Method to extract the crucial dimension of a semiconductor structure. The method consists of: 1.) the determination of a range for each parameter that is to be extracted and generating an electronic spectra database using training spectra and support machine (SVM) training networks to aid in the training of SVMs; 2) making use of the SVMs after training to map measured spectra to produce a matching electronic spectra database; and 3) employing a search algorithm to search for an optimal simulation spectrum within the corresponding electronic spectra database, simulation parameters matching the simulation spectrum being the critical size of the nanostructure that needs to be extracted.

1. Field of the Invention

The invention is related to optical scattering and semiconductor measurement field, and, more specifically it relates to a method of extracting a critical dimension of the semiconductor nanostructure.

2. Description of Related Art

It is important to measure 3D morphology parameters, like feature linewidth, sideswall angles and cycles, to ensure repeatability, operability and the flexibility of the manufacturing process.

In the semiconductor and optical measurement field, an optical scatterometer is the most popular-used instrument to measure important dimensions. Forward optical modeling as well as reverse looking are two methods for measuring the optical scatterometer. This is where you can carry out optical scattering field simulations on the models of geometric nanostructures that are to be studied. Reverse-selecting is the continuous comparison of the measured spectra to the simulation spectraand the parameters of a model corresponding to a simulation spectrum with the greatest degree of similarity are the elements of the nanostructures to be measured. The most popular-used method during reverse seeking of the optical scatterometer is alibrary-matching-based method. This technique involves creating the database of simulation spectra to be used in an underlying structure model to determine the spectrum of that database corresponding with an individual parameter value is sought out for the most comparable simulation spectrum to that model according to an evaluation function. The parameter value of an model is the amount of the simulation spectrum. The spectradatabase may be vast and includes a lot of simulation spectra. It can result in an upward progression of geometrical form in which the database includes many more simulation spectra as well as every spectrum that is independent corresponds to the model as determined by the parameter value. Finally, an evaluation function searches for a simulation spectrum most similar to the model according to the parameter value for a model. The parameter value of a model corresponding to the spectrum of simulation is the structure to be measured. To meet requirements for real-time feature and rapidity in industries, and to implement fast mapping of measured spectra in a large-scale spectra database, a new mapping method needs to be launched, and an old full-librarysearching method needs to be discarded.

Library-mapping-based extraction of geometric parameters of a nano structure includes establishment of a simulation database, and searching in a database. The process of searching in a database involves the search for a spectrum that is most similar to ameasured spectrum within the database, following a set of rules and is the most common type of maximum proximity searching. The traditional methods to solve the problem of maximum proximity include a direct whole-library search method, a k-d treemethod, an analysis of clusters, a local sensitivity hashing and so on. Other than the whole-library search method, however it is not possible to achieve an optimal solution. This is due to the techniques used to solve problems such as the most-similar spectra hunts are not able produce a good global solution. Because spectra have non-obvious characteristics, these methods employ one or more parameters to determine the search. The GPU (Graphic Processing Unit) is utilized to store the most similar simulation spectra within a database. The GPU (Graphic Processing Unit) is an acceleration device specially designed to handle images. It has a faster processing speed for data, and superior data processing capabilities, and more concurrent computation capabilities over CPU. By arranging multiple simulation spectrums into an image matrix that indicates spectra it is possible to quickly map the spectra of the spectrum and determine geometric parameters. However, problems with this method are that, with further expansion ofthe simulation spectra database, more powerful and high-efficient GPUs must be used, which limits scalability of this category of hardware-acceleration-based searching method.

A method to extract a critical dimension from the structure of a semiconductor is an objective of the invention. It can be used to rapidly and precisely extract feature linewidth,height and sidewall angles. This process is simple.

To attain the above goal, in accordance with one embodiment of the invention, there is provided a method for extraction of a critical dimension in a semiconductor nanostructure, the method comprises the steps of: (1) the determination of a rangefor each parameter to be extracted, creating an electronic spectra database using training spectra as well as support vector machine (SVM) training networks to aid in the training of SVMs; (2) employing the SVMs after training to map measured spectrato create a digital spectra database; and (3) employing a searching algorithm to search for an optimum simulation spectrum in the electronic spectra database. The simulation parameters that correspond to the spectrum of simulation being the crucial dimension of the semiconductor nanostructure to be extracted.

Electronic spectra databases may be obtained from a class according to this model by: dividing the ranges of values of each parameter into several subvalue ranges, selecting one subvalue range for each subvalue range, and creating a subparameter mix. Then, using forward optical modeling software to generate a simulation spectrum of the values that are discrete. This simulation spectrum corresponds to all discrete points in every subparameter combination. values.

In a class of this kind, during the process of removing the training spectra, every parameter extracted is associated with a SVM and the number of categories that are contained in an output of the SVM is determined by that of sub-ranges dividedfrom the value range of each parameter to be extracted, each category is identified with an unique number, and is can be used to identify the sub-range. Each sub-range corresponds to a spectra of training set, and each training spectra set represents a sub-range, by randomlychoosing several discrete values within a sub-range, and generating a simulation spectrum for each point on the discrete side using the forward optical modeling program. which is referred to as a simulation spectra collection or the training spectra set is produced, and all of the training spectra containedin training spectra sets that are corresponding to the output ends of the SVM are the training spectra that are that are required by the SVM.

A class in this embodiment has multiple categories at the output of an SVM. Each category represents a subrange of the value range the parameter is attempting to extract. The output of the SVM has a category which indicates a subrange of the value range.

This class describes how to search for the most optimal spectrum for simulation. It involves placing all the electronic spectrum into a matrix with each row being a simulation spectrum, and the simulation spectrumuniquelycorresponding to group parameter computing evaluation function values with each simulation spectra , and the spectrum that is measured in the matrix beginning at the top according to an evaluation function searching using a search algorithm or sorting algorithm for the value of the minimum evaluation function, the simulation spectrum that is the one with the lowest evaluation value being the best simulation spectrum

The advantages of the method of extracting a critical dimension of a semiconductor nanostructure are summarized below. Compared with a library-mapping-based method for extracting feature dimensions in the prior art, the invention is capable ofmapping the measured spectra into a small-range sub-database by adding a process of off-line training using a SVM classifier, and time spent on searching in the sub-database is far less than that on most-similar spectra searching in a large database. Inaddition, by increasing the number of categories within each classifier (namely splitting the value range of each parameter into multiple sub-ranges) A smaller sub-database is possible to be created that further speeds up the extraction of parameters, and the invention also provides a an automated and controlled extraction speed of parameters.

Click here to view the patent on USPTO website.

Get Patents with PatentPC

What is a patent?

Patents are granted by the government in order to protect the invention. It gives the inventor the sole right to develop, utilize and market the invention. Society benefits when new technology is brought to the market. These benefits could be directly realized when people can perform feats previously thought impossible, or indirectly through the economic opportunities that innovation offers (business growth, employment).

Patent protection is sought by a variety of university researchers and drug companies to protect their research and development. Patents can be granted to the physical or abstract nature of a product or process or the method or composition of materials unique to the area. To be granted patent protection, an invention must be valuable or novel, as well as not readily apparent to anyone else in the same subject.

Patents give inventors a chance to be recognized for commercially successful inventions. They provide a reason for inventors to invent. Small-scale businesses and inventors are assured that they will get a return on their investment in technology advancement through patents. They could earn a decent income from their work.

Companies that are able to:

Create and protect innovative products and services;

Enhance the visibility and worth of your products on market

Make your company and products stand out from the rest;

Find business and technical information.

Beware of the possibility of accidentally using third-party proprietary content, or losing valuable information, innovative outputs, or other creative output.

Patents transform inventors’ knowledge into a marketable asset that opens up new possibilities for employment creation and expansion of businesses by licensing or joint ventures.

Small businesses that have patent protection are more appealing to investors who are involved in the commercialization and development of technology.

Patenting can generate new ideas and new inventions. These information may be protected by patents.

Patents can be used to prevent untrustworthy third parties from profiting through the work of inventions.

Commercially successful patent-protected technology revenues can be used to fund technological research and development (R&D) and improve the chances of developing better technology in the near future.

Intellectual property ownership is a way to convince lenders and investors that there are genuine opportunities to market your product. One powerful patent may open the door for numerous financing options. Patents as well as other IP assets are able to be utilized as collateral or security for debt financing. Investors are also able to view your patent assets to boost the value of their company. Forbes and other publications have reported that each patent can add between $500,000 and a million dollars in company valuation.

A well-written business plan is vital for new businesses. It must be built on IP and demonstrate what your service or product stands out. Investors are also impressed if you have IP rights are secure or are on the verge of being secure and if they are supportive of your business strategy.

It is vital to keep an invention secret before applying for patent protection. It is crucial to keep the invention private prior to filing for patent protection. Public disclosure can often make an invention, making it inepatentable. Therefore, prior filing disclosures (e.g. for test-marketing investors, test-marketing, or any other business partners) should only be made after signing a confidentiality agreement.

There are several types of patents. Understanding the different types is crucial to protect your invention. Patents for utility are used to protect the development of new methods and machines. Design patents cover ornamental designs. Utility patents are the best as they protect the owner from copycats as well as other competitors. Most often they are granted for alterations or improvements to existing inventions. Utility patents can also be used to enhance or alter existing inventions. For example, a process patent covers acts or methods of performing an action, while chemical compositions are an assortment of components.

How long does a patent last? Patents for utility last for 20 years from the initial date they were filed, however, their expiration dates can be extended because of delays in the patent office for instance.

Do you want to patent your ideas? Patents are granted only for first-to-file applicants, you need to file quickly – call an attorney for patents at PatentPC to patent your idea now!

A patent search is essential when you’re drafting a patent application. This will enable you to look at other ideas and give you an understanding of them. You’ll be able reduce the scope of your idea. Furthermore, you’ll be aware of the current state of technological advancements in your field of innovation. This will allow you to comprehend the scope of your invention and prepare you to file your patent application.

How to Search for Patents

The first step in getting the patent you want is to conduct a patent search. You can do a google patent search or do a USPTO search. Once the patent application is submitted, the product that is that is covered by the patent application could be called patent-pending, and you can locate the patent application on a public pair. Once the patent office has approved your application, you’ll be able do the patent number lookup to locate the issued patent. Your product will then be patentable. Alongside the USPTO search engine, you can also utilize other search engines such as espacenet as described below. For assistance, consult a patent lawyer. In the US patents are issued by the US trademark and patent office, or the United States patent and trademark office, which is also responsible for examining trademark applications.

Are you interested in finding similar patents? These are the steps you should follow:

1. Brainstorm terms that describe your invention based on the purpose, composition and application.

Write down a brief, but precise explanation of your invention. Don’t use generic terms like “device”, “process” or “system”. Instead, look for synonyms for the terms you selected initially. Next, take note of important technical terms and key words.

Utilize the following questions to help you find the keywords or concepts.

  • What is the objective of this invention? Is it a utilitarian device or an ornamental design?
  • Is the invention a way of making something or performing a function? Or is it a thing or process?
  • What is the nature and purpose of the invention? What is the physical makeup of the invention?
  • What is the goal of the invention
  • What are the terms and phrases in the field of technology used to describe the nature of an invention? To assist you in finding the appropriate terms, use the technical dictionary.

2. These terms enable you to search for relevant Cooperative Patent Classifications at Classification Search Tool. To determine the best classification to your invention, scan the resulting classification’s class Schemes (class schedules). If you don’t get any results using the Classification Text Search, you might consider substituting your words that describe your invention with synonyms.

3. Go through 3. Go over the CPC Classification Definition for the CPC Classification Definition to verify the relevancy of the CPC classification you’ve found. The link to a CPC classification definition is available in the event that the title of the classification has a blue box that includes “D” to the left. CPC classification definitions can aid you in determining the classification’s scope, so you can select the most relevant. These definitions may also include some search tips or other recommendations that could be helpful for further research.

4. The Patents Full-Text Database and the Image Database allow you to search for patent documents that have the CPC classification. By focusing on abstracts and drawings that are representative you can narrow your search for the most relevant patent publications.

5. This collection of patent publication is the best to check for similarity to your invention. Pay attention to the specification and claims. Contact the applicant as well as the patent examiner for additional patents.

6. You can find published patent applications that meet the CPC classification you selected in Step 3. You may also employ the same search strategy that you employed in Step 4 to narrow your search results to only the most relevant patent applications by reviewing the abstracts as well as the drawings for each page. After that, take a close look at the published patent applications with particular attention paid to the claims as well as additional drawings.

7. Find other US patent publications by keyword searches in PatFT and AppFT databases, classification search of non-U.S. patents as described below, and searching for non-patent patent disclosures in the literature of inventions using web search engines. For example:

  • Add keywords to your search. Keyword searches may turn up documents that are not well-categorized or have missed classifications during Step 2. For example, US patent examiners often supplement their classification searches with keyword searches. Think about the use of technical engineering terminology rather than everyday words.
  • Search for foreign patents using the CPC classification. Then, re-run the search using international patent office search engines such as Espacenet, the European Patent Office’s worldwide patent publication database of over 130 million patent publications. Other national databases include:
  • Search non-patent literature. Inventions can be made public in many non-patent publications. It is recommended that you search journals, books, websites, technical catalogs, conference proceedings, and other print and electronic publications.

To review your search, you can hire a registered patent attorney to assist. A preliminary search will help one better prepare to talk about their invention and other related inventions with a professional patent attorney. In addition, the attorney will not spend too much time or money on patenting basics.