TWI532979B - Prober fault identification method - Google Patents

Prober fault identification method Download PDF

Info

Publication number
TWI532979B
TWI532979B TW103139926A TW103139926A TWI532979B TW I532979 B TWI532979 B TW I532979B TW 103139926 A TW103139926 A TW 103139926A TW 103139926 A TW103139926 A TW 103139926A TW I532979 B TWI532979 B TW I532979B
Authority
TW
Taiwan
Prior art keywords
dimension
detected
ratio
gradient difference
value
Prior art date
Application number
TW103139926A
Other languages
Chinese (zh)
Other versions
TW201619583A (en
Inventor
吳秉穎
陳秋旺
Original Assignee
旺矽科技股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 旺矽科技股份有限公司 filed Critical 旺矽科技股份有限公司
Priority to TW103139926A priority Critical patent/TWI532979B/en
Application granted granted Critical
Publication of TWI532979B publication Critical patent/TWI532979B/en
Publication of TW201619583A publication Critical patent/TW201619583A/en

Links

Landscapes

  • Testing Or Measuring Of Semiconductors Or The Like (AREA)

Description

點測機故障判別方法 Point measuring machine fault identification method

本發明係有關於一種點測機故障判別方法,特別是有關於一種用以檢測晶圓上之晶粒的點測機的故障判別方法。 The invention relates to a method for judging faults of a spot measuring machine, in particular to a method for judging faults of a spot measuring machine for detecting crystal grains on a wafer.

隨著發光二極體產業的快速成長,除了發光二極體本身具有的高亮度、高功率、較長壽命等優點,要如何維持發光二極體之品質亦相當重要。因此,發光二極體製造完成後,必須檢測發光二極體之發光特性,以判斷發光二極體之品質是否良好。 With the rapid growth of the LED industry, in addition to the high brightness, high power, and long life of the LED, it is also important to maintain the quality of the LED. Therefore, after the fabrication of the light-emitting diode is completed, it is necessary to detect the light-emitting characteristics of the light-emitting diode to determine whether the quality of the light-emitting diode is good.

發光二極體在製造時會先成長於晶圓之上,再利用雷射切割形成複數LED晶粒於晶圓上。在進行LED晶粒的檢測程序時,是以點測機(prober)的探針(probe)依序接觸每一LED晶粒的兩個電極,不僅可檢測每一LED晶粒的電性,還可同時使受檢測的LED晶粒發光,並透過收光機構(圖未示)感測LED晶粒之發光特性。 The LEDs are grown on the wafers prior to fabrication, and laser-cutting is used to form a plurality of LED dies on the wafer. In the LED chip inspection process, the probes of the probe are sequentially contacted with the two electrodes of each LED die, and not only the electrical properties of each LED die can be detected, but also The detected LED dies can be simultaneously illuminated, and the illuminating characteristics of the LED dies are sensed by a light collecting mechanism (not shown).

在點測的過程當中,由於探針在使用一段時間之後,會於探針之外周緣積卡塵埃或污物,進而影響測試的 可靠度,並使得產品的品質降低。為了解決此問題,目前現有的一種點測機之故障判別方法包含下列步驟:步驟1:訂定一標準值範圍,及一預定判斷機制。步驟2:點測晶圓上預定數量的晶粒,並獲得一組檢測值。步驟3:點測晶圓上預定數量且相鄰已檢測後晶粒的其它晶粒,並獲得另一組檢測值。步驟4:比對二組檢測值,並以二組檢測值的差產生一組核校值。步驟5:統計前述核校值超出該標準值範圍的數量。步驟6:當前述數量符合該判斷機制時,進行點測機故障排除的動作。 During the process of the spot test, since the probe is used for a period of time, it will accumulate dust or dirt on the periphery of the probe, thereby affecting the test. Reliability and reduced product quality. In order to solve this problem, a conventional fault diagnosis method for a spot measuring machine includes the following steps: Step 1: setting a standard value range, and a predetermined judgment mechanism. Step 2: Spot a predetermined number of dies on the wafer and obtain a set of detected values. Step 3: Spotting a predetermined number of other dies on the wafer and adjacent adjacent detected dies, and obtaining another set of detected values. Step 4: Align the two sets of detection values, and generate a set of nuclear calibration values by the difference between the two sets of detection values. Step 5: Count the number of the aforementioned nuclear calibration values beyond the range of the standard value. Step 6: When the foregoing quantity conforms to the judgment mechanism, the action of troubleshooting the spot measuring machine is performed.

然而,上述習知技術會有設定值(即超出該標準值範圍的數量)不容易設定的問題。如果設定值定的太小,在實際測試時,連續發生有設定值數量的晶粒測試為不合格,就會有「需要清針」的需求,但實際上可能只是判斷錯誤,其實並不需要清針,因此將導致測試晶粒時間的延長。如果設定值定的太大,在實際測試晶粒時,可能在連續發生有設定值數量的晶粒測試為不合格的期間,探針上面沾粘的塵埃或污物自動掉落,也不需要清針。由此可知,上述超出該標準值範圍的數量的設定值並不易拿捏。 However, the above-mentioned prior art has a problem that the set value (i.e., the number exceeding the range of the standard value) is not easily set. If the set value is set too small, in the actual test, if the die test with the set number of values is unqualified, there will be a need for "need to clear the needle", but in fact it may only be a judgment error, but it is not necessary. Cleaning the needle will therefore result in an extension of the test die time. If the set value is too large, during the actual test of the die, it may happen that the dust or dirt on the probe is automatically dropped during the period when the die test of the set number of samples is continuously rejected, and it is not necessary. Clear needles. From this, it can be seen that the above-mentioned set value exceeding the range of the standard value is not easy to handle.

因此,如何解決上述習知技術的問題,是目前業界亟欲投入研發資源解決的問題之一。 Therefore, how to solve the above-mentioned problems of the prior art is one of the problems that the industry is currently eager to invest in research and development resources.

本發明提供一種點測機故障判別方法,應用至用以檢測晶圓上的晶粒的點測機。點測機故障判別方法包含: (a)設定晶粒合格條件以及亮度梯度差標準值;(b)使點測機沿著第一維度依序檢測晶粒;(c)若被檢測之晶粒符合晶粒合格條件,則即時測量並計算被檢測之晶粒在第一維度上的第一維亮度梯度差計算值;以及(d)判斷第一維亮度梯度差計算值的絕對值是否超出亮度梯度差標準值,若否,則繼續執行步驟(b)。 The invention provides a method for judging a fault of a spot measuring machine, which is applied to a spot measuring machine for detecting crystal grains on a wafer. The method for judging the fault of the measuring machine includes: (a) setting the die qualification condition and the standard value of the brightness gradient difference; (b) sequentially detecting the die along the first dimension by the spot detector; (c) if the detected die conforms to the die pass condition, then Measuring and calculating a first-dimensional luminance gradient difference calculation value of the detected crystal grain in the first dimension; and (d) determining whether the absolute value of the first-dimensional luminance gradient difference calculation value exceeds the luminance gradient difference standard value, and if not, Then proceed to step (b).

A09~E13‧‧‧晶粒 A09~E13‧‧‧ grain

D1‧‧‧第一維度 D1‧‧‧ first dimension

D2‧‧‧第二維度 D2‧‧‧ second dimension

S100~S114‧‧‧步驟 S100~S114‧‧‧Steps

P‧‧‧弓字型路徑 P‧‧‧ bow type path

W‧‧‧晶圓 W‧‧‧ wafer

第1圖為繪示本發明一實施方式之點測機故障判別方法的示意圖。 FIG. 1 is a schematic diagram showing a method for judging a fault of a spot measuring machine according to an embodiment of the present invention.

第2圖為繪示以本發明一實施方式之點測機故障判別方法檢測晶圓上的晶粒的局部示意圖。 FIG. 2 is a partial schematic view showing the detection of the crystal grains on the wafer by the method for determining the fault of the spot measuring machine according to an embodiment of the present invention.

以下將以圖式揭露本發明之複數個實施方式,為明確說明起見,許多實務上的細節將在以下敘述中一併說明。然而,應瞭解到,這些實務上的細節不應用以限制本發明。也就是說,在本發明部分實施方式中,這些實務上的細節是非必要的。此外,為簡化圖式起見,一些習知慣用的結構與元件在圖式中將以簡單示意的方式繪示之。 The embodiments of the present invention are disclosed in the following drawings, and the details of However, it should be understood that these practical details are not intended to limit the invention. That is, in some embodiments of the invention, these practical details are not necessary. In addition, some of the conventional structures and elements are shown in the drawings in a simplified schematic manner in order to simplify the drawings.

請先參閱第1圖以及第2圖。第1圖為繪示本發明一實施方式之點測機故障判別方法的示意圖。第2圖為繪示以本發明一實施方式之點測機故障判別方法檢測晶圓W 上的晶粒的局部示意圖。 Please refer to Figure 1 and Figure 2 first. FIG. 1 is a schematic diagram showing a method for judging a fault of a spot measuring machine according to an embodiment of the present invention. FIG. 2 is a diagram showing the detection of the wafer W by the fault measuring method of the spot measuring machine according to an embodiment of the present invention. A partial schematic view of the upper die.

如第1圖與第2圖所示,於本實施方式中,點測機故障判別方法主要是應用至用來檢測晶圓W上的晶粒的點測機(圖未示)。點測機故障判別方法包含步驟S100~S114,如下所示。 As shown in FIGS. 1 and 2, in the present embodiment, the spot machine fault discrimination method is mainly applied to a spot measuring machine (not shown) for detecting crystal grains on the wafer W. The tester fault discrimination method includes steps S100 to S114 as follows.

步驟S100:設定晶粒合格條件以及亮度梯度差標準值。 Step S100: setting the die qualification condition and the brightness gradient difference standard value.

上述之晶粒合格條件即針對晶粒的電壓、亮度及波長所設定的預設範圍,而上述之亮度梯度差標準值為2%~4%,較佳地為3%,但本發明並不以此為限,可依據實際需求而彈性地調整。 The above-mentioned die qualification condition is a preset range set for the voltage, brightness and wavelength of the crystal grain, and the above-mentioned standard value of the luminance gradient difference is 2% to 4%, preferably 3%, but the present invention does not To this end, it can be flexibly adjusted according to actual needs.

步驟S102:使點測機根據弓字型路徑P沿著第一維度D1依序檢測晶粒,並沿著第二維度D2換排檢測晶粒。 Step S102: The spot measuring machine sequentially detects the crystal grains along the bow-shaped path P along the first dimension D1, and alternately detects the crystal grains along the second dimension D2.

如第2圖所示,僅就晶圓W上的25個晶粒A09~E13做說明。於本實施方式中,第一維度D1為X維度,而第二維度D2為Y維度。當點測機的探針根據上述之弓字型路徑P進行移動時,會沿著第一維度D1由晶粒A09朝向晶粒A13的方向依序檢測該排所有晶粒。在該排所有晶粒都檢測完畢之後,探針會沿著第二維度D2換排,再沿著第一維度D1由晶粒B13朝向晶粒B09的方向依序檢測,之後依此類推。於實際應用中,上述之第一維度D1與第二維度D2也可以分別是Y維度與X維度,或者為其他自訂之維度。 As shown in Fig. 2, only 25 crystal grains A09 to E13 on the wafer W will be described. In the present embodiment, the first dimension D1 is an X dimension, and the second dimension D2 is a Y dimension. When the probe of the spotting machine moves according to the above-described arcuate path P, all the crystal grains of the row are sequentially detected along the first dimension D1 from the grain A09 toward the grain A13. After all the crystal grains in the row have been detected, the probes are arranged along the second dimension D2, and then sequentially detected along the first dimension D1 from the grain B13 toward the grain B09, and so on. In practical applications, the first dimension D1 and the second dimension D2 may also be the Y dimension and the X dimension, respectively, or other custom dimensions.

步驟S104:若被檢測之晶粒符合晶粒合格條件, 則即時計算被檢測之晶粒在第一維度D1上的第一維亮度梯度差計算值以及在第二維度D2上的第二維亮度梯度差計算值。 Step S104: If the detected crystal grains meet the die qualification condition, Then, the first-dimensional luminance gradient difference calculation value of the detected crystal grain in the first dimension D1 and the second-dimensional luminance gradient difference calculation value in the second dimension D2 are calculated in real time.

實際上,本實施方式之點測機故障判別方法係以晶粒合格條件作為判斷晶粒本身是否合格的第一階段檢測,並以亮度梯度差標準值作為判斷探針(圖未示)是否有異物(例如塵埃或污物)的第二階段檢測。詳細來說,本發明實施方式之點測機故障判別方法先利用晶粒合格條件對每一晶粒進行一般性檢測。若第一階段檢測的檢測結果為「不合格」,則明顯代表晶粒發生嚴重問題。也就是說,在步驟S100中就可先確定晶粒是否為合格的。如果是不合格的晶粒,在步驟S100就會先被紀錄為不良晶粒(Bad Die),在第二階段檢測將會剔除不良晶粒的部份。雖然第一階段檢測的檢測結果為「合格」,有時卻無法確認晶粒本身的測試值的正確或錯誤,因為受到探針上積卡的異物影響,晶粒的測試結果可能不會等於實際值。以發光二極體的產業來說,晶粒除了要確認是否為不良晶粒,還需要根據測試結果來分等級,不同等級的晶粒可以販賣的價錢也會有所不同。若晶粒的測試結果不等於實際值,例如測試結果比實際值差,表示實際上等級比較高的晶粒可能因為測試結果的錯誤而導致晶粒分到較差的等級。因此,本發明實施方式之點測機故障判別方法還進一步基於相鄰的晶粒彼此之間的亮度落差都相似的特性對晶粒進行亮度檢測。若所檢測之晶粒的亮度梯度差計算值的絕對值極大,則表示探針上有 異物積卡。 In fact, the method for judging the fault of the spot measuring machine of the present embodiment adopts the qualified condition of the die as the first stage detection for judging whether the die itself is qualified, and whether the standard value of the brightness gradient difference is used as a determining probe (not shown). The second stage of detection of foreign matter (such as dust or dirt). In detail, the method for judging the fault of the spot measuring machine according to the embodiment of the present invention first performs general detection of each crystal grain by using the qualified condition of the die. If the test result of the first stage test is "failed", it obviously represents a serious problem with the grain. That is to say, in step S100, it can be determined whether the die is qualified or not. If it is an unqualified grain, it will be recorded as a bad die in step S100 and a defective part will be removed in the second stage. Although the test result of the first stage test is "acceptable", sometimes it is impossible to confirm the correct or wrong test value of the die itself. Because the foreign matter of the card on the probe is affected, the test result of the die may not be equal to the actual value. value. In the industry of light-emitting diodes, in addition to confirming whether it is a bad grain, the grain needs to be graded according to the test results, and the price of different grades of grain can be sold. If the test result of the die is not equal to the actual value, for example, the test result is worse than the actual value, it means that the actually higher grade die may cause the die to be poorly graded due to the error of the test result. Therefore, the spot machine fault discrimination method according to the embodiment of the present invention further performs brightness detection on the crystal grains based on characteristics in which adjacent crystal grains have similar luminance differences. If the absolute value of the calculated brightness gradient difference of the detected crystal is extremely large, it means that there is Foreign matter accumulation card.

於本實施方式中,第一維亮度梯度差計算值係由被檢測之晶粒以及其在第一維度D1上至少前一個被檢測之晶粒的第一維亮度比值所計算,並且第二維亮度梯度差計算值係由被檢測之晶粒以及其在第二維度D2上至少前一個被檢測之晶粒的第二維亮度比值所計算。 In this embodiment, the first dimension luminance gradient difference calculation value is calculated from the detected crystal grain and the first dimension luminance ratio of at least the previous detected crystal grain in the first dimension D1, and the second dimension The luminance gradient difference calculation is calculated from the detected crystal grain and its second dimension luminance ratio of at least the previous detected crystal grain in the second dimension D2.

舉例來說,以第2圖中的晶粒C11來說,晶粒C11的第一維亮度梯度差計算值係由晶粒C11以及其在第一維度D1上至少前一個被檢測之晶粒C10的第一維亮度比值所計算,晶粒C11的第二維亮度梯度差計算值係由晶粒C11以及其在第二維度D2上至少前一個被檢測之晶粒B11的第二維亮度比值所計算。由此可知,必須在被檢測之晶粒在第一維度D1上必須有至少兩個以上第一維亮度比值,且在第二維度D2上必須有至少兩個以上第二維亮度比值的情況之下,才能以本實施方式之點測機故障判別方法對被檢測之晶粒進行判斷。因此,於第2圖中,若晶圓W只有25個晶粒A09~E13,則其中僅有晶粒C11~C13、D09~D11、E11~E13可以同時計算出第一維亮度梯度差計算值與第二維亮度梯度差計算值。 For example, in the case of the crystal grain C11 in FIG. 2, the first-dimensional luminance gradient difference calculation value of the crystal grain C11 is obtained by the crystal grain C11 and at least the previous detected crystal grain C10 in the first dimension D1. Calculated by the first dimension luminance ratio, the second dimension luminance gradient difference of the crystal C11 is calculated from the crystal grain C11 and the second dimension luminance ratio of at least the previous detected crystal grain B11 in the second dimension D2. Calculation. It can be seen that there must be at least two first-dimensional luminance ratio values in the first dimension D1 of the detected crystal grains, and at least two second-dimensional luminance ratio values must be present in the second dimension D2. Next, the detected die can be judged by the spot machine fault discrimination method of the present embodiment. Therefore, in Fig. 2, if the wafer W has only 25 crystal grains A09~E13, only the crystal grains C11~C13, D09~D11, and E11~E13 can calculate the first-order luminance gradient difference calculation value at the same time. The calculated value is compared with the second dimension luminance gradient.

詳細來說,被檢測之晶粒的第一維亮度梯度差計算值與第二維亮度梯度差計算值可由下述方式計算而得。 In detail, the first-dimensional luminance gradient difference calculated value and the second-dimensional luminance gradient difference calculated value of the detected crystal grains can be calculated by the following manner.

被檢測之晶粒(以晶粒E13為例)的第一維亮度比值為Lx3_ratio=Lx3/Lx2,被檢測之晶粒在第一維度D1上的前一個被檢測之晶粒(即晶粒E12)的第一維亮度比值為 Lx2_ratio=Lx2/Lx1,Lx3、Lx2以及Lx1分別為被檢測之晶粒以及其在第一維度D1上前兩個被檢測之晶粒(即晶粒E12、E11)的亮度值,則第一維亮度梯度差計算值Ratio_X=(Lx3_ratio/Lx2_ratio)-1。被檢測之晶粒的第二維亮度比值為Ly3_ratio=Ly3/Ly2,被檢測之晶粒在第二維度D2上的前一個被檢測之晶粒(即晶粒D13)的第二維亮度比值為Ly2_ratio=Ly2/Ly1,Ly3、Ly2以及Ly1分別為被檢測之晶粒以及其在第二維度D2上前兩個被檢測之晶粒(即晶粒D13、C13)的亮度值,則第二維亮度梯度差計算值Ratio_Y=(Ly3_ratio/Ly2_ratio)-1。 The first dimension luminance ratio of the detected crystal grains (taking the grain E13 as an example) is Lx3_ratio=Lx3/Lx2, and the detected crystal grains in the first dimension D1 are detected (ie, the grain E12) First dimension brightness ratio Lx2_ratio=Lx2/Lx1, Lx3, Lx2, and Lx1 are the detected crystal grains and the luminance values of the first two detected crystal grains (ie, the crystal grains E12, E11) in the first dimension D1, respectively, and the first dimension The luminance gradient difference calculation value Ratio_X=(Lx3_ratio/Lx2_ratio)-1. The second-dimensional luminance ratio of the detected crystal grains is Ly3_ratio=Ly3/Ly2, and the second-dimensional luminance ratio of the detected crystal grains (ie, the crystal grains D13) of the detected crystal grains in the second dimension D2 is Ly2_ratio=Ly2/Ly1, Ly3, Ly2, and Ly1 are the detected crystal grains and the luminance values of the first two detected crystal grains (ie, the crystal grains D13, C13) in the second dimension D2, respectively, and the second dimension The luminance gradient difference calculation value Ratio_Y=(Ly3_ratio/Ly2_ratio)-1.

於一實施例中,晶粒E13的亮度Lx3與Ly3皆為101,晶粒E12的亮度Lx2為100,晶粒E11的亮度Lx1為98,晶粒D13的亮度Ly2為99,晶粒C13的亮度Ly1為98。根據以上數據可計算出晶粒E13的第一維亮度比值Lx3_ratio=101/100=1.01,晶粒E12的第一維亮度比值Lx2_ratio=100/98=1.0204082,晶粒E13的第二維亮度比值Ly3_ratio=101/99=1.020202,而晶粒D13的第二維亮度比值Ly2_ratio=99/98=1.0102041。因此,晶粒E13的第一維亮度梯度差計算值Ratio_X=(1.01/1.0204082)-1=-1.020%,而晶粒E13的第二維亮度梯度差計算值Ratio_Y=(1.020202/1.0102041)-1=0.990%。於本實施例中,由於晶粒E13的第一維亮度梯度差計算值Ratio_X的絕對值與第二維亮度梯度差計算值的絕對值皆未超過預設為3%的亮度梯度差標準值,因此判定晶粒E13 通過亮度檢測,代表點測機在檢測晶粒E13時,並未有異物積卡於探針上的狀況發生。 In one embodiment, the brightness Lx3 and Ly3 of the die E13 are all 101, the brightness Lx2 of the die E12 is 100, the brightness Lx1 of the die E11 is 98, the brightness of the die D13 is 99, and the brightness of the die C13 Ly1 is 98. According to the above data, the first-dimensional luminance ratio Lx3_ratio=101/100=1.01 of the crystal grain E13, the first-dimensional luminance ratio Lx2_ratio=100/98=1.0204082 of the crystal grain E12, and the second-dimensional luminance ratio value Ly3_ratio of the crystal grain E13 can be calculated. =101/99=1.020202, and the second dimension luminance ratio of the crystal grain D13 is Ly2_ratio=99/98=1.0102041. Therefore, the first dimension luminance gradient difference calculation value of the grain E13 is Ratio_X=(1.01/1.0204082)-1=-1.020%, and the second dimension luminance gradient difference calculation value of the grain E13 is Ratio_Y=(1.020202/1.0102041)-1 =0.990%. In this embodiment, since the absolute value of the calculated value of the first dimension luminance gradient difference Ratio_X of the die E13 and the calculated value of the second dimension luminance gradient difference do not exceed the preset value of the luminance gradient difference of 3%, Therefore, the grain E13 is judged Through the brightness detection, when the spot measuring machine detects the die E13, there is no foreign matter accumulated on the probe.

於另一實施例中,晶粒E13的亮度Lx3與Ly3皆為96,晶粒E12的亮度Lx2為100,晶粒E11的亮度Lx1為98,晶粒D13的亮度Ly2為99,晶粒C13的亮度Ly1為98。根據以上數據可計算出晶粒E13的第一維亮度比值Lx3_ratio=96/100=0.96,晶粒E12的第一維亮度比值Lx2_ratio=100/98=1.0204082,晶粒E13的第二維亮度比值Ly3_ratio=96/99=0.969697,而晶粒D13的第二維亮度比值Ly2_ratio=99/98=1.0102041。因此,晶粒E13的第一維亮度梯度差計算值Ratio_X=(0.96/1.0204082)-1=-5.920%,而晶粒E13的第二維亮度梯度差計算值Ratio_Y=(0.969697/1.0102041)-1=-4.010%。於本實施例中,由於晶粒E13的第一維亮度梯度差計算值Ratio_X的絕對值與第二維亮度梯度差計算值的絕對值皆已超過預設為3%的亮度梯度差標準值,因此判定晶粒E13並未通過亮度檢測,代表點測機在檢測晶粒E13時,發生了異物積卡於探針上的狀況。 In another embodiment, the luminances Lx3 and Ly3 of the crystal grains E13 are all 96, the luminance Lx2 of the crystal grains E12 is 100, the luminance Lx1 of the crystal grains E11 is 98, and the luminance Ly2 of the crystal grains D13 is 99, and the crystal grains C13 are The luminance Ly1 is 98. According to the above data, the first dimension luminance ratio Lx3_ratio=96/100=0.96 of the crystal grain E13, the first dimension luminance ratio Lx2_ratio=100/98=1.0204082 of the crystal grain E12, and the second dimension luminance ratio value Ly3_ratio of the crystal grain E13 can be calculated. =96/99=0.969697, and the second dimension luminance ratio of the crystal grain D13 is Ly2_ratio=99/98=1.0102041. Therefore, the first dimension luminance gradient difference of the crystal grain E13 is calculated Ratio_X=(0.96/1.0204082)-1=-5.920%, and the second dimension luminance gradient difference of the crystal grain E13 is calculated Ratio_Y=(0.969697/1.0102041)-1 =-4.010%. In this embodiment, since the absolute value of the calculated value of the first dimension luminance gradient difference Ratio_X of the die E13 and the calculated value of the second dimension luminance gradient difference have exceeded the preset value of the luminance gradient difference preset to 3%, Therefore, it is determined that the crystal grain E13 does not pass the brightness detection, and the spot detector has a situation in which the foreign matter is accumulated on the probe when the crystal grain E13 is detected.

要說明的是,本實施方式之點測機故障判別方法需要利用上述之第一維亮度梯度差計算值與第二維亮度梯度差計算值進行判斷(而非單純利用亮度值來判斷)的目的,在於此計算步驟可以消除在晶圓W上的不同區域的待測晶粒因製程而產生的差異。 It should be noted that the method for judging the fault of the spot measuring machine of the present embodiment needs to use the above-mentioned first-order luminance gradient difference calculation value and the second-dimensional luminance gradient difference calculation value for judgment (rather than simply using the luminance value to judge). In this calculation step, the difference in the die to be tested in different regions on the wafer W due to the process can be eliminated.

步驟S106:判斷第一維亮度梯度差計算值的絕對 值與第二維亮度梯度差計算值的絕對值是否皆超出亮度梯度差標準值,若否,則執行步驟S102,若是,則執行步驟S108。 Step S106: determining the absolute value of the calculated value of the first dimension luminance gradient difference Whether the absolute value of the value and the calculated value of the second-dimensional luminance gradient difference exceeds the luminance gradient difference standard value, if not, step S102 is performed, and if yes, step S108 is performed.

步驟S108:對點測機進行異物排除和機台錯誤診斷動作。 Step S108: performing foreign object elimination and machine error diagnosis on the measuring machine.

由此可知,當被檢測之晶粒的至少其中一維亮度梯度差計算值的絕對值未超出亮度梯度差標準值(或無法計算出)時,會判定無異物積卡的問題發生,並繼續執行步驟S102。唯有當被檢測之晶粒的兩維亮度梯度差計算值的絕對值皆超出亮度梯度差標準值時,才會判定有異物積卡的問題發生,並接著對點測機進行異物排除(例如,清針)和機台錯誤診斷動作。 Therefore, when the absolute value of the calculated value of the one-dimensional luminance gradient difference of the detected crystal grains does not exceed the standard value of the luminance gradient difference (or cannot be calculated), it is determined that the problem of no foreign matter accumulation card occurs, and continues Step S102 is performed. Only when the absolute value of the calculated value of the two-dimensional luminance gradient difference of the detected crystal grains exceeds the standard value of the luminance gradient difference, it is determined that the problem of the foreign matter accumulation card occurs, and then the foreign matter is excluded from the spot measuring machine (for example, , clear needle) and machine error diagnosis action.

步驟S110:重新測量並計算第一維亮度梯度差計算值與第二維亮度梯度差計算值。 Step S110: Re-measure and calculate the first-dimensional luminance gradient difference calculated value and the second-dimensional luminance gradient difference calculated value.

步驟S112:判斷重新測量並計算之第一維亮度梯度差計算值的絕對值與第二維亮度梯度差計算值的絕對值是否仍超出亮度梯度差標準值,若否,則執行步驟S102,若是,則執行步驟S114。 Step S112: determining whether the absolute value of the calculated value of the first-dimensional luminance gradient difference and the absolute value of the second-dimensional luminance gradient difference calculated value are still beyond the luminance gradient difference standard value, and if not, executing step S102, if yes, Then, step S114 is performed.

步驟S114:判定點測機發生異常情況,並停止點測機以進行異常排除。 Step S114: determining that an abnormal condition occurs in the spot measuring machine, and stopping the measuring machine to perform abnormality elimination.

在對點測機進行異物排除(例如,清針)和機台錯誤診斷動作之後,探針會重新回到被檢測之晶粒,並重新測量並計算被檢測之晶粒的第一維亮度梯度差計算值與第二維亮度梯度差計算值。若重新測量並計算之第一維亮度梯 度差計算值的絕對值與第二維亮度梯度差計算值的絕對值並未超出亮度梯度差標準值,則代表積卡於探針上的異物已在清針的步驟(即步驟S108)中被清除了。相反地,若重新測量並計算之第一維亮度梯度差計算值的絕對值與第二維亮度梯度差計算值的絕對值仍超出亮度梯度差標準值,則代表點測機的探針上的異物即使執行清針的步驟也無法清除,此時就必須判定點測機發生異常情況,並停止點測機。通常來說,解決上述異常情況的方式是直接更換點測機的探針。 After the foreign object removal (for example, needle cleaning) and the machine misdiagnosis action on the spot measuring machine, the probe will return to the detected grain and re-measure and calculate the first dimension brightness gradient of the detected grain. The difference calculation value and the second-dimensional brightness gradient difference calculation value. If the first dimension brightness ladder is re-measured and calculated If the absolute value of the calculated value of the difference value and the absolute value of the calculated value of the second-dimensional brightness gradient difference do not exceed the standard value of the brightness gradient difference, it means that the foreign matter accumulated on the probe is already in the step of clearing the needle (ie, step S108) Was cleared. Conversely, if the absolute value of the calculated value of the first-dimensional luminance gradient difference and the calculated value of the second-dimensional luminance gradient difference are still exceeded and exceed the luminance gradient difference standard value, it is represented on the probe of the spot measuring machine. Even if the foreign matter cannot be cleared even if the foreign matter is executed, it is necessary to judge the abnormality of the spotting machine and stop the measuring machine. In general, the way to solve the above abnormal situation is to directly replace the probe of the measuring machine.

由以上對於本發明之具體實施例之詳述,可以明顯 地看出,本發明的點測機故障判別方法在晶粒通過一般性檢測之後,可再進一步對被檢測之晶粒進行亮度檢測,藉以精確地判斷是否有異物積卡於探針的狀況發生,並解決此狀況造成第一階段檢測結果不準確的問題。並且,相較於習知技術,本發明的點測機故障判別方法可在檢測到異常時即時對探針做清針動作。若清針之後的檢測結果不同,則可清楚得知積卡於探針上的異物在清針後以被清除,並可繼續檢測下一晶粒;若檢測結果相同,則可清楚得知點測機的探針發生了即使執行清針的步驟也無法清除的異常情況,必須停止點測機,以待檢測人員對點測機的探針進行後續更換程序。此外,本發明的點測機故障判別方法是藉由判斷對被檢測之晶粒所計算之亮度梯度差計算值(而非單純判斷亮度值或亮度比值)是否過大,來判斷探針是否有異物積卡,因此可以消除在晶圓上的不同區域的待 測晶粒因製程而產生的差異。也就是說,本發明的點測機故障判別方法中所預設的亮度梯度差標準值可通用至晶圓上的不同區域的待測晶粒,因此並不會產生不容易設定的問題。 It will be apparent from the above detailed description of specific embodiments of the invention It can be seen that the method for judging the fault of the spot measuring machine of the present invention can further detect the brightness of the detected crystal grains after the crystal grain passes the general detection, so as to accurately determine whether the foreign matter is accumulated in the probe. And solve this problem causing inaccuracies in the first phase of the test results. Moreover, compared with the prior art, the spot machine fault discrimination method of the present invention can perform a needle clearing action on the probe immediately when an abnormality is detected. If the detection results after the needle cleaning are different, it is clear that the foreign matter accumulated on the probe is removed after the needle is cleared, and the next crystal grain can be continuously detected; if the detection result is the same, the point can be clearly known. The probe of the measuring machine has an abnormal situation that cannot be cleared even if the step of clearing the needle is performed. The measuring machine must be stopped, so that the tester performs a subsequent replacement procedure on the probe of the measuring machine. In addition, the method for judging the fault of the spot measuring machine of the present invention determines whether the probe has foreign matter by judging whether the calculated value of the brightness gradient difference calculated for the detected crystal grain (rather than simply determining the brightness value or the brightness ratio value) is too large. Accumulation card, thus eliminating the need for different areas on the wafer Measure the difference in grain size due to the process. That is to say, the standard value of the brightness gradient difference preset in the method for judging the fault of the spot machine of the present invention can be universally applied to the die to be tested in different regions on the wafer, and thus does not cause a problem that is not easy to set.

雖然本發明已以實施方式揭露如上,然其並非用以限定本發明,任何熟習此技藝者,在不脫離本發明之精神和範圍內,當可作各種之更動與潤飾,因此本發明之保護範圍當視後附之申請專利範圍所界定者為準。 Although the present invention has been disclosed in the above embodiments, it is not intended to limit the present invention, and the present invention can be modified and modified without departing from the spirit and scope of the present invention. The scope is subject to the definition of the scope of the patent application attached.

S100~S114‧‧‧步驟 S100~S114‧‧‧Steps

Claims (4)

一種點測機故障判別方法,用以檢測一晶圓上的晶粒,該點測機故障判別方法包含:(a)設定一晶粒合格條件以及一亮度梯度差標準值;(b)使一點測機沿著一第一維度依序檢測該些晶粒;(c)若被檢測之該晶粒符合該晶粒合格條件,則即時測量並計算被檢測之該晶粒在該第一維度上的一第一維亮度梯度差計算值以及在一第二維度上的一第二維亮度梯度差計算值;以及(d)判斷該第一維亮度梯度差計算值的絕對值與該第二維亮度梯度差計算值的絕對值是否皆超出該亮度梯度差標準值,若否,則繼續執行步驟(b);其中該第一維亮度梯度差計算值係由被檢測之該晶粒以及其在該第一維度上至少前一個被檢測之該晶粒的第一維亮度比值所計算,並且該第二維亮度梯度差計算值係由被檢測之該晶粒以及其在該第二維度上至少前一個被檢測之該晶粒的第二維亮度比值所計算;其中被檢測之該晶粒的該第一維亮度比值為Lx3_ratio=Lx3/Lx2,被檢測之該晶粒在該第一維度上的前一個被檢測之該晶粒的該第一維亮度比值為Lx2_ratio=Lx2/Lx1,Lx3、Lx2以及Lx1分別為被檢測之該晶粒以及其在該第一維度上前兩個被檢測之該些晶粒的亮度值,則該第一維亮度梯度差計算值Ratio_X=(Lx3_ratio/Lx2_ratio)- 1,被檢測之該晶粒的該第二維亮度比值為Ly3_ratio=Ly3/Ly2,被檢測之該晶粒在該第二維度上的前一個被檢測之該晶粒的該第二維亮度比值為Ly2_ratio=Ly2/Ly1,Ly3、Ly2以及Ly1分別為被檢測之該晶粒以及其在該第二維度上前兩個被檢測之該些晶粒的亮度值,則該第二維亮度梯度差計算值Ratio_Y=(Ly3_ratio/Ly2_ratio)-1。 A spot machine fault discrimination method for detecting a die on a wafer, the spot machine fault discrimination method comprising: (a) setting a die qualification condition and a brightness gradient difference standard value; (b) making a point The measuring machine sequentially detects the crystal grains along a first dimension; (c) if the detected crystal grain meets the die qualification condition, immediately measures and calculates the detected crystal grain in the first dimension a first dimension luminance gradient difference calculation value and a second dimension luminance gradient difference calculation value in a second dimension; and (d) determining an absolute value of the first dimension luminance gradient difference calculation value and the second dimension Whether the absolute value of the brightness gradient difference calculation value exceeds the brightness gradient difference standard value, and if not, proceeding to step (b); wherein the first dimension brightness gradient difference calculation value is determined by the detected crystal grain and Calculating, in the first dimension, a first dimension luminance ratio of the at least one of the detected first crystal grains, and calculating the second dimension luminance gradient difference from the detected crystal grain and at least in the second dimension The second dimension brightness ratio of the previous detected grain Calculating; wherein the first dimension luminance ratio of the detected die is Lx3_ratio=Lx3/Lx2, and the first dimension brightness of the die in which the die is detected in the first dimension is detected The ratio is Lx2_ratio=Lx2/Lx1, and Lx3, Lx2, and Lx1 are respectively the detected luminance of the die and the first two detected crystal grains in the first dimension, and the first dimension luminance gradient Difference calculation value Ratio_X=(Lx3_ratio/Lx2_ratio)- 1. The second dimension luminance ratio of the detected die is Ly3_ratio=Ly3/Ly2, and the detected second dimension brightness ratio of the die of the die in the second dimension is detected. For the case where Ly2_ratio=Ly2/Ly1, Ly3, Ly2, and Ly1 are respectively the detected crystal grain and the brightness values of the first two detected crystal grains in the second dimension, the second dimension brightness gradient difference is The calculated value Ratio_Y=(Ly3_ratio/Ly2_ratio)-1. 如申請專利範圍第1項所述之點測機故障判別方法,其中步驟(d)包含:(d1)若判斷結果為是,則執行步驟(e);以及(e)對該點測機進行一異物排除和機台錯誤診斷動作,其中在步驟(e)之後,該點測機故障判別方法還包含:(f)重新測量並計算該第一維亮度梯度差計算值與該第二維亮度梯度差計算值;(g)判斷重新測量並計算之該第一維亮度梯度差計算值的絕對值與該第二維亮度梯度差計算值的絕對值是否仍超出該亮度梯度差標準值,若是,則執行步驟(h),若否,則繼續執行步驟(b);以及(h)判定該點測機發生異常情況,並停止該點測機以進行異常排除。 The method for judging the fault of the measuring machine according to the first aspect of the patent application, wherein the step (d) comprises: (d1) if the determination result is yes, performing step (e); and (e) performing the measuring machine a foreign object elimination and machine error diagnosis action, wherein after the step (e), the spot machine fault discrimination method further comprises: (f) re-measuring and calculating the first dimension brightness gradient difference calculation value and the second dimension brightness a gradient difference calculated value; (g) determining whether the absolute value of the calculated value of the first-dimensional luminance gradient difference and the absolute value of the calculated value of the second-dimensional luminance gradient difference are still beyond the standard value of the luminance gradient difference, if Then, step (h) is performed, if not, proceeding to step (b); and (h) determining that the spot machine is abnormal, and stopping the measuring machine for abnormality elimination. 如申請專利範圍第1項所述之點測機故障判別方法,其中步驟(b)包含:(b1)使該點測機根據一弓字型路徑沿著該第一維度依 序檢測該些晶粒,並沿著該第二維度換排檢測該些晶粒。 The method for judging the fault of the measuring machine according to the first aspect of the patent application, wherein the step (b) comprises: (b1) causing the measuring machine to follow the first dimension according to a bow-shaped path. The grains are sequentially detected and the grains are detected along the second dimension. 如申請專利範圍第1項所述之點測機故障判別方法,其中該亮度梯度差標準值為2%~4%。 For example, the method for judging the fault of the measuring machine according to the first aspect of the patent application, wherein the standard value of the brightness gradient difference is 2% to 4%.
TW103139926A 2014-11-18 2014-11-18 Prober fault identification method TWI532979B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW103139926A TWI532979B (en) 2014-11-18 2014-11-18 Prober fault identification method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW103139926A TWI532979B (en) 2014-11-18 2014-11-18 Prober fault identification method

Publications (2)

Publication Number Publication Date
TWI532979B true TWI532979B (en) 2016-05-11
TW201619583A TW201619583A (en) 2016-06-01

Family

ID=56509247

Family Applications (1)

Application Number Title Priority Date Filing Date
TW103139926A TWI532979B (en) 2014-11-18 2014-11-18 Prober fault identification method

Country Status (1)

Country Link
TW (1) TWI532979B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112967942A (en) * 2020-08-07 2021-06-15 重庆康佳光电技术研究院有限公司 Wafer testing method and device, computer storage medium and computer equipment

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112967942A (en) * 2020-08-07 2021-06-15 重庆康佳光电技术研究院有限公司 Wafer testing method and device, computer storage medium and computer equipment
CN112967942B (en) * 2020-08-07 2023-03-10 重庆康佳光电技术研究院有限公司 Wafer testing method and device, computer storage medium and computer equipment

Also Published As

Publication number Publication date
TW201619583A (en) 2016-06-01

Similar Documents

Publication Publication Date Title
US8492178B2 (en) Method of monitoring fabrication processing including edge bead removal processing
CN103969267B (en) Cleaning control method of particulate matters on probes of probe card
US6473665B2 (en) Defect analysis method and process control method
KR101021320B1 (en) System and method for process variation monitor
WO2008008817A2 (en) Edge inspection and metrology
US8817251B2 (en) Defect inspection method
KR20170019407A (en) Automatic recipe stability monitoring and reporting
TW201513249A (en) Inspection system, inspection method, and readable recording medium
US7557598B2 (en) Method of inspecting quiescent power supply current in semiconductor integrated circuit and device for executing the method
CN105203941B (en) The method of inspection of wafer sort special pattern and probe card defect
TWI532979B (en) Prober fault identification method
JP2004047542A (en) Chip quality determining method, chip quality determining program, marking mechanism using the program, and fault generation analyzing method of wafer
CN105789076B (en) Point measurement machine fault distinguishing method
KR101593643B1 (en) Device for inspecting LED epi-wafer using photoluminescence imaging
US6885955B1 (en) Semiconductor process yield analysis based on evaluation of parametric relationship
KR20080002044A (en) Method of setting an inspection area
US11656274B2 (en) Systems and methods for evaluating the reliability of semiconductor die packages
JP2013120875A (en) Semiconductor wafer test method
JP2009302246A (en) Selecting method for semiconductor device
US8947118B2 (en) Defect detection in integrated circuit devices
KR20160018365A (en) Method for monitoring the operational state of a surface inspection system for detecting defects on the surface of semiconductor wafers
JP2011017609A (en) Defect inspecting apparatus and defect inspecting method
US20070087593A1 (en) Structure for monitoring semiconductor polysilicon gate profile
TWI789811B (en) Measurement system and measurement method
TW201715627A (en) Defect recognition system