Sticky Accelerometers and Stearic Acid Derivatives


A very good friend was having some problems with a product. I volunteered to help with the problem analysis and resolution by doing some Quantum Chemistry (QC) calculations to determine some molecular properties, Infra Red (IR), and Raman spectra for a suspected chemical culprit. This page describes the work and the results.

Initial Analysis

The suspected culprit was a stearic acid derivative, stearonitrile (an aliphatic nitrile). Stearic acid is an 18 carbon saturated fatty acid found in many products. The basic structure of stearonitrile looks like:


where the gray spheres are Carbon atoms, the white are Hydrogen atoms, and the blue is a Nitrogen. Click on the appropriate file to download the 3D or Rendered files, viewable with Adobe Acrobat or equivalent.


Stearonitrile Calculations

The first thing we did was a quick and dirty geometry optimization and frequency analysis to calculate a rough estimate of some molecular properties and the IR/ Raman spectrum. The Density Functional Method B3LYP was used rather than just straight Hartree-Fock (HF) to improve the quality of the calculation via the inclusion of some electron correlation effects. The basis set used in the B3LYP calculation was a simple STO-3G basis. This calculation was performed just to get a rough comparison of the IR/ Raman spectrum as well as provide some insight into the dipolar nature of the stearonitrile. The results were as follow:


Multipole Molecular Properties:

Dipole moment (Debye):

X= -3.0396 Y= 1.5824 Z= -0.0001 Tot= 3.4269

Quadrupole moment (Debye-Angstrom):

XX= -182.6100 YY= -118.7310 ZZ= -115.4462
XY= 18.6217 XZ= -0.0009 YZ= -0.0004

 Octapole moment (Debye-Angstrom**2):

XXX= -957.2128 >YYY= 0.7875 ZZZ= -0.0003 XYY= -2.1952
XXY= 209.7490 XXZ= 0.0031 XZZ= 10.6218 YZZ= 0.3967
YYZ= -0.0005 XYZ= -0.0049

Hexadecapole moment (Debye-Angstrom**3):

XXXX=-49272.9740 YYYY= -410.214 ZZZZ= -212.2251 XXXY= 2534.1010
XXXZ= 0.0382 YYYX= 8.1837 YYYZ= 0.0002 ZZZX= -0.0083
ZZZY= -0.0014 XXYY= -5982.1565 XXZZ= -5692.2119 YYZZ= -91.8119
XXYZ= -0.0459 YYXZ= -0.0011 ZZXY= -2.6135

These Multipole results indicate Stearonitrile is a highly polar molecule with an oddly shaped charge distribution. This can be seen more clearly in the following 3D rendering of some molecular electrostatic potential isosurfaces (electrostatic potential values of 0.1 (red), 0.0 (gray), -0.1 (blue)):


As before, if you click on the appropriate file to download the 3D or Rendered files, viewable with Adobe Acrobat or equivalent.

Given the obvious nonpolar left side and polar right side of the molecule, one would expect to get surfactant behavior from agglomerations of this species, i.e. preferential association of the nonpolar portions of multiple molecules in a disk or spherical arrangement with the nonpolar ends (with all the Cs and Hs) in the middle and the polar ends (i.e. the C-N) pointing out. This is consistent with the experimental observations..

IR/ Raman Spectrum:

We can see that the B3LYP STO3G predicted spectra differs significantly from the experimental results.While the multipole moments were accurate enough for further analysis, the IR/ Raman spectrum wasn’t.

Derivative Analysis

Normally, the basis set used for IR/ Raman spectra calculations is at least a 6-31G(d) (much larger than the STO-3G basis used above, i.e. 336 basis functions for 6-31G(d) vs 130 for the STO-3G) with a HF or higher level calculation. This level of calculation was going to lead to some problems since we needed the results rather quickly and re-doing the HF, B3LYP (or higher level) calculations on the full stearonitrile with the 6-31G(d) basis was not a quick task. It was also clear that we weren’t particularly interested in reproducing all the various C-H and C-C bond vibrations that form a large manifold of vibrations in the IR/Raman spectrum for the full stearonitrile species. We were primarily interested in the C-N and associated vibrations with a representative sampling of C-H and C-C vibrations near the C-N bond. So, we selected a derivative for stearonitrile – in this case we chose pentanonitrile. Pentanonitrile looks like:


Which we see is just a manageably truncated version of stearonitrile. As before, we did a B3LYP DFT geometry optimization followed by a calculation of the IR/ Raman spectrum. Here’s the electrostatic potential:


and the IR/ Raman results (in red) with the experimental results (in black) from the contaminant:

A resonably good match between the predicted and measured spectra.



Pretty clear confirmation of the aliphatic nitrile nature of the chemical culprit.