How to do a proper prior art search…

Sean Moolman Polymers, Ceramics and Composites 5 April 2006

Why do we do prior art searches?

Page 2 • • Want to know if what we have invented is unique • Can lead to discovery of interesting and useful literature that can inform our own research • Want to determine how close the prior art is to our invention (patentability) This is very important, as patent costs are very high – e.g. filing an international patent application in approx. 10 countries will amount to a total cost of about R1.2 million over a 3-4 year period © CSIR 2006 www.csir.co.za

Page 3

What is classified as ‘prior art’?

• • ANY information in the public domain… Includes video recordings, news paper & magazine articles, speeches, journal papers, patents, etc.

• If it is public ANYWHERE in the world, it is prior art!

(E.g. an old Egyptian parchment in hieroglyphics, held in a library in Tibet = prior art!) © CSIR 2006 www.csir.co.za

Page 4

What is a patent?

• • • Monopoly on a product/process/technique in return for FULL disclosure Patent system = to reward inventor & benefit public Requirements for patentability: 1.

New 2.

Unobvious (to person of ordinary skill in the art – “unexpected”, etc.) 3.

Useful © CSIR 2006 www.csir.co.za

Page 5

Usual problem = proving unobviousness…

• • Hindsight easier than foresight… Look for these things: 1.

Commercial success 2.

Long-felt need without solution 3.

Failure of others 4.

Unexpected results 5.

Superior properties 6.

Professional approval such as in trade publications and texts © CSIR 2006 www.csir.co.za

Main steps in prior art searching

1.

Identify keywords 2.

Set up search terms from keywords 3.

Select databases 4.

Perform search and examination 5.

Summarize and discuss relevant prior art found and give rating for how problematic it is NOTE: Prior art search can never be COMPLETE!!! Can never say with 100% certainty that no prior art exists… Can only try and do as comprehensive a prior art search as possible and capture it properly.

Page 6 © CSIR 2006 www.csir.co.za

STEP 1: Identify keywords

• • • Probably most critical step… Worthwhile to spend some time on this How to identify keywords: 1.

Try and crystallize the two or three MAIN components that make your invention unique (e.g. enzyme; emulsion; crosslink).

2.

Try and think of all the synonyms and similar concepts for each of these and list them.

3.

While searching, you might discover more synonyms & alternative words – incorporate these into your search terms and start again with the searching… Page 7 © CSIR 2006 www.csir.co.za

STEP 2: Set up search terms

• • • Use all the synonyms and alternate words Go from general to specific – have all the synonyms and options in your first search term, and narrow down to very specific words and your core concepts for the last search term This is to ensure that you a) Cover as much ground as possible and try and pick up prior art that you might otherwise have missed (broad) b) See how much prior art there is that is very closely related to your invention (narrow) Page 8 © CSIR 2006 www.csir.co.za

STEP 3: Select databases

• Depends on how much time available & how thorough you want to / need to be… • At the very least it should include: • • • US Patent database ( http://www.uspto.gov/patft/index.html

) European patent database ( http://ep.espacenet.com

) Google Scholar ( http://scholar.google.com

) Remember to search both granted patents & applications!

• For more thorough search, can also include: • Information specialist searches (Dialog, Thomson Innovation, Q Pat (Questel), etc.) • • • • • Page 9 Sciencedirect Ebscohost Scopus Google Etc.

© CSIR 2006 www.csir.co.za

STEP 4: Perform search and examination…

• • • Modify search terms to suit each search engine (e.g. truncation symbols) Example of truncation symbols for different search engines: • • • • • US PTO EPO Sciencedirect Ebscohost Google & Google Scholar $ * !

* No truncation available. Can use ~ immediately preceding a word for synonyms When obtaining large no. of results, check for obvious exclusions and modify search terms accordingly (e.g. using the term cell and getting a lot of results relating to cell phones, then use “-phone*” to try and exclude these…) Page 10 © CSIR 2006 www.csir.co.za

STEP 4: Perform search and examination…

• • • • As soon as you get down to below about 250 results, it becomes manageable to quickly scan the abstracts to see whether the items are relevant You will soon start to see patterns – e.g. a family of very similar patents – then you don’t have to examine each of them in detail Make notes of the relevant ones as you go along & save them somewhere for more detailed study later The number of results you obtain is a vague indication of the patentability (especially the very focused search terms) Page 11 © CSIR 2006 www.csir.co.za

STEP 5: Summarize, discuss, rate

• • • Once your search is complete, examine the relevant prior art items in detail Discuss the reasons why your invention is patentable against each of these items Rate them in terms of how problematic they are according to you (this can e.g. help you and the patent lawyer later to write the patent application in a certain way so as to try and minimize problems / objections from the patent office! It can also help you to identify any further improvements to the invention that can help differentiate from the prior art Page 12 © CSIR 2006 www.csir.co.za