One of the chores that must be done when developing a new technology is “Due Diligence” as applied to intellectual property. In the fabulous world of industrial chemistry, there is an overlapping of the three great magisteria- Business, Chemical Science, and Law. In order to get a new product or process on stream, we must find a line of sight through the many hoops and past the many gatekeepers of those magisteria that can obstruct our path the fame and fortune.
First, allow me to pay homage to two great and wondrous database services- the United States Patent and Trademark Office (USPTO) and Chemical Abstracts Service (CAS). They are the custodians of data generated by some of the most cantankerous and unruly people on earth- lawyers and chemists. Their task is complex and effectively endless.
In actually trying to do a reliable due diligence analysis, a searcher must ascertain that a proposed bit of IP does not conflict with claims in valid patents owned by others. In a chemical IP search, one can divide the claim universe into 2 domains- composition of matter claims and process claims. In the patent, the first sentence will state whether a composition is being claimed or that a process is being claimed. Some patents have both composition claims and process claims.
So here is the problem. Let’s say that you are trying to determine whether or not you have the right to manufacture a particular molecule. This apparently simple question actually deconvolutes to two fundamental questions: 1) Is the composition of matter in the public domain? And, 2) are there claimed processes for it’s manufacture?
There is a third domain and that relates to the use of the composition. However, this does not impinge on the right to manufacture the material, just it’s use.
If the material is claimed as a 1) composition of matter in a valid patent, then you cannot lawfully make it or possess it (for commercial use, though precendence is being set for a bar on R&D use as well) without the permission of the assignee of the patent. Note that the owner of the patent is the assignee, not the inventor(s). If the composition of matter is not claimed, then it is in the public domain, assuming that you did not learn of it under trade secrecy.
So, let’s say then that your target material is in the public domain. Now the question is 2); does your proposed method for its manufacture infringe on claimed methods? This may be a hard or expensive question to answer, and the reason is plain. When you execute a search for IP issues related to a substance, you search path is limited to fields, key words, structures, CASRN’s, etc. that are flagged in the major databases.
A CAS search on a given compound will lead to patent families that mention the compound, its preparation, or its use. But you have no way of knowing whether the patent reference claims the composition, its preparation, or its use. It could very well be that there are no claims pertaining to the compound of interest- it was just cited as an example of some sort.
A CAS search is highly accurate in terms of the focus on a particular compound. However, a USPTO search is not. A chemical search of the USPTO public database (USPTO.gov) is pretty much limited to a search for specific character strings. It is possible to narrow down the scope of a search by concentrating on classification numbers, but I have never been convinced of its thoroughness.
After all of this set up, here is my point. The problem we all face in doing our IP due diligence is that there is no direct means for determining from an indication in a database search report whether or not a composition is in the public domain. A CAS search will not yield a clear yes or no, and the USPTO database search only retrieves hits that have the requested strings. Despite the advances in database technology, the user still has to collect all of the patent citations pertaining to the material and sift through them and interpret the claim language.
Wouldn’t it be useful to the public if an applicant for a US patent were required to collate the claimed compositions for uplink into a database? With such a “field” in a CAS or Beilstein search, you could tell in an instant if the composition was claimed. The same argument holds true for processes. At present, the “retrievability” of claimed art is poor.
Patent attorneys are likely to object along the following arguments: not all patents that you retrieve from a search on CAS or USPTO will be valid. Some patents will have expired naturally, others will have expired for non-payment of fees, and still others will have serious weaknesses that will only be apparent from an examination of the prosecution history as revealed in the file wrapper. Abandonment may be difficult to detect for abstracting services, as would flaws in the prosecution as documented by the wrapper.
Another objection that is unlikely to be openly identified is the matter of clarity. There is may be advantage conferred to assignees when a claim is a bit fuzzy. This may afford some manuevering room during an infringement action, though it might be hard to say who the beneficiary would really be. I would estimate that whomever had the most persuasive attorneys would prevail.
It would be interesting to hear from others about this matter.
Lamentations on Chemistry 
Acutally, Scifinder Scholar can search structures in US patents. But it’s expensive to use.
Hmmm. It sounds like the same feature that I use in the “commercial” version. I can do a structure search and find patents galore, but there is no indication in the SciFinder abstract if the composition is actually claimed or if there may be claimed processes for its preparation or use. That was the thrust of my point. At any given time there is a set of compositions in the literature that are claimed under patent. Over time this set loses compositions due to expiration and gains new members due to new patents.
Asking for disclosure of claimed compositions from a patent applicant is a tall order. The last thing an applicant wants to do is to attract attention to their art. It’s bad enough that part of the deal in obtaining a patent is public disclosure with the requirement of enablement. But it seems to me that the public can reasonably expect that a plain reading of the patent should reveal unambiguously what is claimed. If it is indeed available by a plain reading, then it should be available for collation into a database. Perhaps that should be a benchmark for enablement??
Keep in mind that there are multiple processes to acheive the same end composition – hence the idea of a checkoff box in the Beilstein for a process patent would not uniquely qualify a compoound a patented/unpatented.
What scares me most in doing IP searches is making sure that I use all of the correct search terms, a far bigger problem than it may appear, since the PTO has given the inventor the right to be their own lexicographer. I’m guessing that’s why Dow calls many of their copolymers “interpolymers”, a term that I’ve not seen anywhere else. How many other examples are out their?
It’s a hard problem and I don’t even have a grasp on how to frame the question. What is clear to me is that doing your diligence on a composition or process is way, way too sketchy. Seems to me that the public interest isn’t being taken care of in this regard.
With regard to the second problem (it’s easier), since a patent is to be read and understood by those “skilled in the art”, the inventors should use the language of one “skilled in the art” or suffer the consequences.
As for the problem of multiple process patents to reach the same product, I’m not sure of a solution. Some the the processes are more or less equivalent and are “invented” so as to work around an existing patent or to “extend” the life of an expiring patent. In other cases, the new process is truely revolutionary and should be patented.
I think the difficulty in diligence on the process patents is that as a culture, our language and social orientation is directed to “product” and not process. IUPAC and the ACS have codes on how to identify chemicals, but do not have the same emphasis on codifying processes. Look at polymer nomenclature: the product is described on the basis of the monomer input(s). There is probably a decent sociology dissertation in this topic, although what are the odds of a sociologies wanting to study chemistry nomenclature?
One of the problems with abstracting claimed compositions of matter for data retrieval is the existance of Markush claims. A Markush claim represents a set of members whose number increases geometrically with the number of claimed attirbutes and their ranges. For the abstracting service to generate every permutation of “C3 to C20 hydrocarbyl” in connection with elements from “Group 14-16 of the Periodic table of Elements” is a non-trivial task. To do this would require the generation of large numbers of CASRN’s and names of species that probably have not been made. I can’t see CAS wanting to do that just because a single patent used broad language.
Even without Markush claims, as John Spevacek (Above) perhaps infers (??), polymer nomenclature is a concession to the inherent complexity of polymer composition. The deal with the devil has been to claim whatever arises form the input of certain monomers under certain conditions. Monodisperse homopolymers are easiest to name. Everything else is harder.