WO2022100139A1 - 晶圆测试机台的侦测方法及侦测装置 - Google Patents

晶圆测试机台的侦测方法及侦测装置 Download PDF

Info

Publication number
WO2022100139A1
WO2022100139A1 PCT/CN2021/107441 CN2021107441W WO2022100139A1 WO 2022100139 A1 WO2022100139 A1 WO 2022100139A1 CN 2021107441 W CN2021107441 W CN 2021107441W WO 2022100139 A1 WO2022100139 A1 WO 2022100139A1
Authority
WO
WIPO (PCT)
Prior art keywords
machine
machines
test data
inspected
wafer testing
Prior art date
Application number
PCT/CN2021/107441
Other languages
English (en)
French (fr)
Inventor
王世生
Original Assignee
长鑫存储技术有限公司
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 长鑫存储技术有限公司 filed Critical 长鑫存储技术有限公司
Priority to US17/603,485 priority Critical patent/US20230063456A1/en
Publication of WO2022100139A1 publication Critical patent/WO2022100139A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00Error detection; Error correction; Monitoring
    • G06F11/22Detection or location of defective computer hardware by testing during standby operation or during idle time, e.g. start-up testing
    • G06F11/2247Verification or detection of system hardware configuration
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00Error detection; Error correction; Monitoring
    • G06F11/22Detection or location of defective computer hardware by testing during standby operation or during idle time, e.g. start-up testing
    • G06F11/2205Detection or location of defective computer hardware by testing during standby operation or during idle time, e.g. start-up testing using arrangements specific to the hardware being tested
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00Error detection; Error correction; Monitoring
    • G06F11/22Detection or location of defective computer hardware by testing during standby operation or during idle time, e.g. start-up testing
    • G06F11/2273Test methods

Definitions

  • the present disclosure relates to, but is not limited to, a detection method and a detection device for a wafer testing machine.
  • the wafers are randomly assigned to different test machines for CP test (Circuit Probing; wafer test) and FT test (Final Test; finished product test), and record the CP test and FT test of each wafer. All parameters in the FT test. By counting the recorded parameter values, abnormal test machines can be detected, thereby helping engineers to check and maintain the corresponding test machines.
  • the detection process of abnormal wafer testing machines takes a long time, and the detection accuracy is also low.
  • the present disclosure provides a detection method and a detection device of a wafer testing machine.
  • a first aspect of the present disclosure provides a detection method for a wafer testing machine, the detection method comprising:
  • Each of the to-be-inspected machines is marked according to the counted days when each of the to-be-inspected machines has a significant difference.
  • a second aspect of the present disclosure provides a detection device for a wafer testing machine, the detection device comprising:
  • the storage unit is configured to store the original test data of the same batch of wafers tested by multiple wafer testing machines under multiple test items into the database;
  • a screening unit configured to screen out target test data from the original test data of the database according to preset screening conditions
  • a distinguishing unit configured to perform statistics on the screened-out target test data, so as to distinguish a plurality of the wafer testing machines under each of the test items for comparison machines and machines to be inspected;
  • a comparison unit configured to compare whether there is a significant difference between the target test data of each of the machines to be tested under the corresponding test item and the target test data of the control machine under the same test item within the first predetermined number of days , and count the number of days when each of the machines to be inspected has a significant difference;
  • a marking unit configured to mark each of the machines to be tested according to the counted days when each of the machines to be tested has significant differences.
  • the target test data is screened from the original test data in the database according to the preset screening conditions, and statistics are performed.
  • the control machine is equivalent to the standard machine, that is, the control machine has no abnormality, and then the control machine is used for the control.
  • the machine is the standard to determine whether there is any abnormality in the machine to be inspected.
  • the target test data of each machine to be tested under the corresponding test item compares whether there is a significant difference between the target test data of each machine to be tested under the corresponding test item and the target test data of the control machine under the same test item within the first predetermined number of days. If the target test data has a significant difference, it means that the machine to be tested has an abnormal risk; if there is no significant difference in the target test data of the machine to be tested, it means that the machine to be tested is running well.
  • the detection method of the present disclosure can automatically complete the detection process of the wafer testing machine, the time-consuming of the detection process is shortened. Moreover, the detection method of the present disclosure filters out the target test data from the original test data according to the preset screening conditions, so that the number of target test data used for statistics is large, the error of data analysis is reduced, and the detection is improved. accuracy of results.
  • FIG. 1 is a schematic flowchart of a detection method of a wafer testing machine according to an embodiment of the present disclosure.
  • FIG. 2 is a schematic diagram of a boxplot of a plurality of wafer test stations according to an embodiment of the present disclosure.
  • FIG. 3 is a schematic diagram of the upper and lower quartiles of stable groups according to embodiments of the present disclosure.
  • a wafer testing machine can perform CP testing and FT testing on wafers, and record all parameters of each wafer in CP testing and FT testing. By making statistics on the recorded parameter values, abnormal test machines can be detected, thereby helping engineers to check and maintain the corresponding test machines. Since the data generated by the test machine after testing the wafer is massive, engineers usually store the data in a database.
  • An embodiment of the present disclosure provides a detection method for a wafer testing machine. As shown in FIG. 1 , the detection method for a wafer testing machine according to the embodiment of the present disclosure may include the following steps:
  • Step S110 storing the original test data of the same batch of wafers tested by multiple wafer testing machines under multiple test items in a database
  • Step S120 filtering out the target test data from the original test data in the database according to preset screening conditions
  • step S130 statistics are performed on the screened target test data, so as to distinguish a control machine and a to-be-inspected machine of a plurality of wafer testing machines under each test item;
  • Step S140 compare whether the target test data of each machine to be tested under the corresponding test item and the target test data of the control machine under the same test item have significant differences within the first predetermined number of days, and count each to-be-tested machine. The number of days when the inspection machine has a significant difference;
  • Step S150 marking each machine to be inspected according to the counted number of days in which each machine to be inspected has a significant difference.
  • control machine is equivalent to the standard machine, that is: the control machine does not have abnormal conditions, and then, the control machine is used as the standard to determine whether there is any abnormality in the machine to be inspected. If the target test data has a significant difference, it means that the machine to be inspected has an abnormal risk; if there is no significant difference in the target test data of a machine to be inspected, it means that the machine to be inspected is running well.
  • the detection method of the present disclosure can automatically screen data and perform automatic statistics, that is to say, the detection method can automatically complete the detection of the wafer testing machine. Therefore, the time-consuming of the detection process is shortened.
  • the detection method of the present disclosure selects the target test data from the original test data according to the preset screening conditions, so that the number of target test data used for statistics is large, thereby reducing the error of data analysis and improving the crystallinity. The accuracy of the detection results of the circular test machine.
  • step S110 the original test data of the same batch of wafers tested by multiple wafer testing machines under multiple test items are stored in the database.
  • the wafer testing machine can perform CP test and FT test on the wafer, and record all the parameters of the wafer in the CP test and FT test, wherein: the parameter corresponding to the FT test is temperature, and the CP test Corresponding parameters include current, voltage, inductive reactance, etc., which will not be described in detail here.
  • step S120 the target test data is filtered out from the original test data in the database according to preset filtering conditions.
  • the preset screening conditions may include wafer types and production stages, etc., which will not be described in detail here.
  • the screening process can be automatically performed by the detection system of the wafer testing machine, so that the number of target test data for subsequent statistics is large, thereby reducing the error of data analysis and improving the detection result. accuracy.
  • step S130 statistics are performed on the screened target test data, so as to distinguish a comparison machine and a to-be-tested machine of a plurality of wafer testing machines under each test item.
  • step S130 may include the following steps:
  • Step S1301 sort each wafer testing machine according to the median of the screened target test data of each wafer testing machine, and take the middle 50% of the wafer testing machines as a stable group.
  • sorting each wafer tester according to the median can include the following two steps:
  • step S13011 a boxplot corresponding to each wafer testing machine is drawn according to the screened target test data of each wafer testing machine (as shown in FIG. 2).
  • a boxplot is a statistical chart used to display the dispersion of data groups, which can reflect the characteristics of the original data distribution.
  • the boxplot can also be used to compare the distribution characteristics of multiple groups of data.
  • the process of drawing the boxplot first find the upper edge, lower edge, median and two quartiles of a set of data; then connect the two quartiles to draw the box; And the lower edge is connected to the box to form a boxplot of the data group.
  • the median is located inside the box and will not be described in detail here.
  • Step S13012 sort the medians in each boxplot, so as to complete the sorting of each wafer testing machine, and then take the middle 50% of the wafer testing machines as a stable group, and circle it with a wire frame (as shown in the figure). 3 shown).
  • the number of wafer testing machines is an even number to facilitate the selection of stable groups.
  • the stable group is divided into four equal parts, and the upper quartile P75 and the lower quartile P25 of the stable group can be obtained, which will not be described in detail here.
  • Step S1302 according to the upper quartile P75 and the lower quartile P25 of the stable group, distinguish the control tool and the to-be-inspected tool of the multiple wafer testing tools under each test item.
  • the wafer testing machine is the machine to be tested; if the boxplot of a wafer testing machine The median is higher than or equal to the lower quartile P25 and lower than or equal to the upper quartile P75, then the wafer testing machine is the control machine; if the boxplot of a wafer testing machine If the median is lower than the lower quartile P25, the wafer testing machine is the machine to be inspected.
  • the wafer testing machine is the control machine, and accordingly, the rest of the cases All are machines to be inspected.
  • machine 1, machine 2, machine 3, machine 4, machine 6 and machine 5 are the control machines, and machine 7 and machine 8 are the machines to be inspected, It will not be described in detail here.
  • step S140 compare whether there is a significant difference between the target test data of each machine to be tested under the corresponding test item and the target test data of the control machine under the same test item within the first predetermined number of days, and count the difference.
  • the number of days when each machine to be inspected has significant differences.
  • the first predetermined number of days may be 7 days, that is, each wafer testing machine is detected with one week as a complete cycle.
  • the first predetermined number of days may also be 5 days, 6 days, 8 days, or 9 days, etc., which is not particularly limited here.
  • the t-test method can be used to compare whether there is a significant difference between the target test data of each machine to be inspected and the target test data of the control machine within the first predetermined number of days.
  • the comparison can also be performed by methods such as the f test method or the chi-square test method, which is not particularly limited here.
  • the target test data of a machine to be inspected has significant differences, it means that the machine to be inspected has an abnormal risk; if the target test data of a machine to be inspected has no significant difference, it means that the machine to be inspected has no significant difference Desk is running fine.
  • the t-test method satisfies the following first relationship:
  • x i is each target test data of the control machine within the first predetermined number of days.
  • step S150 each machine to be inspected is marked according to the counted number of days in which each machine to be inspected has a significant difference.
  • step S150 may include the following steps:
  • Step S1501 if the number of days in which a machine to be inspected has significant differences exceeds the predetermined ratio of the first predetermined number of days, the machine to be inspected is marked with a first mark;
  • Step S1502 if a machine to be inspected has a significant difference only in the last second predetermined number of days, a second mark is marked for the machine to be inspected.
  • the first predetermined number of days can be 7 days
  • the value range of the predetermined ratio can be 70% to 90%, that is, if the machine to be inspected has significant performance for 5 to 6 days within a week If there is a sex difference, mark the machine to be inspected first.
  • the value range of the predetermined ratio may also be less than 70% or greater than 90%, which is not particularly limited here.
  • the second predetermined number of days may be 3 days, that is, if the machine to be inspected has significant differences only in the last 3 days within a week, the machine to be inspected is marked with a second mark.
  • the second predetermined number of days may also be 2 days or 4 days, etc., which is not particularly limited here.
  • the machine to be inspected with the first mark is actually in a worse condition than the machine to be inspected with the second mark.
  • the first mark may be "alarm”
  • the second mark may be "warning” , so as to distinguish the operation of the machine to be inspected.
  • the detection method of the embodiment of the present disclosure may further include:
  • the detection status of the wafer testing machine can be periodically sent to the management personnel through emails or scrolling screens, etc., which will not be described in detail here.
  • Embodiments of the present disclosure also provide a detection device for a wafer testing machine, the detection device may include a storage unit, a screening unit, a distinguishing unit, a comparing unit, and a marking unit, wherein:
  • the storage unit is set to store the original test data of the same batch of wafers tested by multiple wafer testing machines under multiple test items into the database;
  • the target test data is screened out from the test data;
  • the distinguishing unit is set to perform statistics on the screened target test data, so as to distinguish the control machine and the test machine of multiple wafer test machines under each test item; compare
  • the unit is configured to compare whether the target test data of each machine to be tested under the corresponding test item and the target test data of the control machine under the same test item within the first predetermined number of days are significantly different, and count each The number of days when the machines to be inspected have significant differences;
  • the marking unit is set to mark each machine to be inspected according to the number of days when the machines to be inspected have significant differences.
  • the detection device may further include a notification unit for notifying the management personnel of the wafer testing machines of the marking results of each machine to be tested.
  • the notification unit may be a mail system, that is, the marking unit is connected to the mail system, and the detection situation of the wafer testing machine is periodically sent to the management personnel by mail, so that the management personnel can grasp the information of the wafer testing machine. specific conditions, and optimize the actual wafer production process accordingly.
  • the target test data is screened from the original test data in the database according to the preset screening conditions, and statistics are performed.
  • the control machine is equivalent to the standard machine, that is, the control machine has no abnormality, and then the control machine is used for the control.
  • the machine is the standard to determine whether there is any abnormality in the machine to be inspected.
  • the target test data has a significant difference, it means that the machine to be tested has an abnormal risk; if there is no significant difference in the target test data of the machine to be tested, it means that the machine to be tested is running well.

Landscapes

  • Engineering & Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Quality & Reliability (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Testing Or Measuring Of Semiconductors Or The Like (AREA)

Abstract

本公开提供一种晶圆测试机台的侦测方法和侦测装置,该侦测方法包括:将多个晶圆测试机台测试到的原始测试数据存储到数据库中;根据预设筛选条件从原始测试数据中筛选出目标测试数据;对筛选出的目标测试数据进行统计,以区分出多个晶圆测试机台中的对照机台和待检机台;比较在第一预定天数内各个待检机台在对应的测试项目下的目标测试数据和对照机台在相同的测试项目下的目标测试数据是否具有显著性差异,并统计出各个待检机台具有显著性差异的天数;根据统计出的各个待检机台具有显著性差异的天数对各个待检机台进行标记。

Description

晶圆测试机台的侦测方法及侦测装置
本公开基于申请号为202011245602.5,申请日为2020年11月10日,申请名称为“晶圆测试机台的侦测方法及侦测装置”的中国专利申请提出,并要求该中国专利申请的优先权,该中国专利申请的全部内容在此引入本公开作为参考。
技术领域
本公开涉及但不限于一种晶圆测试机台的侦测方法及侦测装置。
背景技术
晶圆生产完成后,将晶圆随机分配到不同的测试机台进行CP测试(Circuit Probing;晶圆测试)和FT测试(Final Test;成品测试),并记录下每片晶圆在CP测试和FT测试中的所有参数。通过对记录下的参数值进行统计,即可侦测出异常的测试机台,从而帮助工程师去检查并维护对应的测试机台。
目前,异常晶圆测试机台的侦测过程耗时较长,且侦测的准确性也较低。
发明内容
以下是对本公开详细描述的主题的概述。本概述并非是为了限制权利要求的保护范围。
本公开提供一种晶圆测试机台的侦测方法及侦测装置。
本公开的第一方面提供一种晶圆测试机台的侦测方法,所述侦测方法包括:
将多个晶圆测试机台测试到的同批次晶圆在多个测试项目下的原始测试数据存储到数据库中;
根据预设筛选条件从所述数据库的原始测试数据中筛选出目标测试数据;
对筛选出的所述目标测试数据进行统计,以区分出多个所述晶圆测试机台在各个所述测试项目下的对照机台和待检机台;
比较在第一预定天数内各个所述待检机台在对应的测试项目下的目标测试数据和所述对照机台在相同的测试项目下的目标测试数据是否具有显著性差异,并统计出各个所述待检机台具有显著性差异的天数;
根据统计出的各个所述待检机台具有显著性差异的天数对各个所述待检机台进行标记。
本公开的第二方面提供一种晶圆测试机台的侦测装置,所述侦测装置包括:
存储单元,设置为将多个晶圆测试机台测试到的同批次晶圆在多个测试项目下的原始测试数据存储到数据库中;
筛选单元,设置为根据预设筛选条件从所述数据库的原始测试数据中筛选出目标测试数据;
区分单元,设置为对筛选出的所述目标测试数据进行统计,以区分出多个所述晶圆测试机台在各个所述测试项目下的对照机台和待检机台;
比较单元,设置为比较在第一预定天数内各个所述待检机台在对应的测试项目下的目标测试数据和所述对照机台在相同的测试项目下的目标测试数据是否具有显著性差异,并统计出各个所述待检机台具有显著性差异的天数;
标记单元,设置为根据统计出的各个所述待检机台具有显著性差异的天数对各个所述待检机台进行标记。
本公开实施例所提供的晶圆测试机台的侦测方法及侦测装置,在侦测过程中,首先,对根据预设筛选条件从数据库的原始测试数据中筛选出目标测试数据进行统计,以区分出多个晶圆测试机台在各个测试项目下的对照机台和待检机台,其中,对照机台相当于标准机台,即:对照机台没有出现异常情况,然后再以对照机台为标准来判定待检机台是否有异常情况。
接着,比较在第一预定天数内各个待检机台在对应的测试项目下的目标测试数据和对照机台在相同的测试项目下的目标测试数据是否具有显著性差异,如果待检机台的目标测试数据具有显著性差异,则说明待检机台存在异常风险;如果待检机台的目标测试数据不具有显著性差异,则说明待检机 台运行良好。
最后,在统计出各个待检机台具有显著性差异的天数后,根据统计出的各个待检机台具有显著性差异的天数对各个待检机台进行标记。由此,根据各个待检机台的标记结果即可判断出哪些晶圆测试机台出现了异常情况。
由于本公开的侦测方法能够自动完成晶圆测试机台的侦测过程,从而缩短了侦测过程的耗时。而且,本公开的侦测方法根据预设筛选条件从原始测试数据中筛选出目标测试数据,使得用于统计的目标测试数据的数量较多,减小了数据分析的误差,进而提高了侦测结果的准确性。
在阅读并理解了附图和详细描述后,可以明白其他方面。
附图说明
并入到说明书中并且构成说明书的一部分的附图示出了本公开的实施例,并且与描述一起用于解释本公开实施例的原理。在这些附图中,类似的附图标记用于表示类似的要素。下面描述中的附图是本公开的一些实施例,而不是全部实施例。对于本领域技术人员来讲,在不付出创造性劳动的前提下,可以根据这些附图获得其他的附图。
图1是本公开实施方式晶圆测试机台的侦测方法的流程示意图。
图2是本公开实施方式多个晶圆测试机台的箱线图的示意图。
图3是本公开实施方式稳定组的上四分位数和下四分位数的示意图。
具体实施方式
下面将结合本公开实施例中的附图,对公开实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例是本公开一部分实施例,而不是全部的实施例。基于本公开中的实施例,本领域技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本公开保护的范围。需要说明的是,在不冲突的情况下,本公开中的实施例及实施例中的特征可以相互任意组合。
相关技术中,晶圆测试机台能够对晶圆进行CP测试和FT测试,并记录下每片晶圆在CP测试和FT测试中的所有参数。通过对记录下的参数值进行统计,即可侦测出异常的测试机台,从而帮助工程师去检查并维护对应 的测试机台。由于测试机台在对晶圆进行测试后产生的数据是海量的,所以工程师通常将数据存储在数据库中。
目前,工程师手动随机地从数据库中筛选数据,并对筛选出的数据进行简单的统计运算,以此判别不同测试机台测试相同批次的晶圆的多个测试项目的良品率是否有差异,再根据良品率的差异找出异常的测试机台。
然而,这种方法具有以下几个缺点:①时效性低,不能及时自动地侦测出异常的测试机台,导致发现问题的时间较长;②通过人力随机筛选数据再进行分析,过程繁杂、耗费时间;③所用数据并不是全部数据,使得数据分析结果存在误差,进而导致侦测差异机台的准确度不高。
本公开实施方式中提供一种晶圆测试机台的侦测方法,如图1所示,本公开实施方式晶圆测试机台的侦测方法可包括以下几个步骤:
步骤S110,将多个晶圆测试机台测试到的同批次晶圆在多个测试项目下的原始测试数据存储到数据库中;
步骤S120,根据预设筛选条件从数据库的原始测试数据中筛选出目标测试数据;
步骤S130,对筛选出的目标测试数据进行统计,以区分出多个晶圆测试机台在各个测试项目下的对照机台和待检机台;
步骤S140,比较在第一预定天数内各个待检机台在对应的测试项目下的目标测试数据和对照机台在相同的测试项目下的目标测试数据是否具有显著性差异,并统计出各个待检机台具有显著性差异的天数;
步骤S150,根据统计出的各个待检机台具有显著性差异的天数对各个待检机台进行标记。
其中,对照机台相当于标准机台,即:对照机台并没有出现异常情况,然后,以对照机台为标准来判定待检机台是否有异常情况,例如,如果一待检机台的目标测试数据具有显著性差异,则说明该待检机台存在异常风险;如果一待检机台的目标测试数据不具有显著性差异,则说明该待检机台运行良好。
在统计出各个待检机台具有显著性差异的天数后,再根据统计出的具有显著性差异的天数对各个待检机台进行标记,由此,管理人员再根据各个待检机台的标记结果即可判断出哪些晶圆测试机台出现了异常情况。
因此,相较于现有技术中人工筛选数据和人工统计的方案,本公开的侦测方法能够自动筛选数据、自动统计,也就是说,该侦测方法能够自动完成晶圆测试机台的侦测过程,从而缩短了侦测过程的耗时。
而且,本公开的侦测方法根据预设筛选条件从原始测试数据中筛选出目标测试数据,使得用于统计的目标测试数据的数量较多,从而减小了数据分析的误差,也提高了晶圆测试机台侦测结果的准确性。
下面结合附图对本公开实施方式提供的侦测方法进行详细说明:
在步骤S110中,将多个晶圆测试机台测试到的同批次晶圆在多个测试项目下的原始测试数据存储到数据库中。
如前所述,晶圆测试机台能够对晶圆进行CP测试和FT测试,并记录下晶圆在CP测试和FT测试中的所有参数,其中:FT测试对应的参数为温度,而CP测试对应的参数包括电流、电压、感抗等,此处不再详细描述。
需要注意的是,本公开中假定同批次生产的晶圆在多个测试项目的原始测试数据都是相同的,同时,每种测试参数对应一个测试项目,因此,如果不同晶圆测试机台测试到的同批次晶圆在多个测试项目下的测试数据存在差异,则说明某台或某几台晶圆测试机台出现了异常情况。
在步骤S120中,根据预设筛选条件从数据库的原始测试数据中筛选出目标测试数据。
由于晶圆测试机台能够对多种类型的晶圆、在多个生产阶段的参数进行测试,因此,预设筛选条件可包括晶圆类型和生产阶段等,此处不再详细描述。如前所述,筛选过程可由晶圆测试机台的侦测系统自动进行,使得后续用于统计的目标测试数据的数量较多,从而减小了数据分析的误差,进而也提高了侦测结果的准确性。
在步骤S130中,对筛选出的目标测试数据进行统计,以区分出多个晶圆测试机台在各个测试项目下的对照机台和待检机台。
在示例性实施方式中,步骤S130可包括以下步骤:
步骤S1301,根据筛选出的各个晶圆测试机台的目标测试数据的中位数对各个晶圆测试机台进行排序,并取中间50%的晶圆测试机台作为稳定组。详细介绍,根据中位数对各个晶圆测试机台进行排序,可包括以下两个步骤:
步骤S13011,根据筛选出的各个晶圆测试机台的目标测试数据绘制各 个晶圆测试机台对应的箱线图(如图2所示)。
箱线图是一种用于显示数据组分散情况的统计图,能够反映原始数据分布的特征,同时,利用箱线图还可以进行多组数据分布特征的比较。在箱线图的绘制过程中,首先找出一组数据的上边缘、下边缘、中位数和两个四分位数;然后连接两个四分位数画出箱体;最后将上边缘和下边缘与箱体相连接,即可形成数据组的箱线图。当然,中位数位于箱体内部,此处不再详细描述。
步骤S13012,对各个箱线图中的中位数进行排序,从而完成各个晶圆测试机台的排序,然后取中间50%的晶圆测试机台作为稳定组,并用线框圈出来(如图3所示)。
当然,晶圆测试机台的数量为偶数个,以便于稳定组的选取。如图3所示,在选取稳定组后,对稳定组进行四等分,即可得到稳定组的上四分位数P75和下四分位数P25,此处不再详细描述。
步骤S1302,根据稳定组的上四分位数P75和下四分位数P25区分出多个晶圆测试机台在各个测试项目下的对照机台和待检机台。
例如,如果一晶圆测试机台的箱线图的中位数高于上四分位数P75,则该晶圆测试机台为待检机台;如果一晶圆测试机台的箱线图的中位数高于或等于下四分位数P25、并低于或等于上四分位数P75,则该晶圆测试机台为对照机台;如果一晶圆测试机台的箱线图的中位数低于下四分位数P25,则该晶圆测试机台为待检机台。
也就是说,如果晶圆测试机台的中位数位于下四分位数P25和上四分位数P75组成的区间内,则该晶圆测试机台为对照机台,相应地,其余情况均为待检机台。
由此,对照图3可知,机台①、机台②、机台③、机台④、机台⑥和机台⑤为对照机台,而机台⑦和机台⑧为待检机台,此处不再详细描述。
在步骤S140中,比较在第一预定天数内各个待检机台在对应的测试项目下的目标测试数据和对照机台在相同的测试项目下的目标测试数据是否具有显著性差异,并统计出各个待检机台具有显著性差异的天数。
其中,第一预定天数可以为7天,即:以一周时间为完整周期对各个晶圆测试机台进行侦测。当然,第一预定天数也可以5天、6天、8天或9天 等,此处不作特殊限定。
举例而言,可通过t检验法比较在第一预定天数内各个待检机台的目标测试数据和对照机台的目标测试数据是否具有显著性差异。当然,也可通过f检验法或卡方检验法等方法进行比较,此处不作特殊限定。
例如,如果一待检机台的目标测试数据具有显著性差异,则说明该待检机台存在异常风险;如果一待检机台的目标测试数据不具有显著性差异,则说明该待检机台运行良好。
在示例性实施方式中,t检验法满足如下第一关系式:
Figure PCTCN2021107441-appb-000001
其中,
Figure PCTCN2021107441-appb-000002
为对照机台的目标测试数据的平均值;μ 0为待检机台的目标测试数据的平均值;v为自由度;n为对照机台的目标测试数据的个数;S为对对照机台的各个目标测试数据的标准差,且S满足如下第二关系式:
Figure PCTCN2021107441-appb-000003
式中,x i为对照机台在第一预定天数内的各个目标测试数据。
下面举例对t检验法的判断过程进行详细介绍:
假设待检机台当天检测了三个批次的晶圆,其测量值分别为30、32和40,而对照机台当天检测的晶圆批次有十个,其测量值分别为10、11、10、13、15、25、9、12、10和8,结合第一关系式和第二关系式可知:
Figure PCTCN2021107441-appb-000004
然后,选择显著性水平a,一般选择(0.05,0.01,0.1),再考虑到严谨性,通常选择a=0.01,然后,参照表1中的单侧P(1)对应的数据得到2.821。由于|t|>2.821,所以,认为待检机台的测量值的均值大于对照机台的测量值的均值,即:该待检机台具有显著性差异。
表1 t界值表
Figure PCTCN2021107441-appb-000005
在步骤S150中,根据统计出的各个待检机台具有显著性差异的天数对各个待检机台进行标记。
在示例性实施方式中,步骤S150可包括以下步骤:
步骤S1501,如果一待检机台具有显著性差异的天数超过第一预定天数的预定比率,则对该待检机台打第一标记;
步骤S1502,如果一待检机台只在最近的第二预定天数内具有显著性差异,则对该待检机台打第二标记。
如前所述,第一预定天数可以为7天,而预定比率的取值范围可以为70%~90%,即:如果在一周时间内,待检机台有5天~6天都具有显著性差异,则对该待检机台打第一标记。当然,预定比率的取值范围也可以小于70%或大于90%,此处不作特殊限定。
第二预定天数可以为3天,即:如果在一周时间内,待检机台只在最近的3天具有显著性差异,则对该待检机台打第二标记。当然,第二预定天数也可以为2天或4天等,此处亦不作特殊限定。
因此,打第一标记的待检机台实际上比打第二标记的待检机台的运行情况更恶劣,相应地,第一标记可以为“警报”,而第二标记可以为“警告”,从而对待检机台的运行情况进行区分。
需要注意的是,在对各个待检机台进行标记之后,本公开实施方式的侦测方法还可包括:
将各个待检机台的标记结果通知给晶圆测试机台的管理人员,以便于管理人员掌握晶圆测试机台的具体状况,并据此优化晶圆实际的生产过程。
举例而言,可通过邮件或滚动屏等方式定期将晶圆测试机台的侦测情况发送给管理人员,此处不再详细描述。
本公开实施方式还提供一种晶圆测试机台的侦测装置,该侦测装置可包括存储单元、筛选单元、区分单元、比较单元和标记单元,其中:
存储单元,设置为将多个晶圆测试机台测试到的同批次晶圆在多个测试项目下的原始测试数据存储到数据库中;筛选单元,设置为根据预设筛选条件从数据库的原始测试数据中筛选出目标测试数据;区分单元,设置为对筛选出的目标测试数据进行统计,以区分出多个晶圆测试机台在各个测试项目下的对照机台和待检机台;比较单元,设置为比较在第一预定天数内各个待检机台在对应的测试项目下的目标测试数据和对照机台在相同的测试项目下的目标测试数据是否具有显著性差异,并统计出各个待检机台具有显著性差异的天数;标记单元,设置为根据统计出的各个待检机台具有显著性差异的天数对各个待检机台进行标记
如前所述,在对各个待检机台进行标记之后,可将各个待检机台的标记结果通知给晶圆测试机台的管理人员。相应地,该侦测装置还可包括通知单元,用来将各个待检机台的标记结果通知给晶圆测试机台的管理人员。
举例而言,该通知单元可以为邮件系统,即:标记单元和邮件系统连接,定期通过邮件将晶圆测试机台的侦测情况发送给管理人员,以便于管理人员掌握晶圆测试机台的具体状况,并据此优化晶圆实际的生产过程。
本说明书中各实施例或实施方式采用递进的方式描述,每个实施例重点说明的都是与其他实施例的不同之处,各个实施例之间相同相似部分相互参见即可。
在本说明书的描述中,参考术语“实施例”、“示例性的实施例”、“一些实施方式”、“示意性实施方式”、“示例”等的描述意指结合实施方式或示例描述的具体特征、结构、材料或者特点包含于本公开的至少一个实施方式或示例中。
在本说明书中,对上述术语的示意性表述不一定指的是相同的实施方式或示例。而且,描述的具体特征、结构、材料或者特点可以在任何的一个或 多个实施方式或示例中以合适的方式结合。
在本公开的描述中,需要说明的是,术语“中心”、“上”、“下”、“左”、“右”、“竖直”、“水平”、“内”、“外”等指示的方位或位置关系为基于附图所示的方位或位置关系,仅是为了便于描述本公开和简化描述,而不是指示或暗示所指的装置或元件必须具有特定的方位、以特定的方位构造和操作,因此不能理解为对本公开的限制。
可以理解的是,本公开所使用的术语“第一”、“第二”等可在本公开中用于描述各种结构,但这些结构不受这些术语的限制。这些术语仅用于将第一个结构与另一个结构区分。
在一个或多个附图中,相同的元件采用类似的附图标记来表示。为了清楚起见,附图中的多个部分没有按比例绘制。此外,可能未示出某些公知的部分。为了简明起见,可以在一幅图中描述经过数个步骤后获得的结构。在下文中描述了本公开的许多特定的细节,例如器件的结构、材料、尺寸、处理工艺和技术,以便更清楚地理解本公开。但正如本领域技术人员能够理解的那样,可以不按照这些特定的细节来实现本公开。
最后应说明的是:以上各实施例仅用以说明本公开的技术方案,而非对其限制;尽管参照前述各实施例对本公开进行了详细的说明,本领域技术人员应当理解:其依然可以对前述各实施例所记载的技术方案进行修改,或者对其中部分或者全部技术特征进行等同替换;而这些修改或者替换,并不使相应技术方案的本质脱离本公开各实施例技术方案的范围。
工业实用性
本公开实施例所提供的晶圆测试机台的侦测方法及侦测装置,在侦测过程中,首先,对根据预设筛选条件从数据库的原始测试数据中筛选出目标测试数据进行统计,以区分出多个晶圆测试机台在各个测试项目下的对照机台和待检机台,其中,对照机台相当于标准机台,即:对照机台没有出现异常情况,然后再以对照机台为标准来判定待检机台是否有异常情况。接着,比较在第一预定天数内各个待检机台在对应的测试项目下的目标测试数据和对照机台在相同的测试项目下的目标测试数据是否具有显著性差异,如果待检机台的目标测试数据具有显著性差异,则说明待检机台存在异常风险;如果 待检机台的目标测试数据不具有显著性差异,则说明待检机台运行良好。最后,在统计出各个待检机台具有显著性差异的天数后,根据统计出的各个待检机台具有显著性差异的天数对各个待检机台进行标记。由此,根据各个待检机台的标记结果即可判断出哪些晶圆测试机台出现了异常情况。由于本公开的侦测方法能够自动完成晶圆测试机台的侦测过程,从而缩短了侦测过程的耗时。而且,本公开的侦测方法根据预设筛选条件从原始测试数据中筛选出目标测试数据,使得用于统计的目标测试数据的数量较多,减小了数据分析的误差,进而提高了侦测结果的准确性。

Claims (12)

  1. 一种晶圆测试机台的侦测方法,包括:
    将多个晶圆测试机台测试到的同批次晶圆在多个测试项目下的原始测试数据存储到数据库中;
    根据预设筛选条件从所述数据库的原始测试数据中筛选出目标测试数据;
    对筛选出的所述目标测试数据进行统计,以区分出多个所述晶圆测试机台在各个所述测试项目下的对照机台和待检机台;
    比较在第一预定天数内各个所述待检机台在对应的测试项目下的目标测试数据和所述对照机台在相同的测试项目下的目标测试数据是否具有显著性差异,并统计出各个所述待检机台具有显著性差异的天数;
    根据统计出的各个所述待检机台具有显著性差异的天数对各个所述待检机台进行标记。
  2. 根据权利要求1所述的侦测方法,其中,对各个所述待检机台进行标记,包括:
    如果一所述待检机台具有显著性差异的天数超过所述第一预定天数的预定比率,则对该所述待检机台打第一标记;
    如果一所述待检机台只在最近的第二预定天数内具有显著性差异,则对该所述待检机台打第二标记。
  3. 根据权利要求2所述的侦测方法,其中,所述第一标记为“警报”,所述第二标记为“警告”。
  4. 根据权利要求2所述的侦测方法,其中,所述预定比率的取值范围为70%~90%。
  5. 根据权利要求1所述的侦测方法,其中,对筛选出的所述目标测试数据进行统计,以区分出多个所述晶圆测试机台在各个所述测试项目下的对照机台和待检机台,包括:
    根据筛选出的各个所述晶圆测试机台的目标测试数据的中位数对各个所述晶圆测试机台进行排序,并取中间50%的所述晶圆测试机台作为稳定组;
    根据所述稳定组的上四分位数和下四分位数区分出多个所述晶圆测试机台在各个所述测试项目下的所述对照机台和所述待检机台。
  6. 根据权利要求5所述的侦测方法,其中,根据筛选出的各个所述晶圆测试机台的目标测试数据的中位数对各个所述晶圆测试机台进行排序,包括:
    根据筛选出的各个所述晶圆测试机台的目标测试数据绘制各个所述晶圆测试机台对应的箱线图;
    对各个所述箱线图中的所述中位数进行排序,从而完成各个所述晶圆测试机台的排序。
  7. 根据权利要求6所述的侦测方法,其中,根据所述稳定组的上四分位数和下四分位数区分出多个所述晶圆测试机台在各个所述测试项目下所述对照机台和所述待检机台,包括:
    如果一所述晶圆测试机台的箱线图的中位数高于所述上四分位数,则该所述晶圆测试机台为所述待检机台;
    如果一所述晶圆测试机台的箱线图的中位数高于或等于所述下四分位数、并低于或等于所述上四分位数,则该所述晶圆测试机台为所述对照机台;
    如果一所述晶圆测试机台的箱线图的中位数低于所述下四分位数,则该所述晶圆测试机台为所述待检机台。
  8. 根据权利要求1所述的侦测方法,其中,通过t检验法比较在所述第一预定天数内各个所述待检机台的目标测试数据和所述对照机台的目标测试数据是否具有显著性差异。
  9. 根据权利要求8所述的侦测方法,其中,所述t检验法满足如下第一关系式:
    Figure PCTCN2021107441-appb-100001
    其中,
    Figure PCTCN2021107441-appb-100002
    为所述对照机台的目标测试数据的平均值;μ 0为所述待检机台的目标测试数据的平均值;v为自由度;n为所述对照机台的目标测试数据的个数;
    S为对所述对照机台的各个目标测试数据的标准差,且S满足如下第二 关系式:
    Figure PCTCN2021107441-appb-100003
    式中,x i为所述对照机台在所述第一预定天数内的各个目标测试数据。
  10. 根据权利要求1所述的侦测方法,在对各个所述待检机台进行标记之后,所述侦测方法还包括:
    将各个所述待检机台的标记结果通知给所述晶圆测试机台的管理人员。
  11. 一种晶圆测试机台的侦测装置,包括:
    存储单元,设置为将多个晶圆测试机台测试到的同批次晶圆在多个测试项目下的原始测试数据存储到数据库中;
    筛选单元,设置为根据预设筛选条件从所述数据库的原始测试数据中筛选出目标测试数据;
    区分单元,设置为对筛选出的所述目标测试数据进行统计,以区分出多个所述晶圆测试机台在各个所述测试项目下的对照机台和待检机台;
    比较单元,设置为比较在第一预定天数内各个所述待检机台在对应的测试项目下的目标测试数据和所述对照机台在相同的测试项目下的目标测试数据是否具有显著性差异,并统计出各个所述待检机台具有显著性差异的天数;
    标记单元,设置为根据统计出的各个所述待检机台具有显著性差异的天数对各个所述待检机台进行标记。
  12. 根据权利要求11所述的侦测装置,还包括:
    通知单元,设置为将各个所述待检机台的标记结果通知给所述晶圆测试机台的管理人员。
PCT/CN2021/107441 2020-11-10 2021-07-20 晶圆测试机台的侦测方法及侦测装置 WO2022100139A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US17/603,485 US20230063456A1 (en) 2020-11-10 2021-07-20 Detection method and detection device of wafer testing machine

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011245602.5A CN114461457A (zh) 2020-11-10 2020-11-10 晶圆测试机台的侦测方法及侦测装置
CN202011245602.5 2020-11-10

Publications (1)

Publication Number Publication Date
WO2022100139A1 true WO2022100139A1 (zh) 2022-05-19

Family

ID=81404856

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/107441 WO2022100139A1 (zh) 2020-11-10 2021-07-20 晶圆测试机台的侦测方法及侦测装置

Country Status (3)

Country Link
US (1) US20230063456A1 (zh)
CN (1) CN114461457A (zh)
WO (1) WO2022100139A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115711900A (zh) * 2022-11-30 2023-02-24 安测半导体技术(江苏)有限公司 一种基于神经网络的晶圆测试检测方法

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080155507A1 (en) * 2006-12-22 2008-06-26 Compagnie Industrielle Et Financiere D'ingenierie "Ingenico" Method for auditing compliance of an electronic platform and/or a computer program present on said platform, and device and computer program corresponding thereto
US20120326060A1 (en) * 2011-06-22 2012-12-27 Boaz Kenane Testing method for led wafer
CN105404713A (zh) * 2015-10-26 2016-03-16 威凯检测技术有限公司 一种用于筛选冰箱压缩机制冷量比对测试标准样机的方法
CN105988434A (zh) * 2015-02-04 2016-10-05 中芯国际集成电路制造(上海)有限公司 监测制造机台的方法及其系统
CN106324383A (zh) * 2016-07-29 2017-01-11 北京润科通用技术有限公司 一种机载设备测试方法、系统及仿真控制测试装置
CN106794358A (zh) * 2014-06-20 2017-05-31 华盛顿大学 使用电子射野影像装置(epid)的线性加速器(linac)的验收、调试和持续基准测试
CN107291582A (zh) * 2017-06-13 2017-10-24 常熟天合信电子科技有限公司 基于标准大数据的智能化集成测试方法
CN107679163A (zh) * 2017-09-28 2018-02-09 成都海威华芯科技有限公司 一种单步工序制造因素显著差异分析系统和分析方法
CN111626755A (zh) * 2020-05-29 2020-09-04 歌尔科技有限公司 一种工程样机管理方法、装置及其相关设备
CN111693846A (zh) * 2019-03-12 2020-09-22 新唐科技股份有限公司 芯片测试系统及方法

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5726920A (en) * 1995-09-29 1998-03-10 Advanced Micro Devices, Inc. Watchdog system having data differentiating means for use in monitoring of semiconductor wafer testing line
US6699627B2 (en) * 2000-12-08 2004-03-02 Adlai Smith Reference wafer and process for manufacturing same
US6792386B2 (en) * 2001-12-28 2004-09-14 Texas Instruments Incorporated Method and system for statistical comparison of a plurality of testers
KR100909474B1 (ko) * 2005-08-10 2009-07-28 삼성전자주식회사 웨이퍼 결함지수를 사용하여 국부성 불량 모드를 갖는결함성 반도체 웨이퍼의 검출 방법들 및 이에 사용되는장비들
CN101334414B (zh) * 2007-06-29 2011-11-30 中芯国际集成电路制造(上海)有限公司 用于晶片的缺陷检测机台的匹配方法
JP4931710B2 (ja) * 2007-06-29 2012-05-16 株式会社リコー ウエハにおける良品チップ分類方法、それを用いたチップ品質判定方法、ならびにチップ分類プログラム、チップ品質判定プログラム、マーキング機構及び半導体装置の製造方法
US7856332B2 (en) * 2007-12-04 2010-12-21 International Business Machines Corporation Real time system for monitoring the commonality, sensitivity, and repeatability of test probes
JP2012122765A (ja) * 2010-12-06 2012-06-28 Hitachi High-Technologies Corp 欠陥検査装置
CN108400098A (zh) * 2017-02-08 2018-08-14 上海华岭集成电路技术股份有限公司 验证晶圆测试相关性的方法

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080155507A1 (en) * 2006-12-22 2008-06-26 Compagnie Industrielle Et Financiere D'ingenierie "Ingenico" Method for auditing compliance of an electronic platform and/or a computer program present on said platform, and device and computer program corresponding thereto
US20120326060A1 (en) * 2011-06-22 2012-12-27 Boaz Kenane Testing method for led wafer
CN106794358A (zh) * 2014-06-20 2017-05-31 华盛顿大学 使用电子射野影像装置(epid)的线性加速器(linac)的验收、调试和持续基准测试
CN105988434A (zh) * 2015-02-04 2016-10-05 中芯国际集成电路制造(上海)有限公司 监测制造机台的方法及其系统
CN105404713A (zh) * 2015-10-26 2016-03-16 威凯检测技术有限公司 一种用于筛选冰箱压缩机制冷量比对测试标准样机的方法
CN106324383A (zh) * 2016-07-29 2017-01-11 北京润科通用技术有限公司 一种机载设备测试方法、系统及仿真控制测试装置
CN107291582A (zh) * 2017-06-13 2017-10-24 常熟天合信电子科技有限公司 基于标准大数据的智能化集成测试方法
CN107679163A (zh) * 2017-09-28 2018-02-09 成都海威华芯科技有限公司 一种单步工序制造因素显著差异分析系统和分析方法
CN111693846A (zh) * 2019-03-12 2020-09-22 新唐科技股份有限公司 芯片测试系统及方法
CN111626755A (zh) * 2020-05-29 2020-09-04 歌尔科技有限公司 一种工程样机管理方法、装置及其相关设备

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115711900A (zh) * 2022-11-30 2023-02-24 安测半导体技术(江苏)有限公司 一种基于神经网络的晶圆测试检测方法

Also Published As

Publication number Publication date
US20230063456A1 (en) 2023-03-02
CN114461457A (zh) 2022-05-10

Similar Documents

Publication Publication Date Title
CN109767996A (zh) 晶圆缺陷分析系统及分析方法
CN104867840B (zh) Ye在线检测管控方法
WO2022028102A1 (zh) 测试方法及测试系统
WO2022100139A1 (zh) 晶圆测试机台的侦测方法及侦测装置
CN113406502B (zh) 基于机器学习的缺陷电池筛选方法、设备及介质
CN110596566B (zh) 一种用于ate系统的dpat测试方法
TW442880B (en) Method for automatically classifying the wafer with failure mode
CN115360116B (zh) 一种晶圆的缺陷检测方法及系统
US7937168B2 (en) Automated abnormal machine tracking and notifying system and method
TW201533456A (zh) 晶圓測試數據分析方法
CN105478364B (zh) 一种不良品检测分类方法及系统
JP2007095953A (ja) 半導体装置の選別方法及び半導体装置の選別装置
CN115774185B (zh) 一种车规级芯片dpat检测方法及装置
CN101493494B (zh) 晶圆的测试方法与系统
CN111106028A (zh) 一种半导体芯片测试过程实时监控方法
JPH08124977A (ja) 半導体装置不良解析システム
CN110160918A (zh) 晶圆再验的方法
TW201546467A (zh) 晶圓測試特殊圖案及探針卡缺陷的檢驗方法
JP3550525B2 (ja) 不良モードを有するテストウェハーの自動分類方法
TW202328665A (zh) 分析缺陷的方法
JP2001217169A (ja) データ変動監視方法と監視装置
TWI330304B (en) Debug method for determining failure machines and application system thereof
CN117217476B (zh) 一种产线测试资源自动化配置方法
CN103971194A (zh) 送验材料清单的产生装置及方法
CN111755079B (zh) 一种多晶硅最优原料配比方案确定方法及系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21890670

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21890670

Country of ref document: EP

Kind code of ref document: A1