New coordination complexes-based gas-generating energetic composites

Any correspondence concerning this service should be sent

to the repository administrator:

tech-oatao@listes-diff.inp-toulouse.fr

This is an author’s version published in:

http://oatao.univ-toulouse.fr/26836

To cite this version: Wu, Tao and Sevely, Florent and Julien,

Baptiste and Sodre, Felipe and Cure, Jeremy and Tenailleau,

Christophe and Esteve, Alain and Rossi, Carole New coordination

complexes-based gas-generating energetic composites. (2020)

Combustion and Flame, 219. 478-487. ISSN 00102180

Official URL

DOI :

https://doi.org/10.1016/j.combustflame.2020.05.022

Open Archive Toulouse Archive Ouverte

OATAO is an open access repository that collects the work of Toulouse

researchers and makes it freely available over the web where possible

Combustion

and

Flame

New

coordination

complexes-based

gas-generating

energetic

composites

Tao Wu

a

_{, Florent Sevely}

a

_{, Baptiste Julien}

a

_{, Felipe Sodre}

a

_{, Jeremy Cure}

a

_,

Christophe Tenailleau

b

_{, Alain Esteve}

a

_{, Carole Rossi}

a,∗

a LAAS-CNRS, University of Toulouse, 7 Avenue du colonel Roche, 31400 Toulouse, France

b CIRIMAT, University of Toulouse, CNRS, UT3-Paul Sabatier, 118 route de Narbonne, 31062 Toulouse, France

a

r

t

i

c

l

e

i

n

f

o

Article history: Received 6 April 2020 Revised 27 May 2020 Accepted 27 May 2020 Available online 30 June 2020

Keywords:

Coordination complex Energetic materials Non-toxic gas generator Flame propagation Airbag

a

b

s

t

r

a

c

t

Toreducethesizeinconvenienceofairbagsystemsemployedinhuman-bodyprotectiondeviceswhile maintainingcomparablegasgenerationperformance,anewfamilyofgas-generatingenergetic compos-itesare proposed mixingAl/CuOnanothermite with copper complex (Cu(NH3)4(NO3)2)known for its

propensitytogenerategases(0.03mol/g)suchasN2,O2,N2Othroughexothermicchemical

decomposi-tion(300J/g).Thealuminum(Al)/Copperoxide(CuO)couple,knownasthemostwidelystudied nanoma-terialforthermitereactions,releasingahighenergy,mostlyheat,throughchemicalreaction,isemployed asasourceofheattotriggerandsustainthedecompositionofCu(NH3)4(NO3)2complex.Thiswork

per-mitsdevelopinga newfamily ofgas-generating energeticcomposites thattakes advantageofspecific chemicalandthermalpropertiesofbothmaterials.Wedemonstrateitscapabilitytotunethe pressuriza-tionrate,burnrateand pressurepeakbyvaryingtheAl/CuOoverCu(NH3)(NO2)2massratio.Thepeak

pressureofCu(NH3)4(NO3)2/Al/CuOenergeticcompositesreaches12MPa/g.cm3inaclosevolume,which

is3.3higherthanthatoftraditionalAl/CuOnanothermite.Theyachievemuchlongerhigh-pressure dura-tion(∼30ms).Theyalsoexhibitveryintenseburningwithvelocityreachinghundredsofm/sinopening burningexperiments.Thesenewmaterialsappearverypromisingforgreengasgeneration.

1. Introduction

Automobile airbag systems are the widespread micro-pyrotechnical systems introduced in the earlies 1970s [1] de-signed to slow down themovement ofthe caroccupants within milliseconds to prevent injury during a collision [2,3]. Recently, new applicative markets have emerged with the goal to prevent personal injuriesduring extreme sport (skiing,equitation, motor-cycling), urban two-wheelers or to protect the elderly from hips fracture when falling. For these new applicative domains, smart inflatable protection devices using compressed CO2 bottles are nowavailableonthemarket[4–6].Whenthesensorsdetectafall, the integrated circuitry triggers the actuator, which then inflates thenylonairbagbeforetheimpact[7,8].However,all commercial-izedinflatorsystemsarebulky(50–250cm3_{),heavy(more than1} kg), costly, owing to thefact that a tank withcompressed gasis employedtoperformtheactuation.

Alternatively, gas-generating energetic materials could effec-tively providehigh-energy andhigh-power actuationwithin a

re-∗ _{Corresponding author.}

E-mail address: rossi@laas.fr (C. Rossi).

duced volume (<50 cm3_{) thanks to the large amount of stored} chemical energy they can release quickly. In the context of per-sonalprotection,theyhavetomeetsevererequirementsincluding good performances, moderate temperature (<300°C) [9,10], non-toxicity of gaseous species, reliability [11], safety when handling andduringstorage,andlowenergyignitions.Gas-generating mix-tures developedfor automobileairbag inflation are mostly based on a propellant, such as sodium azide (NaN3

)

andpotassium ni-trate (KNO3) that thermally decomposes into gases and residues when they are ignited andthey have been used fora long time

[12].A series ofchemical reactions produce N2 to fill theairbag andconvertmetallicazidewhichishighlytoxic[13]evenafter re-actiondueto incompletecombustion.Toreducethetoxicity, high nitrogennonazidegascompositionsareproposed andusedin in-flatingpassengerrestraintgasinflatorbags,whereammounium ni-trate,guanidinenitrate,etc,areemployed[14,15].However, guani-dinecompoundisexplosivefromfrictionandthereforedifficultto handle[16].Infact,ammoniumnitratewasusedasagasgenerator toinflateairbagssince2001byTakata;however,morethan16 mil-lionvehicleswererecalledworldwideduetoaserioussafetyissue that high moisture may accelerate the decomposition of ammo-niumnitrateandleadstoviolentexplosionoftheairbags[17,18].

Herein, we consider a new, safeand non-toxicgas-generating materials based on the Werner type complexes composed of a metallic cation (Cu, Fe, Co or Mn) bounded to oxidizing anion ligands (NO−_{, NO}−

3, NO −

2). Numerous complexes have been successfully synthesized and studied by numerous groups such as Cu

(

NH3

)

4

(

NO3

)

2 [19–22], Co

(

NH3

)

6(NO3)

)

2 [23,24], Mg(NH3

)

6(NO3

)

2 [25,26], Ni(NH3

)

6(NO3

)

2 [25], Cd(NH3

)

6(NO3

)

2

[27], becausethey are stableand enable thegeneration of abun-dant environmental friendly gases (N2, O2, N2O) [28] from the thermaldecompositionoftheorganicanionatmoderate tempera-ture(150–220°C).Inaddition,complexcompoundsusuallycontain fuel (ligands such asammonia or nitrogen compounds), oxidizer (likenitratesorperchlorate)andmetalthatmayactasburn mod-ifier.

Amongall publishedcoordinationcomplexes,themoststudied isCu

(

NH3

)

4

(

NO3

)

2 (calledCuC)becauseofits lowdecomposition temperatureandrelativelyeasysynthesisprocess.Despiteapatent that proposesto useaheterometallicWerner complexasagreen gas source for airbag application[29], the applicative interest of CuCasagas-generatingmaterialremainsstalledsinceits exother-mic decomposition is not a self-sustained process, which means thattheentiremassofthecomplexhastobeheatedtoits decom-positiontemperaturetoreleasethegasesrequiringatooimportant externalsourceofheat[20].

ThisstudyproposestoincorporatenanothermitesintotheCuC mixture,toserveasan embeddedsourceofheattodrastically re-ducetheexternalactivationenergyneededtodecomposethe com-plex[11].Incontrasttothecoordinationcomplexinitiationmode, thenanothermiteengagesaself-sustainedreactionuponhotpoint initiation.Inotherwords,afterbeingignitedlocally,the nanother-mite iscapable ofproducingenough heat toquicklyandentirely heatupthemassofCuCtoitsdecompositiontemperature,which decompositionleadstogasreleaseandpressure.

ByvaryingthemassofnanothermiteincorporatedintotheCuC, this study seeks to experimentally find the optimized Al/CuO to CuC mass ratio allowing: (i) the lowering of the power density budget for initiation, (ii) the self-sustained andtotal decomposi-tion ofCuCtomaximize thepressure peak,(iii)providing funda-mental understanding on the interplay ofthe two materials. We will discussthe combustion eventsin such CuC/Al/CuO energetic composites, experimentally evidenced thanks to coupled thermo-chemical analyses using calorimetry coupled with time-resolved massspectrometry. Overall,we believethisworkopens new per-spectivesinthefieldofnon-toxicgasgeneratingmaterialformicro andmini-airbag applicationssince thisnewenergetic composites feature peak pressure of 12 MPa/g per cm3_{, which is 3.3higher} than that of traditional Al/CuOnanothermites,with much longer high-pressuredurations(∼30ms),andwiththeproductionofsafe gasesonly.

2. ExperimentalSection

2.1. Materials

Thealuminumnanopowders(Al)(80nm)werepurchasedfrom Novacentrix andstoredinaglove box.The activeAl content was 69% by mass, calculated from thermogravimetric analysis before use. Copper oxide (∼100 nm),copper nitrate trihydrate and 25% aqueous ammonia, purchased from Sigma-Aldrich were directly usedasreceived.

2.2. SynthesisofCuC

The Cu complex was synthesized according to the literature

[21,22].Inbrief,4.83g(0.02mol)ofcoppernitratetrihydratewas dissolvedin10mLofdistilledwaterfollowedbyadditionof15mL

Table 1

Chemical composition of the different batches of CuC/Al/CuO energetic composites

Samples Al CuO CuC Al/CuO Al/CuC Total (mg) (mg) (mg) (mg) (mg) (mg) Al/CuO 114 286 400 400 Al/CuC 40 360 400 400 CuC/Al/CuO_25 59 71 270 100 300 400 CuC/Al/CuO_50 77 143 180 200 200 400 CuC/Al/CuO_62 86 177 137 248 152 400 CuC/Al/CuO_75 96 214 90 300 100 400 CuC/Al/CuO_82 101 234 65 328 72 400 CuC/Al/CuO_90 107 257 36 360 40 400 CuC/Al/CuO_95 111 271 18 380 20 400

of25% aqueousammoniasolution(0.24mol).The mixedsolution was kept inan icebathwithmoderate stirring foronehour and then withoutstirringforthreehours.The finalproductwas sepa-ratedviavacuumfiltrationanddriedinanovenat70°C.Theyield oftheproductis∼65%(3.2g).

The X-Ray Diffractometer(XRD) patternofas-prepared CuCis well correlated to a published Cu(NH3)4(NO3)2 crystal structure with a pdf#01-070-195.A very smalltrace ofCu2(NO3)(OH) was alsodetectedinthemixture(labeledinblueinSupplementary In-formationFig.S1A).TheFTIRspectrumofsynthesizedcopper com-plex (Fig.S1B)havethesamecharacteristicpeaksastothe litera-ture [21]. Both XRD andFTIR results prove that copper complex was successfullysynthesizedinthiswork.

Scanning ElectronMicroscopy (SEM) was employed to charac-terize thecoppercomplexparticlesandtheresultsshownin Sup-plementary Information Fig. S2 indicate that they are irregularly cylindricalwithan averagesizeabout20µminwidthand60µm inlength.Azoom-inimagealsoshowsthatthesurfaceofparticles isrough,whichcouldbe beneficialforfurtherdepositionofother materials.

2.3. PreparationofAl/CuCenergeticcomposites

Mixtures of Al/CuC were prepared using a classic sonication method where Al nanoparticles were mixed with CuC with an Al/CuC equivalenceratioof1.2inhexanefollowedby20minutes sonication(seeEq.1).ThenAl/CuCenergeticcompositewasstored inhexaneforfutureexperimentations.

4Al

(

s

)

+6Cu

(

NH3

)

4

(

NO3

)

2

(

s

)

→2Al2O3

(

s

)

+ 6Cu

(

s

)

+27H2O

(

g

)

+6NH3

(

g

)

+12N2

(

g

)

+3N2O

(

g

)

(1)

2.4. PreparationofCuC/Al/CuOenergeticcomposites

Different CuC/Al/CuO energetic composites with various mass loadings ofAl/CuOnanothermite(Table1)were preparedwithan equivalenceratioof1.2forbothAl/CuCandAl/CuOinorderto in-vestigate the effectof the nanothermite mass ratio on the pres-surizationandburnrate.Eachcompositewascharacterizedbythe weight percentageofAl/CuOnanothermite, X,rangingfrom25to 95%. Note thatin a X=25% (mixture namedCuC/Al/CuO_25), the mass of Al/CuOthermite andAl/CuC is 100 mgand300 mg, re-spectively.

2.5. ChemicalAnalysis

Themorphology,particlesize,agglomerationwereobservedby SEM using a HitachiS-4800 (cold FEG)operating at 2.0 kV. The chemical composition of thematerials was characterized using a SEIFERT XRD 3000 TT XRD withCu-Ka radiation (

λ

=1.54059 ˚A) fittedwithadiffracted-beamgraphitemonochromator.The diffrac-tionangle(2

θ

)wasscannedfrom10to80°.

Fig. 1. SEM images of Al/CuC (A), CuC/Al/CuO_50 (B) and Al/CuO (C).

2.6. Thermalanalysis

The thermal decomposition of the prepared energetic com-posites was characterized under a ramping heating profile at 10°C/min by: (1) Thermogravimetric/Differential Scanning Calorimetry (TGA/DSC)usinga Mettler ToledoThermogravimetric analyzer over a temperature rangingfrom ambient to 700°C; (2) DifferentialScanningCalorimetryusingaNETZSCHDSC404F3 Pe-gasus device equippedwitha DSC-Cp sensor type S over a tem-perature rangingfromambientto1000°C.Experimentswere per-formed with∼5mgofcompositesinplatinumcruciblesinAr at-mosphere (99.998%pure)ataflow rateof20mL/min.The traces arenormalizedbythemassofenergeticcompositematerial.

2.7. ThermoGravimetric/MassSpectroscopy

The decompositionoftheenergeticcompositeswas performed by mass spectrometry (MS, equipment Pfeiffer Omnistar 1-200 amu) over a temperature range from ambient temperature to 900°Cataheating rateof10°C/min.TheMSapparatusiscoupled with theTGA systemdescribed inthe previous section. All mea-surementsareperformedunderArgonflowinggas.

2.8. Closedbombexperiments

Thepressurereleaseismeasuredinastainlesssteel,high pres-sureresistant cylindricalreactor [30].The dimension ofthe reac-torhasa4mm innerdiameterandis0.7mmlong(totalvolume of∼9mm3_{). Astainless spacer(volumeof∼25mm}3_)wasplaced between the reactor and the pressure sensor. For all other sam-ples, about13 mgof CuC/Al/CuO powderswere packedin a pel-let placedinthebombwhichhasatotalvolume of25mm3_.The packingdensitiesofthesamplefilledinallpelletswerecalculated at around 1.5g/cm3_{. When positioned in the bomb volume, the}

final packingdensity dropsdown to 0.5g/cm3_{.A pressure up to} 100MPacanbemeasuredbyahigh-frequencypressuretransducer (Kistler6215 type)onthebackofthereactor andconnectedtoa Kistlercharge amplifier (5018 type).A schematic and a photo of the systemare shown inSupplementary InformationFig. S3. For ignition,asimpleNiCrwirewasused.

2.9.Openburningexperiments

Thereactionpropagationspeedofcomposites wastestedwith ahome-madeset-up,wherearectangularholdermadeof polycar-bonatewitha trench(30mmlong,1mm wideand1mmdeep) isusedtosupportthecompositematerial.About25mgofpowder waslooselydepositedinthetrench.Forignition,apyroMEMSchip withcoatingofAl/CuOnanolaminateswasusedtoproduceaflame toignitetheenergeticcomposite[31,32].Themeanburnratewas recordedusing ahigh-speedcamera SA3 Photronableto capture thevisible light perpendicular tothe direction offlame propaga-tionofthemixture,ataspeed of75,000framespersecond,with a128×32imageresolution[33].

3. ResultsandDiscussion

3.1. Structuralcharacterizations

The SEM images of Al/CuC energetic materials show that the aluminum nanoparticleswith an average size of 80 nm are well dispersedonthesurfaceofcoppercomplexrods(Fig.1A).TheSEM imagestakenforCuC/Al/CuO_50energeticcomposite(Fig.1B)also show a gooddispersion ofAl andCuO nanoparticlesaround CuC rods.When thenanothermite mass exceeds 50% inCuC/Al/CuO, wenotice someareaswithonly Al/CuOnanothermites(Fig.1B1). As reference,SEM imagesof Al/CuOnanothermitesare presented

Fig. 2. XRD patterns of Al/CuC, CuC/Al/CuO_82, CuC/Al/CuO_75 and CuC/Al/CuO_50.

Table 2

Closed bomb test results of different CuC/Al/CuO compos- ites. Each sample was tested at least three times and mean values are reported.

Sample P max Pressure Packing

(MPa) rise rate Density (kPa/ µs) (g/cm 3 ) Al/CuO 1.8 ± 0.2 11 ± 0.1 0.48 ± 0.1 CuC/Al/CuO_95 3.7 ± 0.2 69 ± 6 0.44 ± 0.1 CuC/Al/CuO_90 3.5 ± 0.1 45 ± 4 0.48 ± 0.1 CuC/Al/CuO_82 4.0 ± 0.3 36 ± 10 0.48 ± 0.1 CuC/Al/CuO_75 5.9 ± 0.2 41 ± 8 0.51 ± 0.1 CuC/Al/CuO_50 No ignition 0.51 ± 0.1 CuC/Al/CuO_25 0.50 ± 0.1

in Fig. 1C. Pictures are plotted witha greater magnification (×5 comparedtootherimages)todistinguishAlandCuOcompounds.

TheXRDpatternsoffivedifferentenergeticcompositesshown inFig.2indicatethecompositionalinformation:CuC, CuOandAl aredetectedinallsamplescontainingnanothermite.

3.2. Pressuremeasurement

All prepared energetic composites were characterized in the closedbombwitha packingdensitykeptataround0.5g/cm3_for all samples (Table 2). We obtained pressure outputs (see typical temporalpressuretracesplottedinFig.3)forallcasesexceptfor:

- PureCuCandAl/CuC,noself-sustainedreactionoccursforthese twosamplesevenwithacontinuouslocalheatingthroughthe NiCrwire.

- CuC/Al/CuO_50andCuC/Al/CuO_25couldnotbecompletely ig-nited.Wesupposethattheheatappliedfromthewirewas in-sufficient to bring these two samples to their decomposition temperaturebefore thewire breaksafter severalmilliseconds. To support this theory, a lighter was used as the heat sup-plier and we observed violent reactions for these two mix-tures after being heated. Interestingly, the moment the plas-ticpelletstartedmelting, ignitionhappened.Itimpliesthat all CuC/Al/CuO energetic composites could be ignited as long as there was enough heat supply to the system. Thus, the igni-tionpathwayemployed inthistest hasits limitforthesetwo compositionsinwhichtheCuCinitiationmodedominates,i.e.

Fig. 3. Temporal pressure traces of CuC/Al/CuO energetic composites and traditional Al/CuO nanothermite as reference.

heating of the whole material at the nominal decomposition temperature.

Amongallothercomposites,thehighestoutputpressure(Table 2)isgeneratedbyCuC/Al/CuO_75with∼6MPa,avaluemorethan three timesofAl/CuOtakenasreferencehere. NotethatAl/Bi2O3 beingamongthethermitecouplesgeneratingthehighestpressure has a peak pressure of only 1.3 times of Al/CuO [30]. Thus, the CuC/Al/CuO_75 composite, significantly outperform Al/Bi2O3 and Al/CuO in termsof pressure generation.Even forCuC/Al/CuO_95, where 5% by weight of Al/CuC was mixed with 95% of Al/CuO, its peak pressure still performs two times that of thetraditional Al/CuOnanothermite.

In addition, the pressurization rates (the slope between two points: the peak pressure point and the point where it reaches 5% of peak pressure on pressure traces in Fig. 3) of these four CuC/Al/CuO energetic composites are much faster than that of Al/CuO. The highest value, ∼70 kPa/

µ

s, is now obtained with CuC/Al/CuO_95.

AnotherinterestingpointhastobenotedfromFig.3:itseems bothCuC/Al/CuO_75andCuC/Al/CuO_82notonlyfeaturerelatively

Fig. 4. SEM images of residues products from Al/CuO (A), CuC/Al/CuO_75 (B) and CuC/Al/CuO_50 (C).

Table 3

Summary of burn rates of different CuC/Al/CuO energetic composites. Each test was done at least 3 times and mean values are reported.

Samples Burn rate Density (m/s) (g/cm 3 ) Al/CuO 36 ± 5 0.8 CuC/Al/CuO_95 40 ± 5 0.7 CuC/Al/CuO_90 64 ± 10 0.7 CuC/Al/CuO_82 79 ± 15 0.9 CuC/Al/CuO_75 141 ± 25 0.8 CuC/Al/CuO_62 135 ± 25 1.0 CuC/Al/CuO_50 3 ± 1 0.8 CuC/Al/CuO_25 No propagation

higherpeakpressurebutalsoshowmuchlongerdurations(30ms) atmaximumpressures thanothersamples(∼5ms)testedinthis work.Itmightbecomeanattributingfactorforairbagapplications. Assummaryofpressuretests,theoptimizedcompositionis es-tablished with 75% loading of nanothermite by weight into CuC. Also note that pressure results are obtained at very low com-paction (only 0.5 g/cm3_{, i.e. ∼8% TMD). These results also show} thatthecompositecanbeadaptedtodifferentscenarios:relatively low peak pressure buthighpressurization rateor relatively high peakpressurebutlowpressurizationratecanbeusedaccordingly.

3.3. Openburningexperiments

ThepropagationspeedsofallCuC/Al/CuOenergeticcomposites were also investigatedandtheresults arereportedin Table3.In thisexperimentthedensities oftheenergeticpowdersinsidethe trenchwerekeptataround0.8g/cm3_{foralltestedsamples.}

Similar to pressuretest, nopropagationwas observed forCuC and Al/CuC composites. Neither does CuC/Al/CuO_25. These facts indicatetheweightpercentageofnanothermitemustreacha

cer-tain level to guarantee a self-propagated system. With 50% (by weight)additionof Al/CuOnanothermite,CuC/Al/CuO_50 demon-stratedaburnrateofaround2m/sbutonly2/3ofthecomposite inthetrenchhaspropagated,suggestingthatabadthermal trans-ferwithinthecompositemightbethecauseofthepoor propaga-tionbehaviorsimilartothepressuretests(hasthelowestthermal conductivity,seeSupportingInformationTableS1).Another possi-blecausemightbethebreakup,crystaltransitionandphase tran-sitionofCuCabsorbing toomuchheat tomaintainsteady propa-gation.

All the other composites exhibit very intense burning (See snapshots in Supporting Information Fig. S4) with steady flame propagations. When the weight percentage of the nanothermite raised to75%, the burn ratereachesthe highestvalue amongall testedsamples,∼141 m/s, 4timeshigherthan traditionalAl/CuO nanothermite giving a burn rate of ∼36 m/s in this experimen-talcondition.Asseeninprevioussection,CuC/Al/CuO_75also fea-tures the highest peak pressure generation. However, further in-creasingthemassloadingofnanothermitedecreasestheburnrate dramaticallyratherthanenhancingit.With90%byweightof nan-othermite addition, the burn rate drops to almost a half of the peakvalue.AnditfinallyreachesthesameburnrateastoAl/CuO when the weight percentage increases to 95%. We deduce that thereisacompromisebetweentheheatproducedbythe nanoth-ermiteandthermalpropertiesvaryingwithnanothermite percent-ageasthe thermalconductivityincrease withnanothermite load-ing: 56 W.m−1_K−1 _{for CuC/Al/CuO_75 against 66 W.m}−1_K−1 _for CuC/Al/CuO_95,thus an increaseof 18%(see Supporting Informa-tionTableS1).

To fill the huge gap between a burn rate of 141 and 3 m/s, CuC/Al/CuOcompositewith62%loadingofnanothermitewas pre-pared and tested afterwards. A burn rate of 137 m/s was ob-tained, slightly slower than that of CuC/Al/CuO_75. Similar to CuC/Al/CuO_50 andCuC/Al/CuO_25, noignition was found inthe

pressuretestofCuC/Al/CuO_62.However,bothCuC/Al/CuO_50and CuC/Al/CuO_62show steadypropagationsafterbeingignited.And CuC/Al/CuO_62 features the second fastestburn rate. Itresonates with the previous discussion that ignition of the composites can only happen when enough heat is supplied in the systemand a certain amount ofnanothermite loading(>62%)is required fora continuousandfastpropagation.

As summary,the composite featuring the fastest burn rate is thosecontaining75%loadingofnanothermitebyweightintoCuC. Italsofeaturesthehighest-pressureoutput.

3.4. Post-combustionproductscharacterization

Reaction products and residues from burn rate tests of all CuC/Al/CuOenergeticcompositeswere collectedandexamined by SEM.AsshowninFig.4,givinganalysisforAl/CuO,CuC/Al/CuO_75 andCuC/Al/CuO_50 samples,therearemainly twopopulations of product particle sizes, demonstratingthe similar combustion be-haviorsbetweenAl/CuOandCuC/Al/CuOcomposites. Onehas rel-ativelylargerdimension witha sphericalshape andanotherwith adimensionassmallas∼50nmbuthighlyaggregated.EDS mea-surement results shownin supporting information Fig. S5-S7 in-dicate that the larger ones are alumina particles with a coating ofcopperparticles,andsmalleronesare coppermetalaggregates (might also contain cuprous oxide particles since oxygen signal wasalsofoundinthesearea).

In addition, there are much less residues collected after the burning of CuC/Al/CuO energetic composites compare to Al/CuO (Fig. S8).The wholetrenchwas coveredby productswithacolor ofdarkredforAl/CuO,however,onlya verythinlayerofresidue withalightgreywereobservedonthetrenchofotherCuC/Al/CuO composites.Thus,mostoftheproductsproducedfromthese com-posites aregaseous species, which isin accordancewiththe fact thattheyoutperformAl/CuOinpressuregenerationtest.

3.5. UnderstandingthereactionsinCuC/Al/CuOenergeticcomposites

To fullyunderstand the combustion eventsof the CuC/Al/CuO energetic composites composite, all samples were thermally and chemically analyzedupon ramping (10°C.min−1) usingcalorimery (TGA/DSC)coupledwithtime-resolvedmassspectrometry.

TheresultsofDSCexperimentsarepresentedinFig.5.TheDSC tracesofCuCandAl/CuOnanothermite are alsoincluded(Fig.5a and 5h)as referencesand details on the CuC thermal decompo-sition (TGA/MS spectra) can be found in Supporting information (Fig.S9 andS10).TheDSCtraceofCuCfeaturestwo endothermic peaks:firstoneat∼120°Cmightcorrespondtotheevaporationof thephysicallyabsorbedwatermolecules.Thesecondendothermat ∼210°Ccouldbe correlatedtothemeltingofthecoppercomplex

[21]and/orthedesorption ofthephysicallyabsorbedammonium. Indeed, a sharpNH3 peak inthe mass spectroscopyspectrum of CuC is observed in the same temperature range (Fig. S10). The exothermicpeaklocatedataround270°Ccorrespondstotherapid decompositionofthecoppercomplex(concomitantwithmassloss seenontheTGAcurveinFig.S9)alongwithreleasesofN2,N2O, H2OandNH3gases(detectedontheMSspectrainFig.S10).Based onthe XRDanalysis(Fig.6)ofthe residuesfromtheDSC experi-ments,theleftover powdersarecopperoxideparticles thatisthe onlysolidproductfromthedecompositionofcoppercomplex.Not onlythe CuCactsa gasgeneratorbutitalsodecomposesinto an oxidizer that further enables nanothermite kind ofreaction with Al.Indeed,whenAlnanoparticlesareaddedintoCuC(Al/CuC,Fig. 5b),asmallexothermicpeakappearsbetween500and600°Cthat islocatedwithintheAl/CuOredoxreactionregionprovingthat Al reactswithCuOoriginatedfromthedecompositionofthecopper complex.

Fig. 5. DSC curves of CuC/Al/CuO energetic composites with different weight per- centage of Al/CuO in argon environment.

However, thedecompositionofCuCneeds tobe initiatedfirst, thenthedecompositionproductCuOcouldreactwithAl nanopar-ticles.Therefore,anextraheatsupplyisrequiredtolaunchAl/CuC reaction, hence the addition of nanothermite into CuC. Different thermal behaviorsareobserved depending onthequantity of in-corporatednanothermite.CuC/Al/CuOcompositeswith50and75% ofnanothermiteinweight(Fig.5c-d)exhibitthreemajor exother-mic peaks:oneatlowtemperature(<270°C)correspondingtothe decomposition of CuC and two at high temperatures (∼575 and 780°C) corresponding to the two-steps redox reaction between Al and CuO with an onset at ∼530°C [34]. Several endothermic peaks are also observed corresponding to the evaporation ofthe physically absorbed water molecules (∼180°C), the desorption of NH3 molecules(∼220°C)andthemeltingofCuC(∼220°C)andAl (∼660°C).

When the quantity of nanothermite is greater than 75% (Fig.s 3e-g), theseexothermic peakscorresponding to Al/CuOredox re-actions remain prominent and all located at around 575 and 780°C. However, the onset decomposition temperature of CuC is shifted toward lower temperatures: 190°C for CuC/Al/CuO_82 against 260°C for pure CuC(see detailed valuesin Table 4). This effect againcould be associatedwith the addednanothermite in the systemhavinga higherthermalconductivitycompareto CuC (TableS1),thusimprovingtheheattransferefficiencyofCuC.

A very small exothermic peak located ataround 450°C is de-tectedforallCuC/Al/CuOcompositeswithdifferentintensities(see arrowsinsupportinginformationFig.S12inclose-upviews).Since there is no more copper complex left at this temperaturebased

Fig. 6. XRD analysis results of residues from DSC measurements

Table 4

Summary of the experimental T onset and 1H for each reaction for pure CuCCs, Al/CuC., Al/CuO, and Al/CuO/CuC. energetic composites with different

Al/CuO vs Al/CuC. mass ratios. The data are all normalized by the mass of the sample mass and have a deviation of ± 5 ᵒ C. Negative values are exothermic processes.

Samples

1: Decomposition of CuC Pre-reaction 2: Al/CuO reaction

T onset ( °C) 1H (J/g) 1H th (J/g) 1H th - 1H (J/g) T onset ( °C) T onset ( °C) 1H (J/g) 1H th (J/g) 1H th - 1H (J/g)

CuC 257 −795 — — — — — — — Al/CuO — — — — — 530 −4071 4072 a _— Al/CuC 265 −1118 — — — 534 −1453 — — CuC/Al/CuO_50 257 −253 −358 105 443 531 −2483 −2763 280 CuC/Al/CuO_75 240 −57 −179 122 451 527 −3196 −3417 222 CuC/Al/CuO_82 192 −16 −129 113 456 532 −2540 −3601 1061 CuC/Al/CuO_90 195 −28 −72 44 445 533 −2390 −3810 1422 CuC/Al/CuO_95 200 −15 −36 21 445 532 −2777 −3941 1164

a theoretical H from a previous publication _{[30] .}

ontheTGA/DSC/MSresultsofcoppercomplexandAl/CuC,andno such peak is observedin Al/CuOorAl/CuC, thisexothermic peak islikelytobeanoutcomefromapre-reactionbetweenaluminum andcopperoxides(thatincludenotonlycommercialCuObutalso CuOoriginatedfromCuC).

All information derived from DSC analysisare summarizedin

Table 4withthe correspondingheats ofreaction (

1H).

Sincethe peakcorrespondingtothedecompositionofCuC(270°CinFig.3a) isshiftedtolowertemperaturesinDSCcurvesofotherCuC/Al/CuO composites,thetotalheatofreactionforthedecompositionofCuC is calculatedbyintegratingthecorresponding exothermicpeak.A temperaturerangefrom400to900°Cisusedforthecalculationof theheatofreactionofAl/CuO. Thedataareall normalizedbythe samplemass.

First, the decompositionstepof CuCremains exothermic with orwithoutthepresenceofnanothermitesinthesystem.However, the intensityofthisexothermicpeak, H, decreasesdramatically withincreasingweightpercentageofnanothermiteinthesystem. Indeed, it is 250 J/g and only 15 J/g for 50% and 95% of Al/CuO mixedwithAl/CuC,respectively.Thisisinaccordancewith calcu-lations showninTable4,wherethecorresponding calculatedH (Hth)ofthedecompositionofCuCisdecreasingwiththe increas-ing ofthe incorporatednanothermite dueto thedecreaseof CuC weightpercentageinsidecomposites.

Second, we note that all experimental H values are much lowerthanthecalculatedones(_Hth).Basedonthecalculation re-sultsonheatofreactionofAl/CuOofallcompositeslistedinTable

4,CuC/Al/CuO_95 was expectedto be theone that generates the highestenergyoutput.However, themaximumheatofreactionis experimentallyreachedwhen75%ofAl/CuOnanothermiteis incor-poratedintotheAl/CuCsystem.ItisalsonotedtheCuC/Al/CuO_75 features thesmallestdifference betweenexperimental and calcu-latedH.Wesuggesttwopossiblereasons:

(1) Bothaluminumandcopperoxidebeingmuchbetterheat con-ductors than CuC, one reason wouldbe simply that the heat produced by the CuC decomposition is quickly absorbed and evacuatedbythenanothermitepresentinthesystem.The ther-mal conductivity of the composite improves with increasing nanothermiteloading(SeeTableS1fordetailedinformation re-gardingthermalpropertiesofeachcomposites).

(2) The second reaction (Al/CuO) can be simply the incomplete-ness of the reactions. Considering CuC has already fully de-composed intoCuO at400 ᵒC, itseffectcan be excludedhere. Since all experiments were conducted under the same con-ditions, human/machine errors should be theoretically identi-cal. Thus, the divergence between calculated and experimen-tal results could be resulted from the mixture itself at this temperature range (>400 _ᵒC). Therefore, residues from DSC measurement were collected and examined by XRD method and the results are presented in Fig. 6. All data is summa-rized in supporting information Table S3. The residues col-lected from Al/CuC and CuC/Al/CuO composites, after being heatedtosuchtemperature,containamixtureofcoppermetal (Cu), cuprous oxide (Cu2O), and/or copper oxide, which

in-dicates an incomplete redox reaction between aluminum and copperoxide originatedfromthethermite anddecomposition ofCuC. By analyzing each curve inFig. 6,the XRD resultsdo support this theory based on the following evidence. Firstly, both CuC/Al/CuO_75 and CuC/Al/CuO_50 have five XRD peaks corresponding to the presence of copper metal (labeled as spade),threemorethantheotherCuC/Al/CuOcomposites.And the peakintensities of CuC/Al/CuO_75 andCuC/Al/CuO_50 are muchstronger, indicatingmorecopper productsandtherefore higherreactioncompletion. Secondly,Cu2Osignalswere found inallCuC/Al/CuOcomposites,indicatingincompleteAl/CuO re-actions. Thirdly,no signal relatedto CuO was detected in the residuesofCuC/Al/CuO_75andCuC/Al/CuO_50,however,itwas foundin allthe other CuC/Al/CuOresidues impliesserious in-completions. Hence, it is reasonable to say that the degree ofcompletion ofAl/CuOredox reactionin CuC/Al/CuO_75and CuC/Al/CuO_50ismuchhigherthantheotherCuC/Al/CuO com-posites.Due to the incompletion, all experimental Hvalues are lower than the corresponding Hth. And a much lower H value was experimentally obtained for CuC/Al/CuO_82, CuC/Al/CuO_90andCuC/Al/CuO_95becauseoftheir significant low degreesofcompletion. Inaccordancewiththe theoretical results,CuC/Al/CuO_75 showsa higherexperimental H than CuC/Al/CuO_50whilesufferingthesameamountofH devia-tion.Moreover, CuC/Al/CuO_75 does outperform all theothers in gas generation and burning rate experiments according to

section3.2and3.3.

Finally, mass spectrometry analyses of all prepared energetic composites upon heating(with aheating rateof10°C.min−1) un-der argonatmosphererevealthatmasslosses(startingat∼250°C and finishing before300°C) are accompanied withthe releaseof differentspecies(N2,N2O, H2OandNH3 gases).AsshowninFig.

7(MSspectraforCuC,Al/CuCandCuC/Al/CuO_50atambient tem-peratureandseveraltemperaturesaround100,220,240,260,270, 280 and300°C) thetype ofgaseous speciesare the same what-everthequantityofAl/CuOincorporatedintotheCuCmaterial.In accordance withthe TGA/DSCresults, theseenergetic composites materials begin to releaseN2Oand N2 at around 250°C and fin-ishbefore300°C.Forallthreecases,watermoleculesaredetected from the beginning to the end, indicating the presence ofwater moleculesinsidethemassspectrometerchamber.Duetothis rea-son, it is difficult to differentiate the source of the H2O signals. Looking atintensitiesofNH3 in allthreegraphs, acommonpath wasnoticedthatNH3reachesitspeakreleaseatatemperature be-tween220-250ᵒC,earlierthanthepeakreleaseofN2OandN2, in-dicatingdesorptionofNH3happenspriortothefulldecomposition ofCuCthatwasverifiedpreviouslybyDSCresults(Fig.5).Thefact that NH3 peaks are shownin all temperature rangeimplies that therearelotsofNH3moleculesattachedon/inthecomposite ma-terials(probablyonCuC)consideringexcessammoniumwasused during the synthesisof CuC.In fact,the signal intensitiesofNH3 andH2Oarestrongerthantheothertwospeciesaccordingtothe MSspectraofCuCshowninsupportinginformationFig.S5.

Altogether theseDSC+MS resultsexperimentally demonstrate that the presence of CuCdoes not affect thechemical nanother-mitereactionalthoughitsignificantlyinfluencestheburningofthe composite.WededucethatatlowheatingratestheCuC decompo-sitionandheatfromthermitereactionworksequentiallybasedon a3sequentialreactionschemistrysummarizedbelow:

Peakat270°C:

2Cu

(

NH3

)

4

(

NO3

)

2

(

s

)

→2CuO

(

s

)

+9H2O

(

g

)

+2NH3

(

g

)

+4N2

(

g

)

+N2O

(

g

)

(2)

Peakat575°C:

2Al

(

s

)

+6CuO

(

s

)

→Al2O3

(

s

)

+3Cu2O

(

s

)

(3)

Fig. 7. Mass spectra of CuC, Al/CuC and CuC/Al/CuO_50 energetic composites at var- ious temperatures in argon environment.

Peakat780°C:

2Al

(

s

)

+3Cu2O

(

s

)

→Al2O3

(

s

)

+6Cu

(

s

)

(4)

Different from DSC experiment where event lasts more than hours, in closebombandburn rateexperiments,thecombustion eventfinishesonatimescaleofmillisecondsandoccursatavery high heating rate [35]. Therefore, combinewith the results from closedbombandburnrateexperiments,herewesuggestthatata high heatingrate, CuC/Al/CuO composites likelyundergoa differ-ent sequentialreaction: (1) Al/CuOigniteslocallyandpropagates; (2)heatgeneratedbyAl/CuOreactionpromotesthedecomposition of CuC; (3) CuO product fromCuC decomposition either partici-pates in Al/CuOreaction or is splashed out by gaseous products beforecontactingwithAl.

4. Conclusions

In this work, for the first time, a coordination complex Cu(NH3)4(NO3)2,was mixedwitha thermiteto build anovel gas generator systemwith versatile andexquisite properties in term of gas generation and associated rising time. The decomposition temperature ofCuCwas locatedataround 270°Caccording toits DSC result. NH3, N2O and N2 were identified to be the gaseous products fromitsdecomposition.Sincewithoutexternalheat,CuC cannotinitiateandsustainitsdecomposition,Al/CuOnanothermite was incorporatedto act as thecomplementary heat source upon initiationofthewholematerial.Theoptimizedweightpercentage ofAl/CuOnanothermiteallowingtheself-sustaineddecomposition of CuC,faster burnrateandmaximumboost onpeak pressureis 75%. Thesenewenergeticcomposites featurepeak pressureof12 MPa/g.cm3_{,whichismuchhigherthat traditionalAl/CuO} nanoth-ermite. Ithas tobe notedthat theincorporation ofnanothermite into thecomplexlowerthe decompositiontemperatureofCuCto 190°C instead of 260°C. We also experimentally proved that by simply tuningtheweightpercentageofincorporatedAl/CuO nan-othermite, different heat ofreactions, pressure outputs andburn ratescanbeobtained.Analysisofresiduesshowedthatthe major-ityofthecombustionproductsisinthegasphaseconsideringonly athinlayerofsoftgreyproductleftafter.

Authorcontributions

T. Wu preparedallsamplesandperformedallcharacterization measurements.F.SevelyassistedT.Wuinexperimentalset-up de-signsandclosebombandopenburningexperiments.C.Tenailleau performed XRD measurements. B. Julien assisted T. Wu in close bomb experiment. J. Cure assisted T. Wu in XRD andMS analy-sis. C.Rossi providedsupport forthemanuscriptpreparation and supervisedtheresearch.

The manuscript was written through contributions of all au-thors. Allauthors havegiven approvalto the final version ofthe manuscript.

FundingSources

C. Rossi received fundingfromthe European ResearchCouncil (ERC) undertheEuropean Union’sHorizon 2020research and in-novationprogramme(grantagreementNo832889-PyroSafe).

Competingfinancialinterests

Theauthorsdeclarenocompetingfinancialinterests.

Acknowledgment

The authors acknowledge support from the European Re-search Council(H2020ExcellentScience)ResearcherAward (grant

832889– PyroSafe)andtheOccitanieRegion/EuropeanUnionfor theirFEDERsupport(THERMIEgrant).Theauthorsalsothank Mo-hammedAizanefromCIRIMATforhisassistanceinTGA/MS analy-sisaswellasDr.Bruno Chaudret,fromLPCNO,forhisinitialidea toexploreWernerComplexasenergeticmaterialin2008[28].

References

[1] R.N. Khouzam , S. Al-Mawed , V. Farah , A. Mizeracki , Next-generation airbags and the possibility of negative outcomes due to thoracic injury, Can. J. Cardiol. 30 (2014) 396–404 .

[2] A . Shaout , C.A . Mallon , Automotive airbag technology: past, present and future, Int. J. Comput. Appl. T. 13 (20 0 0) 159–171 .

[3] Air Bags. https://www.nhtsa.gov/equipment/air-bags .

[4] L. Shamo, This personal airbag could help protect the elderly from hip in- juries. https://www.businessinsider.com/hipair- inflates- prevents- hip- injuries- elderly- wearable- personal- airbag- 2018- 1?IR=T .

[5] Q. Zhang , H.Q. Li , Y.K. Ning , D. Liang , G.R. Zhao , Design and realization of a wearable hip-airbag system for fall protection, Appl. Mech. Mater. 461 (2014) 667–674 .

[6] https://www.abs- airbag.com/fr/systeme- abs.html .

[7] G. Shi , C.S. Chan , W.J. Li , K. Leung , Y. Zou , Y. Jin , Mobile human airbag sys- tem for fall protection using mems sensors and embedded SVM classifier, IEEE Sens. J. 9 (2009) 495–503 .

[8] C.S. Chan , G. Shi , Y. Luo , G. Zhang , W.J. Li , P.H.W. Leong , K. Leung , A hu- man-airbag system for hip protection using mems motion sensors: experimen- tal feasibility results, International Conference on Mechatronics and Automa- tion (2006), pp. 831–836. 25-28 June .

[9] R. Nayak , R. Padhye , K. Sinnappoo , L. Arnold , B.K. Behera , Airbags, Textile Progress, 45, 2013, pp. 209–301 .

[10] H. Potvin , M.H. Back , A study of the decomposition of sodium azide using dif- ferential thermal analysis, Can. J. Chem. 51 (1973) 183–186 .

[11]H.S. Kim , J.H. Kim , J.H. Ku , M.H. Cho , J.K. Cha , J.M. Kim , H.W. Lee , S.H. Kim , Effects of metal oxide nanoparticles on combustion and gas-generating perfor- mance of NaN 3 /Al composite powders ignited using a microhotplate platform,

Adv. Powder Technol (2020) 1023–1031 .

[12] A. Surendran Lathika , S. Suthangathan Paramashivan , B. Karuppudaiyar Ra- masamy , S. Mahadevan , Impact of fuel/oxidizer ratio of NaN 3 and KNO 3 airbag

gas generants on toxic emission and performance, Process Saf. Environ. 133 (2020) 348–357 .

[13] K.R. Olson , I.B. Anderson , N.L. Benowitz , P.D. Blanc , R.F. Clark , T.E. Kearney , S.Y. Kim-Katz , A.H.B. Wu , Poisoning & drug overdose, seventh edition, in: B. Belval, C. Naglieri (Eds.), Poisoning & drug overdose, The McGraw-Hill Com- panies Inc, the United States of America, 2018 .

[14] S. Zeuner , A. Hofmann , R. Schropp , K.-H. Rödig , Guanidine-thermite igniter composition for use in gas generators, TRW Airbag Systems GmbH (20 0 0) US6599380B2 .

[15] P.S. Khandhadia, S.P. Burns, G.K. Williams, High gas yield non-azide gas gener- ants, 1998, US6210505.

[16] U.S. EPA., Provisional peer-reviewed toxicity values for guanidine chloride, U.S. Environmental Protection Agency, Washington, DC, 2014 .

[17] H. Tabuchi, Airbag Compound Has Vexed Takata for Years. https://www. nytimes.com/2014/12/10/business/compound- in- takata- airbags- is- inquirys- focus.html .

[18] A. Lafrance, How Airbags Are Supposed to Work. https://www.theatlantic.com/ technology/archive/2015/06/airbags/396854/ .

[19] T.M. Southern , W.W. Wendlandt , The thermal decomposition of metal com- plexes—XX: Some amine copper(II) nitrate complexes, J. Inorg. Nucl. Ch. 32 (1970) 3783–3792 .

[20]S. Mathew , C.G.R. Nair , K.N. Ninan , Thermal decomposition studies on amine complexes of copper(II) nitrate in solid state, B. Chem. Soc. Jpn. 64 (1991) 3207–3209 .

[21]M. Künzel , J. Selesovsky , R. Matyáš, Characterization of tetraamminecopper salts, 18th New Trends in Research of Energetic Materials, Pardubice, Czech Republic (2015), pp. 664–669 .

[22]M. Künzel , O. Vodochodský, R. Matyáš, Z. Jalový, J. Pachman , J. Maixner , Tetraamminecopper(II) nitrate and its effects on ammonium nitrate(V) central european, J. Energ. Mater. 14 (2017) 169–183 .

[23]M. Liszka-Skoczylas , E. Mikuli , J. Szklarzewicz , J. Hetma ´nczyk , Thermal be- haviour, phase transition and molecular motions in [Co(NH 3 ) 6 ](NO 3 ) 2 , Ther-

mochim. Acta 496 (2009) 38–44 .

[24]W.W. Wendlandt , J.P. Smith , Thermal decomposition of metal complexes—VI some ammine cobalt (II) complexes, J. Inorg. Nucl. Ch. 25 (1963) 985–993 . [25]A. Migdał-Mikuli , E. Mikuli , R. Dziembaj , D. Majda , Ł. Hetma ´nczyk , Thermal

decomposition of [Mg(NH 3 ) 6 ](NO 3 ) 2 , [Ni(NH 3 ) 6 ](NO 3 ) 2 and [Ni(ND 3 ) 6 ](NO 3 ) 2 ,

[26]J.M. Janik , E. Mikuli , A. Migdal-Mikuli , M. Rachwalska , T. Stanek , J.A. Janik , K. Otnes , I. Svare , Calorimetry, proton magnetic resonance and neutron quasielastic scattering studies of [Mg(H 2 O) 6 ](NO 3 ) 2 , Phys. Scripta 28 (1983)

569–572 .

[27]E. Mikuli , M. Liszka , M. Molenda , Thermal decomposition of [Cd(NH 3 ) 6 ](NO 3 ) 2 ,

J. Therm. Anal. Calorim. 89 (2007) 573–578 .

[28]C. Pradère , S. Suhard , L. Vendier , G. Jacob , B. Chaudret , S. Sabo-Etienne , Het- erometallic werner complexes as energetic materials, Dalton. T. 20 (2008) 2725–2731 .

[29] G.K. Lund, J.C. Hinshaw, D.W. Doll, R.J. Blau, Metal complexes for use as gas generants 1997, WO1998006486A2.

[30]L. Glavier , G. Taton , J.-M. Ducéré, V. Baijot , S. Pinon , T. Calais , A. Estève , M. Dja- fari Rouhani , C. Rossi , Nanoenergetics as pressure generator for nontoxic im- pact primers: Comparison of Al/Bi 2 O 3 , Al/CuO, Al/MoO 3 nanothermites and

Al/PTFE, Combust. Flame 162 (2015) 1813–1820 .

[31] A. Nicollet , L. Salvagnac , V. Baijot , A. Estève , C. Rossi , Fast circuit breaker based on integration of Al/CuO nanothermites, Sensor Actuat. A-Phys. 273 (2018) 249–255 .

[32] C. Rossi , Engineering of Al/CuO reactive multilayer thin films for tunable initi- ation and actuation, propellants, explosives, Pyrot. 44 (2019) 94–108 . [33]B. Julien , J. Cure , L. Salvagnac , C. Josse , A. Esteve , C. Rossi , Integration of gold

nanoparticles to modulate the ignitability of nanothermite films, ACS Appl. Mater. 3 (2020) 2562–2572 .

[34]I. Abdallah , J. Zapata , G. Lahiner , B. Warot-Fonrose , J. Cure , Y. Chabal , A. Es- teve , C. Rossi , Structure and chemical characterization at the atomic level of reactions in Al/CuO multilayers, ACS Appl. Ener. Mater. 1 (2018) 1762–1770 . [35]T. Wu , X. Wang , P.Y. Zavalij , J.B. Delisio , H. Wang , M.R. Zachariah , Performance

of iodine oxides/iodic acids as oxidizers in thermite systems, Combust. Flame 191 (2018) 335–342 .