August 4, 2021
Highlight: Order parameter paper makes it into Anniversary collection
Our paper introducing new coordination descriptors and crystal structure fingerprints has been selected for the highlights of the RSC Advances 10th Anniversary collections: Machine learning and artificial neural networks in chemistry. The paper can also be viewed as my most successful article because it attracted most citations per time and author (click onto image on the right). I think that this underscores the importance to work on useful and physically meaningful descriptors when it comes to machine learning etc.
January 11, 2021
New paper: CrystalNN improves coordination number calculations
We have published our paper that introduces a new method (CrystalNN) to calculate coordination numbers (CNs) with Inorganic Chemistry. The new method comes along with a benchmark suite (MaterialsCoord) and MaterialsCoord's application to test different CN methods on many different inorganic materials (elemental, binaries, terneries). Our new method performs similarly well as other estabilished methods, including frequently used approaches such as distance and Voronoi-based methods as well as Jmol, ECoN, and Brunner's method. We hope that CrystalNN and MaterialsCoord will facilitate automatic structural descriptions of inorganic systems and help performing machine learning-based research. Happy reading and happy testing!
January 7, 2021
Back on the map
2.3 years of sweet parental leave have found an end. I enjoyed every day of it, and I can recommend everyone to do the same. To take as much time for your kid(s) as possible. It pays off. For them. For you. For the grand parents and aunties, uncles, cousines, etc. You get to meet so many new people, and, believe me, you acquire new, very helpful skills. Just do it!
But now, I'm also happy to return to the active academic workforce—with new energy and inspiration. I accepted an offer as an academic employee, working remotely for the University of Stuttgart. Specifically, I've teamed up with Niels Hansen to study the thermophoretic behavior of active pharmaceutical ingredients. Stay tuned.
February 7, 2020
New paper: Innovative way to measure crystal structure similarity
We have published our openly accessible paper that introduces new valuable tools and concepts to assess and measure coordination environments and similarity between crystal structures with RSC Advances. To this end, we developed new coordination descriptors (e.g., for T-shapes, square pyramids, hexagonal planar environments) and made them maximally comparable using a Monte Carlo optimization approach. Based on the coordination features, we defined site fingerprints and ultimately structure fingerprints that permitted the comparison of different crystal structures in a quantitative and from a materials science perspective meaningful way. The tools are freely available through the python packages pymatgen and matminer. Happy reading!
March 29, 2019
New paper: High-throughput screening of zeolite database
We have published our high-throughput screening study of finding most likely synthesizable and industrially deployable hypothetical zeolite structures in the database generated by Michael Deem and co-workers in Zeitschrift für Kristallographie. We managed to come up with a short list of 70 candidates from an initial pool of more than 300,000 structures by constructing an efficient filter funnel. The resulting 70 frameworks seem to be particularly interesting synthesis and catalytic characterization targets for experimentalists. Thanks so much to Michael Fischer, Co-Editor of Special Issue "Modelling crystalline microporous materials", for the invitation and handling of our manuscript. Happy reading!
October 31, 2018
Book chapter: The accelerating effects of the Materials Project
Our book chapter presenting the work and accomplishments of the Materials Project is available here. I helped writing section 4 ("User Applications of the Materials Project to Research and Design Problems"), which was big fun. Happy reading!
June 4, 2018
New paper: Uncertainties in predicting nucleation rates
Our paper investigating different sources of uncertainty when predicting solute precipitate nucleation rates via a seeded simulation approach has been published and is openly accessible to everyone here. The paper revisits the case of NaCl nucleation from aqueous solution. We found that precise knowledge of the thermodynamic driving force is important, as is the way how we define the nucleus and its size. We, most certainly, wish you happy reading!
June 1, 2018
Paper: PyCDT—a Python toolkit for modeling point defects in semiconductors and insulators
I forgot to highlight that our paper introducing PyCDT, a Python-based toolkit for modeling point defects in semiconductors and insulators, got published online (gold-route open access) earlier this year (go here for the paper). The tools are openly available on Bitbucket. Happy reading/modeling.
May 29, 2018
New paper: Matminer—an open source toolkit for materials data mining
Our paper that introduces Matminer, a Python-based open source toolkit for materials data mining, got published (go here for the paper). Unfortunately, we could not make it open access. Bummer! But the tools themselves are openly available on GitHub. Happy reading/data-mining.
April 26, 2018
Visualization of Nisses's academic trajectory
Recently, I had to go over my academic trajectory in order to update and reassure my CV entries. That was a perfect opportunity to start a little fun project, in which I visually present my academic achievements and endeavors. You'll find the resulting flick in the sidebar beneath the About section. And, right below this text.
December 3, 2017
New preprint: electrostatic estimation of DFT-NEB diffusion barriers
We have resubmitted our paper about electrostatically estimating ion migration barriers in intercalant host materials using finite size ion models (hence, the name of our method: potential of electrostatics-finite ion size (PfEFIS)). A preprint is already available on ChemRxiv.
November 17, 2017
New paper: order parameters put to (large-scale) use!
We have published a paper about the usefulness of local structure order parameters for computational materials science in Frontiers in Materials. For one, we use the motif recognition capabilities to screen the Materials Project database for element-specific motif occurrences (e.g., relatively speaking, how often is a lithium atom in the database tetrahedrally coordinated vs octahedrally or bcc-like coordinated). Furthermore, we show that the order parameters are reliably implementable in an Interstial Finding Tool (InFiT) and for use as characterizing tools of intercalation diffusion paths. Check out the paper here. It is, for course, open access and our tools publicly available through the pymatgen library.
November 15, 2017
PhD thesis (finally) online
I've finally managed to put my PhD thesis on the document server of the Hamburg University of Technology. Go here if you are interested to read about fluid transport in the nanopores of zeolites with focus on how important the interface between solid and gas is.
May 6, 2016
May 3, 2016
Zeolite discoveries—from a different angle
March 28, 2016
Blog post on the role of databases in Inorganic Chemistry
In the aftermath of a recent conference attendance, I created a blog post that deals with the increasingly important role of databases in Inorganic Chemistry. You find it here.
October 19, 2015
Sometimes cleaning up can be fun. For example today, when I browsed a disk that contains old results from my time at the CECAM. The fun thing was that I found a movie that I've entirely forgotten about: a methane molecule making a diffusional hop in a clathrate hydrate structure. Go here if you want to watch it.
October 14, 2015
Double pack—two new articles
We have published two new articles.
The first paper sheds light on how sodium chloride forms from aqueous solutions. We have used a seeded simulation strategy, and a particular focus was on elucidating the attachment mechanism to the critical nucleus—in a quantitative and validated way. We identified ion desolvation as the limiting resistance in this context because ion diffusion toward the nucleus leads to 100-fold too fast attachment. Everybody can read the article here because we went for open access via the ACS AuthorChoice (thanks again to David Quigley and his ACS Author Rewards at this point—paying nothing to get open access is kinda cool). We have added a video to the article that reflects our seeded simulation approach and monitoring of the nucleus size.
The second article addresses the scientifically and technologically important question how to describe transport of gases and solutes into nanoporous materials. There had been speculations that Fick's laws fail under certain conditions. We show, however, that you are generally fine when using those laws as long as you have the following two pitfalls on your radar. First, the diffusion coefficient can strongly depend on concentration (e.g., cyclohexane in nanoporous glass). And second, strong adsorbate–adsorbent interactions should be treated in a thermodynamically sensible way; that is, lowering the true diffusivity by a factor that corresponds to the adsorption constant (e.g., pyronin B in MOF-5). Experimental evidence was provided by infrared microscopy (first example) and confocal laser scanning microscopy (second example, for which we took data from a study by Han et al). In summary, Fick's laws remain valid! But I guess based on the last article from my PhD I should add: as long as the host size does not approach the nanometer scale. Happy reading here.
November 23, 2014
No fear because now there is PubPeer
As a regular Seven Days in Nature reader, I learned this week about PubPeer. This is a website for discussing scientific articles in an open and easy way. So, you go there, put in the DOI of an article and you can directly comment on anything. The compelling point is that you can do it either with your clear name or anonymously. The former way is, of course, the way we, in science, should always do things: open and transparent. But if we are honest there are a lot of good reasons to have the possibility to comment without letting the other party know who we are. Young researchers, for example, will appreciate this channel of expressing their doubts about any article they come across. This is because they are the experts in their respective fields but they are also often still on the hunt for a permanent position. Clearly, they do not want to step on any (senior scientist's big) toes during this time of their scientific trajectory. Therefore, I think that it is a great post-publishing quality-assurance mechanism to which everybody can contribute.
Small final note: they are already operating since more than a year. So, it is not brand new—but still worth sharing.
November 4, 2014
Caring to sharing
Last week's Nature Editorial on code sharing outlined a new policy of the publisher that is a move in the right direction. It, more or less, requires authors of computational works now to make a statement in the paper about the availability and accessibility of the code used. To this end, Nature introduces a dedicated section on code accessibility for such kind of articles. What it all boils down to is the Editor's decision whether or not a work critically relies on computer code because she/he can ultimately decline the paper if the underlying code is unavailable.
From my own experience, I know that the decision not to share code can simply stem from feelings such as insecurity and immaturity—emotions that are probably very understandable and common among young researchers. In such cases, the new policy can be regarded as taking over a parental position because it helps you remove the last doubts in your decision-making process so to eventually follow the best scientific practice. Hence, I personally encourage the Editors to make as often as possible use of this new “instrument” because it seems to be a good guide.
February 18, 2014
Sinks and sources
Today's zeit.de article on the third-party funding situation in Germany revealed that STEM fields represent the main sinks when it comes to the fraction of external money to the overall financial means available (16–19%). Third-party money comes, for example, from the European Union, industry/private sector, or from the German Research Foundation. The latter is yet the dominant source (34%) to external funding, according to the article.
February 7, 2014
Guest group seminar
February 3, 2014
Blog entry—hopefully not a rare event
I have added a Blog section to reserve this space for short news. Go here for the blog and its first entry about real-world rare event examples.
November 26, 2013
You use it? You help it!
Today I've started donating $3 per month to the Wikimedia Foundation, which runs, for example, Wikipedia.
The donation amount is, of course, not hilariously much but it is a little something to support one of the most important knowledge databases that comes entirely for free.
So, I encourage everybody who can afford to contribute—be it just once or on a regular basis—to do so because virtually everybody makes use of Wikipedia and associated sites.
For example, as a computational scientist there's hardly any working day which I spend without visiting Wikipedia for getting some first clue about some substance that I'd be studying by my simulations or helping me out with some Latex issue during manuscript writing.
I bet you, too, have some good reason to visit Wikimedia's abundant material quite regularly if you read my post.
And, therefore, you have also a pretty good reason to help them staying alive.
If you agree, then don't think anymore.
Take action, contribute (for example, click here), and make a difference.
Have a good one.
November 13, 2013
Added new section: Bits & Bytes.
July 4, 2013
Paper of the year—so far.
My personal paper-of-the-year is Wilfred van Gunsteren's essay about "The Seven Sins in Academic Behavior in the Natural Sciences", which presents different types of misbehavior in order of increasing severity.
The paper really struck a chord in me, now that I have submitted my PhD thesis and, thus, (almost) accomplished a major step in my scientific trajectory (I don't like the word career, as you might have noticed; plus, I still have to tackle the defense so that the step is not fully finalized yet).
When I started thinking about why the paper was so compelling two reasons came instantaneously into my mind.
First, I, at some point in my PhD, found myself in a situation that could have theoretically led to committing van-Gunsteren Sin No. 5: ignoring an error in my own paper which I had realized some time after publication. But I'm not grilling in hell right now, because I started to rerun my simulations with a refined sampling procedure in order to get flawless data. I reassessed whether the main conclusions were affected (they were not; sometimes you simply have to be lucky) and wrote a corresponding "Addition and Correction". I cannot remember exactly how much time it took me to address the whole issue. It was easily 10 working days of only dealing with that very problem but I fear the true block-out time for doing new/current things was even larger. While such times are already hard enough by the doubts arising from having done an error, that pain is soon accompanied by the pressure that you need to proceed. To keep with the biblical analogues: my PhD demons went: "You need to write the next paper. You need to write the next paper." Still, I knew that I had to follow this path because otherwise my conscience would be on fire now and forever.
The second reason why I liked the essay so much is totally pathetic: I enjoyed the German translation a lot, which was written by my dear former colleague Niels Hansen. :-)
Happy Independence Day!
June 5, 2013
Position(s) on solid ground.
Congratulations to my boss, Prof. Baron Peters, for having obtained his tenure position. (That makes my employment also much more stable.)
May 5, 2013
New video online.
The video accompanying my latest publication is now online. You find it [here].
April 22, 2013
I have started my new job at the University of California in Santa Barbara on April 15. I will be working on nucleation problems, which will broaden my research background from a system point-of-view. As concerns the approach, I stay with simulations, thus not going over to measurements.
March 14, 2013
The past months have been quite busy. I finished work on an article which has just been accepted in The Journal of Physical Chemistry C (go [here]) and I have started writing my PhD thesis which will be submitted shortly. So, there is still activity "behind this website" and more activity on the site itself will be observable soon again.