What is MOL-Instincts in short?
MOL-Instincts is a database containing thermo-physico-chemical properties, transport properties, and many other sets of molecular information of pure chemical compounds. More information is available here.
The data and information by MOL-Instincts are not experimental. How can I trust them?
Since reliability is the most critical issue, we spent more than 5 years to collect experimental data available from every possible source, and used them to verify our estimation. Examples of comparison with experimental data are available in the "Accuracy" page (login required). The deviations from experimental data are also provided within MOL-Instincts web search site if the corresponding experimental values are available within our collection. Before putting into the MOL-Instincts database, all the estimates were manually inspected with related experimental data. When the experimental data are unavailable, the trends of other experimental data for similar chemical compounds were used for the inspection.
There are other calculation methods available in the form of software, calculation modules within software, or algorithms in scientific papers. What's the difference, and why should I use MOL-Instincts instead of them?
Running software or calculation modules requires time to learn and execution. It is rare the case that one software or module produces all the necessary data and information we need, so that multiple purchase, learning, and execution are required sometimes. Even if the calculation result is obtained, users wonder if it is accurate enough or not. Verifying the accuracy is rarely straightforward and takes time, too. Sometimes, calculation error occurs without obvious reasons and it takes additional time to find out why. Frequently and more importantly, the calculation results by most of other methods are not accurate enough based on our experiences.
Since MOL-Instincts resolved most of those issues, users can quickly and easily obtain the data and information without spending too much time to search, learn, execute, and verify. In addition, MOL-Instincts also provides the calculation results by other popular methods and compared with available experimental data.
Does MOL-Instincts provide experimental values as well?
MOL-Instincts does not provide experimental values themselves, but it gives the deviation between MOL-Instincts estimated values and refined experimental values, if the experimental data are available within our collection. MOL-Instincts also provides calculated values by other methods. An example screen shot is available in the "Features" page.
What do you mean by 'refined' experimental data mentioned in your website?
We collected the experimental data from many different sources, and we observed that a large portion of the experimental values contains a lot of noises and errors frequently. We therefore spent a significant time to determine if the experimental property values are reliable or not by analyzing multiple experimental data points obtained from many different sources, and considering the trend of the data for similar chemical compounds combined with some theoretical aspects. In some cases, we averaged the multiple points if it satisfies both the experimental trend and the theoretical analysis. The final value is called 'refined' experimental data.
Did you have parameters in the MOL-Instincts estimation models, and determined the parameter values by regression with your collected experimental data? If so, you cannot simply extrapolate to other chemicals in general, unless the parameters have physical meaning and you verified that the determined values of parameters are physically meaningful. What happened during the development of the MOL-Instincts estimation models?
Profound question! It is true that we did have some parameters and we determined them using part of our collected experimental data. But we do know what those parameters mean and most of the determined parameter values were physically meaningful. Please be reminded that we started from fundamental understanding of chemical compounds based on quantum chemistry, and most of parameters were related to the information obtained from the quantum chemical calculation results.
In some small number of cases, however, there were some parameters that we cannot fully assign a physical meaning (other than a degree of importance with respect to certain variables from quantum chemistry and molecular descriptors), e.g., the parameters in the ANN (Artificial Neural Network) model. In this case, we evenly separate the experimental data into three groups, i.e., training group, validation group, and test group. We determined the parameter values using only the training group of experimental data. We used the validation group to verify the trend of the estimation results, and the test group was used for verifying the absolute accuracy. Then, we applied an over-fitting prevention algorithm to determine the final estimations. The extrapolation capability was therefore practically examined by the validation and test steps under the over-fitting prevention constraint.
After this, the final estimation values were manually inspected with related experimental data for further quality services. When the experimental data are unavailable, the trends of other experimental data for similar chemical compounds were used for the inspection.
So, people don't need do the experiment anymore thanks to your development. Is that what you're saying?
We do not claim MOL-Instincts can completely replace the experimentation. Whenever possible, experiment should be carried out as a good double check. We recommend using MOL-Instincts for speeding up your initial task when you have to wait long for the experimental results, or make your work more stable and reliable when you suffer from incorrect data or lack of data. MOL-Instincts could be the best choice when experimental data are unavailable. It is frequently the case that the experimentation is not feasible because the target compound cannot be pure enough due to the separation issues, or it is too toxic or too unstable to perform the experiments. Please refer to the "Limitation of MOL-Instincts" page for further information.
Who needs MOL-Instincts?
Briefly, MOL-Instincts can be used by most of the chemical industries and research institutes, especially for (1) chemical plants and process design, simulation, and optimization, (2) energy efficiency improvement and alternative energy development such as biofuels, (3) reaction engineering, (4) industrial equipment design and optimization, (5) new chemicals development including drugs, cosmetics, flavors, and fragrance. For more information, please have a look at the "Who Needs MOL-Instincts?" page.
Can I use the property data in MOL-Instincts in my process simulation software? If so, how in brief?
Yes, of course. If any chemical compound you need is not available in the built-in database of your process simulation software, using MOL-Instincts property data frequently provides far better results than using the built-in estimation modules of the process simulation software in terms of stability and reliability in our experiences. Please have a look at the "Application Benefits" page for more information. You can simply copy and paste the data as needed, or use the "*.ikc" file automatically generated by MOL-Instincts web search site. A brief explanation on this feature is available at the "Other Features" page.
Do you plan to process additional chemicals and update MOL-Instincts database?
Yes. Based on our current plan, MOL-Instincts database will be updated on a yearly basis with an increase of about 1 million compounds. The update frequency and the increasing number of compounds may change, however, depending on the processing status and the number of CPU cores, as it may increase in the near future.
How are the MOL-Instincts data and information serviced?
You can purchase the usage right for typical period of time to make use of DB.
During your purchasing period, you can use Molinstintcs DB on the unlimited access basis.
For more information or questions, please refer to "contact us" site to get contact point of related ChemEssen staffs.
View Accuracy(Login required)
HOME
