Chemistry at Hamilton

Chemistry

Student Research

HOME
FACULTY

COURSES

RECENT NEWS

STUDENT RESEARCH

SENIOR PROJECTS

SUPERCOMPUTER CONSORTIUM

ALUMNI

CHEMISTRY LINKS

HAMILTON COLLEGE

SHIELDS' LAB GROUP

Due to sheer computer and student power, Shields' lab was able to crank out many important studies in the field of Computational Physical Chemistry, in areas from the calculation of pKa, to the study of the biological polypeptide known to promote resistance to cancer in human mammary cells called AFP, to basic hydrolysis of esters, as well as bio-inorgano-metallic chemistry. Beth, Matt, and Jenn (all pictured below) were the only three undergraduates represented in the International Quamtum Chemistry community this year at the Sanibel Symposeum, where these three students presented posters to the most elite in the field of Physical Chemistry on their summer research with Professor Shields. Dr. Shields claims that any student who completes a summer of research in his lab is free accompany him to poster sessions similar to this, and this has been the second year that Hamilton students have represented the college at the Sanibel Symposium. Click here to see the Sanibel crew. With Steven Feldgus, the resident Dreyfus Postdoc, this summer is sure to offer up some great research and a lot of fun.
Sarah Taylor
Lorena Hernandez and Sam Bono
Gabrielle Markeson
Emma Pokon
Jenn Derby
Matt Liptak
Beth Hayes and Jaime Skiba

Sarah Taylor
Vibrational Analysis of Methane Monooxygenase pK_a Calculations
Pictured Below: Shields Lab Group for the Summer of 2000. Top: Emma Pokon and Abbey Markeson. Middle: Lorena Hernandez, Jenn Derby, Sarah Taylor. Bottom: Matt Liptak, Beth Hayes, Sam Bono, Jaime Skiba. Standing: Professor George Shields.

I spent this summer calculating pK_a values from thermodynamic values with Prof. Shields. I continued the work of Annie Toth on six simple carboxylic acids from the summer of 1999. Relative pK_a calculations, where the pK_a of one acid is determined in relation to another acid, were done first because they allowed us to focus on pK_a values instead of worrying about the correct treatment for the H+ species and water in our thermodynamic cycle. By focusing on relative calculations for six simple carboxylic acids we were able to locate the errors in our calculations as well as get final pK_a values to within an error of .5 pK_a units of experimental values. This knowledge was applied to absolute calculations, using only one acid, yielding pK_a values with the same error and an insight into the correct treatment of water and the H+ species in our thermodynamic cycle. Calculations are currently being performed on chloro-substituted propanoic acids, amines and phenols.

Lorena Hernandez
Sam Bono

AFP PROJECT

During summer 2000, Sam and I have been working with an 8mer peptide present in the third domain of AFP. This 8mer of sequence: EMTPVNPG has been documented as retaining most of the anti-cancer activity of AFP. However, if the 8mer peptide is reduced to a 7mer peptide by cleaving one of its terminal amino acids, it loses all of its biological activity. Based on this information, we have taken a computational approach to assess the conformational differences between the 8mer and the 7mer peptides as to arrive to a conclusion explaining the activity of the 8mer peptides and subsequent inactivity of the 7mers.

Gabrielle Markeson
Pre-Freshman Research Experience

The goal of my research is to check the accuracy of the calculations we perform on computers. I used a list of experimental values from a Cramer paper as a reference to compare the values I would calculate to see how accurate my calculations would be. I compiled a list of all the cations, anions, and neutral molecules, broken down by functional group, and I built the molecules on Spartan. Then I ran the molecules through the HF/6-31G(d), HF6-31+G(d), and HF/aug-cc-pVDZ geometry optimizations. The next step is to run the following solvation calculations on all the molecules in all three geometries: CPCM and PCM calculations in Gaussian, SM5.42R in Gamesol, and SM5.42P in Amsol. After I have all these values I will create a spreadsheet to compare the calculated values to the accepted experimental values. As a result of this research, we will find the most accurate method to perform solvation calculations.

Emma Pokon
Pre-Freshman Research Experience

I started the summer working on the Spartan tutorial program and getting caught up on some background chemistry. After my first week, I started on a project to find the most accurate way to calculate the change in free energy of a gas phase reaction. Working with the NIST site, I found molecules with published DGgas experimental values that had and error value of less than one for the de-protonation reactions. Then I built these molecules and their ions on Spartan. After running ab initio and geometry optimization calculations in Spartan we set up jobs in the Gaussian program to calculate the free energies of the molecules and their ions. Each molecule was run though CBS-QB3, CBS-APNO, and G3 calculations. While I was waiting for these jobs to finish, I continued my search on the NIST site and looked up the references for the DG values I was using for the purpose of possibly finding more good values. I ended up with three groups. The first group consisted of ammonia, methylamine, methanamine (N-methyl), ethylamine, and the ions of these molecules. The second was methane, methyl alcohol, nitrosyl hydride, water, acelyene, ethylene, formaldehyde, hydrogen chloride, propene, isocyanic acid, acetaldehyde, nitrous acid and the ions of these molecules. The third group is benzene, methane (chlorodifluoro-), hydrogen selenide, hydrogen bromide, nitric acid, furan, and their ions. These values were then entered into a spreadsheet along with the experimental values for comparison. The different calculations were compared to each other and against the published values. I still have a job running on Gothics with the CBS-APNO and G3 calculations for the third group of molecules. Once this is done, I will be able to draw a conclusion as to the best way to calculate DG in the gas phase.

I ran into some problems in comparing calculated and experimental values. Although the published values I found have a low error there are also other published values that do not claim to be very accurate and in some cases my calculated values seem to be closer to the published values that have the large error. There were a few instances where the various published values were very different; so I started a second spreadsheet that would help to account for the other published values. A second problem that I had was getting some of the jobs to run. Especially in the last couple of weeks, the computers all seemed to be running low on disk space and memory and didn't have enough room for my jobs. It was therefore necessary to move the jobs to a different computer that had more space. I started on Street but I ended up having to work on Colden, Algonquin, and Gothics before the last few jobs could run.

This summer I definitely learned a lot about working with the computers. I can work well with Spartan and I'm getting to know Guassian. I can write a job in Guassian but I ran into several problems that I had no idea how to deal with. With a little more experience, I could get to know my way around the computers much better. This research project has been an excellent experience and a great introduction to college chemistry.

Jenn Derby

Analysis of Water Dissociation

The dissociation of H2O in water is one of the most fundamental reactions in chemistry. It is important to many chemical systems, including the hydrolysis of cocaine. The structure of H2O in solution and the structures of the ion products are not known for sure, and extensive research has lead to conflicting conclusions. Evaluating different systems of dissociation, analyzing both solvation and gas phase energy calculations and comparing these results to known experimental values can lead to correct structures. The overall goal is to determine the structures of H2O and its ions in solution, which can then be used to model various systems, such as the ester hydrolysis of cocaine. Results will be reported in this poster.

Matthew Liptak
Toward More Accurate pK_a Calculations

I spent this summer calculating pK_a values from thermodynamic values with Prof. Shields. I continued the work of Annie Toth on six simple carboxylic acids from the summer of 1999. Relative pK_a calculations, where the pK_a of one acid is determined in relation to another acid, were done first because they allowed us to focus on pK_a values instead of worrying about the correct treatment for the H+ species and water in our thermodynamic cycle. By focusing on relative calculations for six simple carboxylic acids we were able to locate the errors in our calculations as well as get final pK_a values to within an error of .5 pK_a units of experimental values. This knowledge was applied to absolute calculations, using only one acid, yielding pK_a values with the same error and an insight into the correct treatment of water and the H+ species in our thermodynamic cycle. Calculations are currently being performed on chloro-substituted propanoic acids, amines and phenols.

Beth Hayes and Jaime Skiba

Modeling Ester Hydrolysis

In our search for the transition-state of cocaine, we encountered problems with the accuracy of our calculations. The cocaine transition-state is a large molecule, made up of 42 atoms. This makes for long, expensive calculations. Measuring the energies of smaller molecules, which are modeled to resemble the larger system, facilitates the process of finding the best method to apply to the hydrolysis of cocaine.

Methyl acetate, methyl benzoate and benzyl acetate are three esters that are similar in structure to the cocaine molecule, each containing a phenyl group. Following the thermodynamic cycle for the first step of ester hydrolysis, the Gibb's free energy changes and the activation energies are found. Once these models have been produced, we will determine the level of theory necessary for modeling a large system like the cocaine molecule by comparing our data with experimental data.

Methods: Spartan and Gaussian 98 are the key computer programs that were used to manipulate the molecules. The initial structures of the esters were found using the molecular graphics software Spartan. They were then transported to Gaussian, where they were subjected to geometry optimizations. We completed HF/3-21G optimizations on the structures from Spartan. Then, HF/6-31+G(d) calculations were performed on the HF/3-21G structures. Once the esters had been optimized, CBSQB3 jobs were completed on the ester reactant, transition-state, and the intermediate. The CBSQB3 job begins with a geometry optimization that is very similar to the HF/6-31+G(d) opt. It then proceeds through energy calculations with very high levels of the electron correlation. CBS jobs were also completed on the reactants and the transition-states at varying temperatures, to determine how the energies would vary.

All of the above calculations were done in the gas phase. We want to determine the rate of the reaction in solution. This is accomplished with the help of the thermodynamic cycle and CPCM jobs. CPCM jobs place the individual molecule into the solvent of your choice. You may vary the solvent by varying the dielectric constants of the solvents at different temperatures. The thermodynamic cycle provides equations to determine the energies of the hydrolysis of the esters.

Comparing the results for the methyl benzoate reaction rates and the experimental data that was found may be somewhat deceiving because the computational reaction was in calculated in water and the experimental data outlined ester hydrolysis in four different solvents. Acetone, acetonitrile, methanol-water, and dioxan each affect the rate constant for the reaction.