The initial purpose of the Identify database system is the automatic recognition and comparison of measurement curves. This serves, for example, for quality control and failure analysis. Pure data mining (storing, searching and finding of data) is, of course, the second main application.
How to do data mining with Identify?
One approach of using Identify is to search for similar database measurements or literature data on the base of an input measurement curve. This is illustrated in figure 1, where a Thermomechanical Analysis (TMA) measurement is compared with various database curves. The similar TMA curve was found automatically by Identify, the DSC and TGA curves were in this case found easily by alphabetical sorting of all database measurements shown. It is interesting to see that the glass transition occurs between -70…-60°C (see also the Differential Scanning Calorimetry, DSC, curve) and that the decomposition – which should normally be avoided in DSC or TMA measurements – starts slowly at above 150°C visible in the Thermogravimetry (TGA) curve.
Fig. 1: Identify database search based on a TMA curve of an NR (natural rubber) sample. Overlaid are a similar TMA curve, a DSC and a TGA curve of NR found in the database.
The second approach of data mining – where no similar input measurement is required – is related to the “Manage Libraries/Classes” function (see figure 2). In this case, “NR” was typed in manually in the search field for measurements and literature data, and consequently, the three different NR measurements were found. Restoring of any database measurement, which is simply done via right mouse click, opens the measurement where also the measurement conditions such as temperature program, sample mass, purge gases and sample crucible can be seen in detail.
Fig. 2: Searching within the Identify database using the name filter (indicated by the red circle).
Last but not least, Identify offers also the possibility to filter according to the heating rate, the sample mass and if glass transitions or endothermic and exothermic effects were evaluated. If a string of letters occurs in a measurement name or not, may also be a filter criterion (see figure 3).
Fig. 3: Filter settings within Identify.
The database content of Identify
Shown in figure 4 are the NETZSCH libraries currently containing 1294 entries, which cover various application fields (ceramics, inorganics, metals, alloys, organics, pharma, food, cosmetics and polymers). Available as an option is the KIMW database developed by the Kunststoffinstitut Lüdenscheid, Germany, with DSC curves for 1000 different commercially available polymer grades, where information about the polymer supplier, color and filler material/content is also available.
Fig. 4: Identify database contents: NETZSCH libraries which are always included (1294 entries), the optional KIMW library (1000 entries), and an exemplary library created by a user.
In summary, measurements, associated measurement conditions and evaluations can be easily found and retrieved for all measurements stored in Identify. This knowledge treasured can be helpful prior to and after any measurement!
Please feel free to visit earlier, related blog articles: