Big data turns failed experiments into potential gold mines

When an experiment fails, scientists head back to the drawing board. Often, data from unsuccessful research is shelved and forgotten, if not discarded altogether. But a couple of Japanese chemical companies — Sumitomo Chemical and Mitsubishi Chemical Holdings — have come to the realization that there is value hidden in those reams of information.

Starting in April, Sumitomo Chemical will begin preserving digital records of projects related to the development of electronic materials, regardless of how they turned out. Mitsubishi Chemical will introduce a similar policy for research on highly functional materials in the coming fiscal year.

The companies expect to benefit in at least two ways. First, the data will help researchers avoid wasting time repeating experiments that have already failed. Second, there is always a chance that some overlooked tidbit will lead to a hit product. It would not be the first time.

Old habits

The chemical companies’ teams of scientists conduct all sorts of experiments in pursuit of new materials, which wind up in everything from electronic gadgets to cars. When a test yields the expected results, they save the data for future reference — say, for when they need to work out mass production methods.

When a test is deemed a failure, however, the researchers typically do not bother recording the data, or only jot down some notes.

This is a habit both companies want to break.

At Sumitomo Chemical, the first people to change will be those in the research division for electronic materials. They will trade in their lab notebooks for tablets and enter data from all experiments.

Because everything will be digital, other scientists in the company will be able to search the database and review past cases before embarking on similar experiments.

Mitsubishi Chemical also intends to make its test results shareable. It is now working with an IT company on a prototype data entry system. The plan is to implement the system within fiscal 2017 in some of its functional materials labs. Once any kinks have been ironed out, Mitsubishi Chemical will introduce the system to more labs.

These systems could turn out to be much more than mere time savers. The history of chemistry is filled with chance events and failed experiments that led scientists to major breakthroughs.

Consider the discovery of penicillin. Scottish scientist Alexander Fleming stumbled upon the antibiotic after noticing that his petri dishes had become contaminated with a fungus that killed his bacterial specimens.

U.S. conglomerate 3M did not purposely set out to develop its ubiquitous yellow sticky notes. The Post-it was the result of an unsuccessful effort to develop a powerful adhesive.

Sumitomo Chemical and Mitsubishi Chemical have their own stories to tell.

Sumitomo Chemical did not plan to develop a plastic film that resembles Japanese washi paper. But it ended up selling just such a film, dubbed Wapo, after researchers took notice of another product that was a flop.

Mitsubishi Chemical’s hollow fiber membranes, used for water filtration, are based on a material that was created by chance during the development of a different product.

In the past, scientists could hardly be asked to comb through volumes of old experiment data in search of a golden nugget. But with today’s big data technology, it is much easier to extract important — potentially revolutionary — bits from enormous sets of information.

Sumitomo Chemical and Mitsubishi Chemical both hope to accelerate product development through the effective use of information that would otherwise be buried.

Now is the time to sharpen up research practices. Chemical companies are shifting their focus to lucrative specialty products, decreasing their reliance on petrochemicals, which are sensitive to the price of crude oil and other market conditions.

Sumitomo Chemical spends 160 billion yen ($1.42 billion) a year on research and development, while Mitsubishi Chemical shells out 140 billion yen. Both will probably spend even more money as they step up their specialty projects.

The high cost of R&D is one factor driving a succession of joint ventures and mergers in the chemical sector. Adopting more efficient research methods cannot hurt — and it might just lead to something big.

Enjoy our free content ?
Try our Legend services.

  • Star Level
  • Access to Grendz and the right to READ and SHARE our science techie green pins
  • The right to WRITE and SHARE your OWN science techie green pins
  • WEEKLY mind-blowing e-report with trends and news
  • First to know about new trends and news
  • No ads and no sales pitch
  • A special pin a month, posted by us, highlighting your own business
  • Our iOS app for free
  • Weekly social media promotions of your own pins
  • Customization on which trends categories you want us to follow closer.
  • 24h Support (via e-mail)
  • Cancel anytime your renewal
FREE
The Plain Vanilla Level
$0
LEGEND
A Lot To Gain And Nothing To Lose
$21/mo
Share on Pinterest
More share buttons
Share with your friends










Submit