Simulations using quantum mechanics and molecular force fields prove instrumental in developing ideal biosensors.

Accelrys Image

Modeling allowed rapid screening of many possible replacement molecules for a monolayer of phenanthroline on a gold surface of a sensor. Attaching carbon-60 to only a fraction of the phenanthroline formed a stable layer that retained the needed properties. Source: Accelrys Inc.

Atomistic modeling plays an integral part in the discovery, design, and optimization of new molecules and materials. The computational cost of the calculations and the associated steep learning curve associated limited the applications. Faster computer hardware, more efficient software implementations, and graphical interfaces made these methods more accessible to a broader number of chemists. Modeling has contributed to the development of new pharmaceuticals, structural polymers, catalysts, and coatings.

Today, atomistic modeling is used to study systems with dimensions up to a few 100 nm. Despite being limited to the nanometer scale, these methods make significant contributions to the solution of problems at the engineering scale.

Simulations using quantum mechanics and molecular force fields were instrumental in optimizing the performance of lab-on-a-chip devices for a U.K.-based company that designs and manufactures biosensors. These devices detect the presence of proteins that may be diagnostic for particular conditions.

One such sensor has a chip containing dye molecules that are bound up with a peptide. If the targeted protein is present, it binds to the peptide, freeing the dye, which subsequently binds to the surface of the chip. The dye can be detected by surface enhanced resonance Raman spectroscopy, which is sensitive to small amounts of material.

A dye absorbs more light at specific wavelengths. By tuning the laser used in Raman as close as possible to the spectral maximum, the researchers could increase the sensitivity of the device, which is needed to detect the smallest possible concentrations of proteins.

CASTEP is a software package from San Diego-based Accelrys that predicts the properties of solids, interfaces, and surfaces for a range of properties using the principles of quantum mechanics as implemented by density functional theory (DFT). This method produces reliable results for structural, electronic, and optical properties and was used for predicting the optical spectrum of the dye used in the biosensor.

The researchers had two choices for the laser: an argon ion laser with a wavelength of 458 nm or an Nd:YAG laser at 532 nm. CASTEP predicted the absorbance at 458 nm would be twice that at 532 nm, effectively doubling the sensitivity of the sensor. Since time was of the essence, the engineers made the decision solely on the basis of the calculation. The results were subsequently validated by experiment, confirming the decision.

But it worked fine in the lab!
Modeling also proved instrumental in solving a design problem with a different sensor. In this instance, a molecular monolayer on a gold surface acts as a conductor. The presence of a particular antibody affects the conductivity of the surface, so the antibody can be detected by monitoring conductivity. Initial tests worked fine in the lab, but failed under actual operating conditions.

Molecular dynamics using a force field simulation program called Forcite diagnosed the problem. Force field methods approximate the interactions between atoms using simple analytical expressions, such as spring constant. They can be parameterized to provide results for a wide range of chemistry. Force field calculations are also quite fast ,making them suitable for studies on a long time scale; in this case around 50 psec to 100 psec.

Phenanthroline, known to form a monolayer on gold surfaces, was used for the molecular monolayer. At laboratory conditions, 21 C, the simulations confirmed that phenanthroline would form a stable monolayer. At operating conditions (37 C, human body temperature), however, the simulations predicted excess thermal energy would cause the monolayer to desorb, destroying the sensor.

Simulations, uncovered an alternative. Modeling allowed rapid screening of many more possible replacement molecules than were possible with experimentation alone. The researchers discovered that attaching carbon-60 to only a fraction of the phenanthroline would form a stable layer. This layer retains the electronic properties that make it suitable for use in the sensor, but is massive enough to be stable at 37 C.

The bottom line
By employing modeling, the company solved the problem in less than two weeks at a savings of approximately $712,000.