferometry is widely used in products having high precision requirements, such as ··· Semiconductor wafers, glass panels for liquid crystal displays, etc. Please refer to the first figure for a schematic diagram of a typical non-contact topography measuring device. As shown in the figure, the topography measuring device has a wide-band light source, a collimation lens 20, a 45-degree splitter 30, an imaging eyepiece 40, and an image sensing device. An interference microscope objective set 60, a scanning platform 70 and a computer system 8A. The light supplied from the broadband source 10 is collimated by the collimator lens 2 to form a parallel light, which is projected to the 45-degree beam splitter 30. This parallel light system is reflected by a 45-degree beam splitter 3〇 and is irradiated to the interference microscope objective group. The interference microscope objective set 60 is located above the scanning platform 7〇 and is aligned with the surface of the object to be tested 9〇 carried by the scanning platform 70. The interference microscope set 60 has a microscope objective 62, a mirror 64 and a beam splitter 66. The microscope objective 62 is located above the mirror 64, and the mirror 64 is located above the beam splitter 66. The light from the 45 1274849 degree beam splitter 30 passes through the microscope objective 62 and is incident on the beam splitter 66, which is split into two rays by the beam splitter 66. One of the light rays is incident on the mirror 64, and the other light is irradiated onto the surface of the object to be tested. The two rays of light are respectively reflected by the mirror 64 and reflected by the surface of the object 9 to be measured, and then projected back to the beam splitter 66 to coincide again. It is worth noting that due to the difference in the distance between the two rays (ie, the optical path difference (OPD) of the two rays), interference will occur in the overlapping rays. The light is then projected upwards, through the microscope objective 62 and the 45-degree beam splitter 30, and finally, through the imaging eyepiece 4〇, focused on the image sensing device 50. The magnitude of the optical path difference of a light is subjected to interference microscopy. The effect of the distance between the objective lens set 60 and the viewing platform 70. Therefore, by changing the vertical position of the interference microscope objective set 60 or the scanning platform 70, the optical path difference can be changed, and the image sensing device 5 〇 receives a series of interference images having different optical path differences. Through the analysis of the computer system 80, the light intensity of each pixel position in different interference images is calculated in the series of interference images to form an interference map as shown in the second figure. The interferogram is an ideal interferogram. By calculating the peak value of the wave packet in the interferogram, the position of the zero optical path difference can be derived to confirm the height corresponding to the position of the pixel. In the same way, the height corresponding to the position of the other elements is calculated, and the surface topography of the object to be tested 90 is obtained. Basically, the method of calculating the peak value of the waveform of the interferogram can be roughly divided into two types: phase shifting interferometry (psi)'s vertical scanning interferometry (ver_^cai scani^interferometry, VSI)〇7 1274849 The frequency phase change of the regularity of the interferogram is transmitted through the touch material (4), and the interferogram is set to 1 day, and the surface topography of the object is derived from the leaf. It is worth noting that the surface of the object has a large difference of two degrees (the formed ::= is greater than the wavelength of the light •-), and the phase shift = the surface topography of the object. The vertical scanning interferometry method uses the interference spectrum to be symmetric with respect to the zero optical path difference, and the maximum intensity of the light at the zero optical path difference, analyzes the position of the zero optical path difference, and calculates the surface of the object. Shape: It is worthwhile. The vertical scanning interferometry is not affected by the height difference of the object-surface. However, the accuracy of the measurement is not as good as the phase shift method. The following series are some techniques for estimating the position of the zero optical path difference. U.S. Patent No. 5,633,715 uses the Centroid Approach to calculate the vertical height of the surface to be tested by the position of the centroid of the interferogram at a position of zero optical path difference. Although this method has the advantages of fast speed, Zhanren, and the interference fringe of the right 疋 interferogram at the wave packet, the position of the centroid calculated by the Bay method is inevitable and the position of the zero optical path difference is obvious. The error is generated. In addition, it is inevitable that the noise in the interference spectrum is also calculated in the process of calculating the centroid position by the centroid method. Therefore, the measurement error of this method will be more obvious when the system noise is large or the vertical scanning range of the interference spectrum is large (usually used for objects with a large range of surface height). The method of U.S. Patent No. 5,133,601, in the process of forming an interferogram, is vertically scanned with a scanning step having a phase difference of 9 , to obtain an interference spectrum of 8 1274849. Subsequently, the interference spectrum of the interference spectrum is at the zero line difference, and the intensity of the interference fringe is the largest. The intensity of the light intensity is obtained from the three-point or five-point light intensity data on the interference spectrum to find the interference fringe. The maximum intensity of light contrast (that is, looking for zero-order interference fringe fringe)). Next, in the phase compensation method, on the zero-order interference fringes found here, the position of the zero optical path difference is accurately obtained. " Basically, this method has the following disadvantages: First, the calculation of stripe contrast is huge, and it takes a lot of time; Second, the resistance of this method to noise is not good, once there is obvious near the zero-order interference fringe The noise, the interference fringe obtained by comparing the light intensity is deviated from the zero-order interference fringe, and the position of the zero optical path difference obtained by the subsequent calculation will be an integer multiple of the phase difference of 360 degrees from the actual zero optical path difference position. difference. The method of U.S. Patent No. 5,398,113 uses a Fourier transform (Fourier analysis) and other frequency domain conversion method to transfer the interference information obtained by vertical scanning to a frequency domain to perform a column processing 'and then with a least square approximation method. Estimate the position of the zero optical path difference. It is worth noting that although this method can accurately calculate the surface topography of the object to be tested, the amount of data generated by the frequency domain conversion and the least square approximation method is very large, and it takes a considerable amount of time. As mentioned above, the conventional method in the process of estimating the zero optical path difference, or can not effectively eliminate the influence of noise, is prone to error, or generates a huge amount of data in the calculation process, and takes a lot of time, It is impossible to take into account the accuracy of the measurement and the speed of the calculation. 9 1274849 SUMMARY OF THE INVENTION One object of the present invention is to provide a method of measuring a topography that can effectively reduce the effects of noise in an interference pattern. Another object of the present invention is to provide a method of morphometry measurement that takes into account the accuracy of the measurement while taking into account the need for computational speed. The surface pr〇 file measuring method provided by the present invention is used to accurately and quickly locate the position of the zero optical path difference in the interference spectrum. First, the interference pattern is scanned to find the first-interference fringes corresponding to the maximum light intensity. Then, the first interference fringe and the interference between the first interference fringes are found, and the second interference fringe is found to have the best symmetry of the interference pattern with respect to the second interference fringe. Then, on the second interference fringe, the phase compensation method is used to find the height value corresponding to the peak of the second interference fringe. The invention according to the aforementioned hybrid method, and provides a topographical measuring device. The shape measuring device has a wide county source, a beam splitter, a displacement module, an image sensing module, a sampling module, a spectrum scanning module, a symmetry judging module and a computing module. Among them, the broadband source is - wide county. The light-split beam splits and illuminates the surface of the object to be tested and the reference surface. The displacement module can be used to set the step size and change the _ turn of the light surface to be tested. The shadow mask module is used to collect the image of the wide county economy and the reference φ reflection* to form the image. Moreover, through the image _ _ _ obtained interference image of each element, the division should be at each specific position of the copper object wire, sampling modules to obtain the intensity values of the aforementioned pixels. Moreover, 位移 displacement mode domain age _ body table _ centition _ detachment, take ^ 1274849 position, respectively form a height value group of the surface of the object to be measured specific to the interference spectrum corresponding to the image intensity. The interferogram is statically scanned to find the first-data point corresponding to the maximum light ==> edge and text. The symmetry judging module is located in the data point of the first-beauty point and its surrounding-fixed coffee, and finds the second data point which can make the best symmetry. The calculation module is based on the first-beauty point and its adjacent #-points, and finds the interference value of the interference fringes where the second-point is located on the interference spectrum.
關於本發明之優點與精神可以藉由以下的發明詳述及 所附圖式得到進一步的瞭解。 【實施方式】 請參照第三圖所示，係本發明之形貌測量方法 (surface profile measuring method) —較佳實施例之 流程圖。於步驟A中，同時請參照第四圖所示，寬頻光（例 如白光）係透過一分光鏡66之分束，分別照射一待測物體 表面與一參考面，同時以一固定步幅，調整待測物體表面 與分光鏡之距離，而產生一高度值對應於光強度之干涉圖 譜（如第二圖所示）。 接下來，於步驟B中，掃描此干涉圖譜，找出對應於 敢大光強度之第一資料點P1。同時請參照第三A圖所示， 經由尋找此最大光強度之第一資料π點的位置，即可大致 確認在干涉圖譜中，對應於最大光強度之第一干涉條紋。 然後’於步驟C中’同時請參照第三B圖所示，在第 1274849 一資料點pi及其附近預定數量之資料點pnl，Pn2中，找出 一第二資料點，使干涉圖譜相對於此第二資料點具有最佳 之對稱性。 在選取資料點之過程中，係以第一資料點P1為基準， 並以一定間隔t於第一資料點P1周邊一定範圍内之資料點 中，選取複數個待測資料點Pnl，Pn2。值得注意的是，此 間隔之大小，係足以使第一資料點P1與各個待測資料點 Pnl，Pn2 ’分別位於不同之干涉條紋上。同時，各個待測資 料點Pnl，Pn2最好是其所座落之干涉條紋上，具有最大光 強度之資料點。就一較佳實施例而言，此間隔t之大小係 相當於干涉圖譜上三百六十度相位差之距離。 隨後，就第一資料點P1與選取之待測資料點The advantages and spirit of the present invention will be further understood from the following detailed description of the invention. [Embodiment] Please refer to the third figure, which is a flow chart of the present invention, a flow chart of a preferred embodiment. In step A, at the same time, as shown in the fourth figure, the broadband light (for example, white light) is transmitted through a beam splitter 66 to respectively illuminate the surface of the object to be tested and a reference surface, and adjust at a fixed step. The distance between the surface of the object to be measured and the beam splitter produces an interference spectrum corresponding to the intensity of the light (as shown in the second figure). Next, in step B, the interferogram is scanned to find the first data point P1 corresponding to the intensity of the dare light. At the same time, referring to the third A picture, by finding the position of the first data π point of the maximum light intensity, the first interference fringe corresponding to the maximum light intensity in the interference pattern can be roughly confirmed. Then, 'in step C', please refer to the third B picture, in the 1274849 data point pi and its predetermined number of data points pnl, Pn2, find a second data point, so that the interference spectrum is relative to This second data point has the best symmetry. In the process of selecting the data points, the first data points P1 are taken as the reference, and a plurality of data points Pn1, Pn2 to be tested are selected from the data points within a certain range around the first data point P1 at a certain interval t. It is worth noting that the interval is sufficient for the first data point P1 and the respective data points Pn1, Pn2' to be located on different interference fringes. At the same time, each of the material points Pn1 and Pn2 to be tested is preferably a data point having the maximum light intensity on the interference fringes on which it is located. In a preferred embodiment, the spacing t is equivalent to a distance of three hundred and sixty degrees of phase difference on the interferogram. Subsequently, the first data point P1 and the selected data point to be tested
Pnl，Pn2，分別設定為中心Pc，來計算干涉圖譜之對稱性。 =第三C圖所示，此在計算干涉_之對稱性之步驟，係 就干涉圖譜於中心pe左右_卜定範圍d内之起伏變化， 分別加總。也就是說，就此中心Pc兩侧一定範圍d内，所 有相鄰資料點之光強度差值的麟值，匈加總。然後， 將左右兩側加總後之數值相減，以判斷對稱性。值得注音、 的是，此顧d之大小必須足以涵蓋干涉圖譜中之^個^ 包的寬度。 正/ 透過比較以不同之資料點P1，Pnl，Pn2為中心所計算 之相減後的數值，所得數值最小者所對應之資料點，即係 使干涉條紋具有最佳對稱性之第二資料點P2。 ” 由此觀之，前述步驟C也可說是在第三A圖之第一干 涉條紋及其附近之干涉條財，找出—第二干涉條紋（對 12Pnl, Pn2, respectively set to the center Pc, to calculate the symmetry of the interference spectrum. = As shown in the third C diagram, the step of calculating the symmetry of the interference _ is the sum of the fluctuations of the interference spectrum in the range d around the center pe, respectively. That is to say, for a certain range d on both sides of the center Pc, the ridge value of the difference in light intensity of all adjacent data points is Hungarian. Then, the summed values of the left and right sides are subtracted to determine the symmetry. It is worth noting that the size of this d must be sufficient to cover the width of the ^^ packet in the interferogram. Positive / By comparing the subtracted values calculated by the different data points P1, Pnl, Pn2, the data points corresponding to the smallest value are the second data points that make the interference fringes have the best symmetry. P2. From this point of view, the foregoing step C can also be said to be the interference fringe in the first interference fringe of FIG. A and its vicinity, and find out - the second interference fringe (pair 12)
1274849 應於前述第二資料點P2) 干涉條紋具有最輯雛 於此第二 干涉條紋。 干涉條紋即係零階之 值得注意的是，前述步驟C 次 受到步驟A中所形成之干涉圖雄，复點的方式’會 此取樣密度係受_定步幅之二響。而 例而言，當固定步幅之大 '、&就較佳實施 … 大小係相當於干涉圖譜上士+疮 位差之距離，步驟C所選取之各 曰）十度相 -資斜額^ 铜資料點係分別與第 貝枓點間&四個固疋步幅之距離或其整數倍。 欠隨後’於步驟D中’並請參照第三D圖所示，依據第 :料點P2及其鄰近之資料點pml，pm2,喊pm4，以相位 ，法’ 零階干涉條⑽波峰於干涉圖譜上所對應的 尚度值，亦即零光程差之位置。 值知注意的是，步驟D之選取資料點的方式，亦會受 到步驟A中所形成之干涉圖譜，其取樣密度的影響。二一 較佳實施例而言，當固定步幅之大小係相#於干^圖譜上 九十度相位差之距離，在步驟D中，係依據第二資料點p2 及其乘近四個复料點Pml，Pm2, Pm3, Pm4的資料（如第三]) 圖所示），以下列方程式（1)進行相位補償，以計算第二資 料點P2與實際上零光程差位置之相位差φ。 方程式（1) : Φ =tan-l(2(Ipm2-Ipm3)/(2Ipc-Ipml-Ipm4)。 其中，Ipml、Ipm2、Ipm3與Ipm4分別是第二資料點 P2周圍最近四個資料點Pml，Pm2, Pm3, Pm4所對應之光強度 13 1274849 值；而Ipc則是第二資料點P2所對應之光強度值。 然後，將此相位差Φ，換算為高度值之差距ΔΗ=Φ λ /4；Γ···方程式g)。由此，即可估算出實際上零光程差位 置，所對應之高度值h〇=hp2+Ah···方程式（3)。其中，hp2 係指第二資料點P2所對應之高度值。 明參照第四圖所示，係本發明之形貌測量裝置一較佳 實施例之不意圖。如圖中所示，此形貌測量裝置具有一寬 頻光源 10、一準直透鏡（c〇llimati〇n lens) 2〇、一 45 度分光鏡(Splitter) 30、一成像目鏡4〇、一干涉顯微物 鏡組60 ' —掃描平台70、一位移模組180、一影像感測模 組50、一取樣模組1〇〇、一圖譜掃描模組12〇、一對稱性 判斷模組140與一計算模組16〇。 寬頻光源10所提供之光線係透過準直透鏡2〇，形成 平行光照射至45度分光鏡30。此平行光受到45度分光鏡 30之反射，照射至干涉顯微物鏡組。干涉顯微物鏡組 60係位於掃描平台70上方，並且，對準置放於掃描平台 70上方之待測物體9〇的表面。此干涉顯微物鏡組6〇具有 一顯微物鏡62、一反射鏡64與一分光鏡（beam spHtter) 66。來自45度分光鏡30之光線，係透過顯微物鏡62，照 射至分光鏡66而被分束成兩道光線。此二道光線分別經由 反射鏡64與待測物體之表面90反射後，投射回分光鏡66 再度重合而產生干涉。干涉後之光線向上投射，依序經過 顯微物鏡62與45度分光鏡30，並透過成像目鏡40聚焦 於影像感測裝置50。 位移模組180係可以一定之步幅，改變待測物體90 14 1274849 ^面，分光鏡66之間隔.以調整前述二道光線之光程 差之大小，献’影佩職置5G柯·—且 Γί程Ϊ之干涉影像。此影像細驗5_取之干抑 =’各個晝素係分別對應至待測物體90表面之不同: 取=組刚_以獲取前述各個晝素的強度值。並 =者侧無9〇表面與反射鏡表面（可視為—參考面） 的改變’取樣模組1〇〇係對於待測物體表面之 值對應於影像強度之干涉圖譜。—幸）刀浙成一同度 掃扩照第三人圖所示’圖譜掃描模組12〇係用以 峨爾條 模=同時請參照第三圖所示，對稱性判斷 1140係依據前述圖譜掃描模幻2〇所獲取 於此第—資料㈣及其附近之資料點Pnl，Pn2中 涉圖譜具有最佳對稱性之第二資料财2。同 r Γ =，D _不’計算模組16G係依據對稱性判斷 果、、且所獲取之第二資料點P2，藉以在第二資料點P2 =洛之零階干涉條紋的波峰，於干涉圖譜上所對應的高 ㈣本發明係直接在干涉圖射，找出光 又取帛—資料點p卜此計算過程僅涉及簡單之比 士並=不給耗費過大之计异時間。其次，就一般之干涉圖 曰而.除非疋涉及暗點，干涉圖譜中光強度最大值都非 常接近零階條紋之位置。因此，由步驟β所獲得之第一資 料點Ρ1當不致於與零階條紋之位置有太大之偏離。 透過步驟Β約略確認零階條紋之位置後，於步驟c中， 本發明利用波包之對稱性，找出一使干涉圖譜具有最佳對 稱性之第二資料點Ρ2，以確認確切之零階條紋處。基本 上，在干涉圖譜中，白光干涉所形成之波包的寬度通常不 會太大。而如第三C圖所示，在對稱性之判斷過程中，僅 涉及簡單之加減計算，可以維持理想之計算速度。又，本 發明利用對稱性求取確切之零階條紋的位置，亦可以避免 雜訊對於計算準雜之影響（雜輯於零階敵左右兩側 之干涉_通常有她之影響，祕崎算中係相互 抵銷）。 在步驟D中，本發明係就步驟c所獲致之零階條紋， ^相位補叙方法求取確切之零絲差彳ϋ得注意的 是:透過相位補償之方式所獲得之零光程差位£，其^確 度係可與習知相移演算法概擬。又，由於本發明ς步驟 Β與C中’即已確認零階條紋之位置，因此，不會面 知相移演算法無法應較Α高度差之缺點，同^，亦可 以節省習知相移演算法於相位重建所需花費之時間。、β 以上所述係利用較佳實施例詳細說明本發明，而 制本發明之範圍，而且熟知此類技藝人士皆能明瞭二 而作些微的改變及調整，仍將不失本發明之要所田 不脫離本發明之精神和範圍。 ^ 1274849 【圖式簡單說明】 第一圖係一典型形貌測量裝置之示意圖。 第二圖之係一理想上之干涉圖譜。 第三圖係本發明之形貌測量方法（surface profile measuring method) —較佳實施例之流程圖。 第三A圖係顯示第三圖之步驟B中，於干涉圖譜中選 取第一資料點。1274849 The second data point P2) should have the most interference fringes in this second interference fringe. The interference fringes are zero-order. It is worth noting that the above-mentioned step C is subjected to the interference pattern formed in the step A. The way of the complex point is that the sampling density is affected by the _the stride. For example, when the fixed step size is large, ', & is better implemented... The size is equivalent to the distance of the interference map sergeant + sore difference, the 选取) selected in step C) ^ The copper data points are the distances from the Four Beacons and the four solid steps or their integer multiples. Subsequent 'in step D' and please refer to the third D picture, according to the first: material point P2 and its adjacent data points pml, pm2, shout pm4, phase, method 'zero-order interference bar (10) peak interference The grace value corresponding to the map, that is, the position of the zero optical path difference. It is noted that the manner in which the data points are selected in step D is also affected by the interference spectrum formed in step A, and the sampling density. In the preferred embodiment, when the size of the fixed step is the distance of the phase difference of 90 degrees on the dry spectrum, in step D, the second data point p2 is multiplied by four complexes. The data of the material points Pml, Pm2, Pm3, Pm4 (as shown in the figure)) is phase compensated by the following equation (1) to calculate the phase difference between the second data point P2 and the actual zero path difference position. Φ. Equation (1): Φ = tan-l (2 (Ipm2-Ipm3) / (2Ipc-Ipml-Ipm4). Among them, Ipml, Ipm2, Ipm3 and Ipm4 are the last four data points Pml around the second data point P2, respectively. The light intensity corresponding to Pm2, Pm3, Pm4 is 13 1274849; and Ipc is the light intensity value corresponding to the second data point P2. Then, the phase difference Φ is converted into the difference of height values ΔΗ=Φ λ /4 ;Γ··· Equation g). Thus, the actual zero path difference position can be estimated, and the corresponding height value h 〇 = hp 2 + Ah · · · equation (3). Where hp2 is the height value corresponding to the second data point P2. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Referring to the fourth embodiment, a preferred embodiment of the topography measuring apparatus of the present invention is not intended. As shown in the figure, the topography measuring device has a wide-band light source 10, a collimating lens (c〇llimati〇n lens) 2〇, a 45-degree splitter (Splitter) 30, an imaging eyepiece 4〇, and an interference. The microscope objective set 60' - the scanning platform 70, the displacement module 180, an image sensing module 50, a sampling module 1〇〇, a map scanning module 12〇, a symmetry determining module 140 and a The calculation module 16〇. The light supplied from the broadband source 10 is transmitted through the collimator lens 2 to form parallel light to the 45-degree beam splitter 30. This parallel light is reflected by the 45 degree beam splitter 30 and is incident on the interference microscope objective set. The interference microscope objective set 60 is located above the scanning platform 70 and is aligned with the surface of the object 9 to be tested placed above the scanning platform 70. The interference microscope objective set 6 has a microscope objective 62, a mirror 64 and a beam spHtter 66. Light from the 45 degree beam splitter 30 is transmitted through the microscope objective 62 to the beam splitter 66 to be split into two rays. The two rays are reflected by the mirror 64 and the surface 90 of the object to be measured, respectively, and are projected back to the beam splitter 66 to overlap again to cause interference. The interfering light is projected upwardly, sequentially passes through the microscope objective 62 and the 45 degree beam splitter 30, and is focused on the image sensing device 50 through the imaging eyepiece 40. The displacement module 180 can change the interval between the object to be tested 90 14 1274849 and the beam splitter 66 to adjust the interval of the optical path difference of the two rays, and present the image of the 5G Ke. And Γ Ϊ Ϊ Ϊ 干涉 interference images. This image is examined 5_takes the dry == each element corresponds to the difference of the surface of the object to be tested 90: Take = group just_ to obtain the intensity values of the aforementioned individual elements. And = 9 side of the surface and the mirror surface (can be seen as - reference surface) change 'sampling module 1 〇〇 system for the surface of the object to be measured corresponds to the interference spectrum of the image intensity. - Fortunately, the knife and the same degree sweep the same as the third person's picture shows 'Graphic scanning module 12 〇 is used for the 条 条 = = = at the same time, please refer to the third figure, symmetry judgment 1140 is based on the above-mentioned map scan The second data obtained by the model illusion 2 is obtained from the data- (4) and its data points Pnl and Pn2, which have the best symmetry. The same r Γ =, D _ not 'computation module 16G is based on the symmetry to determine the fruit, and the acquired second data point P2, so that the second data point P2 = Luo zero-order interference fringe peak, interference The corresponding high on the map (4) The invention is directly in the interferogram, to find out the light and take the data - the data point p. This calculation process only involves simple comparison and does not give too much time. Second, as far as the general interferogram is concerned, unless the dark point is involved, the maximum light intensity in the interference spectrum is very close to the position of the zero-order stripe. Therefore, the first information point 获得1 obtained by the step β does not deviate too much from the position of the zero-order stripe. After step Β to confirm the position of the zero-order fringe, in step c, the present invention uses the symmetry of the wave packet to find a second data point 使2 which makes the interference symmetry have the best symmetry to confirm the exact zero order. Stripes. Basically, in the interference spectrum, the width of the wave packet formed by white light interference is usually not too large. As shown in the third C diagram, in the judgment process of symmetry, only the simple addition and subtraction calculation is involved, and the ideal calculation speed can be maintained. Moreover, the present invention uses symmetry to obtain the position of the exact zero-order fringe, and can also avoid the influence of noise on the calculation of the quasi-missing (the interference of the miscellaneous series on the left and right sides of the zero-order enemy _ usually has her influence, Mizusaki The middle lines offset each other). In the step D, the present invention obtains the zero-order fringe obtained by the step c, and the phase complementation method obtains the exact zero-wire difference. Note that the zero-path difference obtained by the phase compensation method is obtained. £, its accuracy can be compared with the conventional phase shift algorithm. Moreover, since the position of the zero-order stripe has been confirmed in the steps Β and C in the present invention, it is not known that the phase shift algorithm cannot be disadvantaged by the height difference, and the conventional phase shift calculation can be saved. It takes time to reconstruct the phase. The above description of the present invention will be described in detail by the preferred embodiments of the present invention, and the scope of the present invention, and those skilled in the art will be able to make a slight change and adjustment without departing from the scope of the present invention. The field does not depart from the spirit and scope of the invention. ^ 1274849 [Simple description of the diagram] The first diagram is a schematic diagram of a typical topography measuring device. The second graph is an ideal interferogram. The third figure is a surface profile measuring method of the present invention - a flow chart of a preferred embodiment. The third A picture shows the first data point in the interference spectrum in step B of the third figure.
第三B圖係顯示第三圖之步驟C中，所選取之第一資 料點及其鄰近之待測資料點。 第三C圖係顯示第三圖之步驟C中，估算干涉圖譜之 對稱性的方法。 第三D圖係顯示第三圖之步驟D中，於零階條紋上， 估算零光程差位置之示意圖。 第四圖係本發明之形貌測量裝置一較佳實施例之示意The third B diagram shows the selected first data point and its adjacent data points to be tested in step C of the third figure. The third C diagram shows the method of estimating the symmetry of the interference spectrum in step C of the third figure. The third D diagram shows a schematic diagram of estimating the position of the zero optical path difference on the zero-order fringe in step D of the third figure. The fourth figure is a schematic representation of a preferred embodiment of the topography measuring device of the present invention.
【主要元件符號說明】 寬頻光源10 準直透鏡20 45度分光鏡30 成像目鏡40 干涉顯微物鏡組60 17 1274849 掃描平台70 電腦系統80 影像感測模組50 取樣模組100 圖譜掃描模組120 對稱性判斷模組140 計算模組160。 位移模組180[Main component symbol description] Broadband light source 10 Collimating lens 20 45 degree beam splitter 30 Imaging eyepiece 40 Interference microscope objective group 60 17 1274849 Scanning platform 70 Computer system 80 Image sensing module 50 Sampling module 100 Graph scanning module 120 The symmetry determination module 140 calculates the module 160. Displacement module 180