Data processing: measuring – calibrating – or testing – Measurement system in a specific environment – Quality evaluation
Reexamination Certificate
2000-11-09
2004-10-26
Barlow, John (Department: 2863)
Data processing: measuring, calibrating, or testing
Measurement system in a specific environment
Quality evaluation
Reexamination Certificate
active
06810344
ABSTRACT:
CROSS-REFERENCE TO RELATED APPLICATION
This application claims benefit of priority under 35 USC § 119 to both Japanese Patent Applications No. 1999-321105, filed on Nov. 11, 1999, and No. 2000-085378, filed on Mar. 24, 2000, the entire contents of which are incorporated herein by reference herein.
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to the field of testing for analyzing defects in semiconductor devices, and, more particularly, to a semiconductor testing method, and a semiconductor testing apparatus for deciding whether a semiconductor device is a good (non-defective) device having no defect or it contains defeats by using IDDQ testing technique, and a program for executing the semiconductor testing method, and the present invention also relates to a semiconductor testing method and a semiconductor testing apparatus for specifying one or more faulty parts involved in the semiconductor device.
2. Description of the Related Art
Recently, the circuit size of a semiconductor chip is enormously increased with an increase of the degree of a microstructure of a semiconductor circuit as a semiconductor device. It is therefore difficult to detect and specify the cause of a fault occurred in the complicated semiconductor circuit.
In particularly, the generation of test vectors to be used for detecting various faults in the semiconductor device with a higher detection rate requires considerable workers and long working time even if the test vectors are generated only by manual or an automatic test pattern generator (ATPG) that generate test vectors.
In addition, it is difficult to detect defects in a semiconductor device having a complicated structure and a large integrated size only by performing functional testing. In order to avoid this drawback, attention is being given to IDDQ testing having a higher detection rate as a new testing technology. This IDDQ testing is a testing method using existing test vectors capable of obtaining a higher detection rate.
By the way, conventionally, when existing test vectors are used in the measurement based on the IDDQ testing, a circuit designer selects the test vectors matched to target circuits for testing. Furthermore, an available IDDQ test vector extraction tool is used as one of recently effective testing methods in order to select effective test vectors.
However, it is considerably difficult to select effective test vectors for performing IDDQ measuring according to the circuit designer's judgment, and the test vectors selected in estimation are not always effective for detecting defects. As a result, an operator cannot select the test vectors efficiently, and much time is thereby spent in measurement using no effective test vectors. In addition, the using of an available IDDQ test-vector extracting tool needs to keep the circumstance for the execution of this tool. That is, firstly, the available IDDQ test-vector extracting tool is selected, and this tool is then installed into a system including storage means such as a hard disk. Furthermore, it is necessary to convert information such as circuits and test vectors into data of a dedicated format that cFan be executed by this tool. This format conversion for the execution of the tool requires much working time. In order to avoid this drawback, it is sometimes necessary to develop a dedicated tool to be used only for this format conversion.
On the other hand, the available IDDQ test-vector extracting tool extracts only test vectors estimated by the execution of a computer simulation. Accordingly, all the estimated test vectors are not always effective. In general, it is often happened that a faulty item detected after actual mass production and before shipping has a limited faulty part in a circuit in the faulty item. On the contrary, because the available IDDQ test-vector extracting tool estimates and generates the test vectors based on the measurement for the entire circuits in a target semiconductor device, the estimated test vectors includes un-necessary test vectors in judgment of the defective item having a limited faulty part. This spends unnecessary much time in the IDDQ measurement. Moreover, in the prior art, the current value corresponding to a test vector is used in the judgment to detect and select a faulty sample. This judgment method decides a sample having a current value that is in excess of a standard current value as a faulty item, and a sample of a current value that is not more than a current-value criteria as a good sample (namely, a non-defective sample as a passed device that is within a manufacture's tolerance level and may be sold to a customer). Accordingly, this conventional judgment method judges a sample as a passed device even if the current value of this sample has a larger current-value change rate, but not in excess of the current value criteria. This conventional judgment method misses to detect a faulty sample correctly.
By the way, there is a case that a manufacture cannot detect any defect in a semiconductor product that has been passed in manufacture's testing (namely, the semiconductor product is within a manufacture's tolerance level and may be sold to a customer), but returned from user as a fault product after shipping. To obtain the method to efficiently specify defects in the faulty product presents an important problem. For example, an emission analysis using the IDDQ testing has been performed. This mission analysis using the IDDQ testing measures a current value output from a target semiconductor circuit using existence test vectors. The emission analysis is then performed in order to detect defective parts in the faulty sample having unique emission by using the test vectors corresponding to the current values that are larger than the current value detected in the passed sample.
In the conventional emission analysis described above based on the IDDQ testing, all the unique emission parts detected in the faulty sample are used for the faulty analysis. This causes a time-wasting in the emission analysis because this emission analysis is performed based on the inefficient test vectors including the analysis for the emission parts that are not generated by the defects.
Recently, there is a possibility to flow the current continuously in the passed sample having a complicated circuit structure. This type of current gives no effect to the operation. On the contrary, the conventional emission analysis using the test vectors corresponding to the current values which are greater than that of the passed sample omits the faulty sample having the current value that is smaller than that of the passed sample from the emission analysis. That is, it is impossible to detect the faulty sample having a larger current value.
As described above, because the conventional emission analysis to analyze faulty samples is not always effective, there is no means to specify the position of a faulty part in the faulty sample having a small current value when compared with the current value flowing through the passed sample. Accordingly, it is difficult to obtain effective test vectors to be used during the emission measurement and to specify the defective part in the faulty sample. Thereby, the operator gives up the execution of the emission measurement using effective test vectors. Further, in the conventional method to specify the faulty part based on the current value data, because the current values between the passed sample and the faulty sample corresponding to a test vector are compared and the emission measurement is performed only when the current value of the faulty sample is greater than that of the passed sample, it has not performed to compare the emission parts in the faulty sample using the difference of current values per address pair indicating the test vectors. In other words, in the prior art there is no attention to consider the difference of the current values per test vector pair for each of the passed sample and the faulty sample. Thereby, the operator often misses to detect the phen
Barlow John
Finnegan Henderson Farabow Garrett & Dunner L.L.P.
Kabushiki Kaisha Toshiba
Lau Tung
LandOfFree
Semiconductor testing method and semiconductor testing... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Semiconductor testing method and semiconductor testing..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Semiconductor testing method and semiconductor testing... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3273920