DTIC ADA553217: Streak Spectrograph Temperature Analysis from Electrically Exploded Ni/Al Nanolaminates

0.65 MB·English

by  Defense Technical Information Center

TSF-29650;NoofPages6
ThinSolidFilmsxxx(2011)xxx–xxx
ContentslistsavailableatScienceDirect
Thin Solid Films
journal homepage: www.elsevier.com/locate/tsf
Streak spectrograph temperature analysis from electrically exploded
Ni/Al nanolaminates
Christopher J. Morris a,⁎, Paul Wilkins b, Chadd May b, Eugene Zakara, Timothy P. Weihs c
aU.S.ArmyResearchLaboratory,2800PowderMillRd,Adelphi,MD20783,UnitedStates
bLawrenceLivermoreNationalLaboratory,7000EastAvenue,Livermore,CA94550,UnitedStates
cDepartmentofMaterialsScienceandEngineering,TheJohnsHopkinsUniversity,3400NorthCharlesStreet,Baltimore,MD21218,UnitedStates
a r t i c l e i n f o a b s t r a c t
Availableonlinexxxx Wereportonelectrically-inducedheatingandmixingofmultilayerednickel/aluminum(Ni/Al)laminates
observedbystreakcameraemissionspectroscopy.Pastexperimentsprobingthekineticenergyofmaterial
Keywords: ejected from the reaction zone indicate that additional kinetic energy originates from Ni/Al samples,
Exothermicmixing presumably fromexothermicmixingbetweenthetwometals.Hereweexaminestreak spectrographs of
Reactivelaminates similarexperimentstodeterminethepresenceofexpectedelementsandtheirtemperatures.Weconducted
Streakcameraspectroscopy
theseexperimentsinroughvacuum,butfoundtheemissiontobedominatedbyargon(Ar)andnitrogen(N)
Plasmatemperatures
lines in addition to the expected emission of Al and Ni, which were also present. Using the spectral
informationofAr,weanalyzedtherelativeintensitiesoffourArpeaksbetween425and455nm,withrespect
totheirexpectedBoltzmanndistributionstoyieldtemperaturesasafunctionoftime.Thesetemperatures
were2.24–2.59eVforAlsamples,and2.93–3.27eVforbothNiandNi/Alsamples,andwerewithinestimates
basedonthemeasuredelectricalenergydeliveredtoeachdevice.ThehigherNi/Alsampletemperatures
seemedtovalidateourpastmeasurementsofincreasedkineticenergyandtheapparentrapidexothermic
mixingbetweenNiandAl,althoughNisamplesyieldedsurprisinglyhightemperaturesaswell.Theseresults
maybeimportantforfuturenanomanufacturingtechniquesinvolvinglocalizedheatingfromreactiveNi/Al
multilayers,wheretheprecisecontrolofspatialtemperaturesmaynecessitateanequallyprecisetemporal
controlofthereaction.
PublishedbyElsevierB.V.
1.Introduction and one way to achieve this control is through more rapid heating
rates.
Thenickel–aluminum(Ni/Al)intermetallicsystemisusefulfora We recently reported on experiments where we electrically
varietyofreactivematerialapplications[1],includingtheinitiationof heated microscale Ni/Al slabs at 1011–1012K/s [7]. In these exper-
subsequentreactions,thermalbatteries[2],andthelocalizedheating iments,electricalcurrentpassedthrough175μmsquare,2.8μmthick
required by many welding and joining applications [3–5]. Reaction structuresofmultilayeredNi/Al,whichwerecoveredwithapolymer
characteristics are well studied at the normal self-heating rates of “flyer” layer. The structures vaporized and rapidly accelerated the
103–106K/s, where the lateral reaction propagation of the Ni/Al flyerlayertoseveralkm/swithin50ns,throughexpansionofgaseous
system is driven by heat diffusion along and between the reactive products.WefoundthatNi/Alconductorlayerscontributedbetween
layers,leadingtopropagationratesashighas10m/sinthecaseof 1.13kJ/gand2.26kJ/gofadditionalkineticenergytoflyervelocities
multilayered Ni/Al [1], or as high as 100m/s for other reactive ascomparedwithconductorscomposedonlyofCu,evaluatedatthe
multilayered systems [6]. At these heating rates and for typical same instantaneous input electrical energy levels. The similarity
metallic or semiconducting material thermal diffusivities, the tem- betweenthisadditionalkineticenergyandthe1.2kJ/gavailablefrom
perature rise extends roughly 10μm into any surrounding material exothermic mixingof Ni/Al [8] suggestedthat the expected mixing
every microsecond. Future nanomanufacturing techniques may contributed directly to increased flyer velocity. We more recently
require more precise control of spatial temperature distributions, comparedNi/Alflyervelocitiestothosefromdevicescomposedonly
ofAlorNi,andfoundsimilarresults[9],withthekineticenergyfrom
Ni/Al samples exceeding that of either Al or Ni. Thus, although the
kinetics of Ni/Al lateral reaction propagation should be too slow to
⁎ Corresponding author at: U.S. Army Research Laboratory, RDRL-SER-L, 2800 cover any significant distance in these timescales, the electrically
PowderMillRd,Adelphi,MD20783,UnitedStates.Tel.:[phone];fax:+1301
forcedheatingofanentiremultilayerslabappearstoleadtoasimilar
3944562.
E-mailaddress:[email protected](C.J.Morris). exothermicrelease.
0040-6090/$–seefrontmatter.PublishedbyElsevierB.V.
doi:10.1016/j.tsf.2011.07.043
Pleasecitethisarticleas:C.J.Morris,etal.,ThinSolidFilms(2011),doi:10.1016/j.tsf.2011.07.043
Report Documentation Page Form Approved
OMB No. 0704-0188
Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and
maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information,
including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington
VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it

does not display a currently valid OMB control number.
1. REPORT DATE 3. DATES COVERED
2011 2. REPORT TYPE 00-00-2011 to 00-00-2011
4. TITLE AND SUBTITLE 5a. CONTRACT NUMBER
Streak Spectrograph Temperature Analysis From Electrically Exploded
5b. GRANT NUMBER
Ni/Al Nanolaminates
5c. PROGRAM ELEMENT NUMBER
6. AUTHOR(S) 5d. PROJECT NUMBER
5e. TASK NUMBER
5f. WORK UNIT NUMBER
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION
U.S. Army Research Laboratory,2800 Powder Mill REPORT NUMBER
Rd,Adelphi,MD,20783
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/MONITOR’S ACRONYM(S)
11. SPONSOR/MONITOR’S REPORT
NUMBER(S)
12. DISTRIBUTION/AVAILABILITY STATEMENT
Approved for public release; distribution unlimited
13. SUPPLEMENTARY NOTES
Thin Solid Films (2011), Preprint, 6 pages
14. ABSTRACT

15. SUBJECT TERMS
16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF 18. NUMBER 19a. NAME OF
ABSTRACT OF PAGES RESPONSIBLE PERSON
a. REPORT b. ABSTRACT c. THIS PAGE Same as 7
unclassified unclassified unclassified Report (SAR)
Standard Form 298 (Rev. 8-98)
Prescribed by ANSI Std Z39-18
2 C.J.Morrisetal./ThinSolidFilmsxxx(2011)xxx–xxx
HerewereportonelectricallyforcedheatingofNi/Alasobserved wire device into the circuit, the high voltage capacitive discharge
by streak camera emission spectroscopy. These experiments were resulted in a current pulse corresponding to a half cycle at 2.35
performedwithnon-reactiveNi,Al,andmultilayeredreactiveNi/Al ±0.02MHz, calculated from pulse widths averaging 213ns. At this
samples,eachwithoutthepolymerflyerlayerusedinpreviouslycited characteristic frequency, calculated skin depths were more than 40
results.Withouttheflyerlayer,anddespitethefactthatexperiments timesthethicknessesofconductorsusedinthisstudy,soweassumed
wereperformedundervacuum,emissionwasdominatedbyresidual skineffectstobenegligible.
argon(Ar)and nitrogen(N).Thus,theemissionwasstimulated by We expected bridge resistance to be a strong function of
electrical arcing through both locally vaporized metal and residual instantaneous temperature induced by electrical current, so we
gasses. Still, the relative intensities of Ar peaks and their expected monitoredvoltagedropusingprobessolderedtostriplinesnearthe
Boltzmanndistributionfunctionsyieldedtemperaturevaluesforeach pointofconnectiontoeachdevice.Ingeneral,themeasuredvoltage
sample as a function of time. The following section describes our dropV wouldbeacombinationofactualbridgevoltagedropV and
m b
fabrication, experimental, and temperature analysis methods in animpedancetermresultingfromanyinductanceinthebridgeL ,
b
detail.
di
V =V +L ð2Þ
2.Methods m b bdt
WefabricateddevicesasdepictedinFig.1a,withtheconductor Thisimpedancetermwouldcausethemeasuredvoltagetoartificially
layerformedofNi,Al,ormultilayeredNi/Al.Thesubstrateineachcase increaseattheonsetofcurrentriseandthenleveloff,leadingtoa
wasquartz.FortheAlsamples,wesputtered20nmoftitanium(Ti) calculatedbridgeresistancethatwouldstartartificiallyhighandthen
followedby 2.7μm of Al.The Ni samplescontained a single 2.7μm decrease with rising current. Because we only observed resistances
layer of sputtered Ni/V (93wt.% Ni, 7wt.% V). The Ni/Al samples which steadily increased from the start of each current pulse, as
containeda20nmTiadhesionlayer,andfiveiterationsofsputtered expected from temperature rise induced by increasing current, we
Ni/V(200nm)followedbyAl(~300nm),finishingwithasixthlayer assumednegligiblebridgeinductancesuchthatVm=Vb.
of Ni/V (~200nm). The total Ni/Al stack was 2.64μm in thickness. The streak spectrograph setup is shown in Fig. 2, where a
Sputteringpowerwasapproximately6.6and3.1W/cm2forNi/Vand plexiglass vacuum chamber held the sample and provided connec-
Al,respectively.Thebasepressurepriortosputteringwaslessthan tionstothefiringcircuitdescribedabove.Thechamberwaspumped
1×10−6Torr,andthesputtergaswasAratapproximately15mTorr. to approximately 10mTorr before each test. An 85mm, f1.4 lens
Weperformedlithographicpatterningusinga2-μm-thickpositive focused the light into a 400μm diameter optical fiber, and another
photoresistlayertodefinethe175μmsquarebridgeregions,shown 105mmf1.8lenscollimatedthelighttowardsa1200mm−1grating.
inFig.1b.AwetetchremovedallexposedconductorlayersexceptTi A second 105mm f1.8 lens focused the light into a streak camera
(Aluminum Etchant Type A, Transene Company, Inc., Danvers, (EG&G model L-CA-24) which was connected to a 16-bit mono-
Massachusetts). An HF solution diluted 50:1 with deionized water chrome, 4096×4096pixel CCD camera (Spectral Instruments, 800-
etchedtheremainingTiforthosesampletypescontainingthislayer. 175).
Finally,aresiststripperremovedthephotoresist(PRS-3000,JTBaker Usingthe1200mm−1grating,wecapturedseveralstreakimages
Inc.,Phillipsburg,NewJersey),whilekaptontapemaskedindividual from 405, 532, and 655nm laser sources, a comb generator, and a
devicecontactpads.Wedicedthewafersintoindividualsamplesand mercuryarclamp,allofwhichallowedustocalibratetheoutputof
removedthekaptontapemaskingtoexposedevicecontacts. thegratingandtemporalresolutionofthestreakcamera,estimatethe
Theexperimentinvolvedaseriescircuitconsistingofabridgewire opticalresolution,andcorrectforsomeopticalpathinefficiencies.The
device,a0.12μFcapacitorchargedto1200V,aswitch,a5mΩcurrent usablewavelengthrangewasapproximately140nmat22.05pixels/
viewing resistor (2ns rise time), and low inductance strip line nm, and the temporal range was 350ns at 10.74pixels/ns. By
connecting all components. Without the bridge wire device, the centering the grating at 450nm and 532nm over two separate
responseofcurrentiinthecircuitwas experiments, we could span the wavelength range of 380nm to
600nm.BasedonaGaussianfittoseverallaserlines,weestimated
d2i Rdi 1 theopticalresolutionofthissetuptobe2.0nm.Finally,althoughwe
+ + i=0 ð1Þ
dt2 Ldt LC didnothavereadyaccesstoacalibratedbackgroundlightsourceto
correctfordifferencesindetectorsensitivitytodifferentwavelengths,
whereL and R representthetotal seriesinductanceand resistance,
respectively.Byfittingtheresponseofthecircuittothesolutionof
400µm
Eq.(1),wedeterminedinductanceandresistancevaluesof3.7nHand
optical
38mΩ,respectively,correspondingtoanundampednaturalfrequen-
fiber Grating
cy of 7.6MHz. With added damping from the insertion of a bridge
Lines
depict
light path
To
high-V
firing
Vacuum Streak
circuit
Sample chamber Camera
CCD
Fig.2.Streakspectrographtestsetup.Vacuumchamberwaspumpedtoapproximately
Fig.1.Fabricationprocessflowandmicroscopeimageofcompleteddevice. 10mTorr.
Pleasecitethisarticleas:C.J.Morris,etal.,ThinSolidFilms(2011),doi:10.1016/j.tsf.2011.07.043
C.J.Morrisetal./ThinSolidFilmsxxx(2011)xxx–xxx 3
wedidsubtracta darkcurrentbackgroundfromeachimagewhich 3.Resultsanddiscussion
accountedforabout0.2%ofeachimagepixel'sdynamicrange.Wealso
corrected for optical path inefficiencies along the time axis using a 3.1.Spectroscopicresults
radial dual-Gaussian correction function that forced streak images
from continuous laser sources to be constant over time. We then Fig.3showsatypicalstreakspectrograph,obtainedfromaNi/Al
performed this same correction to all subsequent captured streak laminate using the 600mm−1 grating. The initial time of zero was
images. defined as the point at which the average intensity across all
Inadditiontothedatagatheredwiththe1200mm−1grating,we wavelengthswashalfitsmaximum,thereforecorrespondingclosely
acquiredapartialsetofdatausinga600mm−1grating,centeredat tothetimeofburstandmaximumintensity.Similartorelatedstudies
532nm.Thissetupallowedawavelengthrangewitheachexperiment of other energetic materials over longer timescales [14], a typical
ofapproximately380to680nm,withanopticalresolutionof3.1nm. captured image featured an initial period of nearly broadband
To investigate the apparent temperature values associated with emission, with superimposed spectral peaks which quickly grew in
bridgebursts,wecomparedtherelativeintensitiesofvariouspeaks, intensity. Such broadband emission has been observed for other
each at wavelength λ, with the expected Boltzmann distribution reactive,Al-basednanothermitematerials[15],aswellaswithlaser
equation. This equation yields a linear relationship between the ablation of Al/Teflon films [13]. The multilayered Ni/Al samples
logarithmofthepeakintensityIandtheenergyE oftheexcitedstate generally exhibited an overall brightness that was higher and
i
associatedwiththepeak[10], persistedlongerthaneitherAlorNisamples.
Figs. 4, 5, and 6 show streak spectrographs of Al, Ni, and Al–Ni
(cid:1) (cid:3) bridgesat10,50,and250nsfollowingtheonsetofemission,orburst,
ln Iλ =−Ei +C ð3Þ utilizingthe1200mm−1grating.Althoughperformedundervacuum,
gkAki kT the similar emission in each case suggests elements from residual
gassesinthevacuumchamber.Weattributedmanyofthecommon
peakstoArandnitrogen(N),whichlikelycamefromthesurrounding
atmosphere at 10mTorr, a common pressure range at which Ar
InEq.(3),gkisthepartitioncoefficientassociatedwiththeexcited plasmasareusedinmicrofabricationprocesses[16].GiventhatArwas
level, Aki is the transition probability, the non-subscript k is usedduringthesputterdepositionprocess,itwasalsopossibleforAr
Boltzmann'sconstant,andCisaconstantassociatedwiththeground to originate from the films themselves, which for Ni films at the
state. The slope of this linear relationship is equal to −1/kT. This sputteringconditionsusedcanbeontheorderof0.3%byatomicratio
method relies on the assumption that emission intensities are not [17].
affectedbyself-absorption,whichmayormaynotapplyduringthe Emission peaks indicating Al and possibly Ni were also present.
earlystagesofarapidlyexpandingcloudofvaporizedmaterial.The Specifically, Fig. 4 shows the prominent doublet at 394.4/396.1nm
method also relies on intensities being proportional to actual line indicating vapor phase atomic Al, which was not present in the Ni
radiance,andthereforegenerallyrequiresdetectorcalibrationacross samples as expected. Also labeled are Ar-I and Ar-II lines at 407.2,
therangeofwavelengthsused. 413.2, 415.9, 420.1, 427.8 and 433.1nm. The broad doublet near
WeselectedfourArlinesshowninTable1fortheBoltzmannplots, 450nmwasactuallyacombinationofseveralpossiblelines:N-IIat
representing a 30nm wavelength range near the center of each 444.7nm, Ni-I at 447.0nm (for Ni and Ni/Al samples only), Ar-I at
spectrum.Weconsideredthisrangetobesufficientlysmalltowaive 447.5 and 451.1nm, Ar-II at 448.2nm, and N-III at 451.1nm. The
therequirementfordetectorcalibrationacrossallwavelengths.The clusterofpeaksjustpast460nmwaslikelyfromAr-IIat459and461,
partition and transition parameters in Table 1 are from [11], and and at least five N-I and N-IV peaks were between 460.1 and
energyvaluesarefrom[12].Usingthesevaluesandintensitiesfrom 462.4nm.Lastly,Ar-IIat498nmisalsolabeled.Fig.5showsthesame
digitizedstreakimages,weusedEq.(3)tocalculatetheslopefroma lines at 50ns following the onset of emission. Some lines became
least squares linear fit, from which we calculated temperature as a more prominent including a clear shift in the Al sample line near
functionoftimeforeachsampletype. 460nm, which we have labeled Al-II at 466.7nm. Fig. 6 at 250ns
If we were to iteratively correct the intensity of each peak for followingtheonsetofemissionshowsthebestevidenceofresidual
blackbodyemissionatthecalculatedtemperature,theactualemission gas signatures, where many of the Ni sample lines clearly coincided
intensity at 425nm would be 79% of that at 455nm, according to withAlsamplelines.TheboxedlabelsinFig.6indicatethefourArlines
Plank's equation evaluated at a temperature of 3eV. Eq. (3) shows
that a factor of0.79 on intensitywouldresultin a −0.23offset on
calculated vertical axis values, which we found to affect calculated
temperaturesbyamaximumof3.9%.Inrelatedworkinvestigatingthe
emissionoflaserablatedAl/Teflonfilms,time-resolvedthroughthe
use of a streak camera, the observed broadband emission was not
attributedtoclassicalblackbodyradiation,becauseitdidnotchange
withincreasinglaserirradianceaswouldbeexpected[13].Therefore,
we did not attribute our emission to that of a classical blackbody
source,anddidnotperformcorrectionsbasedonPlank'sequation.
Table1
SpectroscopicparametersofArlines.
Spectrum Wavelength(nm) Energyofupperlevel(eV) gkAki(108s−1)
Ar-II 427.8 21.35 0.320
Ar-II 433.1 19.61 2.30
Ar-I 451.1 3.028 0.0118 Fig.3.StreakspectrographofaNi/Albridgefiredat1.2kV,viewedwitha600mm−1
Ar-II 454.5 19.87 1.884
grating.
Pleasecitethisarticleas:C.J.Morris,etal.,ThinSolidFilms(2011),doi:10.1016/j.tsf.2011.07.043
4 C.J.Morrisetal./ThinSolidFilmsxxx(2011)xxx–xxx
Fig.4.Ni/Al,Al,andNisampleemissionat10nsfollowingtheonsetofemission,over
Fig.6.Ni/Al,Al,andNisampleemissionat250nsfollowingtheonsetofemission,over
wavelengthsof380to520nm,withmajorpeaksidentified. wavelengthsof380to520nm,withmajorpeaksidentified.Thethreeboxedlabelsrefer
tothefourArlinesusedfortemperatureanalysisbycomparisontointensitiespredicted
bytheBoltzmannequation.
thatarelistedinTable1andareusedfortemperatureanalysisinthe
followingsection.
Fig.7showsemissionat100nsfollowingtheonsetofemissionin 100nsfollowingburst.TheR-squaredvaluewas0.998,andtheslope
thewavelengthrangeof450nmto610nm.Whilemanyofthelines correspondedtoatemperatureof3.07eV,or35,650K.Ingeneral,the
inFigs.4–6wereattributedtoAr,Fig.7providesclearevidencethat linearfitqualitywasquitehigh,withR-squaredvaluesrangingfrom
nitrogenemissionwasalsopresent,becausetheprominent,common 0.964to0.999.Althoughtherewereatleastsixteenothersignificant
peaksnear570nmcoincidedbestwithN-IIlinesat566.7,567.6,568, Arpeaksbetween425and465nm,allexceptthreeofthefourArlines
568.6, and 571.1nm. Therefore, although some Ar could have specified were too close in wavelength to significant Al, Ni, and N
originatedfromthecommondepositionprocessusedtodepositthe lines,andourexperimentalsetupcouldnotresolvethem.Oneofthe
films,residualgassesinthetestchamberwerethemostlikelysource four lines used, 451.1nm, may also have been coincident with a
fortheobservedNemission.LinesfromAl-IIIat569.6and572.3nm moderatelysignificant line fromdoubly ionized nitrogen. However,
werealsopresentintheAlsampleinFig.7,butlesssointheNi/Al weusedthislinewithspectroscopicparametersfromArinorderto
sample.Fig.7alsoprovidesevidenceofvaporphaseNi-Iat508.1nm, include enough energy level variation to yield reasonable approxi-
presentinboththeNiandNi/Alsampletraces. mations of slope and therefore temperature. Recognizing that the
Althoughseverallineswereobservedbetween450and550nm, intensity of this particular line may have been influenced by N
wedidnotattributethemtoaluminumoxide,whichnormallyoccurs emission,Fig.8showsthatevenafactoroftwoinintensitydifference
in processes involving highly-reactive Al and any presence of at the 3eV level would only lead to a difference in slope of
atmospheric oxygen [14,18,19]. The fact that any lines in this approximately12%.Therefore,wesafelyassumedthattheintensity
wavelengthrangewerealsopresentintheemissionfromNisamples ofthislinewasprimarilyfromAremission.
indicates that our timescales were not long enough to allow Fig.9showstemperatureasa function oftimepredictedbythe
convectivemixingbetweenAland atmospheric oxygen.Due tothe BoltzmannequationforAl,Ni,andmultilayeredNi/Alsamples,with
multilayered nature of the Ni/Al samples, however, very little datafromtwosamplesofeachtypeinordertoshowtypicalvariation
convectionwouldberequiredforexothermicmixingtooccur. fromsampletosample.Thetemperatureswerequitehigh,withall
samples starting at 3.1–3.4eV, and then either rising or falling to
3.2.Temperatureanalysis relatively constant values after 150ns. The final temperatures at
300ns were 2.30–2.54eV (26,700 to 29,500K) for Al, and 3.19–
Tocalculatetemperatures,Fig.8showsatypicallinearfitoffourAr 3.33eV for both Ni/Al and surprisingly, Ni. Fig. 10 also shows
line intensities and spectroscopic parameters from Table 1, plotted temperatures calculated from experiments using the 600mm−1
according to Eq. (1). This particular plot is from a Ni/Al sample at grating,wheretheaveragingeffectofloweropticalresolutionledto
Fig.5.Ni/Al,Al,andNisampleemissionat50nsfollowingtheonsetofemission,over Fig.7.Ni/Al,Al,andNisampleemissionat100nsfollowingtheonsetofemission,over
wavelengthsof380to520nm,withmajorpeaksidentified. wavelengthsof450to610nm,withmajorpeaksidentified.
Pleasecitethisarticleas:C.J.Morris,etal.,ThinSolidFilms(2011),doi:10.1016/j.tsf.2011.07.043
C.J.Morrisetal./ThinSolidFilmsxxx(2011)xxx–xxx 5
19 4
18 3.8
17 3.6 Ni
A) 16 3.4
λn( I /gk 1145 T (eV) 23..823
l 13 Ni/Al
2.6
12
2.4
11
2.2 Al
10
0 5 10 15 20 25 2
E(eV) 0 50 100 150 200 250 300
i time (ns)
Fig.8.LinearfittoaplotoffourArlineintensitiesandspectroscopicparametersin
Fig.10.TemperaturevstimepredictedbytheBoltzmannequationforonesampleeach
Table1vsenergyaccordingtoEq.(1),foraNi/Alsampleat100nsfollowingtheonset of Al, Ni, and multilayered Ni/Al, each obtained with a 600mm−1 grating. The
of emission. The R-squared value was 0.998, and the slope corresponded to a
intensitiesofseverallinesfromboththeAlandNi/Alsampleswereunavailableatearly
temperatureof35,650K,or3.07eV.
timesduetodetectorsaturation,sodatacoveringthesetimeintervalsarenotshown.
less scatter in the data and yielded temperatures generally within shows the cumulative energy through each of the three samples
0.3eVofthecalculationsusingthe1200mm−1grating.Theintensity which produced the streak spectrographs in Figs. 4–6. The point of
of emission was approximately four times brighter than with the maximumresistanceforeachsampleoccurredatt=0,corresponding
1200mm−1grating,sothesimilartemperaturecalculationsindicate with the point of maximum voltage and bridge burst [20], and
that any non-linearity in detector response at higher signals was therefore correspondingclosely with t=0 in the spectroscopic and
consistentoverthewavelengthsstudied.Asindicatedinthecaption, temperatureplots.Assumingallenergywentintoheatingthebridges
thehigherintensitiesofseverallinesfrombothAlandNi/Alsamples defined by their original dimensions, 2.78mJ, 5.57mJ, and 4.02mJ
saturatedthedetectoratearlytimes. wererequiredtoatomize(i.e.,heat,melt,andvaporize)theAl,Ni,and
Other lines from which we could have calculated temperatures Ni/Alsamples,respectively.AsindicatedbytheinsetofFig.11,theAl
include those from N, Ni, or Al. However, the prominent N lines andNi/Alsampleswerelikelyatomizedatt=0,andtheNisample
coveredtoonarrowanenergyrangetoyieldgoodslopeapproxima- reacheditsatomizationenergyat12.2ns.
tions.Similarly,thefewNilinesannotatedinthefigureswerevery The cumulative energies into each sample were enough to fully
closeinenergylevels,andexceptfortheNilineannotatedinFig.7, ionizeandgeneratethetemperaturesobserved.TheenergyintotheAl
theywerenotdistinctfromotherArandNlines.TherewereseveralAl samplewasenoughtoionizeitby14.5ns,andthepresenceofionized
lines,buttheonlydistinctoneswereat394.4and396.2nm,whichare Al agrees qualitatively with Figs. 4 and 5 showingthe Al-II peak at
at nearly identical energy levels. Using four other Al lines at 452.9, 466.7nm, presentat 50ns but not at 10ns. The energy into the Al
466.7,569.7,and572.2nm,whichspanenergylevelsbetween13and samplewasalsosufficienttodoublyionizeitat71ns.Fig.6showsan
20eV,alinearfitforAlat250nsyieldedanR-squaredvalueof0.41 Al-IIIpeakat453nm,whileFig.7showsAl-IIIat569.6and572.3nm.
and a temperature of 2.56eV. This temperature was close to the By300ns,approximately72%oftheAlsamplecouldbetriplyionized
temperatures shown in Figs. 9 and 10 at 250ns, and although this with 2.4eV/atom remaining to contribute to the temperature of
levelofsimilaritywasnotobservedforallAlsamplesatalltimes,the 2.4eV calculated in Figs. 9 and 10. The Ni sample reached its first
fact that similar temperatures were obtained using measurements ionizationlevelat50ns,andby300ns,76%ofthesamplecouldbe
and parameters for a different element suggests that temperatures doubly ionized with 3.4eV/atom remaining to correspond to
obtainedthroughanalysisofArspectrallineswereaccurate. calculated temperatures at that time. Finally, the Ni/Al sample
Wealsocalculatedenergyinputtoeachsamplebyintegratingthe reached the first ionization energy of its Al and Ni components at
productofmeasuredcurrentandvoltagethrougheachbridge.Fig.11 8.3nsand23ns,respectively,andthesecondionizationenergyofits
40
4
35 Ni
3.8 Ni
3.6 30 Ni/Al
Ni/Al
3.4 mJ) 25 Al
T (eV) 3.32 Energy ( 1250 3456 atomization
2.8 2
10
1
2.6
0
5
2.4 -15 -7.5 0 7.5 15
2.2 0
Al -50 0 50 100 150 200 250 300
2 time (ns)
0 50 100 150 200 250 300
time (ns) Fig.11.ElectricalenergydissipatedvstimeinAl,Ni,andNi/Albridges,calculatedfrom
integrationofcurrentandvoltagemeasurementsthrougheachsample.Insetshows
Fig.9.TemperaturevstimepredictedbytheBoltzmannequationfortwosampleseach detail of energies dissipated at t=0, corresponding closely to theoretical energies
ofAl,Ni,andmultilayeredNi/Al,eachobtainedwitha1200mm−1grating. requiredforatomizationofeachsampletype.
Pleasecitethisarticleas:C.J.Morris,etal.,ThinSolidFilms(2011),doi:10.1016/j.tsf.2011.07.043
6 C.J.Morrisetal./ThinSolidFilmsxxx(2011)xxx–xxx
Al component by 52.3ns. By 300ns, 85% of the remaining Ni rangeof3.1–3.4eVwithinthefirstfewnanosecondsoftheonsetof
component could be doubly ionized leaving 3.0eV/atom to corre- emission,withAlsamplescoolingtoroughly2.46eVafter300ns,and
spond to calculated temperatures at that time. Of course, the Ni/Al samples remaining around 3.27eV. Nickel samples yielded
prediction of how energy distributed between temperature and temperatureswhichfirstincreased,thendecreasedtothesamelevel
ionized specie proportions is not necessarily valid, but we present as those of Ni/Al samples. Although these temperature levels were
these values to demonstrate that sufficient energy was supplied to high,theywerereasonableforlocalizedplasmatemperaturesgiven
each sample to generate the species observed and temperatures thelevelofelectricalenergysuppliedtoeachdevice.
calculated. Plasma temperatures in the range of 1eV up to several Further testing should aim to limit the interaction of metallic
hundred eV are commonly calculated from electrically exploding sampleemissionandelectricalarcingthroughresidualatmospheric
wiresandothershortdurationplasmadischarges[21–24]. elements. Doing so will better isolate individual Al and Ni spectral
ThehighertemperaturesofNi/AlcomparedwithAlmayprovide lines,andinsurethattemperaturesbetterreflecttherapidlyheated
evidence of additional energy from exothermic mixing of Ni/Al. metallic species. Still, results reported here support the hypothesis
However,calculatedNitemperatureswereactuallyhigherthanthose that Ni/Al exothermic mixing is possible from material which is
of multilayered Ni/Al at early times, and cooled to similar levels at electricallyheatedatextremelyshorttimescales.
latertimes.Thisresultwassurprisinggiventhattheoverallemission
ofNisampleswaslowerthaneitheroftheAlorNi/Alsamples,but
giventheconsistentlyhighcorrelationoflinearfitstotheBoltzman Acknowledgements
equation,thecalculatedeffectivetemperaturesappeartobeaccurate.
Itispossiblethatthehighercalculatedtemperaturewassimplythatof The authors gratefully acknowledge Dr. Kevin McNesby at U.S.
surroundingArplasma,andnottheNiitself,perhapsduetohigherNi ArmyResearchLaboratoryforhelpfuldiscussionsonanalysisofstreak
resistivity causing more current to flow through the surrounding emission spectroscopy. Professor T.P. Weihs acknowledges the
ionizedatmosphere. financial support of the Office Naval Research through Grant No.
Regardless of temperature, the presence of atomic Al and Ni N00014-07-1-0740.
spectral lines indicate that atomic species of these elements were
present. Although there was also evidence of singly and doubly
ionizedAl,thereweresignificantlevelsofAlatomsexcitedfromtheir References
groundstate,suggestingthattheywereavailabletoexothermically [1] A.J.Gavens,D.V.Heerden,A.B.Mann,M.E.Reiss,T.P.Weihs,J.Appl.Phys.87
mixwithNi.Tobettershowevidenceofexothermicbehaviorbetween (2000)1255.
NiandAlundertheseconditions,weplanfuturetestswithsamples [2] M.Ding,F.Krieger,J.Swank,C.McMullan,G.Chen,J.Poret(Eds.),2008NDIA
Power&EnergyWorkshop,Warren,MI,2008.
covered by a flyer layer as in [7]. We expect this layer to limit the [3] A.J.Swiston,T.C.Hufnagel,T.P.Weihs,Scr.Mater.48(2003)1575.
stimulation of Ar and N emission, and enhance emission from [4] J.Wang,E.Besnoin,A.Duckham,S.J.Spey,M.E.Reiss,O.M.Knio,M.Powers,M.
vaporized conductor materials. Although it is possible that varying Whitener,T.P.Weihs,Appl.Phys.Lett.83(2003)3987.
[5] B.Boettge,J.Braeuer,M.Wiemer,M.Petzold,J.Bagdahn,T.Gessner,J.Micromech.
amountsofArmaybetrappedinthefilmsfromthesputterdeposition
Microeng.20(2010)064018.
process, temperature estimates would better represent the films [6] A.S.Rogachev,Russ.Chem.Rev.77(2008)21.
themselves instead of any surrounding atmosphere. Also, better [7] C.J.Morris,B.Mary,E.Zakar,S.Barron,G.Fritz,O.Knio,T.P.Weihs,R.Hodgin,P.
calibration of the detector would allow quantification over a wider Wilkins,C.May,J.Phys.Chem.Solids71(2010)84.
[8] V.M.Sandakov,Y.O.Esin,P.V.Gel'd,Russ.J.Phys.Chem.1020(1971)45.
wavelength range, as well as allow temperature determination [9] C.J.Morris,P.Zakar,EugeneWilkins,C.May,T.P.Weihs,inProc.27thAnnualArmy
through alternative methods such as pyrometry. We hope to carry ScienceConference,Orlando,FL.
[10] I.Reif,V.A.Fassel,R.N.Kniseley,Spectrochim.Acta,PartB28(1973)105.
out these tests in the near future, but the present set of results
[11] J.R. Fuhr, W.L. Wiese, L.I. Podobedova, D.E. Kelleher (Eds.), CRC Handbookof
suggests that exothermic mixing of Ni/Al is possible from rapidly ChemistryandPhysics,91stedition,CRCPress,2010.
heatedandvaporizedmaterial. [12] G.Norlén,Phys.Scr.8(1973)249.
[13] R.W.Conner,D.D.Dlott,J.Phys.Chem.A114(2010)6731.
[14] J.R.Carney,J.S.Miller,J.C.Gump,G.I.Pangilinan,Rev.Sci.Instrum.77(2006)
4.Conclusions 063103.
[15] D.S.Moore,S.Son,B.Asay,PropellantsExplos.Pyrotech.29(2004)106.
Thegoalofthisstudywastovalidatepastresultssuggestingthat [16] D.Boonyawan,S.Davydov,B.Yotsombat,N.Chirapatpimol,T.Vilaithong,Rev.Sci.
Instrum.73(2002)754.
Ni/Alexothermicmixingcanbeelectricallyforcedtooccuroververy [17] H.F.Winters,E.Kay,J.Appl.Phys.38(1967)3928.
shorttimescales.Timeresolvedstreakspectroscopyofmultilayered [18] P.Lynch,H.Krier,N.Glumac,J.ThermophysHeatTransfer24(2010)301.
and single component bridges indicated that emission from multi- [19] N.Glumac,H.Krier,T.Bazyn,R.Eyer,Combust.Sci.Technol.177(2005)485.
[20] R.Baksht,A.Pokryvailo,Y.Yankelevich,I.Ziv,J.Appl.Phys.96(2004)6061.
layeredNi/Alsamplesexhibitedanoverallbrightnessthatwashigher
[21] P.G.Burkhalter,C.M.Dozier,D.J.Nagel,Phys.Rev.A15(1977)700.
andpersistedlongerthaneitherAlorNicontrolsamples.Theatomic [22] A.Y. Faenov, S.Y. Khakhalin, A.A. Kolomensky, S.A. Pikuz, A.I. Samokhin, I.Y.
Al and Ni spectral lines in these experiments indicate that atomic Skobelev,J.Phys.D:Appl.Phys.18(1985)1347.
[23] X.Zhou,Y.Li,J.Wang,Z.Huang,IEEETrans.PlasmaSci.29(2001)360.
species of these elements were present, along with weaker lines
[24] G.S.Sarkisov,S.E.Rosenthal,K.R.Cochrane,K.W.Struve,C.Deeney,D.H.McDaniel,
indicating ionic species. We calculated temperatures to be in the Phys.Rev.E71(2005)046404.
Pleasecitethisarticleas:C.J.Morris,etal.,ThinSolidFilms(2011),doi:10.1016/j.tsf.2011.07.043