Image of NO (A, J=7.5) after a collision of two molecular beams, one having NO seeded in He colliding with a beam on NO seeded in Ar. The NO in booth beams is excited into the A electronic state and the j=0 rotational state.  The He in one beam collides with the NO(A, j=0) in the Ar beam (small circle on left of image) and the Argon in one of the beams collides with NO(A, j=0) in the He beam (large ring on right of the image).


Collisional Energy Transfer Studied Under Single Collision Conditions

Ion imaging allows one to measure the differential cross section of a single quantum state of a collision process.  The process could be either inelastic collisional energy transfer or reactive collisions. These studies are very sensitive tests of scientists’ ability to calculate the potential energy function between molecules and their ability to calculate dynamics upon that potential energy surface.

Initial experiments of inelastic collision were of ground state molecules having collision with atoms (NO, CO, H2, D2, HCL with He, Ne, Ar and Kr).  These studies validated intermolecular potential energy surfaces and provided important information for modeling of combustion kinetics.  Studies utilize laser ionization of a specific quantum state of a molecule and the projection of those ionized molecules onto an imaging detector to determine the velocity and direction of the scattering.  Because a laser is used for the ionization, one can use the polarization properties of the laser to learn more information about the collision. In particular, one can learn how the angular momentum vector of product molecule is aligned in space (is the molecule spinning like a frisbee, a cartwheel or a propeller) as well as the orientation of the angular momentum vector (is the molecule rotating clockwise or counterclockwise with respect to the velocity vector of the molecule).  The data set below was the first measurement of the orientation of the angular momentum vector of a product of a reaction and it was found to vary with the angle of scattering and the rotational state of the scattered product.

Images of specific quantum states of NO (jf) product after scattering NO(j=.5) from Argon and collecting the images with circularly polarized light. The patterns indicate that the product molecules in specific quantum states scattering at specific angles have a propensity to rotate either clockwise or counterclockwise and this propensity alternates more quickly with low rotational state products than with high rotational state products.

These ground state collisional studies also led to a new way to cool molecules to a few millikelvin temperature, best demonstrated by collision of Kr atoms form Kr atoms. When two molecular beams are crossed at 90o and the two atoms (or molecules) that collide weigh the same, then one of the possible scattering velocities that are available for elastic scattering that conserve both energy and angular momentum is zero. Think of playing a game of pool: the cue ball and the pool ball always scatter at 90o.  If you were to run the pool shot backwards, you would see that scattering at 90 degrees can lead to one stationary ball (cold) and one moving ball (hot). The spot at the bottom of the image below on the left is of millikelvin temperature Kr atoms.

More recently we have pre-excited the molecule to be collided with to measure collisional energy transfer of electronically excited state molecule or highly vibrationally excited state molecules.

By electronically exciting NO2 just below the energy at which it dissociates the energy initially deposited into the electron is transferred into the ground electronic state and converts the energy into vibrational energy.  A single collision with an Ar atom is then used to deposit enough energy into the vibrationally excited state molecule to dissociate the molecule.  The NO (j) that is produced by the collision is then imaged onto a detector and the translational distribution measured. By measuring the velocity distribution of all the energetically possible quantum state one can reconstruct the collisional energy transfer function of the vibrationally excited state molecule.

Image of NO(j=5.5) formed from collision of Ar with vibrationally excited NO2. The NO2 was excited 50 cm-1 below the dissociation energy. Analysis revealed a double exponential decay but theory over predicted the amount of energy transfer in the second exponential by a factor of 2. This study was the first study of the collision induced dynamics of an ephemeral state of the NO2 molecule.

By electronically exciting the NO molecule to the A state at the crossing of a molecular beam on and atomic beam on, we were able to measure the differential cross section, alignment and orientation of product NO(A) state molecules. Even though the NO(A) state has only a 200 nsec lifetime sufficient collisions occurred between the excitation laser and the detection laser (400 nsec) to record the data.  Below is a sample of the data for different laser polarizations used to determine the orientation of the NO after the collision.

NO(A, J=4) product of the collision between NO(Aj=0) and a Ne atom. Images are different for different circular polarizations of the detection laser indicating orientation of the angular momentum vector. This was the first direct measurement of a four-vector correlation between the initial velocity vector, initial angular momentum vector, the final velocity vector and the final angular momentum vector.

Future studies will utilize high resolution infrared lasers to excite a very narrow velocity distribution of the molecular beam and we will study extremely high velocity resolution scattering.

This demonstrates the capabilities in the laboratory to measure transient processes with unprecedented detail.  We have not only measured this sort of differential cross section data we have also aligned the NO(A) to the direction of the collision and compared collisions of aligned NO molecules where the angular momentum vector is aligned parallel or perpendicular to the direction of the collision. This work is being done in collaboration with researchers Matt Costen and Ken McKendrick, both of Heriot Watt University in Scotland.

We utilize a crossed molecular beam machine with velocity mapped ion imaging ion optics to make the measurements.


A partial list of recent collaborators includes Dr. David Parker, University of Nijmegen; Dr. James Valentini, Columbia; Dr. Joe Cline, University of Nevada, Las Vegas; Dr. Steven Stolte, Univ. of Amsterdam; Dr. Paul Houston, Cornell; Dr. Ahren Jasper, Argonne; Drs. Thomas Sharples, Mathew Costen, and Kenneth McKendrick, Harriot Watt University, Scotland; and Dr. Javier Aoiz, Madrid.

PI: David W. Chandler

Sandia has developed a massively parallel DNS capability for turbulent reactive flows coupled with detailed chemical reactions and molecular transport. The DNS code, S3D, is used to perform DNS code of fundamental “turbulence-chemistry” interactions in combustion at Sandia and by researchers worldwide, notably at KAUST, U. of New South Wales, U. Southampton, Tokyo Tech U., and UNIST in South Korea, among others. S3D is also used as a CFD testbed for computer science developments in advanced programming systems and in data science with in situ analytics and visualization. In addition to the physical insights gleaned from fundamental research on turbulence-chemistry interactions in combustion, the DNS benchmarks are used by the international modeling community to develop predictive models used in engineering CFD to design fuel efficient, clean burning engines for ground transportation, propulsion, and power generation. Notably, we use predictive models to enhance the performance of light- and heavy-duty direct injection gasoline and diesel internal combustion engines with fuel efficiencies upwards of 60%, while minimizing emissions of nitric oxides and soot; to utilize high hydrogen content fuels as a clean energy carrier to reduce CO2 emissions in fuel and operational flexible power generation; to explore the use of ammonia/hydrogen/nitrogen blends as a carbon-free energy carrier for power generation, and to provide reliable flameholding in scramjet applications.

Experimental turbulent combustion research is focused on revealing and understanding the interactions between fluid dynamics, molecular transport, and combustion chemistry in flames. Many aspects of the complex interaction between fluid flow and chemistry can be explored non-intrusively using state-of-the-art laser-based optical diagnostics. Complementary diagnostics are applied to a variety of flames in the Advanced Imaging Research Laboratory (AIL) and the Turbulent Combustion Laboratory (TCL).

The goal of this research is to provide fundamental understanding of transport-chemistry interactions and to accelerate science-based predictive capabilities that can guide design, operation, and fuel formulation for practical combustion devices. Therefore, CRF experimental research is conducted in close collaboration with computational scientists and turbulent combustion modelers at Sandia and around the world. Apparatuses are often designed for fundamental studies of ‘building block’ flows with well-defined boundary conditions, such as unsteady laminar flames, turbulent jet and counterflow flames, and relatively simple flames stabilized by bluff-body recirculation or swirl. The strong coupling among experiment, theory, modeling, and simulation provides a powerful approach to understanding complex combustion processes.

Example turbulent flame configurations used for fundamental studies of turbulence-chemistry interactions. Different flow configurations and fuels provide a progression of complexity with respect to fluid dynamics and chemical kinetics. Systematic experimental studies of these flames using laser diagnostics are combined with theory and modeling to advance our understanding of the coupling between transport and chemistry.


Simultaneous single-shot LIF imaging of OH and CH2O in a turbulent partially premixed dimethyl ether/air jet flame with increasing amounts of localized extinction as the Reynolds number increases from left to right. Intermittent localized extinction and re-ignition involve highly nonlinear turbulence-chemistry interactions that are among the most challenging processes to model in turbulent combustion. Source: Frank, J. H., Advances in imaging of chemically reacting flows. J. Chem. Phys. 2021, 154 (4), 040901.


Based on experiments published in (1) Coriton, B.;  Im, S.-K.;  Gamba, M.; Frank, J. H., Flow Field and Scalar Measurements in a Series of Turbulent Partially-Premixed Dimethyl Ether/Air Jet Flames. Combust. Flame 2017, 180, 40-52, and (2) Coriton, B.;  Zendehdel, M.;  Ukai, S.;  Kronenburg, A.;  Stein, O. T.;  Im, S.-K.;  Gamba, M.; Frank, J. H., Imaging measurements and LES-CMC modeling of a partially-premixed turbulent dimethyl ether/air jet flame. Proc. Combust. Inst. 2015, 35 (2), 1251-1258.

PIs: Jonathan H. Frank, Robert S. Barlow

Chemical reactions give rise to nascent products, termed non-equilibrium (NE), whose energy distribution does not correspond to that of either the reactants or the bath. Such NE product ensembles have different effective reactivity relative to their thermal equivalents, thereby impacting important macroscopic quantities in a complex reactive system. NE products can have enough energy to dissociate promptly prior to thermal stabilization, or incipient adducts can react with other species while collisionally cooling. Large temporal and/or spatial gradients in composition or temperature generated at the continuum scales also create ephemeral NE ensembles by inducing fast transport processes. These arise because the characteristic timescales of energy transfer of the various molecular energy modes (translation, rotation, vibration) differ substantially from each other and because steep spatial gradients can transport hot and cold ensembles short distances of just a few tens of mean free paths away from each other. The interplay of transport and NE populations is a fundamental scientific phenomenon rooted in the Boltzmann equation. The emergence and effect of molecular NE caused by large gradients generated at the continuum macro-scale requires the formulation of new phenomenological non-equilibrium kinetic models and a multiscale data-driven framework to bridge myriad scales: from the atomic to the mesoscopic and up to the continuum scale. The challenge is to maintain the microscopic fidelity across these scales in a computationally tractable manner.

Our research objective is to understand how NE chemical processes affect the behavior of complex systems, and to create multi-scale bridging models with the assistance of data-driven machine learning (ML) methods to describe them. In particular, we are studying NE processes arising due to large temporal and/or spatial thermal and composition gradients which induce fast transport, and how these effects compare to homogeneous reaction kinetics. A variety of chemistry-induced NE effects in homogeneous systems at the microscale have been studied and shown to have significant impact on oxidation kinetics in a variety of contexts. These studies have demonstrated the effect of NE product energy distributions on subsequent reactions’ rate coefficients and branching ratios. They have also shown how NE reaction cascades develop and thermalize under different conditions.

We have begun construction of a new multi-scale data-driven framework comprised of the following elements: (1) reaction cross-sections, derived from ab initio chemistry calculations (TST/QCT/ME), to be used as input to mesoscale simulations; (2) a mesoscale molecular method, Direct Simulation Monte Carlo (DSMC), newly extended to statistically treat nonequilibrium transport processes coupled with NE chemistry in the continuum regime; and (3) direct numerical simulation (DNS), to calculate turbulent reacting continuum flows, with an extension to treat nonequilibrium effects using a multi-temperature phenomenological model derived from coarse-graining microscopic models. The DNS approach solves the continuum near-equilibrium Navier-Stokes (NS) equation, resolving turbulence and chemical scales from microns-cm and from ns-ms on petascale supercomputers. DNS is used to simulate turbulent flows and shocks that provide realistic time-varying thermal and composition gradients across which NE occurs. DNS is also used to validate NE phenomenological models derived from mesoscopic DSMC results in complex flows.


Combustion systems are characterized by a complex interplay between convection and diffusional transport processes and chemical reaction rates, particularly in turbulent flows, wherein strong, time-varying gradients in both chemical composition and temperature couple with chemical reactions. These systems are practically relevant and scientifically challenging to investigate and understand, due to their wide range of relevant spatial and temporal scales. The CRF program has provided decades of leadership in combining experiments and modeling of turbulent combustion through advanced diagnostics development, leading-edge modeling and simulation, and collaborations within the greater combustion research community, as exemplified by the International Workshop on Measurement and Combustion of Turbulent Flames. Current work in the interactions of chemistry and transport focuses on connecting continuum flow to the molecular level, building from detailed studies of elementary non-reactive collisions to modeling the connection of spatial temperature gradients to non-equilibrium chemistry.