Sandia National Laboratories March/April 2002 volume 24, No. 2

aser-induced breakdown spec-troscopy (LIBS) has been applied to

the in situ detection of chemical elementsin the boiler of a power generator inMaui, Hawaii. The effort was part of aDOE BioPower Program demonstrationproject headed by Hawaiian Commercialand Sugar Company (HC&S) to test thefeasibility of using fiber cane as a boilerfuel. The Sandia team of Linda Blevins,Doug Scott, and Howard Johnsen and col-laborators Lee Jakeway of HC&S, ScottTurn of the Hawaii Natural EnergyInstitute, and Bryan Jenkins and RobWilliams of the University of California,Davis, found that co-firing fiber cane withdifferent fuels mitigated its tendency toform boiler deposits. This agrees with thefindings from experiments performed atthe CRF Multi-Fuel Combustor laboratoryin support of the demonstration project.

HC&S currently burns bagasse, thefibrous residue left after sugar is extractedfrom sugar cane together with coal andfuel oil to provide steam and electricityfor the sugar factory and the local electriccompany. Fiber cane (see Figure 1)potentially has better fuel characteristicsthan sugar cane and could be an attractiverenewable energy alternative to fossilfuels. However, because fiber cane,unlike sugar cane, would be burneddirectly without first extracting sugar, thealkali metals and other inorganics that areusually leached from bagasse duringsugar extraction could increase boilerdeposition.

In the LIBS technique, a high-energypulsed laser ionizes the gases and parti-cles, and the ionic and atomic emissionfrom the hot plasma is recorded. For thecurrent tests, a newly acquired system

consisting of an intensified charge-cou-pled device (ICCD) camera coupled to anechelle spectrometer was used. Theechelle spectrometer employs two grat-ings to allow detection of ultraviolet, visi-ble, and near-infrared light in a singlemeasurement. The LIBS systememployed a Nd:YAG laser with a pulsewidth of 10 ns and pulse energy of 350mJ. Spectra were collected at 5 Hz with adelay time of 10 µs and a gate width of50 µs. For this 2-week field trial, 1000-point ensemble averages representingabout 3 minutes of data were recorded forabout 8 hours per day.

A three-foot-long, water-cooled, nitro-gen-purged optics probe was designed andbuilt for the project. The probe was

inserted into the boiler near the first bankof superheater tubes. Multimode fiberoptics delivered light from the receivingoptics to the ICCD, which was kept cooland clean in the power plant controlroom. Magnesium, silicon, titanium, iron,calcium, aluminum, sodium, and potassi-um were detected during the field trial.Potassium is a prime contributor to boilerdeposition, and its LIBS-measured con-centration was used to predict the likeli-hood that boiler deposits would form.

The CRF team took LIBS data fromthree fuel combinations: coal, bagasse/coal, and fiber cane/bagasse/coal. Fueloil was added intermittently for boiler sta-bilization. For the co-firing cases, thefuel flows were estimated to be about

Figure 1. Harvested and chopped fiber cane drying on an airstrip between adjacent growing sugar

cane fields. Drying prepares the fiber cane for combustion in a power generation boiler. Sugar mak-

ing was bypassed, so there were questions about the fiber cane’s propensity to form boiler deposits.

Answers are being provided with LIBS data.

COMBUSTION RESEARCH FACILITY NEWS 1

Laser-Induced Breakdown Spectroscopy (LIBS) Used to EvaluateFiber Cane as Biomass Fuel in Hawaiian Power Generation Boiler

L

Sandia National Laboratories March/April 2002 volume 24, No. 2

2 COMBUSTION RESEARCH FACILITY NEWS

85% biomass by mass. The fiber caneconstituted about 18% of its biomass fuelblend. Figure 2 shows a typical LIBSspectrum collected during fiber cane co-firing. LIBS signals were quantified usinglaboratory calibrations performed atSandia after the field trial. In addition toLIBS, the team applied a variety of otherdiagnostics including a pitot tube, severaldeposition tubes, two continuous emis-sions monitors, an extractive wet-chem-istry impinger, and various thermocou-ples. Samples of fuels and effluents werealso collected for analysis. The resultinginformation comprises the most extensivedata set ever collected during a biomasscombustion field demonstration.

Figure 2. LIBS spectrum for fiber cane, bagasse, and coal co-firing in Hawaiian Commercial and Sugar

Company's traveling grate boiler. The spectrum is a 1000-shot, three-minute ensemble average. Key

species’ peak locations are identified with detection wavelengths expressed in nanometers. The

echelle spectrometer allows single-shot light collection between 200 nm and 900 nm.

Offgas SensorsInstalled at Two USSteel Plants

Sandia’s Steel Team (Sarah Allendorf,David Ottesen, Bob Green, BenChorpening, Doug Scott, ShaneSickafoose, and Gary Hubbard) have suc-cessfully completed the installation oftwo real-time, offgas sensors for processcontrol in commercial steelmaking. Thefirst sensor prototype is a double-ended,tunable-diode-laser-based device thatcontinuously measures the compositionand temperature of gases emitted duringelectric arc steelmaking, and is installedat The Timken Company, in Canton,Ohio. The second system is a single-ended sensor that detects infrared emis-sion from the offgas produced duringbasic oxygen steelmaking, and is installedat U.S. Steel’s Edgar-Thomson plant out-side Pittsburgh, Pennsylvania. Bothinstruments were installed duringNovember, 2001, and are still functioningwith occasional monitoring of their per-formance via long-distance computercontrol.

Wen Hsu and Sarah Allendorf have been named new CRF Department Heads. Sarah heads up the

Combustion Chemistry group and Wen takes over the Diagnostics and Remote Sensing Group from

Bob Gallagher, who moves on to Industrial and Combustion Processes. Sarah recently finished up a 6-

year stint as leader for the steel sensors project. Wen comes to the CRF from the Systems Research

Department at Sandia California's Center for Exploratory Systems and Development.

CRF Names NewDepartment Heads

Sandia National Laboratories March/April 2002 volume 24, No. 2

COMBUSTION RESEARCH FACILITY NEWS 3

The reacting flow and diagnostics portion of

the Department of Energy, Chemical Sciences

program at the Combustion Research Facility

was reviewed on March 4-6, 2002.

The participants included (Standing left to

right) Dr. Chiping Li (Naval Research

Laboratory), Professor Phil Johnson (SUNY,

Stony Brook), Dr. William Kirchhoff (Department

of Energy, Chemical Sciences), Professor Mike

Heaven (Emory University), Dr. Allan Laufer

(Department of Energy, Chemical Sciences),

Professor Ann Karagozian (UCLA), Dr. Charles

Romine (Department of Energy, MICS), Bob

Carling, Wen Hsu, Sarah Allendorf, Bill McLean,

Larry Rahn, and Dr. Heinz Pitsch (Stanford

University).

Dr. Ton van Mol of the Netherlands Organization for Applied Scientific Research (TNO)

shown with staff members Tony McDaniel (left) and Mark Allendorf (right). Dr. van Mol

is conducting experiments using a high-temperature flow reactor facility at the CRF to

determine the oxidation kinetics for tin precursors used in the formation of tin oxide

films on float glass. The project involves collaboration with PPG Industries, Inc. in

Pittsburgh, PA.

Friends and colleagues gathered atthe CRF on January 31 to share rem-iniscences of Rich Palmer whopassed away after a long illness.Rich started his 30-year Sandiacareer in Albuquerque after earning aPh.D. in Physics from PrincetonUniversity. He came to the CRF in1972 with an established reputationas a laser scientist and quickly addedmanagement and computer skills tothat reputation. He worked in anumber of different capacities at theCalifornia site, before returning tothe CRF in 1998. While Rich wasremembered in an astonishingly dif-ferent number of ways by those pres-ent, it was his courage that seemed totouch everyone.

Rich Palmer

Belated kudos to Chuck Mueller whose presentation of his paper "GlowPlug Assisted Ignition and Combustion of Methanol in an Optical DirectInjection Diesel Engine" won an award for Excellence in Oral Presentationat the SAE 2001 World Congress held in Detroit.

SAE Presentation Award

Sandia National Laboratories March/April 2002 volume 24, No. 2

4 COMBUSTION RESEARCH FACILITY NEWS

Direct Transition State Theory Analysis ExplainsTemperature Dependent Branching in NH2 + NO Reaction

hermal De-NOx is a noncatalyticprocess used to remove nitrogen

oxides from flue gas by the addition ofammonia. While the process is complex,the overall mechanism is exquisitely sen-sitive to the branching of the NH2 + NOreaction between two product channels (a)and (b):

NH2 + NO N2 + H2O (a) NNH + OH (b)

This sensitivity arises from the fact thatreaction (b), followed by the sequence,

NNH N2 + HH + O2 OH + OO + H2O OH + OH,

is chain branching, whereas reaction (a) ischain terminating.

The branched chain character of theprocess is important for predicting a num-ber of its observed properties. The chainbranching is also sensitive to the lifetimeof the NNH radical.

In earlier work, Jim Miller and PeterGlarborg (Technical University ofDenmark) constructed a satisfactorychemical kinetic model for Thermal De-NOx by fitting the branching fraction,α(T), to the available experimental dataand employing an NNH lifetime that wasconsistent with independent experimentalmeasurements and theoretical predictions.

However, the temperature dependenceof α(T) used in the Miller-Glarborg modelwas not consistent with the most reliabletheoretical prediction of this parameter.

Extending their earlier theoreticalanalysis, Stephen Klippenstein and Jim

Miller of the CRF, together with LarryHarding of Argonne National Laboratory,have now provided an a priori theoreticalexplanation (from first principles withoutdata fitting) for the temperature depend-ence of this branching. A schematic dia-gram of the potential energy surface gov-erning this reaction is provided in Figure1. The branching ratio is effectivelydetermined by the ratio of the transitionstate partition function for trans-HNNOH

NNH + OH (ts8) to that for thecis-trans isomerization of HNNOH (ts3,ts4).

The barrierless nature of the reverseNNH + OH recombination presents diffi-culties for standard implementations oftransition state theory.

Figure 1. The reaction coordinate diagram for the reaction of NH2 with

NO. The dashed lines indicate the barrierless transitions that yield NNH +

OH from each of the HNNOH conformers.

Figure 2. Comparison of recent theoretical predictions of the branching

ratios with experiment and with prior theoretical results. The solid line

represents the reference case with the NNH + OH reaction endothermic-

ity reduced to 0.9 kcal/mol and the ts3, ts4 barrier height reduced by

4 kcal/mol.

T

In recent work, Klippenstein andHarding have developed a novel imple-mentation of transition state theory specif-ically designed to provide accurate a pri-ori theoretical predictions for such barri-erless reactions.

This implementation directly couplesab initio quantum chemical simulationswith the transition-state partition functionevaluations, and includes a variable defi-nition of the reaction coordinate. Thisdirect variable reaction coordinate transi-tion state theory analysis was applied tothe partition function for ts8, with the

interaction energies determined at a highlevel of electronic structure theory(CAS+1+2/cc-pvdz).

This analysis for ts8 was coupled withmore standard evaluations for the othertransition states including ts3 and ts4, butagain employing state-of-the-art ab initioquantum chemical properties, to yieldestimates for α(T). A reduction by 1.6kcal/mol in the endothermicity of channela, and by 4.0 kcal/mol in the energies ofts3 and ts4, yields theoretical predictionsfor α(T) that are in quantitative agreementwith the experimental measurements and

the Miller-Glarborg model, as illustratedin Figure 2. Importantly, such revisionsin the thermochemistry are within theerror bounds of quantum chemical esti-mates. Thus, experiment, theory, and theMiller-Glarborg model are now in com-plete accord for this key reaction inThermal De-NOx. This work has beenpublished in Faraday Disc., 119, 207-222, (2001).

Sandia National Laboratories March/April 2002 volume 24, No. 2

RDX Decomposition Studies Point to the Formation of NonvolatileResidue as an Emergent Phenomenon

The capability to accurately predictthe response of a system containing ener-getic compounds RDX or HMX toimproper storage or to exposure, such as afire, is invaluable from a safety stand-point. A prediction requires the existenceof a robust model thatcan answer severalfundamental questionspertaining to the ‘dam-aged’ energetic materi-al. The Behrens Cycleis a scheme developedat the CRF to makesuch predictions (seeCRF News, Vol 23,No.2). The schemeattempts to account forall the products andrates by dividing ther-mal decompositioninto five cycles: solidphase, NO cycle,Nonvolatile Residue(NVR), Nitroso, andCH2O/NO2. Mostrecently, SeanMaharrey, DeneilleWiese-Smith, and RichBehrens have devoteda major effort tostudying the NVRcycle for RDX andfound that this cyclehas the characteristicsof an emergent phe-nomenon.

The NVR reaction cycle initiateswithin the solid-phase reaction cycle andinvolves reactions of secondary productsfrom the RDX decomposition. It domi-nates when the gaseous decompositionproducts are confined under high-pres-

sure. The NVR reaction-cycle becomesthe main rate-controlling reaction cycle asthe amount of NVR increases.

The NVR reaction cycle is responsi-ble for what has been called the ‘autocat-alytic’ behavior often observed in thermal

decomposition experi-ments. Autocatalyticbehavior is oftenattributed to reactionintermediates formedwithin a sequentialreaction mechanismthat act to catalyze thealready existing reac-tion without beingconsumed by the reac-tion. The overallNVR process, howev-er, also involves a pre-cursor formation stepand subsequentdecomposition upondepletion of RDX.

The precursor stepinvolves thermal initi-ation of surface activesites that readily pro-mote the formation ofthe NVR. A multistepnucleation andgrowth processbegins with sinteringof the individualRDX crystals into abulk agglomerate.Then transport and

Figure 1. This series of optical micrographs shows the NVR formation process. At the top left (a)

is the starting RDX sample. At ~ 3 % RDX decomposition (b), the RDX crystal has started to sinter.

At ~ 30% decomposition (c), the RDX particles have been completely covered by the sintered sur-

face, and the black spots have grown in size and number. Isolated black spots show the appear-

ance of grainy, orange patches on their surface, while close-lying black spots have developed

grainy, orange threads connecting them. In (d), the individual orange threads are shown to have

combined to form the red, translucent NVR.

COMBUSTION RESEARCH FACILITY NEWS 5

The CRF News is Published bimonthly by theCombustion Research Facility, Sandia NationalLaboratories, Livermore, California, 94551-0969.

Director: William J. McLean, Mail Stop 9054

Editor: Howard Lentzner, Email: hllentz@sandia.gov

Graphic Artist: Daniel Strong

reaction of RDX on this sintered surfaceresults in localized decomposition sites onthe surface. Secondary reactions on thedecomposition sites produce either elon-gated NVR ‘strands’ between close lyingdecomposition sites or NVR patches onisolated sites, followed by the growth ofthese isolated strands and patches into thebulk, amorphous NVR. Figure 1 depictsthe overall decomposition through thecomplex reaction pathways that couplethe NVR reaction cycle to the other reac-tion cycles.

This complex coupling of the solid-phase reaction cycle, the NVR reactioncycle and subsequent decomposition ofthe NVR, renders the ‘autocatalytic’ labelan insufficient description of the process.An emergent phenomena, as described in

the nonlinear system science community,is a process whereby a closed system, leftto itself, will seek its own path to adesired final product or an intermediate,which can itself alter the emergent phe-nomena process. Factors that tend toinfluence the emergent phenomenainclude initial environmental factors, suchas heat and mass transport between thesystem and its environment beforebecoming closed, initial state of ‘damage’to the system and distribution of this‘damage’ throughout the system, past his-tory of the system before onset of theemergent phenomena, and formation ofintermediates within the system by theemergent phenomena that can create new‘feedback’ loops within the system andinfluence or even alter the emergent phe-

nomena process. Scientists like NobelLaureate physicist Bob Laughlin believethat emergent phenomena are prevalent inthe mesoscale, or the region where matteris from 10 to10,000 angstroms: biggerthan a molecule, but smaller than a livingcell.

Thus, the overall NVR process, nucle-ation and growth through the solid-phasereaction cycle, ‘autocatalysis’ through theNVR reaction cycle, and NVR reaction/decomposition by the coupling pathways,when taken together, describe a complexprocess that is more accurately describedby the emergent phenomena process thanthe simpler ‘autocatalysis’ label.