• Nenhum resultado encontrado

Braz. J. Phys. vol.30 número3

N/A
N/A
Protected

Academic year: 2018

Share "Braz. J. Phys. vol.30 número3"

Copied!
10
0
0

Texto

(1)

Plasma-deposited a-C(N):H Films

D.F. Franeshini

InstitutodeFsia,Universidade FederalFluminense

AvenidaLitoraneas/n,Niteroi,R.J.,24210-340,Brazil

Reeived15February,2000;reeivedinnalformon22Marh,2000

Thegrowth behaviour, lmstruture and mehanial properties of plasma-deposited amorphous

hydrogenated arbon-nitrogen lms are shortly reviewed. The eet of nitrogen-ontaining gas

additiontothedepositiontothehydroarbonatmospheresusedisdisussed,onsideringthe

mod-iations observed inthe hemial omposition growth kinetis, arbon atom hybridisation and

hemialbondingarrangementsofa-C(N):Hlms. Theoverallstruturebehaviourisorrelatedto

thevariationofthemehanialproperties.

I Introdution

In the last ten years there has been a great interest

in the study of plasma-deposited amorphous

hydro-genated arbon-nitrogen (a-C(N):H) lms. The main

reason for the initial interest was the suggestion, by

LiuandCohen[1℄,thatthehypothetialompound

-C

3 N

4

wouldhavemehanialpropertiessimilartothat

of rystalline diamond. Sine that suggestion, muh

work has been done on the searh for arbon nitride

synthesis, eortthat wasreentlyreviewed[2,3℄. The

aimoftheworkonplasma-depositeda-C(N):Hwas

ob-viously not to produe rystalline arbon nitride, but

to study thesupposedbeneial eetsof nitrogen

in-orporation on the mehanial, eletrial and optial

propertiesofa-C:Hlms.

Theresearhona-C(N):H lmsstartedevenbefore

LiuandCohen'ssuggestion,withthestudyofnitrogen

eletroni dopingofa-C:HlmsreportedbyJonesand

Stewart[4℄. Afterthat,followedthepioneeringworksof

Han andFeldman [5℄andAmirandKalish[6℄,mainly

onsidering optial and eletrial properties, and the

work by Kaufman et al [7℄ on Infrared Spetrosopy

of a-C(N):H lms, whih stated important strutural

eets of nitrogen inorporation into amorphous

hy-drogenatedarbonlms. Furtherresearhontheeld

showedthatsomeaspetsofthenitrogeninorporation

proessin a-C:H lms plaeobstalesto the

ahieve-mentoftheexpetedbenets. Therstoneisthe

lim-ited nitrogen uptake observed in a-C(N):H lms. No

more than about 20 at.% N ould be ahieved up to

now in plasma deposited a-C(N):H lms. Theseond

oneisthatnitrogeninorporationresultsinastrong

de-reaseinthesp 3

arbonatomfration[8℄,whihisthe

mainresponsiblefora-C:Hlmrigidity. Andnallythe

preferentialbondingofhydrogenatomstothenitrogen

atoms, whihmakesdiÆultarbon-nitrogenextended

networkformation,andaddsterminatinggroupstothe

amorphousnetwork.

Despite suh limitations, plasma-deposited

a-C(N):H lms were found to be used in a number of

appliations. The stress redution indued by

nitro-geninorporation[9℄andonsequentadhesion

improve-ment,allowedthedevelopmentof a-C(N):H

antiree-tiveoatingsforGe-basedinfrareddetetors[10℄. Itwas

foundalsothatnitrogenaneletroniallydopea-C:H

lms,andstronglydereasethedefetdensity,enabling

itsuse asasemiondutor material [11℄. Nitrogen

in-orporation was found also to derease the threshold

eletrieldineletron-eldemissionproess[12℄,

mak-ingpossibletheuseofa-C(N):Hlmsasaoveroaton

emissiontipsinat-paneldisplaydevies[13℄.

Inasimpleway, harda-C:Hlmstruture maybe

pituredbyanover-onstrainedrandomnetwork

om-posedbysp 3

andsp 2

-hybridisedarbonatomsand

hy-drogenatoms. Thismeansthatthemeanoordination

number of the atoms forming the network is greater

than the ideal geometrial value [14℄. This high

de-gree of overonstraining in a-C:H lms is due to the

preseneofalargefrationof sp 3

atoms,highly

ross-linked,whihresultsintheobservedrigidity,aswellas

inthedevelopmentofahighinternalompressivestress

duetohemialbonddistortion. So,in thestudyof

a-C(N):Hlms struturespeial attention must be paid

to the struture hanges that aet the network

on-netivity, suh as hemial omposition, arbon atom

hybridisation,andhemialbondingsheme.

The proesses involved in plasma deposition of

amorphous hydrogenated arbon lms are quite

om-plex. Besides the hydroarbon plasma hemistry

de-tails, eah dierent speies from the plasma interat

withthegrowinglayerinadierentway[15℄. Atrst,

wehavethepositivelyhargedions,mainlythe

arbon-arryingones,that areextratedfrom theplasmaand

(2)

aeleratedtowardsthesubstratebythenegatively

bi-ased substrateeletrode. Theyare theresponsible, by

energy deposition, forthe ativation of C-Cbond

for-mation, among other onsequenes. Carbon-arrying

slow neutralradials ontributemanly to lm growth

itself, bystiking to dangling bonds at thelm

grow-ing surfae. Hydrogen fastions andslowneutralsan

be involved in dangling bond reation or saturation,

among other proesses. Nitrogen addition to the

de-position atmosphere may, besidesaltering the plasma

hemistry, also alter the surfaeproess ating in the

growinglayer,aswillbedisussedin thispaper.

The aim of the present work is to shortly review

aspets onerned to the nitrogen inorporation

pro-ess,lmgrowthkinetis,andmodiationofa-C(N):H

lmstrutureandmehanialproperties. Noreferene

will be made on eletrial or optial properties of

a-C(N):Hlms. Furtherrefereneonthissubjetmaybe

found in a number of papers [4,5,6,11,16,17℄. In

se-tionII aredisussed resultsonthe eets of

nitrogen-ontaininggas addition to thedeposition atmosphere,

on lm hemial omposition and growthkinetis. In

setionII,Themain struturalhangesobservedin

a-C(N):Hlms,asdeterminedbyseveralharaterisation

tehniques,aredisussedinsetionIII.Inaddition,the

struturehangesare relatedto theobserved

mehan-ialproperties. SetionIII summarisestheoverall

dis-ussion.

II Chemial Composition and

Growth Kinetis

Several plasma deposition methods were used in the

studyofa-C(N):Hlms. Mostoftheworkwasdoneon

lmsdepositedbyonventionalradio-frequenyPlasma

Enhaned Chemial Vapour Deposition (rf-PECVD)

[18℄, in a hydroarbon / nitrogen-ontaining gaseous

mixture,withthesubstrateplaedontherf-biased

ele-trode [19-22℄. Inthis method theapaitiveoupling

of therfpowerto theathode allowsthedevelopment

of an averaged-in-time DC negative bias at the

pow-eredeletrode,knownasself-biaspotential(V

B ). The

potentialV

B

extrats the ions in the plasma, and

a-eleratethem towardsthelmgrowingsurfae. There

were somereportsonlmsdepositedbytheDC

glow-disharge method [23℄, and also a variation of the rf

ow-disharge, in whih the substrate is plaed in a

negativelyDC-biasedeletrodeparallelandopposedto

therf-poweredone[7℄.

In order to ahieve higher plasma ionisation and

dissoiation, some authors used high density plasma

soures to study a-C(N):H lms. The Eletron

Cy-lotron Resonane- Mirowave (ECR-MW) plasma

soure[24,25℄,andthehelialresonator plasmasoure

[26℄ were used to generate a high density plasma,

to ether with rf-biasing of the substrate eletrode, in

ordertoextrat theionsfrom theplasma. Avariation

oftherf-glow-dishargemethod,inwhihasteady

mag-neti eld is perpendiularly imposed to the rf-biased

eletrode surfae, was used to deposit a-C(N):H lms

with a gaseous mixture largelydiluted in helium [11℄.

Prodution of a-C(N):H lms was also performed by

using highly ionised plasma soures [27℄ or ion beam

soures[28℄,inordertodepositorassistthedeposition

proess . Theeets ofnitrogen-ontaininggasonthe

hemialompositionoflmsdepositedbyseveral

teh-niques, and preursor atmospheres may be viewed on

Table1. Inthistablearedisplayedthemaximum

nitro-gen ontent and thehydrogen ontentrange obtained

for the whole variation of the nitrogen-ontaining gas

partial pressure. Foreasieromparison of the results,

the table also displays the deposition tehnique used

andthemaindepositionparameters.

Asshownintable1,themaximumnitrogenuptakes

are about20 at %,and were obtainedusing aetylene

- nitrogen atmospheres. On the other hand, results

using other gasmixtures hardlyreah15 at. %

nitro-gen. Despite this dependene onthe preursorgas, it

seemsthatthemaximumnitrogenontentdoesnot

de-pendonthedepositiontehniqueused. Itisinteresting

to notealso that themaximumnitrogenuptakeis not

diretly orrelated to the nitrogen inorporationyield

of the partiular nitrogen-ontaininggases. It is

eas-ily seen by omparing results from referenes [9℄ and

[20℄inthetable,whihwereobtainedbyusingCH

4 -N

2

and CH

4 -NH

3

mixtures. Both works obtained about

thesamemaximumnitrogenuptakes,despitethegreat

dierenesinthenitrogenatomifrationin the

depo-sition atmosphere at the maximum nitrogen

inorpo-ration . Thus, proess other then hemial reativity

may drive the nitrogen inorporation. In whih

on-erns to the hydrogen ontent some authors report a

ontinuouslydereasingbehaviouruponnitrogen

inor-poration,whileothersreportarelativelyinsensitive

be-haviour,showing alsodierent levelsof hydrogen

on-tent.

Theobservedrelativelylowupperlimit for the

ni-trogenontentin a-C(N):H lmshasbeenasribedto

thestrongdereaseonthelmdepositionrateupon

ni-trogenpreursoradditiontothedepositionatmosphere,

asshownonFig.1. Inthisgureareshownplotsof

de-position rate againstthenitrogen ontent, normalised

to that of the nitrogen-free lm, for several preursor

gas mixtures [22℄. All theplots showa learderease

onthedepositionrateagainstnitrogenontent. Inthe

gure, the deposition rate plots for CH

4 -N

2 or NH

3 ,

and C

2 H

2 -CH

3 NH

2

gaseousmixturesseemto followa

very similar dependene with nitrogen ontent. This

ommon dependene showsan about 13 at. % N

ap-parentvanishingofdepositionrate. Ontheotherhand,

theplotorrespondingtotheC

2 H

2 -N

2

mixturefollows

alearly dierentdependene,showingthe deposition

ratefallathighernitrogenontents,allowingthe

(3)

Table1. Chemialompositionanddepositiondetailsfora-C(N):Hlmsdepositedbyseveraltehniques.

Ref. Depositionmethod Hydroarbon Nitrogenpreursor Substrate Pressure Maximum H

(partialpressure bias(-V) (Pa) NContent Content

range) inthelm inthe

(at.%) lm

7 rfPECVD/DCbias C5H10/Ar N2 380 3.5 11

-(0-95%)

11 magnetield CH

4

/He N

2

- 13 15 12-21

enhanedrfPECVD (0-80%)

16 rfPECVD C2H2 N2 - 7 20 30-38

(0-70%)

9 rfPECVD CH

4

N

2

370 8 11 12-17

(0-50%)

20 rfPECVD CH4 NH3 370 8 13 16-20

(0-25%)

21 rfPECVD C

2 H

2

CH

3 NH

2

350 3 12 12-16

(0-40%)

22 rfPECVD C2H2 N2 300 3 22 11-13

(0-90%)

23 DCPECVD C

2 H

2

N

2

500 20 14 19-37

29 rfPECVD C

6 H

6

N

2

500 13 10 22-27

(0-80%)

29 rfPECVD C6H6 NH3 500 13 8 27

(0-80%)

52 ECR-MW/rf bias CH

4

/Ar N

2

120 - 11

-(0-63%)

Figure1. Variationoftherelativedepositionrateasa

fun-tionofthenitrogenontentinthelm,forlmsdeposited

from CH

4 -N

2

[9℄, CH

4 -NH

3 [20℄, C

2 H

2 -CH

3 CH

2

[21℄, and

C

2 H

2 -N

2

mixtures[22℄.

Althoughdataonplasmahemistryof

hydroarbon-nolearonnetionis madewith thelmgrowth

pro-ess. Thehangesobservedintheoptialemission

spe-traofhydroarbon-N

2

mixturesaretheinreaseinthe

CNandN-derivedemissionlines,inparallelwitha

de-rease in the intensity of the CH emission lines [31℄.

Sinethesehangesannotbythemselvesbeorrelated

withthelmsurfaeproess, littleinformationanbe

extratedfromthiskindofanalysis.

Two fators were identied as being the ause for

the observed strong derease on the deposition rate

[22℄. The rst one is the erosion of arbon atoms by

energeti nitrogen ions like N

2

+ that omes from the

plasma. This eet wasstudied in detail by Hammer

andGissler[32℄. Theyfoundthatlow-energy(150eV)

N +

2

ion bombardment of a hydrogen-free amorphous

arbon lm resultedin arbon atom removal at rates

as high as 0.5 Catom perN +

2

ion, mainly as CN and

C

2 N

2

moleules. Thisproesswasidentiedasa

hem-ial sputtering proess, sine the observed sputtering

rateisfarhigherthanthatexpetedbyphysial

sput-teringproess. AsimilarproesswasobservedbyHong

andTurban [33℄ whenstudying the ething proess of

a-C:HlmsbyN

2

ECRplasmas. Theyfoundthat

rf-biased, previously grown a-C:H lms were eroded by

theN

2

plasma,beingevolvedin this proessthesame

kindoffragmentsastheobservedin N +

2

ion

bombard-mentofhydrogen-freearbonlms. Theseondfator

istheevaporationofN

2

(4)

andarbonatomsinorporationmayshowadisordered

nature in a-C(N):H lm growth, nitrogen evaporation

is verylikely to ourwhen the lm nitrogen ontent

inreases.

BothfatorsweretakenintoaountbyTodorovet

al[34℄inordertomodelion-beamdepositeda-CN

x lm

growth,for apartiular C/Nionarrivalratio. Inthis

work,besidesarbonatomhemialsputteringarbon

removal by N+ ions and N

2

evaporation, the authors

also inluded N atom sputtering by C +

ions, within

a Monte-Carlo simulation of the ollision proess

fol-lowing ion subsurfae penetration. With this model

theyouldtresultsofthehemialompositiondepth

prole,but no attemptwasmade tomodeldeposition

rates.

Asimplestatistialmodelfora-C(N):Hlmsgrowth

kinetis and nitrogen inorporation was reently

pro-posed [35℄. This model inorporates the main

ef-fets aused by the plasma nitrogenated speies on

the growth kinetis: The hemial sputtering of

ar-bonatomsbynitrogenions,andtheevaporationofN

2

moleules. This was ahievedby onsideringonly two

speiesarrivingatthelmsurfaewithomplimentary

probabilities: a fully aggregating \C" speies,

repre-senting arbon-arryingions and radials, and a \N"

speies, representing the N +

2

ions extrated from the

plasma. In themodel, when a\N" speies fall overa

\C" atomin thedeposit,the interation betweenand

inoming \N" speies and the deposit is desribed by

theinterationparameterq. Thisparameterrepresents

theprobabilityof aN atom to removeaCatom from

thelm,being(1 q)theprobabilityofaggregationof

the\N"speies. Whena\N"speiesfallsoveranother

a N atom, both leavethe deposit asan N

2

moleule.

N atom removalbya \C"speieswasnotonsidered,

sineformosta-C:Hlmdepositionproessesthemain

hannel for material aggregationat thelm surfaeis

thestikingofarbonarryingslowradials. Byusing

aninterationparameterqequalto0.25(veryloseto

thearbonremovalratereportedinreferene[19℄),this

modelwasfoundtotverywellthedepositionrate

re-sultsobtainedfromplasmadepositionusing

aetylene-nitrogen mixtures reported in referene [22℄. In

ad-dition, the obtained maximum nitrogen uptake as a

funtion of the interation parameter showed that no

morethanabout33at .%N ouldbe inorporatedin

a-C(N):H lms, even onsidering no hemial

sputter-ingbutonlyN

2

evaporation. Moreover,notouldbe

obtainedfordeposition rateexperimental resultsfrom

depositionatmospheresotherthanC

2 H

2 -N

2 .

Thelowernitrogenuptakeassoiatedwithasharper

dereaseonthedepositionrateuponnitrogenshownon

gure 1 (for the CH

4 -N

2 or NH

3 and C

2 H

2 -CH

3 NH

2

mixtures) was asribed to the role of hydrogen in the

deposition proess [22℄. Aording to Jaob [15℄, the

mainroleofhydrogenina-C:Hlmdepositionatroom

higher the hydrogen ontent in the deposition

atmo-sphere, lower would be the surfae sites available to

lm growth by slowradials. Fig. 2showsthe

hydro-gen atomifrationin thedepositionatmosphere, asa

funtion ofthenitrogenpreursorgaspartialpressure,

for the lm deposition experiments shown on Fig. 1.

The ranges shown orrespondto thenon-zero

deposi-tionraterange. Itis easytonote thatwhenusing the

C

2 H

2 -N

2

mixture,thedepositionatmospherehydrogen

frationisfarlowerthanthatshownbytheother

mix-tures. Furtherimprovementoftheabovedisussed

sta-tistial model for a-C(N):H lm growth inorporated

suh an eet, by bloking axed fration of surfae

sites to lm growth [36℄. With this modiation, the

modelwasableto t thetwokindsof deposition rate

urvesshownin Fig. 1.

Figure2. Hidrogenontentinthedepositionatmospheres,

forthegaseousmixturesindiatedintheplot,indenon-zero

depositionraterange.

III Film struture and

mehan-ial properties

As it was mentioned in the introdution, the main

struturalaspetonerningamorphousarbonlmsis

thearbonatomhybridisation. Thepreseneofalarge

frationofsp 3

arbonatomsina-C:Hlms,withahigh

ross-linkingof the amorphousnetwork,givesrise the

rigidityshownbythislms. So,therstonerninthe

(5)

state ofarbonatoms. Unfortunatelythere areonlya

fewreportsofsuhkindofanalysis.

Seth and Babu [37℄ reported a study on the

arbon sp 3

fration of a-C(N):H lms deposited

from butadiene-nitrogen and butadiene-ammonia

at-mospheres, as determined by 13

C-NMR spetrosopy.

They reported that nitrogen inorporation resulted

in an inreased C sp 3

fration for ammonia-derived

lms, whereasnot remarkablehanges were found for

nitrogen-derivedlms.

The above nding was not orroborated by later

studies. Table 2 summarises results on the

hybridi-sation state of arbon atoms in plasma deposited

a-C(N):Hlms. There areshownresultsobtainedbythe

omparison of theof the

and

peaksof the

CK-edgeEletronEnergyLoss(EELS)spetraoflms

deposited byrf-PECVD [8℄(alsoshownin Fig. 3)and

magnetially-enhanedrf-PECVD[11℄.Inaddition,are

also shown results obtained from the ombination of

the EELS spetra in the plasmon region and Auger

Eletron Spetra(AES) takenfrom lmsdeposited by

ECR-MWontorf-biasedsubstrates[24℄. Togetherwith

the sp 3

fration results,thedepositiondetails are also

shown.

Figure3.VariationoftheCatomsp 3

frationasafuntion

of nitrogen ontent for lmsdeposited CH4-N2 and

CH4-NH3 mixtures(referene[8℄).

Table2. Thesp 3

arbonatomfrationvariationuponnitrogeninorporation.

Ref. Depositionmethod Maximumnitrogen sp 3

Catom

ontent(at%) frationrange(%)

8 rf-PECVD 11 50-10

CH4/NH2

8 rf-PECVD 11 50-25

CH

4 /NH

3

24 ECR-MW 4.6 41-15

CH

4 /N

2

11 Magnetially enhaned 15 35-20

PECVD

CH

4 /N

2 /He

Asdisplayedintable2nitrogeninorporationinall

ases results in aderease of the sp 3

C atom fration

upon nitrogeninorporation. Nosensitivityonthe

ni-trogen preursor gas was observed. Even the use of

a high density plasma soure ould not avoid thesp 3

fration derease. This strong derease in the arbon

sp 3

frationonthenitrogen ontentwasalsoobserved

uponnitrogenontentinhighlytetrahedralamorphous

arbon(t-a-C)lms[38℄.

Ramanspetrosopyangiveomplimentary

infor-mation aboutsp 2

arbonatom arrangementon

amor-phous arbonlms. As in a-C:H lms, the main

fea-tures of theRaman spetraof a-C(N):H lms are the

so-alled D and G bands present in graphiti arbon

materials. Information onsp 2

Catom strutureis get

from Ramanspetraby theanalysis of the integrated

band intensityratioI

D /I

G

,andthepeakpositionsand

lution by tting two gaussian lines to it. Mariotto

et al. [39℄ studied the Raman spetra of a-C(N):H

lmsdeposited byrf-PECVD in methane-nitrogen

at-mospheres. They found a ontinuous inrease of the

I

D /I

G

bandratiouponnitrogen inorporation,andan

inreaseof the G band peak position towardsthat of

rystallinegraphite. Thisbehaviourwasasribedtoan

inreasein thesizeornumberofthegraphiti lusters

presentin the lm,in aordanewith the

interpreta-tionstatedby Dillonand o-workers[40℄,in thestudy

of dierent kinds of amorphous arbons, heat-treated

at inreasing temperatures. Similar onlusions may

bedrawnbyomparisonto resultsreportedbyTamor

et al. [41℄, on the study of a-C(H) plasma-deposited

lmsat severalself-biasvoltages.

Fig. 4 shows a typial variation of Raman

(6)

show-methane-ammoniaatmospheres[20℄. Themain results

ofthetting proedure(I

D /I

G

ratio andthepeak

po-sitions !

G

of theG band) are displayed ontable 3as

funtions of the nitrogen atomi fration in the lm.

This table learly shows the inrease on I

D /I

G and

G

uponnitrogeninorporation. Ramanspetrataken

from a-C(N):H lms deposited by dierentdeposition

tehniquesandpreursorgasmixtures[42℄alwaysshow

thiskindofbehaviour.

Figure4. Ramanspetratakenfromlmsdeposited using

CH4-NH3 forseveralnitrogenontents(referene[20℄).

Table3. TheRamanbandintensityratioI

D /I

G andG

bandpeakpositionasfuntionsofthenitrogenontentin

thelm(Ref. 20).

Nitrogenontent(at%) ID/IG !G (m 1

)

8 0.73 1545

5.5 0.83 1549

7.9 0.98 1555

11 1.07 1557

IfweanalyseRamanSpetrosopyresultstogether

withtheabovearbonatomhybridisationstudies,one

an say that nitrogen inorporation into a-C:H lms

maygiverisetoaninreasingpreseneoflusteredSp 2

arbonatoms. Suh struturemodiationwouldgive

risetoadereaseintheamorphousnetwork

onnetiv-ity,assuggestedbyRobertson[43℄.

Infrared(IR)Spetrosopy has been used to probe

loal hemial bonds onguration in a-C(N):H lms.

lms are the C-H band, at about 2925 m 1

, the

N-H strething band, at about 3500 m 1

, the C N

strethingband,at aboutand abroadband plaedin

the 1400-1600m 1

wave-number range. By

ompar-ingspetratakenfrom 14

Nand 15

N-inorporateda-C:H

lms,Kaufmanetal. [7℄showedthethisbroadbandis

infatanIRobservationoftheIR-forbiddenRamanD

and Gbands. Thisis madepossiblebythe symmetry

breaking introduedby thepresene ofpolararbon

-nitrogenbondsinthelms. Thepreseneofthisband

learly shows that there are N atoms bonded to sp 2

arbonlustersina-C(N):H lms.

Fig. 5 shows a typial IR spetra evolution upon

nitrogen inorporation into a-C:H lms. The gure

shows, besides the ontinuous inreasein the Raman,

NH, andCNband intensities,aontinuousdereasein

the CH band intensity, whih almost vanishes for the

higher nitrogenontents, showingthat hydrogen

pref-erentiallybondstonitrogen. Asanumberofworkson

a-C(N):Hlmdepositionshowednosensitivevariation

of hydrogen ontent upon nitrogen inorporation, we

mayexpetthatmostoftheNatomswouldbepresent

asterminal bonds in the amorphousnetwork. This is

alsotheasefortheCN bonds. So,wemayexpet

that nitrogenatoms,atleastasagreatpart,mustnot

beontributingtothenetworkonnetivity.

Figure5. Infrared spetratakenfromlmsdeposited using

(7)

Furthersupportforthisonlusionisgivenbya

X-RayPhoto-EletronSpetrosopy(XPS) study on

hy-drogen inorporation into a-CN

x

sputtered lms [44℄.

The XPS results showed that hydrogen interrupts

C-N bond formation, and also evideniated preferential

hydrogenationofN atoms.

Thisis alsotheaseof the 13

C NMRspetrosopy

studyarriedoutbyLaMannaetal. [45℄ona-C(N):H

lms deposited by rf-PECVD on the grounded

ele-trode. Contrary to most NMR studies on amorphous

arbonlms,whihingeneralshowbroadbands

assoi-ated tosp 2

andsp 3

-hibridizedatoms, thisstudyfound

verythin linessuperimposedto thebroadones,whih

were assoiated to ordered strutures. Theyould t

the obtained spetrato strutures similar to nitrogen

ontaining fused aromati rings terminated by NH

2

groups. Suh a struture is very similar to sp 2

ar-bon lusters bonded to mostlyhydrogenated nitrogen

atoms.

The formation of network-terminating bonds as a

onsequene of nitrogen inorporationfound also

par-allel from other observations, not diretly related to

atomi oreletronistruture. Dopler-broadening

ob-servationofpositronannihilationina-C(N):Hlms[46℄

showedthat nitrogeninorporationinreasestheopen

volumefrationinthelm,whatwewouldexpetfrom

the breaking of network interonnetions. Hydrogen

thermaleusionstudiesshowed,forthesame

nitrogen-inorporatedlms, lowtemperature(about200 Æ

C)

ef-fusionpeaks. Astheompata-C:Hlmsusuallyshows

onlyhightemperature(about600 Æ

C)hydrogeneusion

peaks,thelowtemperaturepeaks presentina-C(N):H

lmswereassignedtothepreseneofopenvoidsinthe

lm[47℄.

Theinueneofnitrogeninorporationonthe

stru-ture modiation of amorphous arbon lms has also

beensubjetfortheoretialinvestigation. Although

al-most totallyfousedonhydrogen-freearbon-nitrogen

materials, those works may give some insight to the

disussionofhydrogen-ontaininglms.

Weih et al. [48℄ reported a moleular dynamis

simulation study on the formation of arbon-

nitro-gensolids. Theystudiedthenitrogeninorporationon

amorphous arbon within several density ranges, and

found that the presene of nitrogen indues the

on-tinuous inrease of the sp 2

C atom fration in all the

ases. Hu et al. [49℄ studied by semiempirial

meth-ods the nitrogen inorporation eets in arbon

lus-ters formed by diamond ells, and found that for N

ontents greater than about 12 at. % a transition to

the sp 2

stateis likelyto our. Both ndings at least

support the possibility that the sp 2

C atom inrease

observed in a-C(N):H lmsmay be due to a hemial

bondingeet.

Another approah is to onsider the eets of ion

bombardment. It iswell known that fastionispeies

bombardmentof lm growingsurfae plays akeyrole

in the formation of C atoms tetrahedralbonding. As

it wasdisussed in session II, theinidene of arbon

or nitrogen fastspeies over alm ontainingarbon

andnitrogen,maygiverisetohemialsputtering. As

pointedoutby Martonet al. [50℄,damageby ion

im-patmayatasasourefortrigonalbondformationin

arbon-nitrogen lms,and thus may be playing arole

ina-C(N):Hstruture.

Asdisussedabove,thestrutureofa-C(N):Hlms

maybeviewed asaover-onstrained,stressedrandom

network. Sothemeasurmentoftheintemalstress,

be-sidesbeingofpratialinterest,maygiveadiagnostis

of the degreeof overonstrainig, and thus of the

on-netivityoftheamorphousnetwork.

Nitrogen inorporation into a-C:H lms result, as

a general rule in the redution of the internal stress

[22,42,51℄. Fig. 6showstheinternalompressivestress

variationforthreeseriesofa-C(N):H lms

rf-PECVD-deposited from dierent gaseous mixtures [9,20,22℄.

The three urves show a lear stress derease upon

nitrogen inorporation, at least aboveaertain

nitro-genontent. Thisbehaviourisin loseorrespondene

to the abovedisussed struture hanges in a-C(N):H

lms. The inreaseonthe sp 2

fration,the lustering

ofsp 2

arbonbonds, andtheintrodutionof terminal

sitesin thenetwork,allofthemmayontributetothe

dereaseofthenumberofonstrainsin theamorphous

network,andthus tostressrelaxation.

Figure 6. Variation of the internal ompressive stress for

lmsderivedfromCH4-N2,CH4-NH3andC2H2-N2gaseous

mixtures,asafuntionofnitrogenontent.

Butisalsolearinthegurethatthestressrelease

isaomplishedindierentwaysforeahofthe

(8)

series-haveshownalmostthesameCsp 2

fration

vari-ation [8℄. The diereneobservedin thetheirinternal

stress variation was asribed to dierent variations in

thehydrogenontentofthetwoseriesofsamples[9,22℄.

The hydrogen ontentmayalso explain the

dierenti-ated behaviour of the C

2 H

2 -N

2

series, whih shows a

relatively large nitrogen ontent range with no stress

variation,andthelowerhydrogenontentsofthethree

series(Seetable 1).

Although the internal stress derease has been

always observed upon nitrogen inorporation in

a-C(N):H lms, the mehanial hardness has shown a

spreadintheobservedbehaviour. Someworksreported

no remarkable hanges in the hardnessupon nitrogen

inorporation,atthesametimethat astrongderease

is observed in theinternal stress[9,10,29℄. Other ones

reported ontinuous hardness derease upon nitrogen

inorporation [16,19℄, as it would be expeted from a

dereasednetworkonnetivity. Dierenthardness

be-haviours-onstantordereasinguponNinorporation

- as the deposition parameters hanged were also

ob-served[16℄.

The reasonfor the observed disrepanies may be

duetothedierentmethodsemployedin thehardness

measurements. Mostoftherstworksonthe

mehan-ialproperties of a-C(N):H properties used miro-

in-dentation methods. In this methods, sine the depth

of indentation may be omparable to thin lm

thik-ness, thehardnessvalues measuredmay be inuened

bythesubstratehardness,reduingthevalidityofthe

omparisonbetweenexperiments.

More reent studies used nano-indentation

teh-niques for hardness measurement of a-C(N):H lms.

In this ase a hardness derease has been always

ob-served. This wastheaseofa-C(N):H lmsdeposited

byECR-MWplasmasinCH

4 -N

2

atmospheresreported

by Chanet al. [52℄. Hauert et al. observedthe same

behaviourforlmsdepositedfrom

pentadiene-nitrogen-argon mixtures by DC-biased rf-PECVD [53℄. In this

ase,earliermiro-indentationhardnessmeasurements

haveshown essentiallyonstanthardness, in lms

de-positedbythesametehniqueandgaseousmixture[10℄.

Fig. 7showsthehardnessvariationfora-C(N):Hlms

rf-PECVDdepositedinC

2 H

2 -N

2

atmospheres,plotted

together with the stress variation [22℄. The observed

variationisratherstep-likethanaontinuously

dereas-ingone. Butthehardnesslearlyfallswhenthestress

showdereasingbehaviour.

Atomi Fore Mirosopy was used to investigate

the frition oeÆient and surfae roughness of

a-C(N):Hlmsobtainedrf-plasmadeompositionofCH

4

-NH

3

mixtures [54℄. The frition oeÆient was found

tobeinsensitivetonitrogeninorporation,beinginthe

0.2range. Thesurfaeroughnesswasfoundtoinrease

byafatoroftwofora11at %nitrogenontent. The

roughnessinreaseasuponnitrogeninorporationwas

latedtothedetailsoflmgrowthkinetis[36℄.In

refer-ene[36℄arandomdepositionmodelfora-C(N):Hlm

growthwasused toalulatesurfaeroughnessvalues.

A inreasingbehaviourwasfound forinreasing

nitro-genontent,asaonsequenebetweengrowthandthe

erosiondrivenbyN +

2

ionbombardment.

Figure7. Variationofthenanohardnessandinternal

om-pressivestressasfuntionsofthenitrogenontent(referene

22).

IV Summary

The hemial omposition, growth kinetis, struture

and mehanial omposition hanges on the nitrogen

inorporationproessarereviewed. With minor

dier-enes, a-C(N):H lmsdepositedby severaltehniques,

using arange of deposition parametersand preursor

gasmixtures,showedthesameoverallbehaviour.

Nitrogen inorporation in the deposition

atmo-sphereauseseveredereaseonthelmdepositionrate,

whihlimitsthemaximumnitrogenuptaketoabout20

at%N.Thebehaviourofthelmgrowthkinetisseems

to bedrivenby theompetition betweenerosion

(ar-bonatom hemialsputteringby N

2

fastions andN

2

moleules evaporation)and aggregationproesses,

be-ingmodiedbythehangesinthehydrogenuxtothe

growinglayer.

The main struture hanges observedin a-C(N):H

deposition,thedereaseoftheCatomspfration,

lus-tering oftheCsp 2

atoms,andpreferential

(9)

the mehanial properties of the lms hange

drasti-ally,asisshownbythestrongdereaseobservedinthe

mehanial hardness and internal ompressive stress.

Besidesthestruturehanges,thehydrogenontent

it-self wasfoundto alterthemehanialpropertiesalso.

Referenes

[1℄ A.Y.LiuandM.L.Cohen-Siene245,841(1989).

[2℄ D.Marton,K.J.Boyd,J.W.Rabalais,I.J.Mod.Phys.

B9,3527(1995).

[3℄ S.Muhl,J.M.Mendez,DiamondRelat.Mater.8,1809

(1999).

[4℄ D.I.Jones,A.D.Stwart,Phil.Mag.B46,423(1982).

[5℄ H.X.Han,BernardJ.Feldman,SolidStateComm.65,

921(1988).

[6℄ O.Amir,R.Kalish,J.Appl.Phys.70,4958(1991).

[7℄ J.H. Kaufman,S.Metin, D.D. Saperstein, 39(1989),

13053,Phys.Rev.B39,13053(1989).

[8℄ D.F. Franeshini, F.L. Freire Jr, S.R.P Silva, Appl.

Phys.Lett.68,2645(1996).

[9℄ D.F. Franeshini,C.A.Ahete, F.L.FreireJr.,Appl.

Phys.Lett.60,3229(1992).

[10℄ S.Metin, J.H.Kaufman,D.D.Saperstein,J.C. Sott,

J.Heyman,E.Haller,J.Mater.Res.9,396(1994).

[11℄ S.R.P. Silva, J. Robertson, G.A.J. Amaratunga, B.

Raferty,L.M.Brown,J.Shwan,D.F.Franeshini,G.

Mariotto,J.Appl.Phys.81,2026(1997).

[12℄ G.A.J.Amaratunga,S.R.P.Silva,Appl.Phys.Lett.68,

2529 (1996).

[13℄ E.J. Chi, J.Y. Shim,H.K.Baik, H.Y. Lee, S.M. Lee,

S.J.Lee,J.Va.Si.Tehnol. B17,731(1999).

[14℄ J.C. Angus,F.Jansen,J.Va.Si.Tehnol.A6,1778

(1988).

[15℄ W.Jaob,ThinSolidFilms326,1(1998).

[16℄ J.Shwan,W. Dworshak,K.Jung, H.Ehrardt,

Dia-mondRelat.Mater.3,1034(1994).

[17℄ O.Stenzel,M.Vogel,S.Ponitz,R.Petrih,T.

Wallen-dorf, C.V. Borzyskowski, F.Rozploh, Z. Krasilnik,

N.Kalugin,Phys.StatusSolidA140,179(1993).

[18℄ J.W. Zou, K. Reihelt, K. Shmidt, et al., J. Appl.

Phys.65,3914(1989).

[19℄ P. Wood, T. Wyedeven, O. Tsuji, Thin Solid Films

258,151(1995).

[20℄ F.L. Freire Jr., D.F. Franeshini, Thin Solid Films

293,236(1997).

[21℄ M.M. Laerda, D.F. Franeshini, F.L. FreireJr., G.

Mariotto,DiamondRelat.Mater.6,631(1997).

[22℄ L.G. Jaobsohn, F.L. Freire Jr., M.M.Laerda, D.F.

Franeshini,J.Va.Si. Tehnol.A17,545(1999).

[23℄ A. Grill, V. Patel, Diamond Films Tehnol. 2, 61

(1992).

[24℄ S.Bhattaharyya,C.Valle,C.Cardinaud,O.Chauvet,

[25℄ H.Saitoh,T.Inque,S.Ohshio,Jpn.J.Appl.Phys.37,

4983(1998).

[26℄ J.H.Kirn,D.H.Ahn, Y.H. Kim,H.K. Baik,J.Appl.

Phys.82,658(1997).

[27℄ J. Shwan, V. Batori, S. Ulrih, H. Ehrardt, S.R.P.

Silva,J.Appl.Phys.84,2071(1998).

[28℄ H.W.Song,F.Z.Cui,X.M.He,W.Z.Li,J.Phys.

Con-dens.Matter6,6125(1994).

[29℄ K-RLee,K.Y.Eun,J.SRhee,Mat.Res.So.Symp.

Pro.356,233(1995).

[30℄ S.F.Durrant,N.Maral,S.G.Castro,R.C.G.Vinhas,

M.A.Bia de Moraes, J.H. Niola, Thin Solid Films

259,139(1995).

[31℄ K.J. Clay, S.P.Speakman, G. s

J. Amaratunga,S.R.P.

Silva,J.Appl.Phys.79,7227(1996).

[32℄ P.Hammer,W.Gissler,DiamondRelat.Mater.5,1152

(1996).

[33℄ J. Hong, G. Turban, Diamond Relat. Mater. 8, 572

(1999).

[34℄ Todorov-J.Va.Si.Tehnol.A12,3192(1994).

[35℄ F.D.A. Aar~ao Reis, D.F. Franeshini, Appl. Phys.

Lett.74,209(1999).

[36℄ F.D.AAar~aoReis,D.F.Franeshini,Phis.Rev.E61,

3417(2000).

[37℄ J.Seth,A.J.I. Ward,V. Babu, Appl.Phys.Lett. 60,

1957(1992).

[38℄ V.S. Veerasamy, J. Yuan, G.A.J. Amaratunga, W.I.

Milne,K.W.R.Gilkes, M. Weiler,L.M. Brown, Phys.

Rev.B48,17954(1993).

[39℄ G. Mariotto, F.L. Freire Jr., C.A. Ahete, hin Solid

Films241,255(1994).

[40℄ R.O.Dillon, J.A.Woolam, V. Katkanant, Phys.Rev.

B29,3482(1984).

[41℄ M.A. Tamor, W.C. Vassell, K.R. Carduner, Appl.

Phys.Lett.58,592(1991).

[42℄ F.L.FreireJr.,Jpn.J.Appl.Phys.136,4886(1997).

[43℄ J.Robertson,Phys.Rev.Lett.68,220(1992).

[44℄ S.Souto,F.Alvarez,Appl.Phys.Lett.70,1539(1997).

[45℄ J.LaManna,J.Bradok,Wilking, S.H.Lin,B.J.

Feld-man,SolidStateComm.109,573(1999).

[46℄ F.L.FreireJr.,D.F. Franeshini,C.A.Ahete, Phys.

StatusSolidB192,493(1995).

[47℄ D.F.Franeshini,F.L.FreireJr.,W.Beyer,G.

Mari-otto,DiamondRelat.Mater.3,(1993),(1993).

[48℄ F.Weih,J.Widany,Th.Frauenheim,Phys.Rev.Lett.

78,3226(1997).

[49℄ J.Hu, P. Yang, C.M. Lieber, Phys. Rev.B57, 3185

(1998).

[50℄ D.Marton,K.J. Boyd,J.W. Rabalais, Y.Lishitz, J.

Va.Si.Tehnol.A16,455(1998).

(10)

[52℄ Wai-ChungChan,Man-KeungFung,Kai-HoLai,Igor

Bello,Shuit-TongLee,Chun-SingLee,J.Non-Cristall.

Solids124,180(1999).

[53℄ R.Hauert,A.Glisenti,S.Metin,J.Goitia,J.H.

Kauf-man, P.H.M. vanLoosdreht, A.J. Kellok, P.

Ho-mann,R.L.White,B.D.Hermsmeier,ThinSolidFilms

268,22(1995).

[54℄ R.Prioli,S.I.Zanette,A.O.Caride,D.F.Franeshini,

F.L.Freire,J.Va.Si.Tehnol.A14,2351 (1996).

[55℄ S.R.P. Silva,G.A.J. Amaratunga,J.R. Barnes, Appl.

Imagem

Figure 1. Variation of the relative deposition rate as a fun-
Figure 2. Hidrogen ontent in the deposition atmospheres,
Figure 3. Variation of the C atom sp 3
Figure 4. Raman spetra taken from lms deposited using
+3

Referências

Documentos relacionados

lms were grown on < 100 > silion using the neutral luster beam deposition tehnique). These lms were highly rystalline and

the tree level) of Fierz-Pauli higher derivative theory. T o begin with we analyze the poles to

The observed harateristis are explained using small signal a iruit analysis.. It is shown that the theoretial urve generated using the a iruit analysis gives

oupied moleular orbital (H), the next lower oupied level (H-1), their energy dierene value () and their relative. ontribution dierene to the LDOS ()

The stati and spherially symmetri solution of the eld of a harged point.. partile

the present higher-derivative model the magneti eld restores hiral symmetry broken initially on.. the

The expliit form of the multipliative and onformal anomalies in losed oriented hyperboli.. manifolds

The ratio of positron/eletron uxes originated in nulear spallation reations in the Earth`s magne-.. tosphere