Author Archives: Kristine Vorwerk

Interesting talks and workshops at the ACS National Meeting

In addition to presenting my work at the poster session as Dr. Kennerly already mentioned, I attended a wide variety of talks and events at the ACS National Meeting. Some of the talks were related to my field of research, such as one by Dr. Joseph Subotnik. That said, one of my favorite things to do at ACS is always learning about new areas of chemistry, so I went to quite a few talks out of sheer curiosity. One I particularly enjoyed discussed modifying the surface coatings of solar cells to allow dust to be washed off more easily, making them more useful in desert areas. I had never realized how significant of a problem dust was for solar technology, so it was fascinating listening to someone who had devoted significant amounts of time to fixing this problem.

Since I am graduating this year and searching for a job, I also participated in some of career building workshops offered at the conference. In particular, I attended one which was designed to allow undergraduates to network with professionals from all areas of chemistry. There were quite a few people representing academia at this event, and I think it would have been a bit more helpful if there were more from industry and perhaps also the government. That said, I met with a representative from ACS who directed me to some helpful resources for recent graduates, so it was a worthwhile event.

Overall the conference was fun and instructive and definitely worth the trip.

1 Comment

Filed under conference report

Running the NewtonX Tutorial

After I got the tests to run, I started working my way through the tutorial.  Specifically, I am trying part 4.  I managed to get to step 18 before I ran into problems.  In step 18, we are trying to generate the inital conditions for NewtonX to use.  To do this you run a line of code:

$NX/ >initcond.log &

This line of code immediately spits out a message which is a number in brackets followed by another number.  The numbers change each time.  It then prints a message saying is dying now.  At this point, the console freezes and does not respond until you abort with ctrl-c.  It then spits out an exit code, “Exit 255.” (Which I have been unable to find more information on).

To try to figure out what was wrong, I examined the initcond.log file.  So far I have gotten two different messages which seem to alternate with no real reason.  The first occurs early in the run and says:

Cheking input files

Cheking geometry lines

Searching for qvector...

qvector does not exist

Creating qvector...

It appears to die while trying to make the missing qvector.  The qvector it is referring to used to be a file in version 1.1 that contained the quantum vibrational number, but it isn’t referenced at all in 1.4.

The second error happens later in the run and says:

Starting at Tue Nov 22 08:21:30 PST 2016

Vertical energies with Gaussian 09

Starting .  $g09root/g09/bsd/g09.profile;$g09root/g09/g09 at Tue Nov 22 08:21:30 PST 2016

Finished .  $g09root/g09/bsd/g09.profile;$g09root/g09/g09 with ERROR at Tue Nov 22 08:21:30 PST 2016

Finished with ERROR at Tue Nov 22 08:21:30 PST 2016

This error seems to be referring to the same issue with NewtonX inserting an extra /g09/ which we were seeing earlier.  However, when I tried using the same set of variable definitions which worked in the tests, it did not change anything.  Also, I poked around in the file and found the same qvector error as earlier but instead of crashing after creating the qvector, it continued successfully.  So it seems like that error is one that it can sometimes get past.

So far I haven’t been able to figure out when one error occurs versus the other.  There doesn’t seem to be any particular logic to it.




Filed under computing

Running the NewtonX Tests

First you need to install NewtonX.  To do this, you request a download link from the NewtonX site, follow the link, and download the binary file recommended (it will be called something like “NX-1.4.0-3-source+binary.tgz”).  Then copy the file to comet, and untar and unzip it by typing tar -zxf NX-1.4.0-3-source+binary.tgz.  This corresponds to section 9.1 in the documentation.  Do not go further and do 9.2; 9.2 is only for if you want to install the program manually.

After installing NewtonX, I spent a while trying to get the built in tests to run.  Eventually I came up with this procedure:

First we have to set up the variables so NewtonX can find Gaussian.  We do this through the following lines of code:

export NX=/home/kvorwerk/NX-1.4.0-2/bin
export g09root=/share/apps/compute/gaussian/09.E.01/g09
source $g09root/bsd/g09.profile
export g09root=/share/apps/compute/gaussian/09.E.01

The last line is necessary to avoid an error by which NewtonX adds an extra /g09/ into the path where it searches for gaussian.  Also note the line setting the NX variable would need to be modified to match the path where you installed NewtonX.

Next we make a directory to run the tests in.  This can be anywhere.  Then we select the tests to run by typing:


This shows a list of tests.  We want to run the gaussian09 one and the NewtonX ones; that is, 18, 19, 24, and 25.  To actually run the tests we type:

$NX/ >test.log

This creates a bunch of subdirectories and a file called test.log.  The test.log file contains basic output, and more output is in the sub directories.  When I ran the tests, the test.log file told me that:

18 finished, but with a message saying:

Files /home/kvorwerk/nx_test/TEST_NX/MD-SICH4-G09-TDDFT-NAD/RESULTS/dyn.out and   /home/kvorwerk/NX-1.4.0-2/bin/../test-nx/STANDARD-RESULTS/MD-SICH4-G0    9-TDDFT-NAD/RESULTS/dyn.out differ.  It seems to be important differences. Something may be wrong in the installation.

The files it refers to are easy to find.  However, I have been unable to find the file or script which compared them.

19 finished normally

24 finished with a message saying:

Files /home/kvorwerk/nx_test/TEST_NX/MD-ANALYTICAL-MODEL1/RESULTS/dyn.out and   /home/kvorwerk/NX-1.4.0-2/bin/../test-nx/STANDARD-RESULTS/MD-ANALYTIC    AL-MODEL1/RESULTS/dyn.out differ.  It seems to be only minor numerical differences. Probably installation is OK.

I do not think this error is worth worrying about.

25 finished normally.

The issue with test 18 may be due to us using a more recent version of gaussian (09.E) than the developers used to create the test.  Hopefully we will be able to find the comparisson they use and learn if this is indeed the case.


1 Comment

Filed under computing

A Look at Various Methods of Computationally Computing Absorbance Spectra

Yesterday was my senior seminar presentation.  I chose a paper entitled Computing the Absorption and Emission Spectra of 5-Methylcytidine in Different Solvents, which examined the absorbance spectra — and therefore also the vertical excitation energies — of a molecule called 5-methylcytidine in different solvents.  The authors (Martinez-Fernandez et. al.) use a range of solvation models and a mix of QM/MM/MD methods to find their spectra, and it is a fairly interesting look at different ways to computationally solve the problem.


In the paper, the absorption and emission spectra of 5-methylcytidine in water, acetonitrile, and tetrahydrofuran are studied both experimentally and computationally.  This molecule is chosen because it is the nucleoside associated with 5-methylcytosine, a derivative of the C base linked to UV caused mutations in DNA1,2.  The nucleoside is used instead of 5-methylcytosine itself because the sugar ring affects the absorbance and thus must be included for the results to be accurate to those of living systems.

When studying 5-methylcytidine, the paper seeks to answer two main questions: how the molecule’s spectra can be most accurately modelled computationally, and what insights into the solvent’s effects on the spectra can be found.  To answer these questions, the results of three computational methods were compared to experimental spectra.  In the first, the molecule was modelled using a static quantum mechanical approach, with the solvent being treated using a PCM model.  In the second, a static mixed quantum mechanical/molecular modelling approach is used, where the solvent-molecule interaction is initially simulated and optimized to a low energy point using MM, and then the absorption/emission energies are found using QM.  In the third, a molecular dynamics simulation is run, from which a sample of the different conformations of molecule and solvent are chosen.  The absorption/emission energies are then found at each of these points using MM single point calculations.  In each case, the absorption spectra obtained were stick spectra.  Over each peak, a Gaussian was fit to simulate vibrational effects and provide complete spectra.  Note for the emission spectra, only the first two methods were used as the third was computationally infeasible.

The computationally obtained absorption spectra all successfully predicted the three main peaks shown in the experimental spectra at 278 nm, 242 nm, 200 nm, though in some cases the peaks were blue shifted due to the lack of vibrational effects in the calculations.  Of the three computational methods, the third was most successful at predicting the experimental results, as it closely matched the experimental and included some absorbance between the two higher wavelength bands which the other models missed.  This was thought to likely be due to the increased accuracy of this method at representing the solvent.  The spectra also mirrored the solvent shifts seen in the experimental data, with acetonitrile and tetrahydrofuran having for the most part similar effects on the absorbance.  Their spectra relative to that of water were red shifted for the band at 278 nm and of higher intensity but similar energy for the other bands.

The computational emission spectra were consistent with those found experimentally for water.  The spectra of the other two solvents were similar to each other both experimentally and computationally.  For both solvents, the shift in the spectra from that of water were overestimated using both computational methods.  It was concluded that the main errors in both absorption and emission spectra were due to the lack of accounting for vibrational effects.

The differences in the absorption spectra for the solvents were concluded to be due to the hydrogen bonding ability of the solvent.  The HOMO and LUMO of the molecule in acetonitrile and tetrahydrofuran were very similar in shape and energy, despite the difference in polarity3 between the solvents.  This accounted for their similar absorbance, and suggested that solvent polarity is not the main factor in the solvent effect.  In water, hydrogen bonding dramatically changed the shapes of the HOMO and LUMO orbitals, thus changing their energies and the wavelengths which could be absorbed.  Therefore, hydrogen bonding was concluded to be the main cause of the solvent effects observed.


The paper can be found at:

Matrinez-Fernandez, L.; Pepino, A.; Segarra-Marti, J.; Banyasz, A.; Garavelli, M, Importa, R. Computing the Absorption and Emission Spectra of 5-Methylcytidine in Different Solvents: A Test-Case for Different Solvation Models.  Journal of Chemical Theory and Computation.



1 Comment

Filed under Paper Review