Professional Documents
Culture Documents
Matematicko-fyzikln fakulta
DISERTAN PRCE
Jan Kluso
Praha 2013
Chtl bych vyjdit podkovn vem, kdo svoj pomoc pispli k tto
disertan prci. Podkovn pat na prvnm mst koliteli doc. Mgr. Pavlu
Kudrnovi, Dr., kter moji prci po celou dobu doktorskho studia z odbornho
hlediska vedl. Vm si jeho ochoty a ptelskho pstupu, s nimi se vdy vnoval
aktuln eenm otzkm. Velkou zsluhu na pedkldan prci m tak prof.
RNDr. Milan Tich, DrSc. Jako konzultantovi mu vdm pedevm za dlouh
diskuze nad dosaenmi vsledky. Osobn velmi oceuji jeho zjem o studovanou
problematiku, kter byl i pro m v prci motivac. Zvltn dk pat Mgr. Zdeku
Hubikovi, Ph.D., vedoucmu oddlen nzkoteplotnho plazmatu na Fyziklnm
stavu Akademie vd R, v.v.i., za poskytnut nvrhu k sestaven pulznho zdroje a
dal dleitou spoluprci. Z ostatnch koleg z fyziklnho stavu, kte se na moj
prci podleli, bych chtl jmenovat alespo Mgr. Petra Virostka, Ph.D. Podkovn
bych chtl vyjdit tak prof. Romanu Schrittwieserovi z Universitt Innsbruck za
pozvn na svoje pracovit a umonn proveden men na tamjm
experimentlnm zazen. Nechtl bych zapomenout na ptelsk prosted
v laboratoi a kamardskou spoluprci pi een vech bnch kol po celou dobu
studia. Nakonec bych chtl podkovat tm, kte m v prci podporovali mimo jej
odbornou strnku.
Prce byla soust nkolika grantovch projekt. Byla podporovna
vzkumnm zmrem Ministerstva kolstv, mldee a tlovchovy R
MSM0021620834, granty 202/08/H057, 202/09/0800, P205/11/0386 Grantov
agentury R, granty 143307, 135207, 120510 a 604612 Grantov agentury UK a
programem CEEPUS II AT-0063.
i
Prohlauji, e jsem tuto disertan prci vypracoval samostatn a vhradn
s pouitm citovanch pramen, literatury a dalch odbornch zdroj.
ii
Nzev prce: Impulsn plazmatick systmy
Abstract: The doctoral thesis deals with the experimental study of the low-
temperature plasma designated for the preparation of thin films using the method of
physical sputtering. It describes measurements which were realized on two types of
devices on the low-pressure plasma jet sputtering system and on the planar
magnetron. The plasma diagnostics in the discharge excited in the pulse regime
represent the main theme of the thesis. The discharge in the low-pressure plasma jet
was studied by means of an electrostatic Langmuir probe, in the case of the planar
magnetron, mass spectroscopy with energetic as well as time resolution was used. To
understand the results achieved in the pulse generated plasma good knowledge of the
basic continuous DC regime of the discharge excitation is essential. Therefore a
considerable part of the thesis deals purely with the continuous DC discharge. In the
case of the low-pressure plasma jet, the attention was focused also on the
characteristic property of this plasma source, which is the flow from the jet.
iii
Obsah
Obsah 1
2. Teoretick vod 7
3. Experimentln systm 25
1
6.3. Elektronov koncentrace .................................................................. 72
8. Zvr 94
Literatura 98
2
1. Motivace a cle prce
3
Motivace a cle prce
4
Motivace a cle prce
5
Motivace a cle prce
6
2. Teoretick vod
Obr. 2.1. Vvoj stejnosmrnho elektrickho vboje s rostoucm proudem. V obrzku jsou rozlieny
ti zkladn oblasti. Pevzato z [Roth, 2000].
7
Teoretick vod
8
Teoretick vod
9
Teoretick vod
Obr. 2.2. Struktura stejnosmrnho doutnavho vboje s prbhy: (a) intenzity svtla, (b) elektrickho
potencilu, (c) elektrickho pole v podlnm smru, (d) hustoty elektronovho a iontovho proudu, (e)
elektronov a iontov koncentrace a (f) prostorovho nboje. Pevzato z [Raizer, 1991].
10
Teoretick vod
11
Teoretick vod
v tomto ohledu proto uspodn, kdy dut katoda zrove slou jako vstup
pracovnho plynu, co je ppad plazmov trysky. Prbh elektrostatickho
potencilu uvnit dut katody lze zskat eenm Poissonovy rovnice v jedn
dimenzi:
ne e r
2 (r ) n ne r
e
1 exp (2.5)
0 0 k B Te
Obr. 2.3. Schematick znzornn proces v dut katod vlevo a een jednodimenzionln
Poissonovy rovnice pro vlcovou geometrii dut katody vpravo, pevzato z [Hubika, 2007] a
[Niedrist, 2011].
12
Teoretick vod
2.3. Magnetron
13
Teoretick vod
potom vykonvaj EB drift, tj. ve smru kolmm k obma polm krou nad katodou
po trajektorich ve tvaru stoen roubovice. Tento pohyb je popsn cyklotronn
frekvenc c, Larmorovm polomrem rL a driftovou rychlost gyranho stedu vD:
mv
rL (2.6)
qB
qB
c (2.7)
m
EB
vd (2.8)
B2
E oznauje intenzitu elektrickho pole, B magnetickou indukci a v sloku rychlosti
stice kolmou k driftov rychlosti. Napt na katod u magnetronu typicky dosahuje
velikosti okolo U = 500 V a magnetick pole m hodnotu v du destek mT.
V takovch podmnkch in gyran polomr u elektron nkolik mm, zatmco pro
ionty se pohybuje v du metr. Pohyb iont tedy magnetick pole v podstat
neovlivuje. Elektrony vzan nad povrchem katody zsadnm zpsobem zvyuj
pravdpodobnost ionizace pi srce s elektronem. Zven hustota plazmatu v
oblasti nad katodou vede ke zven potu na ni dopadajcch iont, co pin vy
rychlost odpraovn a s n spojenou vy depozin rychlost. Navc vysok mra
ionizace umouje hoen magnetronovho vboje za nich tlak a nich napt
na katod.
Obr. 2.4. Princip magnetronu: schematick znzornn ezu planrnm magnetronem s permanentnm
magnetem.
14
Teoretick vod
I kU n (2.9)
kde k je konstanta mrnosti, do jej hodnoty se promt materil katody, druh a tlak
pracovnho plynu i geometrie vboje viz. [Rossnagel, 1987, Ellmer, 2007].
Exponent n vyjaduje mru zachycen plazmatu nad katodou, nabv hodnot z
intervalu 0 a 10. Pi vysokch hodnotch n pracuje magnetron pi tm
konstantnm napt a chov se tedy jako zdroj proudu.
Parametrem, kter m vrazn vliv na strukturu a vlastnosti rostoucch
tenkch vrstev, je tok nabitch stic, zejmna kladnch iont na substrt. Pmou
cestou ke zven iontovho toku je piloen uritho pedpt na substrt. Toto
een ale v mnoha ppadech nen vhodn, protoe dopadajc ionty zskvaj pli
vysok energie a ve struktue vrstvy se objevuj defekty. Doshnout znanch hodnot
koncentrace relativn nzkoenergetickch iont v mst substrtu lze prostednictvm
modifikace magnetickho pole v tzv. nevyvench (unbalanced) magnetronech
[Window, 1986, Savvides, 1986]. Zmnou v uspodn magnet se nkter
magnetick silokivky neuzavraj mezi stedem a okrajem tere, ale vedou a
k substrtu. Podl nich pak pronikaj nabit stice.
V bnch uspodnch vrazn pevauj ionty inertnho pracovnho plynu
nad ionty atom odprench z tere. Pitom nkdy je teba doshnout ionizace u co
nejvt sti materilu z tere. Primrnm dvodem je skutenost, e tok nabitch
stic lze na rozdl od neutrl snadno kontrolovat. Toho lze vyut k zen depozice
do pesn definovanch pozic nebo k usmrnn energie dopadajcch stic.
Za elem generovn plazmatu s vysokm podlem ionizovan sloky
deponovanho materilu bylo vyvinuto nkolik specilnch metod. Jedna skupina
tchto metod je zaloena na ptomnosti sekundrnho vboje, kter zvyuje hustotu
plazmatu v prostoru mezi terem a substrtem. Sekundrn vboj pedstavuje
napklad induktivn vzan vboj generovan cvkou umstnou u katody
[Rossnagel, 1993] nebo mikrovlnn vboj zaloen na elektron cyklotronov
rezonanci [Strak, 2012]. Jinm zpsobem je zven ionizace zvltn geometri
katody, kter vede k dokonalejmu zachycen elektron ne u klasickho
magnetronu. Pkladem takov varianty je magnetronov vboj v dut katod
[Klawuhn, 2000]. Zven mry ionizace lze doshnout tak aplikovnm vych
vkon, co je ppad ji zmnn obzvl vznamn techniky Hipims (High
15
Teoretick vod
16
Teoretick vod
3
I[mA]
I II III
Vfl 1
Vpl
0
-10 -5 0 5
U[V]
Obr. 2.5. Voltamprov charakteristika vlcov sondy. Nameno v magnetronovm vboji pi
tlaku p = 10 Pa a vbojovm proudu I = 100 mA. Dlka sondy lp = 3 mm, prmr dp = 45 m.
17
Teoretick vod
me d 2 I pe
f
4
(2.13)
e 3 ne A p 2 dU 2p
18
Teoretick vod
Te,eff
2
2 f d
(2.15)
3k B 3k B f d
Z Druyvesteynovy formule dle pmo vyplv vztah pro stanoven elektronov
koncentrace:
4 me d 2 I pe
ne
e3 Ap 2 dV p2
d (2.16)
e U p V pl
I pi I pi 0 1 (2.17)
k B Te
k B Ti Te
I p V pl neA p vi ve neA p
1
(2.18)
4 2 mi me
19
Teoretick vod
k B Te
I pe neAp ve exp
1
< 0: (2.20)
4
r p2
rs2
neAp vi s
1 r
I pi 1 erf erf (2.21)
4 rp rs2 r p2 rs2 r p2
r p2
rs2
neAp ve s
1 r
> 0: I pe
rp 1 erf 2 erf
(2.22)
4 rs r p2 rs2 r p2
neAp vi exp
1
I pi (2.23)
4
rp vyjaduje polomr sondy, rs polomr stnc vrstvy a = Te/Ti je anizotermick
parametr plazmatu. Chybov funkce erf(x) je zde definovna vztahem:
exp t dt
2
erf ( x) 2
(2.24)
x
20
Teoretick vod
(2.22) pro rzn tlouky vrstvy vetn limitnho ppadu (2.25) jsou zakresleny na
obr. 2.6.
2,5
rs >> rp
rs = 20rp
2,0 10rp
5rp
1,5 3rp
Ipe [(mA) ]
2
2rp
1.5rp
1,0
2
0,5
0,0
0 1 2 3 4 5
Up-Vpl [V]
Obr. 2.6. Prbhy druh mocniny elektronovho proudu na sondu podle piblen OML pro rzn
tlouky nbojov vrstvy vetn limitnho ppadu nekonen tlust vrstvy. Vypoteno pro parametry:
ne = 51016 m-3, Te = 0,15 eV, rp = 25 m, lp = 3 mm.
21
Teoretick vod
22
Teoretick vod
23
Teoretick vod
2E
1/ 2
2 E k 1 / 2
f E (r, E k ) 4 3k f v r, (2.29)
m m
Akoli tedy men prbh iontovho signlu v zvislosti na hodnot vstupnho
potencilu energetickho filtru Va nelze pmo ztotonit s precizn definovanou
energetickou rozdlovac funkc, v literatue se pro zjednoduen standardn pouv
oznaen iontov energetick rozdlen nebo energetick profil. Stejn oznaen je
pouvno i v tto prci.
24
3. Experimentln systm
25
Experimentln systm
Obr. 3.1. Vakuov schma experimentlnho systmu. V obrzku jsou zakresleny zkladn funkn
sousti systmu vetn jejich pipojen k aparatue.
Obr. 3.2. Schematick znzornn plazmov trysky (barevn uprosted), zapojen elektrickho obvodu
pro pulzn buzen tryskovho vboje (vlevo) a elektronickho obvodu pro men s elektrostatickou
Langmuirovou sondou (vpravo).
26
Experimentln systm
Svisle shora byla do hlavn komory vedena plazmov tryska. Pipojen trysky
k vakuov ndob bylo realizovno pes linern posuv, co umoovalo pohyb
trysky podl vlastn osy v rmci intervalu o dlce l = 10 cm. Tryska je schematicky
znzornna a barevn vyznaena na obr. 3.2. V principu je tvoena nerezovou
trubikou o prmru d = 6 mm stc do titanovho tere, mdnmi bloky
zajiujcmi chlazen a keramickm krytem. Titanov ter m rovn tvar trubiky,
jej rozmry in: vnj prmr dOUT = 8 mm, vnitn prmr dIN = 5 mm a dlka
l = 37 mm. istota titanu pouitho na vrobu tere byla 99,995 %. Prv titanov
ter zaujm roli dut katody, ve vnitnm prostoru probh vboj v dut katod a
dochz k odpraovn materilu ze stn, kter me bt nsledn deponovn na
povrch substrtu. Ponvad prakticky veker vkon dodvan ze zdroje byl
disipovn uvnit katody, bylo nezbytn katodu chladit. Z toho dvodu byl titanov
ter obklopen mdnmi bloky, ktermi protkala voda. Mezi bloky a terem nebyl
umstn dn izolan prvek, a tedy protkajc voda byla na stejnm potencilu
jako katoda. Takov een bylo mon s ohledem na nzkou vodivost bn vody
z vodovodnho potrub. Vysok napt ze zdroje bylo na katodu pivdno pes
stabilizan odpor (o velikosti R = 25 nebo R = 110 . Pouit stejnosmrn
zdroj Advanced Energy MDX 500 s uzemnnm kladnm plem poskytoval
maximln hodnoty napt, proudu a vkonu, Umax = 1,2 kV, Imax = 500 mA a
Pmax = 500 W. Nerezov ndoba reaktorov komory byla uzemnn a zastvala tak
funkci anody. Pitom nejbli s n vodiv spojen objekt pedstavoval nerezov
stolek substrtu. Vka stolku pod osou byla nastaviteln, stolek a st trysky se
teoreticky mohly dotkat.
Pro dosaen pulznho reimu vboje bylo teba pivst na katodu vysok
napt s pulznm prbhem. Mezi katodu a stejnosmrn zdroj vysokho napt byl
proto zapojen spnac obvod napovch pulz. Tento obvod byl zkonstruovn pmo
v laboratoi na zklad nvrhu Dr. Zdeka Hubiky z Fyziklnho stavu AV R.
Elektronick schma obvodu je po stech zakresleno na obrzcch obr. 3.3 -
obr. 3.6. Prvkem v obvodu, kter spn vysok napt, je dvojice paraleln
zapojench IGBT tranzistor BUP 314 viz. obr. 3.3. Technologie IGBT umouje
27
Experimentln systm
28
Experimentln systm
Obr. 3.4. Budic obvod pro spnac IGBT tranzistory zen externm TTL signlem.
Obr. 3.5. Zdroj stabilizovanho stejnosmrnho napt 15 V pro spnac IGBT tranzistory.
29
Experimentln systm
30
Experimentln systm
31
Experimentln systm
500
400
|Ip| [A]
300
t = 50s
200
100s
150s
100 300s
500s
0
-8 -6 -4 -2 0 2
Up [V]
Obr. 3.8. Pklad vvoje sondov charakteristiky v tryskovm vboji buzenm pulznm naptm o
frekvenci f = 1 kHz a std S = 10 % pro nkolik vybranch as od zatku pulzu na katod.
32
Experimentln systm
33
Experimentln systm
2m 1
s m m 2m
t FS s ext sm sd (3.1)
e V p Vext V0 Vax en 2eV 2eVte Vdyn
ax
34
4. Men rychlosti proudn plazmatu a neutrlnho
plynu v systmu nzkotlak plazmov trysky
35
Men rychlosti proudn
realizovno pomoc statick Pitotovy trubice, tzv. Prandtlovy sondy. Byly zskny
radiln profily rychlosti v nkolika vzdlenostech od trysky i pro nkolik rznch
prmr trysky. Tvar profil odpovd gaussovskmu prbhu, v nejblim okol
trysky je parabolick. Ve zvolench podmnkch dosahovalo proudn
nadzvukovch rychlost blcch se a k hodnotm Machova sla M = 2. Dlka
oblasti, ve kter nabvalo proudn nadzvukovho charakteru, nkolikrt pesahovala
teoretickou hodnotu pro laminrn reim. Z toho vyplv, e proudn nen ist
laminrn. Pi proudn z trysek malch prmr (d = 2mm) se vyskytovaly rzov
vlny. Bylo tak provedeno porovnn situace, pi zapnutm a vypnutm vboji.
V probhajcm vboji byly nameny mrn ni hodnoty rychlosti proudn, co
bylo vysvtleno ohvnm plynu. Ob ve uveden prce byly uskutenny
v podmnkch vysokch prtok pracovnho plynu v du jednotek Pam3/s a
relativn vysokch tlak. Nkter dal vsledky tkajc se proudn v systmu
plazmov trysky byly dosaeny prostednictvm potaovch model viz. nap.
[Selezneva, 2001]. Experimentln studium interakce plazmovho kanlu se
substrtem u systmu nzkotlak plazmov trysky buzen RF polem bylo
publikovno v [cha, 1996].
Obr. 4.1. Vvoj plazmovho kanlu u st trysky pi zvyovn prtoku pracovnho plynu a tlaku p
v aparatue. Obr. nalevo: = 0,3 Pam3/s, p = 30 Pa, obr. napravo: = 1,2 Pam3/s, p = 70Pa.
Pevzato z [Tich, 1994].
36
Men rychlosti proudn
Obr. 4.2. Schma experimentlnho uspodn pro men rychlosti proudn plazmatu a neutrlnho
plynu v systmu nzkotlak plazmov trysky pomoc Langmuirovy sondy a Pitotovy trubice vetn
obvodu pro pulzn buzen vboje.
37
Men rychlosti proudn
bhem aktivn sti periody pulznho buzen vboje. Jestlie vn trysky nen
ptomno elektrick pole, je pohyb iont i neutrl po oputn trysky ovlivovn
pouze vzjemnmi srkami a difz. Oba efekty vedou ke zpomalovn pohybu
stic podl osy trysky. Kladn ionty mohou bt detekovny elektrostatickou
Lagmuirovou sondou pipojenou na zporn potencil. Kad napov pulz na
katod zpsob s uritm zpodnm proudov impuls na sond. Pokud je men
asovho prbhu proudu na sond synchronizovno s pulznm buzenm tryskovho
vboje, lze as detekce impulsu na sond povaovat za dobu letu iont od st trysky
k sond. Prbh rychlosti je potom mon zskat jako asovou derivaci uraen
drhy po proveden vce men v rznch vzdlenostech od trysky.
Uveden postup je pmoar v ppad, kdy sondou zaznamenvme
dostaten vrazn a pedevm krtk a ostr pulzy (vzhledem k period buzen
vboje). Ve skutenosti vak pulzy ideln tvar nemaj. Typicky proud na sond
prudce narst, doshne maxima a pozvolna zanik. S pibvajc vzdlenost od st
trysky je maximum m dl tm mn zeteln, co je zejm zpsobeno srkami a
difznm rozpadem plazmatu. Do prbhu pulzu se tak promtaj pechodov jevy
souvisejc s bnou hranou napt na katod. Prakticky bylo teba zvolit co nejni
opakovac frekvenci a stdu, pi n byl vboj stabiln. Pitom nebylo douc
vrazn zvyovat napt na katod, aby byly zachovny podmnky, ve kterch m
bt vboj studovn. S ohledem na tyto skutenosti byly zvoleny nsledujc
parametry vboje: napt na katod U = 500 V, budic frekvence f = 400 Hz, stda
S = 1 % (dlka aktivn sti periody Tp = 25 s). V mench, kdy nebyla explicitn
studovna zvislost na prtoku plynu, byla nastavena hodnota Ar = 30 sccm. Stedn
hodnota iontovho proudu za danch podmnek byla piblin I = 75 mA. Prbh
iontovho proudu ukazuje obr. 4.3. Jednotliv kivky v grafu odpovdaj rznm
vzdlenostem sondy od st trysky. Polohy maxim v grafu na obr. 4.3 lze povaovat
za nejpravdpodobnj doby letu iont od trysky na sondu. Po odeten tchto
hodnot byla zkonstruovna zvislost drhy uraen ionty na ase, jej derivac je ji
hledan rychlostn prbh. Derivace nebyla potna pmo z namench dat,
hodnoty byly nejprve proloeny polynomem druhho stupn. Zpracovn dat
znzoruje obr. 4.4. Z nj je patrn, e parabolick zvislost odpovdajc
rovnomrn zpomalenmu pohybu je skuten dobrou aproximac.
38
Men rychlosti proudn
0,5
h = 27mm
32mm
0,4 37mm
42mm
0,3 47mm
52mm
57mm
I [mA]
0,2
0,1
0,0
-0,1
0 100 200 300 400 500 600 700
t [s]
Obr. 4.3. Typick prbh iontovho proudu na sondu v ase pi postupn zmn polohy sondy v ose
trysky. Souadnice h udv svislou vzdlenost od st trysky. Parametry vboje: p = 20 Pa,
= 30 sccm, U = 500 V, f = 400 Hz, S = 1 %.
60 125
50
100
40
75
h [mm]
v [m/s]
30
50
20
25
10
0 0
0 100 200 300 400 500 25 30 35 40 45 50 55 60
t [s] h [mm]
Obr. 4.4. Zpracovn namench asovch prbh proudu na sondu z obr. 4.3. Drha iont v ase
byla proloena parabolou vlevo, rychlost byla potom spotena jako asov derivace a vynesena proti
vce sondy pod tryskou - vpravo.
39
Men rychlosti proudn
Zskan prbhy rychlost v ose trysky pro rzn tlaky jsou zobrazeny na
obr. 4.5. Graf odpovd prtoku argonu Ar = 30 sccm, pi kterm vykazuje tryskov
vboj stabiln chovn, a v experimentech s plazmovou tryskou byl volen nejastji.
U jednotlivch bod je vyznaena chyba odhadnut ve vi 10 %. Hlavn zdroj
chyby pedstavuje proveden aproximace drhy iont v ase polynomem druhho
stupn. Z grafu je patrn, e rychlost proudn podl osy trysky monotnn kles a je
vrazn ovlivnna tlakem v aparatue. Zatmco pi tlacch v du nkolika destek Pa
nepesahuje rychlost proudn hodnotu 100 m/s, pi tlacch v du jednotek Pa
nabv hodnot nkolikrt vych. Nejvy rychlosti se vyskytuj nejble k st
trysky. Za nzkch tlak dochz k velik absolutn zmn rychlosti v rmci
menho intervalu vzdlenost od trysky, nap. pro p = 2 Pa: klesne mezi h = 32 mm
a h = 57 mm o v > 250 m/s. V menm oboru tlak a vzdlenost od trysky m
pokles rychlosti se vzdlenost piblin odmocninov charakter. To odpovd
pouitmu piblen rovnomrn zpomalenho pohybu.
p = 2Pa
500 5Pa
10Pa
20Pa
400
50Pa
100Pa
300
v [m/s]
200
100
0
25 30 35 40 45 50 55 60
h [mm]
Obr. 4.5. Prbh rychlosti proudn plazmatu v ose trysky pro rzn hodnoty tlaku pracovnho plynu
v aparatue zskan z men iontovho toku. Parametry vboje: Ar = 30 sccm, U = 500 V,
f = 400 Hz, S = 1 %.
40
Men rychlosti proudn
0,7 0,6
h = 27mm Ar = 100sccm
0,6 32mm 0,5 75sccm
37mm
0,5 50sccm
42mm 0,4
30sccm
0,4 47mm
0,3
52mm
I [mA]
I [mA]
0,3
57mm 0,2
0,2
0,1
0,1
0,0 0,0
-0,1 -0,1
0 100 200 300 400 500 0 100 200 300 400 500
t [s] t [s]
Obr. 4.6. asov prbh iontovho proudu na sondu pi vych prtocch pracovnho plynu. Vlevo
situace pi p = 10 Pa a Ar = 100 sccm pro rzn pozice sondy pod tryskou, vpravo vvoj s postupn
rostoucm prtokem v nejmen nastaviteln vzdlenosti od st trysky h = 27 mm pi tlaku p = 10 Pa.
41
Men rychlosti proudn
300
Ar = 100sccm
75sccm
250 50sccm
30sccm
200
v [m/s]
150
100
50
0
25 30 35 40 45 50 55 60
h [mm]
Obr. 4.7. Prbh rychlosti toku plazmatu v ose trysky pi rznch prtocch pracovnho plynu.
p = 10 Pa. Ve vech uvedench bodech byl pozorovn pouze pomalej nbh iontovho proudu.
12 1000
h = 27
10 32
37 800
8 42
47 600
6
52
v [m/s]
I [mA]
4 57
400
2
200
0
-2 0
0 100 200 300 400 500 25 30 35 40 45 50 55 60
t [s] h [mm]
Obr. 4.8. asov prbhy iontovho proudu na sondu v ose trysky pi p = 3 Pa a Ar = 100 sccm
(vlevo) s vyhodnocenm rychlosti proudn plazmatu v ose (vpravo). Za tchto podmnek byl ve vech
mstech men pozorovn rychlej nbh iontovho proudu viz. text.
42
Men rychlosti proudn
1 2
pT p S v (4.1)
2
kde je hustota plynu a v jeho rychlost. Pro stanoven rychlosti proudn tak sta
znt rozdl mezi celkovm a statickm tlakem. Pi men vtch rychlost,
dosahujcch hodnot okolo rychlosti zvuku, je teba pout piblen pro adiabatick,
izoentropick proudn idelnho plynu. V tomto piblen lze vyjdit celkov tlak
nsledovn:
1v 2 1 1 2 1
pT p S 1 p S 1 M (4.2)
2 p 2
S
kde = cp/cv je pomr specifickch tepel danho plynu oznaovan jako adiabatick
exponent. Pro argon je pi teplot T = 300 K = 1,67. Ve vztahu (4.2) je zavedeno
Machovo slo M = v/va, jen udv rychlost v nsobcch rychlosti zvuku va. Pro
rychlost zvuku plat:
p S
va RT (4.3)
Statick Pitotova trubice je obvykle konstruovna jako dv koaxiln veden
trubice s mrn odlinm prmrem. Vnitn trubice je na konci orientovanm proti
smru toku plynu oteven a na opanm konci uzaven, vnj trubice je naopak na
stran vstupujcho plynu uzaven, ale v bon stn m otvory. Ve vnitn trubici je
tak mon mit celkov tlak a ve vnj statick tlak, ppadn prostednictvm
diferencilnho manometru pmo rozdl obou tlak. Detailnj popis proveden
Pitotovy trubice lze najt nap. v [Klopfenstein, 1998]. Pro ne uveden men
rychlosti proudn z plazmov trysky byla vyrobena Pitotova trubice podle vlastnho
nvrhu tak, aby co nejlpe vyhovovala pouit v aparatue. Nkres trubice je
znzornn na obr. 4.9. Pitotova trubice byla realizovna dvojic nerezovch trubiek
vedench vedle sebe. Vzhledem ke vnjmu prmru trubiek 1,2 mm bylo
zajitno pln postaujc prostorov rozlien. Tyto dv trubiky byly zavedeny do
6 mm irok nerezov trubiky, piem jedna z nich prochzela skrz a na opanm
konci opt vystupovala ven a druh byla na vnj trubiku napojena. Vstupn st
Pitotovy trubice o dlce 35 mm byla zahnuta kolmo ke zbytku trubice, jeho dlka
inila 450 mm. Vnj nerezov trubika byla peletna a prostednictvm rouben
Swagelok s teflonovm tsnnm pipojena k prub aparatury. Toto proveden
umoovalo pohyb trubice pi udren vakua v aparatue v du 10-4 Pa. Tlo trubice
43
Men rychlosti proudn
44
Men rychlosti proudn
500 500
p = 10Pa I = 0mA
p = 10Pa, vboj 300mA
400 p = 50Pa 400
500mA
p = 50Pa, vboj
300 300
v [m/s]
v [m/s]
200 200
100 100
0 0
0 10 20 30 40 50 60 0 10 20 30 40 50
h [mm] h [mm]
Obr. 4.10. Srovnn vsledk men s Pitotovou trubic pi vypnutm a zapnutm vboji. Vlevo
situace pro dva rzn tlaky a vbojov proud I = 300 mA, vpravo porovnn situace bez vboje
s vbojem o dvou rznch proudech pi tlaku p = 10 Pa. Ar = 30 sccm.
800 p = 1Pa
10Pa
700
30Pa
600 100Pa
500
v [m/s]
400
300
200
100
0
0 10 20 30 40 50 60
h [mm]
Obr. 4.11. Prbh rychlosti proudn neutrlnho plynu v ose trysky uren Pitotovou trubic pro tyi
vybran tlaky. Ar = 30 sccm.
45
Men rychlosti proudn
800
700
600
500
v [m/s]
400
va = 322m/s
300
200
100
0
0 20 40 60 80 100
p [Pa]
Obr. 4.12. Zvislost rychlosti proudn plynu z trysky na tlaku v bod umstnm 7 mm od st trysky
v jej ose. Ar = 30 sccm. V grafu je vyznaen tlak, pi kterm dochz k pechodu mezi
podzvukovm a nadzvukovm reimem proudn.
46
Men rychlosti proudn
okruhem, zskme piblinou hodnotu tlaku v trysce pIN ~ 20 Pa. Potom by mlo
dojt k ustaven nadzvukovho proudn pi tlacch v aparatue nich ne p = 10 Pa.
Reln vak k uritmu ohevu plynu pi prchodu tryskou dochz, a proto je
kritick hodnota tlaku pro pechod do nadzvukovho reimu vy. Pestoe hodnota
kritickho tlaku zskan z uvedenho teoretickho modelu na zklad nepln
znalosti parametr vboje me bt brna vcemn jen jako odhad, je ve velmi
dobr shod s vsledkem men na obr. 4.12.
Graf na obr. 4.13 znzoruje rst rychlosti proudn se zvyujcm se
prtokem pracovnho plynu. Zvislost byla studovna do maximlnho prtoku
Ar = 100 sccm, vy prtoky se bn u systmu nzkotlak plazmov trysky
nepouvaj. Ve srovnn s ve uvedenou tlakovou zvislost je zjitn nrst
rychlosti s prtokem jen mrn. Pomal rst rychlosti s prtokem pi konstantnm
tlaku znamen, e se relativn rychle zvtuje plocha pn k ose trysky, kterou plyn
prochz, tedy nastv roziovn kanlu proudn.
500
400
300
v [m/s]
200
100
h = 7mm
17mm
0
0 20 40 60 80 100 120
Ar [sccm]
Obr. 4.13. Zvislost rychlosti proudn plynu z trysky na prtoku pracovnho plynu ve dvou
vzdlenostech pod tryskou v jej ose. p = 10 Pa.
47
Men rychlosti proudn
500
400
300
v [m/s]
200
100 p = 10Pa
0
0 10 20 30 40 50 60
h [mm]
300
250
200
v [m/s]
150
100
50 p = 30Pa
0
0 10 20 30 40 50 60
h [mm]
100
80
60
v [m/s]
40
20
p = 100Pa
0
0 10 20 30 40 50 60
h [mm]
Obr. 4.14. Srovnn vsledk rychlosti proudn dosaench menm Langmuirovou sondou
(erven) a Pitotovou trubic (ern). Grafy odpovdaj tem vybranm tlakm, postupn seshora
p = 10 Pa, p = 30 Pa a p = 100 Pa,Ar = 30 sccm.
48
Men rychlosti proudn
49
5. Sondov diagnostika tryskovho vboje
v kontinulnm stejnosmrnm reimu
500 p = 0.85Pa
5Pa
10Pa
400
300
I [mA]
200
100
0
200 210 220 230 240 250 260 270
UKat [V]
51
Sondov diagnostika tryskovho vboje v kontinulnm stejnosmrnm reimu
r [mm]
0 10 20 30 40 50 60 70 80
0
10
20
h [mm]
30
40
50
Obr. 5.2. Znzornn bod v prostoru pod tryskou, kde byla uskutenna sondov men. Rovina je
urena radiln souadnic r a axiln souadnic h. Sted st trysky m souadnice r = 0, h = 0. Osa
trysky je vyznaena erven.
52
Sondov diagnostika tryskovho vboje v kontinulnm stejnosmrnm reimu
1,5
1,0
0,5
Vpl [V]
0,0
h = 10mm
h = 32mm
h = 50mm
-0,5
-10 0 10 20 30 40 50 60 70 80 90
r [mm]
Obr. 5.3. Radiln prbh potencilu plazmatu ve tech rznch axilnch vzdlenostech. p = 1 Pa,
I = 300 mA, U ~ 250 V, Ar = 30 sccm.
53
Sondov diagnostika tryskovho vboje v kontinulnm stejnosmrnm reimu
1,5
1,0
0,5
Vpl [V]
0,0 r = 0mm
10mm
25mm
50mm
-0,5
0 10 20 30 40 50
h [mm]
Obr. 5.4. Axiln prbhy potencilu plazmatu pro stejn podmnky jako u obr. 5.3.
54
Sondov diagnostika tryskovho vboje v kontinulnm stejnosmrnm reimu
1,0
17
0,5
0,0
-10 0 10 20 30 40 50 60 70 80 90
r [mm]
Obr. 5.5. Radiln prbh elektronov koncentrace v rznch axilnch vzdlenostech. Elektronov
koncentrace byla vyhodnocena z elektronovho saturovanho proudu na zklad piblen OML.
p = 1 Pa, I = 300 mA, U ~ 250 V, Ar = 30 sccm.
3
r = 0mm
10mm
25mm
50mm
2
ne [10 m ]
-3
17
0
0 10 20 30 40 50
h [mm]
Obr. 5.6. Axiln prbhy elektronov koncentrace ve vybranch radilnch vzdlenostech pro stejn
podmnky jako u obr. 5.5.
55
Sondov diagnostika tryskovho vboje v kontinulnm stejnosmrnm reimu
Obr. 5.7. Vvoj elektronov koncentrace v objemu plazma-chemickho reaktoru pi pracovnm tlaku
p = 1 Pa. Nejble ke trysce bylo men provedeno 7 mm od st na ose trysky. Podmnky jsou stejn
jako u obr. 5.5.
56
Sondov diagnostika tryskovho vboje v kontinulnm stejnosmrnm reimu
ne [10 m ]
-3
-3
17
16
2
1
0 0
0 10 20 30 40 50 -10 0 10 20 30 40 50 60 70 80 90
h [mm] r [mm]
Obr. 5.8. Ukzka porovnn hodnot elektronov koncentrace vyhodnocench pomoc rznch metod:
z elektronovho proudu v potencilu plazmatu - IVpl, z proloen kvadrtu elektronovho saturovanho
proudu - IeSqr a z integrlnho vyhodnocen na zklad Druyvesteynovy formule - Int. Nalevo axiln
prbhy v ose, napravo radiln prbhy ve vce h = 32 mm. Podmnky jako u obr. 5.5.
57
Sondov diagnostika tryskovho vboje v kontinulnm stejnosmrnm reimu
1
10
h = 7mm
12mm
0
35mm
10 50mm
]
-3/2
EEPF [(eV)
-1
10
-2
10
-3
10
0,0 0,5 1,0 1,5 2,0 2,5
[eV]
Obr. 5.9. EEPF v ose trysky v nkolika vybranch vkch. p = 1 Pa, I = 300 mA, U ~ 250 V,
Ar = 30 sccm.
1
10
0
10
]
-3/2
EEPF [(eV)
-1
10
-2
10
10
-3 80
0,0 70
60
0,5 50
1,0 40 ]
30 m
[e 1,5 m
V] 20 r[
2,0 10
2,5 0
Obr. 5.10. Radiln prbh EEPF ve vce h = 10 mm pod tryskou. Podmnky men stejn jako u
obr. 5.9
58
Sondov diagnostika tryskovho vboje v kontinulnm stejnosmrnm reimu
0,3
Te [eV]
Te, eff
0,2
0,0
-10 0 10 20 30 40 50 60 70 80 90
r [mm]
Obr. 5.11. Radiln prbhy elektronov teploty v nkolika rznch vkch pod tryskou Te, Fit I
oznauje elektronovou teplotu vyhodnocenou z proloen druh derivace sondov charakteristiky, Te,eff
efektivn hodnotu elektronov teploty. I = 300 mA, U ~ 250 V, Ar = 30 sccm.
0,5
r = 0mm
25mm
0,4
Te, eff
0,3
Te [eV]
0,2
0,0
0 10 20 30 40 50
h [mm]
Obr. 5.12. Axiln prbhy elektronov teploty v ose trysky a v radiln vzdlenosti r = 25 mm.
Oznaen veliin a podmnky men stejn jako u obr. 5.11.
59
Sondov diagnostika tryskovho vboje v kontinulnm stejnosmrnm reimu
60
Sondov diagnostika tryskovho vboje v kontinulnm stejnosmrnm reimu
1,5 1,5
p = 1Pa
3Pa
1,0 1,0
Vpl [V]
Vpl [V]
0,5 0,5
0,0 0,0
-0,5 -0,5
0 2 4 6 8 10 12 0 20 40 60 80 100
p [Pa] r [mm]
Obr. 5.13. Zvislost potencilu plazmatu na tlaku. Vlevo vvoj v ose trysky pro h = 50 mm pi
postupnm zvyovn tlaku, vpravo radiln prbh ve vce h = 10 mm pro p = 1 Pa a p = 3 Pa.
I = 300 mA, U ~ 250 V, = 30 sccm.
3
ne [10 m ]
-3
17
2
17
1
1
0 0
0 2 4 6 8 10 12 0 10 20 30 40 50
p [Pa] h [mm]
Obr. 5.14. Zvislost elektronov koncentrace na tlaku. Vlevo vvoj v bod r = 0 mm, h = 50 mm,
vpravo prbh v ose trysky pro p = 1 Pa a p = 3 Pa. Podmnky jako u obr. 5.13.
61
Sondov diagnostika tryskovho vboje v kontinulnm stejnosmrnm reimu
1
0,5 10
Fit Ie
0,4
Fit I
Te,eff 10
0
]
-3/2
0,3
EEPF [(eV)
Te [eV]
-1
10
0,2
p = 1Pa
-2
10 3Pa
0,1
5Pa
10Pa
-3
0,0 10
0 2 4 6 8 10 12 0,0 0,5 1,0 1,5 2,0 2,5
p [Pa] [eV]
Obr. 5.15. Vlevo tlakov zvislost elektronov teploty v rozmez jednotek Pa v bod r = 0 mm,
h = 50 mm. Zobrazeny jsou vsledky dan temi rznmi metodami vyhodnocen. Vpravo EEPF pro
tyi vybran hodnoty pracovnho tlaku ve stejnm mst. .I = 300 mA, U ~ 250 V, = 30 sccm.
62
Sondov diagnostika tryskovho vboje v kontinulnm stejnosmrnm reimu
1
10
0
10
]
-3/2
EEPF [(eV)
-1
10
-2
10
-3 80
10 70
0,0 60
0,5 50
40
1,0 30 ]
1,5 20 m
m
[e 10 r[
V] 2,0
0
2,5
Obr. 5.16. Radiln prbh EEPF ve vzdlenosti h = 32 mm pod tryskou pi tlaku p = 10 Pa.
I = 300 mA, U ~ 250 V, = 30 sccm.
63
6. Sondov diagnostika tryskovho vboje v pulznm
stejnosmrnm reimu
64
Sondov diagnostika tryskovho vboje v pulznm stejnosmrnm reimu
65
Sondov diagnostika tryskovho vboje v pulznm stejnosmrnm reimu
0 3
-200
2
UKat [V]
I [A]
-400
1
-600
0
-800
-100 -50 0 50 100 150 200 250
t [s]
Obr. 6.1. asov prbh napt na katod (ern) a vbojovho proudu (mode) pi pulznm buzen
s frekvenc f = 1 kHz, stdou S = 10% a naptm na zdroji U = 450 V. as t = 0 s odpovd zatku
pulzu. Na kivce napt v neaktivn sti periody je patrn zvlnn o frekvenci 100 kHz. Jde o
nedouc penos ze zdroje vysokho napt, kter vak vsledky provedench men neovlivuje.
2,0
f = 1kHz
5000Hz
250Hz
1,5 100Hz
1,0
I [A]
0,5
0,0
0,0 0,5 1,0
t [ms]
Obr. 6.2. Porovnn asovho prbhu vbojovho proudu v aktivn sti periody pro vybran budic
frekvence. S = 10 %, U = 450 V.
66
Sondov diagnostika tryskovho vboje v pulznm stejnosmrnm reimu
67
Sondov diagnostika tryskovho vboje v pulznm stejnosmrnm reimu
1,5
1,0
V pl[V]
0,5
0,0
-0,5
0
20
r[mm]
40
60
80
250 300
150 200
100 50 100
0
t[s]
Obr. 6.3. Prbh potencilu plazmatu v ase od zatku napovho pulzu na katod a v radiln
souadnici ve vzdlenosti h = 30 mm pod tryskou. Pulzn buzen: U = 450 V (I ~ 215 mA), f = 1 kHz,
S = 10 %. Pracovn podmnky: p = 1,5 Pa, Ar = 80 sccm. Zachycena je pouze st periody.
2,0 2,0
r = 1mm
21mm t = 25s
1,5 1,5 50s
41mm
66mm 100s
1,0 106mm 1,0 125s
Vpl [V]
200s
Vpl [V]
0,0 0,0
-0,5 -0,5
0 200 400 600 800 1000 -20 0 20 40 60 80 100 120
t [s] r [mm]
Obr. 6.4. ezy grafem na obr. 6.3: vlevo asov prbh potencilu plazmatu v rznch vzdlenostech
od osy a vpravo radiln prbh v nkolika vybranch asech v rmci jedn periody pulznho buzen.
68
Sondov diagnostika tryskovho vboje v pulznm stejnosmrnm reimu
2,0 p = 1.25Pa
2Pa
3Pa
1,5 4Pa
5Pa
Vpl [V]
1,0
0,5
0,0
0 200 400 600 800 1000
t [s]
Obr. 6.5. asov prbh potencilu plazmatu pi nkolika rznch pracovnch tlacch. Zmeno v ose
trysky, h = 30 mm. Parametry vboje: U = 450 V (I ~ 195 mA), f = 250 Hz, S = 10 %, Ar = 80 sccm.
6.2. EEPF
69
Sondov diagnostika tryskovho vboje v pulznm stejnosmrnm reimu
0
10
]
-3/2
EEPF [(eV)
-1
10
t = 50s
-2
10 100s
150s
200s
250s
-3
10
0,0 0,5 1,0 1,5
[eV]
Obr. 6.6. Vvoj EEPF ve vybranch asech v rmci jedn periody pulznho buzen v bod r = 0 mm,
h = 30 mm. U = 450 V (I ~ 185 mA), f = 1 kHz, S = 10 %, p = 1 Pa, Ar = 80 sccm. Konci
napovho pulzu odpovd as t = 100 s.
70
Sondov diagnostika tryskovho vboje v pulznm stejnosmrnm reimu
1
10
0
10
]
-3/2
EEPF [(eV)
-1
10
-2
10
1000
800 0,0
600 0,5
400
t[
1,0 V]
s
200 [e
]
1,5
Obr. 6.7. Vvoj EEPF v ase od zatku pulzu na katod. Sonda v bod r = 0 mm, h = 30 mm.
U = 450 V (I ~ 195 mA), f = 500 Hz, S = 10 %, p = 1 Pa, Ar = 80 sccm. Konci napovho pulzu
odpovd as t = 200 s.
Na obr. 6.8 je ukzno srovnn EEPF pro rzn budic frekvence v ase
t = 100 s po zatku pulzu. Protoe stda byla ve vech ppadech zvolena shodn
S = 10 %, je u jednotlivch frekvenc rozdln doba neaktivn sti periody, je
pedchz zatku pulzu. Zobrazen EEPF se v podstat nijak neli. To znamen, e
del neaktivn fze se, alespo v pozorovanm oboru frekvenc, neprojev
v pomalejm nbhu aktivn fze periody.
1
10
0
10
]
-3/2
EEPF [(eV)
-1
10
-2
10 1kHz
500Hz
250Hz
-3
100Hz
10
0,0 0,5 1,0 1,5 2,0 2,5
[eV]
Obr. 6.8. EEPF v ase t = 100 s pi rznch frekvencch buzen. Sonda byla umstna v bod
r = 0 mm, h = 30 mm. U = 450 V (I ~ 200 mA), S = 10 %, p = 1 Pa, Ar = 80 sccm.
71
Sondov diagnostika tryskovho vboje v pulznm stejnosmrnm reimu
3
4E+17
n e[10 m ]
-3
3E+17
17
2
2E+17
1
1E+17
0
0
20
40 1000
r[m
800
60
m]
600
80 400
200 ]
0
t[s
100
Obr. 6.9. Elektronov koncentrace v ase a v radilnm smru pro h = 30 mm. Zobrazeno v linernm
mtku. U = 450 V (I ~ 215 mA), f = 1 kHz, S = 10 %, p = 1 Pa, Ar = 80 sccm.
72
Sondov diagnostika tryskovho vboje v pulznm stejnosmrnm reimu
obr. 6.10 a obr. 6.11 pedstavuj ezy grafem na obr. 6.9 pro nkolik vybranch
radilnch souadnic a asovch okamik. Ve vnitnm prostoru dut katody dochz
se zatkem napovho pulzu ke skokovmu nrstu elektronov koncentrace.
Generovan plazma je v dsledku proudn z trysky odneno do objemu reaktoru,
kde dle jako dohasnajc plazma tak difunduje smrem ke stnm. Prbh a tempo
rstu elektronov koncentrace proto zvis na axiln i radiln vzdlenosti od st
trysky. V principu se asov vvoj elektronov koncentrace chov podobn jako
asov vvoj iontovho proudu tekoucho na sondu pi men rychlosti proudn
plazmatu popsanm v kapitole 4.1. V oblasti nejintenzivnjho proudn pod stm
trysky je pozorovn rychl a strm nbh. Ped dosaenm maxima se rychlost rstu
postupn sniuje. Ve vtch vzdlenostech od st a zejmna dle od osy je naopak
zejm zpodn nrstu elektronov koncentrace. Nrst potom probh pozvolna,
nevrazn maximum se nachz a pomrn hluboko v neaktivn sti periody.
V radilnch vzdlenostech pesahujcch r = 80 mm, kde ji je sonda ble ke stn
ne k trysce, je ped hlavnm maximem zeteln jet mrn nbh trvajc mn ne
100 s. Ten me pedstavovat reakci nbojov vrstvy na plazma pichzejc
v pulzu z katody. Pi delm pulzu se na vysok hodnot elektronov koncentrace
stabilizuje (viz. ne), dle zstv konstantn a po odeznn pulzu nsleduje
monotnn pokles. Samotn ambipolrn difzi pi nemnn elektronov teplot
odpovd exponenciln pokles v ase, v grafu na obr. 6.10 se semilogaritmickm
mtkem by jej vystihovala pmka. Pozorovan zvislost ovem kles pomaleji.
Hlavn pinou je ptomnost slabho kontinulnho vboje na pozad, kter udruje
uritou minimln hladinu koncentrace. V mench vzdlenostech od trysky bylo
pozorovno, e pokles probh ve dvou fzch, resp. piblin po 200 s nastv jeho
zpomalen (lpe to vystihuje obr. 6.14, kde jsou zobrazeny i kivky odpovdajc
nim frekvencm). Takov efekt ji byl zjitn i v pulznm magnetronovm vboji,
nap. v [ada, 2011, Anders, 2009]. Je spojovn s vskytem metastabilnch atom
s dlouhou dobou ivota.
73
Sondov diagnostika tryskovho vboje v pulznm stejnosmrnm reimu
18
10
r = 1mm
6mm
26mm
46mm
10
17
66mm
86mm
106mm
ne [m ]
-3 16
10
15
10
0 200 400 600 800 1000
t [s]
Obr. 6.10. asov prbh elektronov koncentrace ve vzdlenosti h = 30 mm pod stm trysky
s parametrem radiln souadnice r. Kivky v grafu pedstavuj ezy grafem na obr. 6.9 zobrazen
v semilogaritmickm mtku.
4
t = 50s
100s
150s
3 200s
300s
500s
ne [10 m ]
-3
2
17
0
-20 0 20 40 60 80 100 120
r [mm]
Obr. 6.11. Radiln prbh elektronov koncentrace pro h = 30 mm v nkolika vybranch asech od
zatku napovho pulzu na katod. Kivky v grafu pedstavuj ezy grafem na obr. 6.9.
74
Sondov diagnostika tryskovho vboje v pulznm stejnosmrnm reimu
17
10
-3
16
10
0 200 400 600 800 1000
t [s]
Obr. 6.12. asov prbh elektronov koncentrace v radiln vzdlenosti r = 1 mm od osy plazmov
trysky s parametrem axiln souadnice h. U = 450 V (I ~ 215 mA), f = 1 kHz, S = 10 %, p = 1 Pa,
Ar = 80 sccm.
75
Sondov diagnostika tryskovho vboje v pulznm stejnosmrnm reimu
18
10
17
10
ne [m ]
-3
13Pa
7,5Pa
10
16 5Pa
4Pa
3Pa
2Pa
15
1Pa
10
0 200 400 600 800 1000
t [s]
18 18
10 10
17 17
10 10
ne [m ]
ne [m ]
-3
-3
16 16
10 10
15 15
10 10
0 200 400 600 800 1000 0 200 400 600 800 1000
t [s] t [s]
Obr. 6.13. asov vvoj elektronov koncentrace pi rznch tlacch pracovnho plynu a budicch
frekvencch. Jednotliv grafy odpovdaj rznm frekvencm: nahoe f = 1 kHz, dole vlevo f = 250 Hz
a vpravo f = 500 Hz. Zmeno v bod r = 0 mm, h = 30 mm. U = 450 V, S = 10 %, Ar = 80 sccm.
76
Sondov diagnostika tryskovho vboje v pulznm stejnosmrnm reimu
16
10
15
10
0 200 400 600 800 1000 1200 1400
t [s]
Obr. 6.14. Prbh elektronov koncentrace v ase pro tyi vybran budic frekvence. Zmeno v bod
r = 0 mm, h = 30 mm. U = 450 V (I ~ 200 mA), S = 10%, p = 1 Pa, Ar =80 sccm.
77
Sondov diagnostika tryskovho vboje v pulznm stejnosmrnm reimu
0,3
Te [eV]
0,2
0,1
0,0
0 200 400 600 800 1000
t [s]
Obr. 6.15. Prbh elektronov teploty v ase pro tyi vybran budic frekvence. Zmeno v bod
r = 0 mm, h = 30 mm. U = 450 V (I ~ 200 mA), S = 10 %, p = 1 Pa, Ar = 80 sccm.
78
7. Studium magnetronovho vboje hmotnostn
spektroskopi s energetickm rozlienm
79
Hmotnostn spektroskopie s energetickm rozlienm
obr. 3.2). Jako pvod argonu byla vyuita plazmov tryska, kontinuln prtok
argonu v provedench experimentech dosahoval Ar = 30 sccm a Ar =100 sccm.
6
10
5
10
4
I [c/s]
10
3
10
2
10
0 20 40 60 80 100
m [amu]
Obr. 7.1. Hmotnostn spektrum pozitivnch iont kontinulnho stejnosmrnho vboje v magnetronu
s titanovm terem v argonu. Energie pro prchod filtrem nastavena na hodnotu E = 2,8 eV, sonda
spektrometru orientovna proti magnetronu a umstna ve vzdlenosti d = 100 mm. I = 150 mA
(U ~ 220 V), p = 5,5 Pa, Ar = 50 sccm.
7
10 40 2+
Ar 2.7
40 +
Ar
6 41 +
2.8
10 ArH
48 +
Ti
40 +
5 Ar2
10
2.65
I [c/s]
4 2.75
10
2.85
3
10
2
10
0 1 2 3 4 5
E [eV]
Obr. 7.2. Energetick spektrum pro nejvraznj ionty podle hmotnostnho spektra na obr. 7.1.
V grafu jsou vyznaen hodnoty energi, v nich dosahuj kivky maxima. Nejsou zobrazeny kivky
40
pro vechny izotopy Ar a Ti, u dvakrt ionizovanch iont Ar2+ in skuten energie dvojnsobek
hodnoty uveden na ose.
80
Hmotnostn spektroskopie s energetickm rozlienm
81
Hmotnostn spektroskopie s energetickm rozlienm
82
Hmotnostn spektroskopie s energetickm rozlienm
7
10 40 +
Ar p = 1Pa
6
2Pa
10 3Pa
5Pa
5
10 10 Pa
I [c/s]
4
10
3
10
2
10
1
10
0 2 4 6 8 10
E [eV]
40
Obr. 7.3. Energetick spektrum iont Ar+ generovanch v kontinuln buzenm magnetronovm
vboji pi rznch pracovnch tlacch. Meno ve vzdlenosti d = 50 mm od stedu tere. I = 300 mA
(U ~ 250 V), Ar = 50 sccm.
6
10 48 +
Ti p = 1Pa
2Pa
5
10 3Pa
5Pa
10 Pa
4
10
I [c/s]
3
10
2
10
1
10
0 2 4 6 8 10
E [eV]
Obr. 7.4. Energetick spektrum iont 48Ti+. Podmnky men jsou uvedeny v popisku obr. 7.3.
40
U rozdlen iont Ar+ se chvost s vymi energiemi objevuje pouze pi
nejnich studovanch tlacch, s rostoucm tlakem velmi rychle zanik. stice s
vy energi, kter vznikaj obma uvedenmi mechanizmy, jsou zhy
termalizovny. Pi tlaku p = 10 Pa in stedn voln drha piblin l = 1 mm,
zejm existuje jen minimln pravdpodobnost, e si na drze mezi magnetronem a
spektrometrem stice uchov vy energii (pi vzjemn vzdlenosti d = 50 mm je
pravdpodobnost, e stice projde beze srky P exp d / l 2 1022 ).
83
Hmotnostn spektroskopie s energetickm rozlienm
84
Hmotnostn spektroskopie s energetickm rozlienm
3,25
40 +
Ar
48 +
Ti
EAV [eV]
3,00
2,75
0 2 4 6 8 10
p [Pa]
Obr. 7.5. Stedn energie v zvislost na tlaku vyhodnocen z rozdlen energi na obr. 7.3 a obr. 7.4
7 4
10 40 2+ 10 40 +
Ar p = 1 Pa Ar2 p = 1Pa
6
2Pa 2Pa
10 3Pa 3Pa
5Pa 5Pa
5
10 10Pa 10 Pa
I [c/s]
I [c/s]
3
10
4
10
3
10
2 2
10 10
0 4 8 12 16 20 0 2 4 6 8 10
E [eV] E [eV]
4
10 +
ArH p = 1Pa
2Pa
3Pa
5Pa
10 Pa
I [c/s]
3
10
2
10
0 2 4 6 8 10
E [eV]
40
Obr. 7.6. Energetick spektrum iont Ar2+ (vlevo nahoe), 40Ar2+ (vpravo nahoe) a41(ArH)+ (dole).
Podmnky men jsou uvedeny v popisku obr. 7.3.
40
Dvakrt ionizovan ionty Ar2+ zskaj na danm potencilovm rozdlu
oproti ostatnm sledovanm iontm dvojnsobnou energii. Z grafu na obr. 7.6 vlevo
nahoe je patrn, e celkov mnostv iont 40Ar2+, tj. vetn termalizovanch, se pi
vych tlacch razantn zmenuje. Primrnm zdrojem tchto iont je interakce
elektronu a jednou ionizovanho argonovho iontu, piem hodnota druh ionizan
85
Hmotnostn spektroskopie s energetickm rozlienm
4 -1
10 10
I [c/s]
I / Itot
3 -2
10 10
2 -3
10 10
1 -4
10 10
0 2 4 6 8 10 0 2 4 6 8 10
E [eV] E [eV]
48
Obr. 7.7. Porovnn energetickch rozdlen iont Ti+ pi dvou rznch hodnotch vbojovho
proudu. Vlevo jsou vyneseny absolutn hodnoty detekovanho signlu, vpravo hodnoty po
normalizaci. Meno ve vzdlenosti d = 50 mm od stedu tere, p = 1 Pa, Ar = 50 sccm.
48
Obr. 7.7 ukazuje ovlivnn energetickch rozdlen iont Ti+ velikost
vbojovho proudu. Zachycena je situace pro proud I = 300 mA (U ~ 250 V) a
I = 150 mA (U ~ 220 V). Pro lep srovnn jsou zobrazena souasn normalizovan
rozdlen, kde jsou hodnoty vztaeny k ploe pod pslunou kivkou grafu.
Vbojov proud v podstat pedstavuje tok kladnch iont na katodu, a je proto
pirozen, e k nmu pmo mrn narst signl na detektoru. Po normalizaci jsou
rozdlen pro oba proudy prakticky toton. Hlavnm dvodem, pro se rozdlen
neli, je skutenost, e akoli v jednom ppad dosahuje proud dvojnsobn
hodnoty ne ve druhm, napt na katod se zmn jen mrn. To vyplv z vlastnost
magnetronovho vboje. Prv napt na katodovm spdu je primrnm zdrojem
energie stic obsaench v plazmatu. Zvislost energi iont na dodanm vkonu
v obdobnch podmnkch se zabvala prce [Hippler, 2005]. Jej vsledky jsou
v dobr shod s vsledky ve uvedench men.
86
Hmotnostn spektroskopie s energetickm rozlienm
87
Hmotnostn spektroskopie s energetickm rozlienm
od dcho signlu, bylo vdy shodn s dobou zznamu. Vtina men nebyla
provedena pro celou periodu, nicmn vdy byl sledovn signl pichzejc z aktivn
sti periody a zatku a konce neaktivn fze. Energetick rozdlen byla
promovna s krokem 0,2 eV, rozsah mench energi byl vdy stanoven na
zklad pedbnho hrubho men.
Zapojen magnetronu v rmci experimentlnho uspodn pro pulzn buzen
vboje se shodovalo se zapojenm pouitm pi mench s plazmovou tryskou
popsanm v kapitole 6 (pulzn spna byl z dvodu stabilizace vboje pemostn
rezistorem o odporu R = 6,2 ka dal rezistor o odporu R = 110 byl zapojen do
srie). Zdroj pracoval v reimu konstantnho napt. Jeho velikost byla nastavovna
v rozmez U = 400 V a U = 500 V, pi zvolen std S = 10 % se potom velikost
stednho vbojovho proudu pohybovala okolo hodnoty ID ~ 100 mA. Tvar
generovanch napovch pulz na katod a vbojovho proudu bhem aktivn sti
periody je uveden na obr. 7.8. Konstantn prtok argonu inil stejn jako pi
experimentech s kontinuln buzenm vbojem Ar = 30 sccm a Ar =100 sccm.
2,5
0
2,0
-200 1,5
UKat [V]
1,0
-400
I [A]
0,5
-600
0,0
-800 -0,5
-100 -50 0 50 100 150 200 250 300
t [s]
Obr. 7.8. asov prbh napt (ern) a proudu (mode) u magnetronovho vboje pi pulznm
buzen s frekvenc f = 500 Hz, stdou S = 10% a naptm na zdroji U = 400 V. Zobrazena je jen st
periody.
88
Hmotnostn spektroskopie s energetickm rozlienm
7
40 + 10
Ar
6
10
5
10
I [c/s]
4
10
3
10
10
2 350
300
0 250
2 200
4 150
100
E[ 6 50 s]
eV 8 t [
] 0
10
7
48 + 10
Ti
6
10
5
10
I [c/s]
4
10
3
10
10
2 350
300
0 250
2 200
4
6 150
8 100 ]
10 s
E[
eV 12 50 t[
] 14 0
16
7
40 2+
10
Ar
6
10
5
10
I [c/s]
4
10
3
10
10
2 350
0 300
250
4 200
8 150
12 100
E[
eV 50 s]
] 16 t [
0
20
40
Obr. 7.9. asov vvoj energetickch rozdlen iont Ar+, 48
Ti+ a 40
Ar2+ pi frekvenci pulznho
buzen f = 1 kHz a std S = 10 % zaznamenan v mst detektoru. as t = 0 s odpovd nbn
hran napovho pulzu na katod. d = 50 mm, p = 1 Pa, U = 400 V (I ~ 110 mA), Ar = 50 sccm.
89
Hmotnostn spektroskopie s energetickm rozlienm
10
1 3,5 8
40 +
Ar t = 25s
0
100s
10
150s 6
200s 3,0
-1
10 350s
6
-2
10
2,5
2
-3
10
-4
10 2,0 0
0 2 4 6 8 10 0 50 100 150 200 250 300 350 400
E [eV] t [s]
1
10 48 +
Ti t = 25s
2,0
100s 3,5
0
10 150s
200s 1,5
-1
10 350s
1,0
-2 3,0
10
6
-3 0,5
10
-4
10 2,5 0,0
0 2 4 6 8 10 12 14 0 50 100 150 200 250 300 350 400
E [eV] t [s]
1
10 40 2+
7 4
Ar t = 25s
100s
0
10 150s 3
200s 6
10
-1
350s
2
I / Itot
-2
6
10
5
-3
1
10
-4
10 4 0
0 4 8 12 16 20 0 50 100 150 200 250 300 350 400
E [eV] t [s]
40
Obr. 7.10. V levm sloupci normalizovan energetick rozdlen iont Ar+, 48
Ti+ a 40
Ar2+ ve
vybranch asech. Vpravo prbhy stedn energie a celkov hodnoty signlu v ase. Vyhodnoceno
z graf na obr. 7.9.
40
Vsledky men energetickho rozdlen iont Ar+, 48
Ti+ a 40
Ar2+ v ase
pro frekvenci f = 1 kHz jsou zachyceny v grafech na obr. 7.9 a obr. 7.10. V grafech v
levm sloupci na obr. 7.10 jsou vykreslena rozdlen v nkolika vybranch asech,
piem zobrazen data byla na rozdl od obr. 7.9 normalizovna na stejnou plochu
pod jednotlivmi kivkami. Tyto vsledky ukazuj rozdl v energetickch
rozdlench iont v aktivn a neaktivn fzi periody. V neaktivn fzi prakticky
nebylo pozorovno pro dn iont prodlouen energetickho rozdlen k vym
energim, zhy po dosaen nzkoenergetickho maxima nsleduje v rozdlench
prudk pokles a zastoupen na vych energich se bl k nule. Bhem aktivn fze
90
Hmotnostn spektroskopie s energetickm rozlienm
91
Hmotnostn spektroskopie s energetickm rozlienm
vboje. V neaktivn fzi ani bhem pechodu mezi fzemi tedy nevznikaj (ppadn
jen v zanedbatelnm mnostv) ionty s vymi energiemi ne ionty z aktivn sti
pulzu. Neaktivn fze pedstavuje fzi dohasnajcho plazmatu, kdy ji vznikl ionty
ztrcej svoji pvodn energii a zanikaj.
V grafech na obr. 7.11 jsou zobrazeny asov prbhy vyhodnocench
stednch energi iont zrove pro ti rzn situace. Srovnn je provedeno pro
frekvence f = 100 Hz, f = 500 Hz a f = 5 kHz, piem pro frekvenci f = 100 Hz
dosahovalo napt na zdroji hodnoty U = 500 V a stda inila S = 5 %, pro dal dv
frekvence bylo nastaveno napt o velikosti U = 400 V a stda S = 10 %. Ve vech
ppadech byla ovem shodn stedn hodnota vbojovho proudu ID ~ 100 mA.
Zajmav efekt byl pozorovn v prbzch iontovch energi pi frekvenci
f = 500 Hz. Krtce po zatku aktivn fze dochz k prudkmu nrstu stedn
energie iont 48Ti+, zatmco u stedn energie iont 40Ar+ se naopak objevuje pokles.
Toto pozorovn ukazuje na interakci argonovch iont s vy energi s odprenmi
atomy titanu, pi kter probh penos nboje. Uveden jev byl vzn pouze na
frekvence okolo f = 500 Hz. U dvou zbvajcch prbh v grafech, kter odpovdaj
vy a ni frekvenci, dochz pro vechny sledovan ionty k pravidelnmu zven
stedn energie po celou dobu aktivn fze periody bez dalch doprovodnch zmn.
Z graf v pravm sloupci obr. 7.10 lze odest, e pi frekvenci f = 1 kHz se dan
efekt jet slab projevuje. Men byla se stejnm vsledkem zopakovna pro rzn
vzjemn vzdlenosti magnetronu a spektrometru. Pravdpodobnm vysvtlenm
zjitnho chovn je, e pi frekvencch v oblasti kolem f = 500 Hz se z tere
efektivnji odprauj atomy titanu, jejich koncentrace v plazmatu je vy. Na
ionizaci titanovch neutrl se potom vznamnou mrou podlej ptomn ionty
argonu s vy energi. Ve vsledcch men se toto projevuje na zatku pulzu, kdy
jsou spektrometrem detekovny nejrychlej stice, jak ji bylo diskutovno ve.
Stedn energie iont u kivek na obr. 7.11 odpovdajcch frekvenci
f = 100 Hz je znateln vy ne ve zbvajcch dvou ppadech. Zde se nejedn o
ovlivnn budic frekvenc, ale naptm ve vboji. Zvten katodovho spdu
znamen urychlen iont na vy energie. Pomalej pokles stednch energi
v ppad frekvence f = 100 Hz je zpsoben del aktivn st periody. V plazmatu je
v podstat bhem aktivn fze akumulovna vt celkov energie a jej nsledn
uvolnn probh del dobu.
92
Hmotnostn spektroskopie s energetickm rozlienm
40 +
3,5 Ar
3,0
2,0
0 200 400 600 800 1000
t [s]
3,5
3,0
EAV [eV]
2,5
48 +
Ti
2,0
0 200 400 600 800 1000
t [s]
40 2+
3,5 Ar
3,0
EAV [eV]
2,5
2,0
0 200 400 600 800 1000
t [s]
Obr. 7.11. Vvoj stednch energi t studovanch iont v ase pro ti vybran frekvence pulznho
buzen vboje. p = 1 Pa, Ar = 50 sccm, d = 100 mm. Pro frekvence f = 500 Hz a f = 5 kHz
U = 400 V a S = 10 %, v ppad f = 100 Hz U = 500 V a S = 5 %.
93
8. Zvr
94
Zvr
95
Zvr
96
Zvr
97
Literatura
98
Coburn, J.W., A System for Determining the Mass and Energy of Particles Incident
on a Substrate in a Planar Diode Sputtering System, Rev. Sci. Instrum., 41, 1219-
1223, 1970
Craig, S., G.L. Harding, Effects of argon pressure and substrate temperature on the
structure and properties of sputtered copper films, J. Vac. Sci. Technol., 19, 205-
215, 1981
ada, M., Z. Hubika, P. Admek, P. Ptek, H. chov, M. cha, L. Jastrabk,
Investigation of RF and DC plasma jet system during deposition of highly
oriented ZnO thin films, Surf. Coat. Tech., 174-175, 627-631, 2003
ada, M., Z. Hubika, V. Kuliovsky, P. Admek, J. Olejnek, P. Boh, Plasma
diagnostics in the pulse magnetron sputtering system used for deposition Ti-C:H
thin films, Surf. Coat. Tech., 200, 3861-3867, 2006
ada, M., P. Virostko, . Kment, Z. Hubika, Comparative Study of Total Power
Density at a Substrate in Pulsed DC Magnetron and Hollow-Cathode Plasma Jet
Sputtering Systems, Plasma Process. Polym., 6, S247-S252, 2009
ada, M., Z. Hubika, P. Admek, J. Kluso, L. Jastrabk, Time-resolved plasma
parameters in the HiPIMS discharge with Ti target in Ar/O2 atmosphere, Surf.
Coat. Tech., 205, S317-S321, 2011
Druyvesteyn, M.J., Zeitschr. f. Fyzik, 64, 781, 1930
Ellmer, K., Magnetron Sputtering of transparent zinc oxide: relation between the
sputtering parameters and the electronic properties, J. Phys. D: Appl. Phys., 33,
R17-R32, 2000
Ellmer, K., R. Wendt, K. Wiesemann, Interpretation of ion distribution functions
measured by a combined energy and mass analyzer, International Journal of
Mass Spectrometry, 223-224, 679-693, 2003
Ellmer, K., Magnetron Discharges for Thin Film Deposition v Low Temperature
Plasma, Fundamentals, Technologies, and Techniques, Vol. 1, ed. R. Hippler,
H. Kersten, M. Schmidt, K.H. Schoenbach, Willey-VCH, Berlin, 2007
Godyak, V.A., Piejak, R.B., Alexandrovich, B.M., Measurements of electron energy
distribution in low-pressure RF discharges, Plasma Sources Sci Technol., 1, 36-
58, 1992
Helmersson, U., M. Lattemann, J. Bohlmark, A.P. Ehiasarian, J.T. Gudmundsson,
Ionized physical vapor deposition (IPVD): A review of technology and
applications, Thin Solid Films, 513, 1-24, 2006
99
Hippler, R., S. Wrehde, V. Strak, O. Zhigalov, H. Steffen, M. Tich, M. Quaas, H.
Wulff, Characterization of a Magnetron Plasma for Deposition of Titanium
Oxide and Titanium Nitride Films, Contrib. Plasma Phys., 45, 348-357, 2005
Hubika, Z., Hollow cathodes and plasma jets for thin film deposition v Low
Temperature Plasma, Fundamentals, Technologies, and Techniques, Vol. 1, ed.
R. Hippler, H. Kersten, M. Schmidt, K.H. Schoenbach, Willey-VCH, Berlin,
2007
Hubika, Z., M. Chichina, A. Deyneka, P. Kudrna, J. Olejnek, H. chov, M.
cha, L. Jastrabk, P. Virostko, P. Admek, M. Tich, Low pressure plasma-jet
systems and their application for deposition of ceramic thin films, Journal of
Optoelectronics and Advanced Materials, 9, 875-880, 2007
Hubika, Z., P. Virostko, M. Tich, M. ada, P. Admek, J. Olejnek, A. Deyneka,
O. Churpita, V. Valvoda, L. Jastrabk, Deposition of BaxSr1-xTiO3 thin Films by
Double RF Hollow Cathode Plasma Jet System, Contrib. Plasma. Phys., 48,
515-520, 2008
Ivanov, I., P. Kazansky, L. Hultman, I. Petrov, J.E. Sundgren, Influence of an
external axial magnetic field on the plasma characteristics and deposition
conditions during direct current planar magnetron sputtering, J. Vac. Sci.
Technol. A, 12, 314-320, 1994
Johnsen, R., A. Chen, M.A. Biondi, Three-body association reactions of He+, Ne+,
and Ar+ ions in their parent gases from 78 to 300 K, J. Chem. Phys., 73, 1717-
1720, 1980
Jousten, K., Handbook of Vacuum Technology, Willey-VCH, Weinheim, 2008
Kadlec, S., C. Quaeyhaegens, G. Knuyt, L.M. Stals, Energy distribution of ions in an
unbalanced magnetron plasma measured with energy-resolved mass spectrometry,
Surf. Coat. Tech., 89, 177-184, 1997
Kazemeini, M.H., A.A. Berezin, N. Fukuhara, Formation of thin TiNxOy films by
using a hollow cathode reactive DC sputtering system, Thin Solid Films, 372,
70-77, 2000
Kelly, P.J., R.D. Arnell, Magnetron sputtering: a review of recent developments and
applications, Vacuum, 56, 159-172, 2000
Kickuth, R., Plasma Technology, Process Diversity + Sustainability, BMBF, Bonn,
2001
100
Klawuhn, E., G.C. DCouto, K.A. Ashtiani, P. Rymer, M.A. Biberger, K.B. Levy,
Ionized physical-vapour deposition using a hollow cathode magnetron source for
advanced metallization, J. Vac. Sci. Technol. A, 18, 1546-1549, 2000
Klopfenstein, R., Air velocity and flow measurements using a Pitot tube, ISA
Transactions, 37, 257-263, 1998
Kluso, J., P. Kudrna, A. Kolpakov, I. Pickov, Z. Hubika, M. Tich,
Experimental study of the discharge in the low pressure plasma jet sputtering
system, Contrib. Plasma Phys., 53, 10-15, 2013a
Kluso, J., P. Kudrna, M. Tich, Measurement of the plasma and neutral gas flow
velocities in a low-pressure hollow-cathode plasma jet sputtering system,
Plasma Sources Sci. Technol., 22, 015020, 2013b
Kolev, I., A. Bogaerts, Numerical study of the sputtering in a dc magnetron, J. Vac.
Sci. Technol. A, 27, 20-28, 2009
Kouznetsov, V., K. Mack, J.M. Schneider, U. Helmersson, I. Petrov, A novel pulsed
magnetron sputter technique utilizing very high target power densities, Surf.
Coat. Tech., 122, 290-293, 1999
Krack, J., Elektrick vboje, SNTL, Praha, 1964
Kudrna, P., Sondov diagnostika nzkoteplotnho plazmatu, Doktorsk disertan
prce, Univerzita Karlova v Praze, 1997
Kudrna, P., J. Kluso, S. Leshkov, M. Chichina, I. Pickov, Z. Hubika, M. Tich, A
Study of Plasma Parameters in Hollow Cathode Plasma Jet in Pulse Regime,
Contrib. Plasma Phys., 50, 886-891, 2010
Langmuir, I., H.M. Mott-Smith, The theory of collectors in gaseous discharges, Phys.
Rev., 28, 727-763, 1926
Leshkov, S., P. Kudrna, M. Chichina, J. Kluso, I. Pickov, Z. Hubika, M. Tich,
Spatial Distribution of Plasma Parameters in DC-Energized Hollow Cathode
Plasma Jet, Contrib. Plasma Phys., 50, 878-885, 2010
Little, P. F, A. von Engel, The hollow-cathode effect and the theory of glow
discharges, Proceedings of the Royal Society of London. Series A, Mathematical
and Physical Science, 224, 209-227, 1954
Martiovit, V., Zklady fyziky plazmy, Univerzita Komenskho, Bratislava, 2006
Massey, J.T., S.M. Canon, Constricted Discharges in the Rare Gases. I.
Spectroscopic and Electrical Measurements, J. Appl. Phys., 36, 361-372, 1965
101
McLeod, P.S., L.D. Hartsough, High-rate sputtering of aluminium for metallization
of integrated circuits, J. Vac. Sci. Technol., 14, 263-265, 1977
Miina, M., L.R. Shaginyan, M. Maek, P. Panjan, Energy resolved ion mass
spectroscopy of the plasma during reactive magnetron sputtering, Surf. Coat.
Tech., 142-144, 348-354, 2001
Musil, J., P. Baroch, J. Vlek, K.H. Nam, J.G. Han, Reactive magnetron sputtering of
thin films: present status and trends, Thin Solid Films, 475, 208-218, 2005
Niedrist, R., Plasmadiagnostik an einer Titan-Sputterquelle, Diplomov prce,
Universitt Innsbruck, 2011
Novk, M., M. cha, V. Kapika, L. Jastrabk, L. Soukup, Z. Hubika, M. Klma, P.
Slavek, A. Brablec, J. Phys. IV France, 7, 331-339, 1997
Olthoff, J.K., R.J. Van Brunt, S.B. Radovanov, J.A. Rees, R. Surowiec, Kinetic-
energy distributions of ions sampled from argon plasmas in a parallel-plate,
radio-frequency reference cell, J. Appl. Phys., 72, 115-125, 1994
Passoth, E., J.F. Behnke, C. Csambal, M. Tich, P. Kudrna, Y.B. Golubovskii, A.
Porokhova, Radial behaviour of the electron energy distribution function in the
cylindrical magnetron discharge in argon, J. Phys. D: Appl. Phys., 32, 2655-
2665, 1999
Penning, F.M., J.H.A. Moubis, Proc. Kon. Ned. Akad. Wetensch, 43, 41-56, 1940
Petrov, G.M., C.M. Ferreira, Numerical modeling of the constriction of the dc
positive column in rare gases, Physical review E, 59, 3571-3582, 1999
Pfau, S., M. Tich, Langmuir probe diagnostics of low-temperature plasmas v Low
Temperature Plasma, Fundamentals, Technologies, and Techniques, Vol. 1, ed.
R. Hippler, H. Kersten, M. Schmidt, K.H. Schoenbach, Willey-VCH, Berlin,
2007
Raizer, Y. P., Gas Discharge Physics , vol. 1, Springer, Berlin, 1991
Rossnagel, S.M, H,R. Kaufman, Charge transport in magnetrons, J. Vac. Sci.
Technol. A, 5, 2276-2279, 1987
Rossnagel, S.M., J. Hopwood, Magnetron sputter deposition with high levels of
metal ionization, Appl. Phys. Lett., 63, 3285-3287, 1993
Roth, J.R., Industrial Plasma Engineering, vol.1, Institute of Physics Publishing,
2000
Savvides, N., B. Window, Unbalanced magnetron ion-assisted deposition and
property modification of thin films, J. Vac. Sci. Technol. A, 4, 504-508, 1986
102
Schrittwieser, R., Course on plasma diagnostics. Text k pednce na Leopold-
Franzens Universitt Innsbruck, 2004
Schrittwieser, R., C. Ionita, A. Murawski, C. Maszl, M. Asandulesa, A. Nastuta, G.
Rusu, C. Douat, S.B: Olenici, I. Vojvodic, M. Dobromir, D. Luca, S. Jaksch, P.
Scheier, Cavity-hollow cathode sputtering source for titanium films, J. Plasma
Physics, 76, 655-664, 2010
Selezneva, S.E., M.I. Boulos, Supersonic induction plasma jet modeling, Nuclear
Instruments and Methods in Physics Research B, 180, 306-311, 2001
Sellers, J., Asymetric bipolar pulsed DC: the enabling technology for reactive PVD,
Surf. Coat. Tech., 98, 1245-1250, 1998
Sigmund, P., Theory of Sputtering. I. Sputtering Yield of Amorphous and
Polycrystalline Targets, Phys. Rev., 184, 383-416, 1969
Soukup, L., M. cha, F. Fendrych, L. Jastrabk, Z. Hubika, D. Chvostov, H.
chov, V. Valvoda, A. Tarasenko, V. Studnika, T. Wagner, M. Novk,
Copper nitride thin films prepared by the RF plasma chemical reactor with low
pressure supersonic single and multi-plasma jet system, Surf. Coat. Tech., 116-
119, 321-326, 1999
Sderstrm, D., Modelling and Application of the Hollow Cathode Plasma,
Disertan prce, Uppsala Universitet, 2008
Strak, V., A.P. Herrendorf, S. Drache, M. ada, Z. Hubika, M. Tich, R. Hippler,
Highly ionized physical vapour deposition plasma source working at very low
pressure, Appl. Phys. Lett., 100, 141604, 2012
cha, M., Z. Hubika, M. Tich, M. Novk, L. Soukup, L. Jastrabk, J.F. Behnke, V.
Kapika, K. Kapoun, M. er, The Interaction of the Supersonic Plasma-Jet with
the Substrate in the RF Plasma-Chemical Reactor, Contrib. Plasma Phys., 36,
605-611, 1996
Tich, M., M. cha, L. Brdo, L. Soukup, L. Jastrabk, K. Kapoun, J. Tou, Z.
Mazanec, R.J. Soukup, A Study of the Gas Flow in the RF Low-Pressure
Supesonic Jet Plasma Chemical System, Contrib. Plasma Phys., 34, 765-772,
1994
Tich, M., Z. Hubika, P. Virostko, I. Pickov, M. cha, M. ada, J. Olejnek, O.
Churpita, L. Jastrabk, P. Admek, P. Kudrna, J. Kluso, S. Leshkov, M.
Chichina, . Kment, Langmuir Probe Measurements of Spatial Plasma Profiles
103
and Temporal Dependences in a DC-Energized Hollow-Cathode Plasma Jet
System, J. Plasma Fusion Res. Series, 8, 1277-1282, 2009
Tich, M., Z. Hubika, M. cha, M. ada, J. Olejnek, O. Churpita, L. Jastrabk, P.
Virostko, P. Admek, P. Kudrna, S. Leshkov, M. Chichina, . Kment, Langmuir
probe diagnostics of a plasma jet system, Plasma Sources Sci. Technol., 18,
014009, 2009b
Toader, E.I., W.G. Graham, C.M.O. Mahony, P.G. Steen, Characterization of high-
density, direct-current reflex discharge plasma source operating in Ar and N2,
Rev. Sci. Instrum., 73, 2974-2980, 2002
Virostko, P., Z. Hubika, P. Admek, M. ada, J. Olejnek, M. Tich, M. cha,
Measurement of Plasma Parameters in Low Temperature High Density Hollow
Cathode Plasma Jet Working in Magnetic Field, Contrib. Plasma Phys., 46, 445-
450, 2006
Virostko, P., Z. Hubika, M. ada, P. Admek, . Kment, M. Tich, L. Jastrabk,
Electrical Probe Diagnostic of the Hollow Cathode Plasma Jet System for
Deposition of TiOx Thin Films, Contrib. Plasma Phys., 48, 527-533, 2008
Vlek, J., P. Kudlek, K. Burcalov, J. Musil, High-power pulsed sputtering using a
magnetron with enhanced plasma confinement, J. Vac. Sci. Technol. A, 25, 42-
47, 2007
Vojvodic, I., S:B. Olenici, C. Ionita, S. Jaksch, P. Balan, B. Rasul, P. Scheier, R.
Schrittwieser, Titanium thin films sputtered by a cavity hollow cathode
discharge on highly oriented pyrolytic graphite, Proc. 28th Int. Conf. on Plasma
and Ionized Gases, Praha, 657-660, 2007
Waits, R.K., Planar magnetron sputtering, J. Vac. Sci. Technol., 15, 179-187, 1978
Window, B., N. Savvides, Charged particle fluxes from planar magnetron sputtering
sources, J. Vac. Sci. Technol. A, 4, 196-202, 1986
Yushkov, G.Y., A. Anders, Origin of the Delayed Current Onset in High-Power
Impulse Magnetron Sputtering, IEEE Transactions on Plasma Science, 38,
3028-3034, 2010
Zhang, J., J. Sun, D. Wang, X. Wang, A novel cold plasma jet generated by
atmospheric dielectric barrier capillary discharge, Thin Solid Films, 506-507,
404-408, 2006
104
Publikace autora disertan prce
A5. Kluso, J., P. Kudrna, M. Tich, Measurement of the plasma and neutral gas
flow velocities in a low-pressure hollow-cathode plasma jet sputtering system,
Plasma Sources Sci. Technol., 22, 015020, 2013
105
C) Publikace ve sborncch konferenc
C5. J. Kluso, P. Kudrna, M. Tich: Measurement of the gas flow velocity in the low
pressure hollow cathode plasma jet sputtering system, v WDS11 Proceedings of
contributed papers: Part II- Physics of Plasmas and Ionized Media, ed. J.
afrnkov a J. Pavl, Praha, Matfyzpress, 125-130, 2011
106
C7. R. Perekrestov, P. Kudrna, J. Kluso, M. Tich, Ti/TiO2 Thin Films Deposition
by Means of Hollow Cathode Plasma Jet, v WDS12 Proceedings of contributed
papers: Part II- Physics of Plasmas and Ionized Media, ed. J. afrnkov a J.
Pavl, Praha, Matfyzpress, 3843, 2012
107
Seznam pouitch symbol
plocha
prez trysky
P, AS plocha sondy, plocha nbojov vrstvy okolo sondy
B magnetick indukce
C elektrick kapacita
d prmr, prmr sondy, vzdlenost
dIN, dOUT vnitn a vnj prmr
e = 1,60210-19 C elementrn nboj
E intenzita elektrickho pole, energie
Eb povrchov vazebn energie
Ei ionizan energie
Es subliman energie
Et transmisn energie pro prchod energetickm analyztorem plazma monitoru
f frekvence
fE energetick rozdlovac funkce
fv rychlostn rozdlovac funkce
h vka, vertikln vzdlenost od st trysky
i hustota elektrickho proudu
I elektrick proud
I intenzita signlu detekovanho plazma monitorem (etnost)
ID elektrick proud na detektoru plazma monitoru
Ip, Ipi, Ipe, sondov proud, jeho iontov a elektronov sloka
j, j+, j- hustota toku, hustota toku kladnch, resp. zpornch nosi nboje
j0 hustota toku primrnch elektron
kB = 1,3810-23 JK-1 Boltzmannova konstanta
l dlka, dlka sondy, stedn voln drha
m hmotnost
mi, me hmotnost iont, resp. elektron
M Machovo slo
n koncentrace
108
ni, ne koncentrace iont, elektronov koncentrace
p tlak
pIN tlak uvnit trysky
pS, pD, pT statick, dynamick a celkov tlak v teorii Pitotovy trubice
P vkon
P pravdpodobnost srky
q elektrick nboj
r polomr
r radiln vzdlenost od osy trysky
ra vektor prostorov souadnice
rL Larmorv polomr
rp charakteristick rozmr sondy, polomr sondy
rs polomr nbojov vrstvy kolem Langmuirovy sondy
R elektrick odpor
R = 8,31 Jmol-1K-1 univerzln plynov konstanta
s drha
sext, sen, sm, sd drha iontu jednotlivmi stmi plazma monitoru
S erpac rychlost, stda
t as
tD, tTD doba zznamu a celkov doba zznamu pi asov rozliench mench
s plazma monitorem
tFS doba letu iont spektrometrem
T teplota, perioda
Te elektronov teplota
Tm teplota tn
Te,eff efektivn hodnota elektronov teploty
Tp dla aktivn sti periody
U elektrick napt
UO extrakn napt na vstupu plazma monitoru
Up napt na sond, prrazn napt
UZ zpaln napt
v rychlost
va rychlost zvuku
vD driftov rychlost
109
V elektrostatick potencil
Vfl plovouc potencil
Vpl potencil plazmatu
Vext, Vax, Vte, Vdyn potencil na danch prvcch plazma monitoru
110
Piloen publikace
3. Kluso, J., P. Kudrna, M. Tich, Measurement of the plasma and neutral gas flow
velocities in a low-pressure hollow-cathode plasma jet sputtering system,
Plasma Sources Sci. Technol., 22, 015020, 2013
111
Contrib. Plasma Phys. 50, No. 9, 886 891 (2010) / DOI 10.1002/ctpp.201010150
c 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1 Introduction
The thin lm deposition can be effectively pursued by methods, which use low temperature plasma. These
methods enable deposition of thin layers with various physical and chemical properties. As a result the low
temperature plasma deposition has a wide range of industrial applications. Very often plasma-aided technologies
are used in engineering for coating objects by hard protective, wear-resistant, corrosion resistant or low friction
layers. Plasma deposition plays another important role in microelectronics where materials with specic optical
or electrical properties are deposited (e.g. in the process of manufacturing of integrated circuits). Deposition
process can be carried out also onto thermal sensitive substrates e.g. onto polymer substrates. This is facilitated
by the fact that the low temperature plasma used in plasma technology is not in thermal equilibrium. While the
effective temperature of the neutral gas is very often at the level of the room temperature, electron temperature
reaches the value in the order of electron-volts.
Among a number of technologies for the plasma deposition which have been developed the low pressure
hollow cathode plasma jet sputtering system is a very promising one. Physically it is based on the hollow cathode
discharge that has been described in the literature [1]. Since a well dened plasma channel is formed the plasma
jet systems are especially suitable for deposition on substrates with complicated shapes [2]. This is considered
to be their main advantage. The low pressure hollow cathode plasma jet sputtering system has been already used
for physical vapour deposition of many different materials. For example we can mention tribological layers of
TiO2 [3], TiN [4], piezoelectric layers [5] of ZnO or ferroelectric layers of SrTiO3 (STO) and Bax Sr1x TiO3
(BSTO) ceramics [6].
Big inuence on the deposition process has the excitation of the discharge. In some cases the operation in
pulsed regime can be especially convenient. Its main importance is in deposition of non-conducting layers where
it prevents arcing. Further, by using the pulsed regime a signicant reduction of the heat load onto the substrate
Corresponding author: E-mail: pavel.kudrna@mff.cuni.cz, Phone: +420 221 912 225, Fax: +420 284 685 095
c 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
112
Contrib. Plasma Phys. 50, No. 9 (2010) / www.cpp-journal.org 887
can be achieved. Simultaneously, during the pulse-on time the plasma density can be considerably higher than in
continuous regime at the same average discharge power. That promotes formation of more dense and stable lm
structure. Since the ion/neutral ux to the substrate is controlled too, the substrate ion bombardment is limited
and the comprehensive stress in the lm is reduced. For further information see e.g. [7].
2 Experimental setup
The experimental setup is schematically depicted in Fig.1. The stainless steel vacuum apparatus is constructed for
UHV conditions. It is continuously pumped by oil-free pumping system consisting of two turbomolecular pumps
backed by a piston pump. As the vacuum apparatus includes two chambers - the main chamber which serves as the
reactor vessel and the differentially-pumped mass-spectrometer chamber - a turbomolecular pump is connected
to each chamber. The ultimate pressure of this system reaches the order of 105 Pa. During the experiment the
working pressure was held at the value from several units of Pa to several tens of Pa. For measuring pressure in
this range a membrane vacuum meter was used. Accurate pressure was adjusted by means of gate-valve between
pump and main chamber. The plasma jet is led into the main chamber vertically. Its basic element is a water
cooled nozzle made of pure Ti which works as a hollow cathode and which is reactively sputtered. It has the
inner diameter of 4 mm and the length of 40 mm. Our experimental setup enables feeding two different gases
into the reactor continuously; their ow is controlled by two separate mass ow controllers. During presented
measurements only argon continuously owed through the hollow cathode with the ow rate of 30 sccm.
The pulsed voltage was applied directly to the nozzle. For creation of power pulses the negative DC power
supply was connected to battery of 33 capacitors with capacity of 500 F which were being discharged via
the IGBT switch controlled by the TTL generator. To obtain better stability of the pulsed discharge a 3.8 k
resistor was connected across the switch. The voltage of the DC power supply was adjusted in all experiments
at the minimal value for which the discharge was stable (it was approximately 400 V). All measurements were
performed at voltage pulses repetition frequency of 1 kHz and the pulse width 100 s (duty cycle 10 %) or 50 s
(duty cycle 5 %). In these conditions the mean discharge current varied between 150 and 250 mA. Instantaneous
potential of the nozzle was measured by oscilloscopic probe, instantaneous discharge current by a Hall sensor.
The time dependence of the nozzle potential and the discharge current is shown in Fig. 2. The increase of the
nozzle potential towards 0 V after switch off the voltage pulse is probably caused by positive charge in the cathode
region of the discharge.
Ar Stainless steel cover
25 Isolation amplifier
Cooling water 200k ISO 100CP
- 20k D/A
+
3.8k
Pulse
switch PA41
ADLINK DAQ 2010
Copper blocks
PC+
Insulating cover
Ti nozzle
- + Langmuir probe
U R A/D
Plasma jet
-
Negative DC power supply Substrate, grounded wall, + TL071
reference electrode Current-voltage converter
Fig. 1 Scheme of experimental setup for Langmuir probe measurements of plasma parameters in the low temperature hollow
cathode plasma jet system in pulsed regime. For enhancing the stability of the discharge a 3.8 k resistor is connected across
the pulse switch. (Color gure: www.cpp-journal.org).
www.cpp-journal.org
c 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
113
888 P. Kudrna et al.: Hollow cathode plasma jet in pulse regime
3 Probe measurements
Plasma parameters were studied by means of Langmuir probe diagnostics. The cylindrical probe was realized
by a tungsten wire with diameter of 30 or 50 m. The probe length of 3 mm ensured sufcient spatial resolution
and maximal probe current in the order of several mA which was well measurable by the used electronics.
The probe had horizontal orientation and was mounted on horizontally movable feedthrough. Considering the
dimensions of main chamber probe measurements could be performed up to the distance of 130 mm far from the
vertical axis of the plasma jet. The vertical distance between the nozzle end and the probe was 20 mm. Probe
characteristics were acquired by electronic circuit shown in the right part of Fig. 1. Measurements were controlled
by PC with Adlink DAQ 2010 card with A/D converter maximal sampling rate of 2 MSamples/s and resolution
of 14 bits. Output voltage on D/A converter was isolated by optically-coupled isolation amplier (Burr-Brown
ISO100) and connected to high voltage operational amplier (Apex PA41). This amplier, powered by oating
electrochemical supply a battery of galvanic cells set the probe voltage. Probe current was converted to
voltage using an operational amplier TL071 and measured by A/D converter of the DAQ card. As a reference
electrode we have used the grounded substrate table and reactor vessel as depicted in Fig. 1. All voltages shown
in Figs. 2 through 6 are therefore referenced to the grounded main vacuum chamber.
Fig. 2 The discharge voltage and current vs. time. Since Fig. 3 Time evolution of Langmuir probe characteristics
a 3.8 k resistor is connected across the switch certain a during one period. The highest probe current is measured
negative voltage is present on the nozzle during the pulse- at the end of the 100 s pulse. (Color gure: www.cpp-
off time. (Color gure: www.cpp-journal.org). journal.org).
During measurements the probe voltage was changed in 20 mV steps and for each voltage value the time
dependence of the probe current was recorded at the maximum sampling rate of the A/D converter starting at the
leading edge of the cathode voltage pulse. Consequently, at 1 kHz pulse frequency approximately 2000 probe
current values were collected at a particular probe voltage within the discharge repetition period of 1 ms. That
was repeated for each value of probe voltage in between 2 V and 2 V with reference to grounded walls of the
discharge chamber, i.e. for 200 probe bias voltages. In such a way we obtained a 2D array of 2000200 probe
current values in dependence on time and voltage. By swapping the subscripts corresponding to time and voltage
the set of 2000 current-voltage characteristics was obtained with the delay from the beginning of the discharge
pulse as a parameter. Several characteristics representing the time evolution during one discharge period are
shown in Fig. 3. Such a big number of characteristics had to be evaluated automatically. We used our own
software written in Agilent VEE. In evaluation of a characteristic by this software at rst the oating and plasma
potentials were determined. Plasma potential Vpl was determined from the maximum of the rst derivative. Then
the electron density ne was evaluated according to the Langmuir theory for collisionless movement of charge
carriers in a thick sheath, so called orbital motion limit [8, 9]. Using this theory we can assume that in the region
of the saturated electron current the dependence of the square of electron current Ip on probe potential Vp is
linear:
2q03 A2p 2
Ip2 (Vp ) = n Vp (1)
2 me e
c 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.cpp-journal.org
114
Contrib. Plasma Phys. 50, No. 9 (2010) / www.cpp-journal.org 889
where q0 and me are the electron charge and the electron mass, Ap is the probe area. Therefore the obtained data
were tted by straight line and from its slope the electron density was calculated. The electron temperature was
evaluated from the slope of electron retarding current with the help of formula:
q0 d
Te = ln(|Ie |) (2)
kB dVp
where kB is the Boltzmann constant and Ie is the electron component of the probe current.
Fig. 4 The time dependence of the electron density at different pressures (left) and different probe positions. The pulse
repetition frequency 1 kHz, pulse width 100 s (duty cycle 10 %). To enhance the readability of the graphs only each tenth
point is depicted. (Color gure: www.cpp-journal.org).
Left graph in Fig. 4 shows the typical course of the electron density in semilogarithmic scale during one
period for several different pressures. The distance between the probe and the vertical axis of the plasma jet was
approximately 20 mm. The ow rate of the working gas was kept constant; we gradually increased the pressure
in the reactor by closing the gate-valve between main chamber and pump. The character of the time dependence
is similar for all measurements. While during the pulse-on time the electron density is rising up, it is falling down
immediately after switching off the voltage pulse. For 100 s on-time no saturation of the electron density was
observed. Maximal values of the electron density are in the order of almost 1018 m3 . They are considerably
higher than the values obtained in continuous DC regime during earlier measurements [10]. The course of the
decay of the electron density does not correspond with the pure ambipolar diffusion to the walls because that
should at constant electron temperature yield an exponential decline. The observed slower dependence could
be demonstration of the background continuous DC discharge that coexists with the afterglow pulse-generated
plasma. The time dependence of the electron density at different radial distances is shown in the right panel of
Fig. 4. Faster decay of the electron density near the central axis could be indicative of features connected with the
ow of the working gas. In bigger distance from the plasma jet axis the beginning of the increase of the electron
density is shifted in time up to several tens of microseconds from the beginning of the on phase. The delayed
onset of the electron density with respect to the radial distance from the plasma jet axis is diffusion related.
Fig. 5 represents the time evolution of the plasma potential. The temporal plasma potential changes are
relatively very small, e.g. in comparison with the changes of plasma potential in rf mode of the discharge exci-
tation [11]. After the beginning of the pulse the plasma potential decreases, approximately in 50 s it reaches its
www.cpp-journal.org
c 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
115
890 P. Kudrna et al.: Hollow cathode plasma jet in pulse regime
minimal value and then it grows again. That may be connected with the re-arrangement of the hollow cathode
discharge inside the nozzle during the rise of the discharge current; similarly as described below for the electron
temperature.
Fig. 5 The temporal dependences of the plasma potential for 100 s pulse (left) and 50 s pulse (right). In the central axis
position the probe was situated 30 mm below the plasma jet, in the edge position the horizontal distance between the central
axis and the probe was about 50 mm. (Color gure: www.cpp-journal.org).
The falling edge of the discharge pulse causes a very sharp increase of plasma potential. We suppose that this
effect is caused by the releasing of a relatively big amount of space charge which shields the cathode during the
pulse-on time. For all pressures the plasma potential has slightly positive value with decreasing tendency down
to zero farther from the central axis. That demonstrates that almost all the discharge voltage is concentrated on
the cathode fall within the hollow cathode and contributes to the intense sputtering of the inner nozzle surface.
The plasma exiting the nozzle is more or less afterglow plasma that is not further heated by the applied discharge
voltage. The observable increase of plasma potential in the afterglow phase of the discharge farther from the
discharge axis might be connected with the onset of the background DC discharge.
Fig. 6 The temporal dependence of the electron temperature. (Color gure: www.cpp-journal.org).
Results obtained from the evaluation of the electron temperature depicted in Fig. 6 conrm this discharge
mechanism. The double-peak structure on the temporal course of the electron temperature during the active pulse
is probably due to the re-arrangement of the hollow cathode discharge inside the nozzle caused by the initial
rise of the discharge current. The duration of the rst peak on the electron temperature dependence namely
corresponds to the duration of the current rise in Fig. 2 (approximately 50 s). That hypothesis is also supported
by the right panel in Fig. 6 where the discharge pulse was shortened down to 50 s; the second peak is then barely
visible. The absolute value of the electron temperature remains even in the active discharge pulse very low - in
the order of several tenths of electron-volt. In these conditions no signicant dependence on the pressure or on
c 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.cpp-journal.org
116
Contrib. Plasma Phys. 50, No. 9 (2010) / www.cpp-journal.org 891
the probe position was found. The noise in the course of the electron temperature was caused by the evaluation
process and has no physical origin. The slight rise of the electron temperature after the end of the active pulse
that is discernible especially with the shorter discharge pulse is most probably connected with the onset of the
background DC discharge.
5 Conclusion
Langmuir probe measurements showed that during the pulse-on time the electron density in the investigated
hollow cathode plasma jet system can rise up to the value of 1018 m3 . That is much higher density than that
obtained in earlier measurements in continuous regime of DC discharge excitation in similar system [10]. The
measured decay of the electron density is inuenced by the weak continuous DC discharge, which is necessary for
keeping the discharge stable. The electron temperature remains very low within the whole discharge period; the
difference in the electron temperature between the pulse-on and the pulse-off time being approximately 0.05 eV.
That demonstrates that the plasma exiting the nozzle is not further heated by the applied discharge voltage. The
plasma potential rst descends then rises during the pulse-on time. A detectable rise of the plasma potential
immediately after the end of the pulse can be explained by the release of a relatively big amount of space charge
which shields the cathode during the pulse-on time.
Acknowledgements This work was partially nancially supported by Czech Science Foundation, grants 202/09/0800,
202/07/0044, 202/03/H162, by AS CR, grant KAN 101120701 and by Grant Agency of Charles University, projects 120510,
135207 and 143307. This work is part of the research plan MSM 0021620834 that is nanced by the Ministry of Education,
Youth and Sports of the Czech Republic.
References
[1] L. Bardos and H. Barankova, Surf. Coat. Tech. 133134, 522 (2000).
[2] M. Tichy, M. Scha, L. Bardos, L. Soukup, L. Jastrabk, K. Kapoun, J. Tous, Z. Mazanec, R. Soukup, Contrib. Plasma
Phys. 34, 765 (1994).
[3] L. Bardos and H. Barankova, Surf. Coat. Tech. 146147, 463 (2001).
[4] L. Soukup, M. Scha, F. Fendrych, L. Jastrabk, Z. Hubicka, D. Chvostova, H. Schova, V. Valvoda, A. Tarasenko,
V. Studnicka, T. Wagner, M. Novak, Surf. Coat. Tech. 116119, 321 (1999).
[5] M. Cada, Z. Hubicka, P. Adamek, P. Ptacek, H. Schova, M. Scha, L. Jastrabk, Surf. Coat. Tech. 174175, 627 (2003).
[6] Z. Hubicka, M. Chichina, A. Deyneka, P. Kudrna, J. Olejncek, H. Schova, M. Scha, L. Jastrabk, P. Virostko,
P. Adamek, M. Tichy, Journal of Optolectronics and Advanced Materials 9, 875 (2007).
[7] M. Cada, Z. Hubicka, V. Kuliovsky, P. Adamek, J. Olejncek, P. Bohac, Surf. Coat. Tech. 200, 3861 (2006).
[8] I. Langmuir, H. M. Mott-Smith, Phys. Rev. 28, 72 (1926).
[9] S. Pfau, M. Tichy, Langmuir Probe Diagnostics of Low-Temperature Plasmas, in R. Hippler, H. Kersten, M. Schmidt,
K. H. Schoenbach, Eds., Low Temperature Plasmas, Fundamentals, Technologies and Techniques, Vol. 1, WILEY-
VCH Verlag GmbH & Co KGaA, Weinheim, (2008).
[10] M. Tichy, Z. Hubicka, M. Scha, M. Cada, J. Olejncek, O. Churpita, L. Jastrabk, P. Virostko, P. Adamek, P. Kudrna,
J. Kluson, S. Leshkov, M. Chichina, S. Kment, J. Plasma Fusion Res. Ser. 8, 1277 (2009).
[11] J. W. Bradley, H. Backer, P. J. Kelly, R. D. Arnell, Surf. Coat. Tech. 142144, 337 (2001).
www.cpp-journal.org
c 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
117
Contrib. Plasma Phys. 53, No. 1, 10 15 (2013) / DOI 10.1002/ctpp.201310002
Key words Hollow cathode, plasma jet sputtering system, Langmuir probe.
This article reports on experimental study of the discharge in the low pressure plasma jet sputtering system
operated in the continuous DC regime of the discharge excitation. Presented measurements were focused on
the description of the working pressure impact on the basic discharge characteristics. The pressure depen-
dence of the gas ow velocity from the nozzle was examined by means of Pitot tube. Further an approximate
measurement of the discharge current spatial distribution was realized as well. The second half of the article
presents results from Langmuir probe measurements under the conditions of low pressures in the range of units
of pascals. These conditions are of special importance in some particular applications especially in preparation
of thin layers with high crystallinity. Used experimental setup enabled to perform Langmuir probe diagnostic
with spatial resolution both in the plasma jet axis direction and in the direction perpendicular to it.
1 Introduction
The deposition of thin layers is a very important application of the low temperature plasma. Besides hard, wear-
resistant, or low friction coatings which are nowadays already commonly produced in industry, also layers with
specic electrical or optical properties can be achieved. Some of them cannot be prepared in other ways. From the
point of view of the control and reproducibility of the deposition processes, knowledge of the relations between
external discharge parameters and local plasma parameters has fundamental importance.
The low pressure hollow cathode plasma jet sputtering system [1] has become a well established tool for the
thin lm deposition. Its most characteristic and signicant property is the formation of a well dened plasma
channel what makes the plasma jet sputtering system especially convenient for the deposition on substrates with
complicated shapes, e.g. inside cavities, inner walls etc. [2]. Like other methods based on the utilisation of the
low temperature plasma due to the plasma anizothermicity the plasma jet is applicable for deposition on thermal
sensitive substrates, e.g. polymers. To the long series of materials which has been successfully prepared by
means of the plasma jet sputtering system TiO2 [3], TiN [4], AlN [5], CrN [6], Cu3 N [7], Al2 O3 [5], LiCoOx [8],
ZnO [9] or SrTiO3 (STO) and Bax Sr1x TiO3 (BSTO) ceramics [10] belong.
2 Experimental setup
The scheme of the experimental setup is depicted in Fig. 1. The plasma jet is inserted from top along the vertically
oriented axis of the cylindrical chamber with diameter of 30 cm and height of 30 cm. The chamber which serves as
the plasma chemical reactor is UHV constructed, pumped by oil-free pumping system down to ultimate pressure
in the order of 106 Pa. The plasma jet is terminated by a titanium nozzle of a cylindrical shape with the inner
diameter of 5 mm and the length of 37 mm. Inside the nozzle the hollow cathode discharge can be ignited when
the working gas is being fed through it and simultaneously the negative voltage is applied to it. As the nearest
grounded object there is situated a stainless steel substrate holder below the plasma jet so it should act as the
Corresponding author. E-mail: Pavel.Kudrna@mff.cuni.cz, Phone: +420 221 912 225, Fax: +420 284 685 095
c 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
118
Contrib. Plasma Phys. 53, No. 1 (2013) / www.cpp-journal.org 11
anode. Nevertheless as it will be stated further in dependence on the discharge parameters large part of the
discharge current can ow to chamber walls. The used power supply was operated in the current stabilized mode.
Because of comparatively high power dissipated inside the hollow cathode the nozzle was surrounded by copper
blocks cooled by owing water. In all experiments the ow of the argon was set to the value of 30 sccm.
Fig. 1 Scheme of the experimental setup with the Pitot tube for gas ow velocity measurements and electronic circuit for
Langmuir probe measurements.
For the measurement of the gas ow velocity from the nozzle an in-house made Pitot tube was used. It was
manufactured from a couple of stainless steel tubes with inner diameter of 1.2 mm guided side by side. One of
the tubes had the inlet oriented in the ow direction whereas the second one was at the end closed and the inlet
was drilled in the wall of the tube below the end perpendicular to the ow direction. According to the Bernoullis
law we were able to determine the gas ow velocity from the difference of the pressures at the inlets of these
tubes. It was measured by means of a differential pressure transducer with the accuracy of 5 % of the 0.2 Torr full
scale.
Right part of the Fig. 1 shows the electronic circuit for the Langmuir probe measurements. The probe itself
was realized by a tungsten wire with the diameter of 50 m and the length of 1.53 mm. The probe holder was
made of Degussit tubes with outer diameter 2 mm. Because of high sputtering rate of titanium from the nozzle
and therefore also fast growth of electrically conductive layer on the ceramic probe holder no contact between the
probe tip and the holder had to be ensured. For this reason a small ceramic tube was slid to the space in between
the outer holder tube and the probe wire about one mm far from the holder tube edge to separate them. Moreover
before every measurement the probe was connected to positive voltage of 70 V to be cleaned by heating due to
the high electron current. The experimental setup enabled us to perform experiments with a very good spatial
resolution. The radial movement of the Langmuir probe was accomplished by the linear motion feedthrough
operated by a step motor. The position of the plasma jet with reference to the probe in the axial direction was
accomplished by moving the whole nozzle system by a manually operated motion feedthrough.
119
12 J. Kluson et al.: Hollow cathode plasma jet
The typical pressure range in which the low pressure plasma jet system is operated covers the interval from
units of Pa up to approximately 100 Pa. Higher pressure is not commonly used since it brings difculties with
the controllability of the discharge and the precise settings of the discharge parameters. On the other hand lower
pressures around 1 Pa are used often if achievement of high crystallite surfaces is required [13,14]. The Langmuir
probe measurements presented in this paper are focused especially to this pressure region.
Fig. 2 Pitot tube measurement of the gas ow velocity Fig. 3 Current owing to the substrate holder in dependence
from the nozzle. The pressure dependence was measured on the working pressure. The total discharge current (dis-
at the axis 7 mm below the nozzle outlet. charge operated in the current stabilized mode) was 150 mA.
Working pressure inside the experimental chamber directly impacts the gas ow from the plasma jet. From
the gas ow velocity measurements shown in Fig. 2 realized by a Pitot tube it is apparent that in dependence on
the working pressure the plasma jet can be operated in subsonic as well as in supersonic regime. From the mass
conservation law it can be shown that in case of the adiabatic ow of a monoatomic gas the transition from the
subsonic to the supersonic regime occurs when the working pressure sinks under roughly one half of the pressure
inside the nozzle [15]. A theoretical study of the pressure inside the nozzle was presented in [16]. According
to this study the pressure inside the nozzle rises directly proportionally with the gas ow and with the square
root of the thermodynamic gas temperature. In the limit case when all the delivered power is absorbed to the
cooling medium it gives for our experimental conditions the pressure of approximately 20 Pa. Within the error of
measurement the results in Fig. 2 are in a good agreement with this theoretical prediction. The axial dependence
of the gas ow velocity was also measured by moving the nozzle axially with respect to the Pitot tube. The
measured velocity decrease downstream along the plasma jet axis is supposed to be due to the combined effect
of the mutual collisions between the particles and the diffusion process.
Results of experiments aiming at describing the working pressure impact to the plasma behavior are presented
in Fig. 3. Here we intended to analyze the spatial distribution of the discharge current. For that purpose we used
the circular shaped stainless steel substrate holder with the diameter of 15 cm. The nozzle outlet was positioned
5 cm above the holder so the distance to the grounded walls of the vacuum vessel was more than two times larger.
The substrate holder was isolated from the surrounding walls except for the connection to the ground via an
ammeter. In the commonly used pressure range of tens of Pa just a very slight current signal was detected to
the holder, see Fig. 3. In other words almost all the discharge current was owing to the chamber walls. The
process responsible for this effect is supposedly the ambipolar diffusion. For the pressures around 80 Pa and
higher the signicance of the current to the holder in respect to the total discharge current started to increase and
at the pressure of 150 Pa almost all the discharge current ew to the holder. Certain part of the discharge current
owing to the holder was observed also at low pressures under 20 Pa. The course of the current to holder in this
region can be well explained by the results from gas ow velocity measurement stated above. Due to greater
mean free path of the electrons and consequently higher energies which they attain the probability to reach the
substrate increases.
c 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.cpp-journal.org
120
Contrib. Plasma Phys. 53, No. 1 (2013) / www.cpp-journal.org 13
Fig. 4 Course of the electron density in the radial direction 32 mm below the nozzle outlet (left panel) and in the axial
direction at the plasma jet axis (right panel). Measurements in the axial direction were performed at low pressure of 0.7 Pa.
Results obtained from three different methods of electron density determination are compared here.
Two graphs in Fig. 4 show radial and axial prole of the electron density. The maximal concentrations in the
order of 1017 m3 are achieved just in the nearest vicinity of the nozzle outlet, outwards within approximately
15 mm the electron density rapidly decreases to the values more than one order lower. Nevertheless as it is
apparent from the picture the absolute values of the electron density are in the range of several Pa strongly
dependent on the pressure. While as it was presented in [17] in the pressure range of tens of Pa the electron
density decreases with the pressure due to faster recombination the dependence at low pressures seem to be
inverse. It reveals that at lower pressures higher energetic particles are present in the reactor chamber volume
and the ionization process takes place there. Higher concentration of neutral buffer gas increases the probability
of collisions with these particles and then higher electron density is observed. Steeper radial decay of electron
density at higher pressures can be caused by faster recombination of molecular ions, which occurrence increases
with the pressure due to higher probability of three body reactions.
To prove the accuracy of the data we evaluated the electron density by three different ways. It was determined
from the electron current at the plasma potential, from the integral of the EEDF using Druyvesteyn formula and
from the square of the electron current in the electron saturated region according to so called orbital motion limit
model [1921]. Correspondence of the results produced by all three methods within the 20 % error limit we
consider as satisfactory.
Operating the discharge at lower pressures results in changes in the electron energy distribution. Mean free
path of the particles is getting longer so they can be accelerated to higher energies. Left panel of Fig. 5 presents
the comparison of EEPFs measured at 0.7 Pa and at 15 Pa. EEPFs were determined according to Druyvesteyn
formula from the numerical computation of the second derivative of theprobe characteristics. To remove the
dependency on electron density the EEPF was normalized to 1: 0 f () d = 1. While clearly Maxwellian
EEPF corresponds to the pressure of 15 Pa, in case of the pressure lower than 1 Pa the appropriate EEPF can be
described rather as a bi-Maxwellian one, i.e. it consists of two distinct parts with different temperatures. This
effect is typical for Ar discharges and is connected with the transition from the non-local discharge energy balance
at low pressures to the collisional electron heating mode at higher pressures, see e.g. [22].
www.cpp-journal.org
c 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
121
14 J. Kluson et al.: Hollow cathode plasma jet
Fig. 5 Left panel: Comparison of electron energy probability function for two working pressures. r = 0 mm, h = 32 mm.
Right panel: Radial prole of the electron temperature 32 mm below the nozzle outlet evaluated from the probe characteristics
by three different methods. Working pressure p = 1 Pa.
In the region of low energies up to about 0.3 eV presented EEPFs are distorted due to certain smoothing during
the evaluation process. Right panel of Fig. 5 shows the radial course of electron temperature 32 mm below the
nozzle outlet. Depicted is the temperature of the main colder electrons group that was evaluated (i) from the
slope of the electron current part of the probe characteristic and (ii) from the slope of the second derivative of
the total probe current in the electron retarding region; both in semilogarithmic scale. In addition, the effective
electron temperature was calculated from the mean electron energy. The effective electron temperature is slightly
higher than the temperature of the cold electron group because of the contribution of higher energetic electrons.
However, in general it can be stated that the electron temperature is very low: in the range of several tenths of
eV. Performed measurements revealed that despite of the obvious presence of electrons with higher energies the
considerable majority of the electrons loses their energy within the nearest surroundings of the nozzle outlet.
Fig. 6 Radial prole of the plasma potential on the left (h = 32 mm) and axial prole of the plasma potential at the plasma
jet axis on the right. Working pressure p = 1 Pa.
Radial and axial course of the plasma potential in Fig. 6 witnesses just a very weak electric eld in the space
between the plasma jet and the substrate. The movement of electrons in the radial direction outwards the axis is
decelerated. We envisage that it is the consequence of ambipolar diffusion. In the axial prole almost no distinct
changes were observed. But here we have to take into consideration the fact that the plasma jet was axially moved
while the probe position remained xed. Thus with the change of the distance between the probe and the nozzle
the distance from the nozzle to the substrate holder was altered as well. Nevertheless it is apparent that the high
negative potential connected to the nozzle is concentrated within its nearest vicinity and does not signicantly
penetrate outside.
c 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.cpp-journal.org
122
Contrib. Plasma Phys. 53, No. 1 (2013) / www.cpp-journal.org 15
5 Conclusion
By means of two independent experiments the working pressure impact on the discharge in the plasma jet system
was investigated. Pitot tube measurements showed that the system can be operated both in supersonic as well
as in subsonic regime; depending on the working pressure. Since the diffusion process depends on the pressure
the discharge current spatial distribution changes considerably with the pressure. Further by Langmuir probe
measurements the radial and axial distributions of the main plasma parameters were examined in the conditions
of low pressures in the range of units of pascals. In dependence on pressure we found two different forms of
the electron energy probability function and differences in the magnitude and radial behaviour of the electron
density. Comparatively low electron temperature/effective temperature as well as very low plasma potential was
found within the space accessible for measurements. That suggests that most of the applied dc voltage remains
on the cathode fall inside the nozzle.
Acknowledgements This work was partially nancially supported in frame of projects 202/07/0044, 202/09/0800, P205/
11/0386 and 202/08/H057 afforded by the Czech Science Foundation, by the CEEPUS project CIII-AT-0063-07-1112 and
the research program MSM0021620834 of the Ministry of Education, Youth, and Sports of the Czech Republic, by the Grant
Agency of the Charles University, grant No.120510, and by EURATOM.
References
[1] L. Bardos and H. Barankova, Surf. Coat. Tech. 133134, 522 (2000).
[2] M. Tichy, M. Scha, L. Bardos, L. Soukup, L. Jastrabk, K. Kapoun, J. Tous, Z. Mazanec, and R. Soukup, Contrib.
Plasma Phys. 34, 765 (1994).
[3] L. Bardos and H. Barankova, Surf. Coat. Tech. 146147, 463 (2001).
[4] L. Bardos, H. Barankova, and S. Berg, Vacuum 46, 1433 (1995).
[5] H. Barankova and L. Bardos, Surf. Coat. Tech. 120-121, 704 (1999).
[6] H. Barankova, L. Bardos, and L.E. Gustavsson, Surf. Coat. Tech. 188, 703 (2004).
[7] L. Soukup, M. Scha, F. Fendrych, L. Jastrabk, Z. Hubicka, D. Chvostova, H. Schova, V. Valvoda, A. Tarasenko, V.
Studnicka, T. Wagner, and M. Novak, Surf. Coat. Tech. 116119, 321 (1999).
[8] Z. Hubicka, M. Cada, I. Jakubec, M. Bludska, Z. Malkova, B. Trunda, P. Ptacek, J. Pridal, and L. Jastrabk, Surf. Coat.
Tech. 174, 632 (2003).
[9] M. Cada, Z. Hubicka, P. Adamek, P. Ptacek, H. Schova, M. Scha, and L. Jastrabk, Surf. Coat. Tech. 174175, 627
(2003).
[10] Z. Hubicka, M. Chichina, A. Deyneka, P. Kudrna, J. Olejncek, H. Schova, M. Scha, L. Jastrabk, P. Virostko,
P. Adamek, and M. Tichy, Journal of Optolectronics and Advanced Materials 9, 875 (2007).
[11] P.D. Kelly and R.D. Arnell, Vacuum Tech. 56, 159 (2000).
[12] S. Craig and G.L. Harding, J. Vac. Sci. Technol. 19, 205 (1981).
[13] V. Stranak, M. Quaas, H. Wulff, Z. Hubicka, S. Wrehde, M. Tichy, and R. Hippler, J. Phys. D: Appl. Phys. 41, 055202
(2008).
[14] J. Musil, D. Herman, and J. Scha, J. Vac. Sci. Technol. A 24, 522 (2006)
[15] Gas Flow, in Handbook of Vacuum Technology, edited by K. Jousten, WILEY-VCH Verlag GmbH & Co KGaA,
Weinheim, 2008.
[16] Z. Hubicka, Hollow cathodes and plasma jets for thin lm deposition, in Low Temperature Plasma Physics, Funda-
mental, Technologies and Techniques, edited by R. Hippler, WILEY-VCH Verlag GmbH & Co KGaA, Weinheim,
2007.
[17] S. Leshkov, P. Kudrna, M. Chichina, J. Kluson, I. Pickova, P. Virostko, Z. Hubicka, and M. Tichy, Contrib. Plasma
Phys. 50, 878 (2010)
[18] M. Tichy, Z. Hubicka, M. Scha, M. Cada, J. Olejncek, O. Churpita, L. Jastrabk, P. Virostko, P. Adamek, P. Kudrna,
J. Kluson, S. Leshkov, M. Chichina, and S. Kment, J. Plasma Fusion Res. Ser. 8, 1277 (2009).
[19] I. Langmuir and H.M. Mott-Smith, Phys. Rev. 28, 72 (1926).
[20] S. Pfau and M. Tichy, Langmuir Probe Diagnostics of Low-Temperature Plasmas, in R. Hippler, H. Kersten, M.
Schmidt, and K.H. Schoenbach, Eds., Low Temperature Plasmas, Fundamentals, Technologies and Techniques, Vol. 1,
WILEY-VCH Verlag GmbH & Co KGaA, Weinheim, (2008).
[21] F.F. Chen, Plasma Sources Sci. Technol. 18, 035012 (2009)
[22] V.A. Godyak, R.B. Piejak, and B.M. Alexandrovich, Plasma Sources Sci. Technol. 1, 36 (1992)
www.cpp-journal.org
c 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
123
IOP PUBLISHING PLASMA SOURCES SCIENCE AND TECHNOLOGY
Plasma Sources Sci. Technol. 22 (2013) 015020 (7pp) doi:10.1088/0963-0252/22/1/015020
Abstract
We present results from two mutually independent measurements of plasma and neutral gas
flow velocities in a plasma jet deposition system. Operation of a hollow-cathode discharge
within the plasma jet nozzle in the pulse regime enables the simple use of a classical Langmuir
probe for the plasma flow velocity measurement. In this method, we assume that the plasma is
generated solely inside the nozzle during the power impulse and we measure the time of flight
of the ions along a known distance between the nozzle end and the probe. The plasma velocity
at the plasma jet axis is then determined by differentiation of the dependence of the distance
covered by ions on time. As the second method the well-known Pitot tube is used for
measurement of the neutral gas velocity. By comparison of both methods we have
experimentally proved that the neutral gas flow velocity is almost unaffected by the presence
of the plasma, i.e. it does not substantially depend on whether the discharge is switched on or
off. The results of both methods correspond well; detected differences are qualitatively
explained. It is documented that the plasma jet can be operated both in a subsonic and in a
supersonic regime. We present the dependences of the plasma and neutral gas flow velocity on
the distance from the cathode, on the pressure in the reactor chamber, and on the flow rate of
the working gas.
(Some figures may appear in colour only in the online journal)
1. Introduction its use for deposition onto substrates of larger areas introduces
difficulties. The discharge plasma in the system has already
The low-pressure plasma jet sputtering system was developed been studied in many different experimental arrangements and
more than two decades ago and has now become an approved with various operation parameters (see, e.g., de Araujo et al
tool for thin film deposition with many industrial applications. 2006, Schrittwieser et al 2010). The experiments with the
The main part of the system is the nozzle that carries the system were dedicated in the first place to the deposition of
flow of the working gas, typically argon. The nozzle acts special materials with utilization in various fields of industry.
as a hollow cathode, and the discharge that burns in its Among the most important materials that were successfully
inner volume is excited either by a dc or an RF power prepared by means of the system are the following: TiO2
source. The most significant property of the system is the (Bardos and Barankova 2001), TiN (Soukup et al 1999), AlN
formation of a plasma jet channel that makes the system (Barankova and Bardos 1999), CrN (Barankova et al 2004),
especially convenient for the deposition of materials onto Al2 O3 (Barankova and Bardos 1999), LiCoOx (Hubicka
substrates with more complex geometry. Conversely, the et al 2003), ZnO (Cada et al 2003) or SrTiO3 (STO) and
plasma jet system is intended for local surface treatment and Bax Sr 1x TiO3 (BSTO) ceramics (Hubicka et al 2007a).
0963-0252/13/015020+07$33.00 1 2013 IOP Publishing Ltd Printed in the UK & the USA
124
Plasma Sources Sci. Technol. 22 (2013) 015020 J Kluson et al
+
switch
Insulating cover
Differential -
U
+
10k
pressure Ti nozzle
transducer 90V
Oscilloscope
pt Pitot tube Langmuir probe
ps
Figure 1. Scheme of the low-pressure hollow-cathode plasma jet sputtering system with the Pitot tube and Langmuir probe used for the flow
velocity measurements and the power supply circuit for the pulse excitation of the discharge. The grounded vacuum chamber is not shown.
In our recent work we focused particularly on the -processessecondary emission of electrons from the
Langmuir probe diagnostics of a plasma jet discharge (Kudrna cathode surface due to the positive ion impactthe ionization
et al 2010). Quantities characterizing the plasma jet, e.g. is performed by photo-ionization and by the presence of so-
the plasma potential, electron density, electron temperature or called pendulum electrons, high-energy electrons oscillating
electron energy probability function (EEPF), were evaluated in the inner volume of the nozzle. These pendulum electrons
in a broad range of experimental conditions. In this paper, are accelerated at a rather high energy in the cathode fall,
we present measurements of the flow velocity of the neutral which is formed around the cathode inner walls and can ionize
gas and of the plasma jet leaving the nozzle. Although it is the gas atoms several times along their passage across the
evident that the flow velocity has a considerable influence diameter of the cathode. The energy lost by an electron due
on the plasma channel structure and, consequently, on the to the ionization process is regained inside the cathode fall
deposition process, it has not been studied more precisely at the opposite side of the cathode/nozzle. One electron can
until now. One of several experimental works concerning therefore generate several ionelectron pairs and the result is
this phenomenon is represented in Tichy et al (1994), where high ionization efficiency of the hollow cathode.
a study of the formation of the plasma jet channel in an During the discharge the inner surface of the nozzle is
RF plasma jet system was performed by digital processing being sputteredas a result of the impact of the positive ions
of the discharge photographs. The measurements presented at the cathode surfaceand the released material is carried
there were performed at comparatively high working gas flow downstream by the gas flow and deposited on the surface of a
rates (up to several hundred sccm) and the dependences on substrate that is placed below the nozzle. The electronically
the flow rate and on the pressure inside the reactor could controlled power switch enables operation of the system in a
not be separated. More results concerning the gas flow pulse mode. In order to achieve well-defined conditions of the
from the plasma jet nozzle were obtained with the help of experiment, the system is constructed for ultra-high vacuum
numerical simulations (see, e.g., Selezneva and Boulos 2001). (UHV) conditions and its ultimate pressure reaches the order
We examined the flow velocity along the plasma jet axis in of 106 Pa. The apparatus is continuously pumped by a turbo-
the whole range of working pressures commonly used in the molecular pump backed by an oil-free mechanical pump. The
system independently of the gas flow rate. typical working pressure of the system ranges in value from
several units of Pa up to several tens of Pa. Argon was used as
2. Experiment the working gas, and in the experiments its flow rate was set
to Ar = 30 sccm if not explicitly stated elsewhere.
2.1. Experimental setup
2.2. Measurement of the ion flux velocity using the pulsed
The main vacuum chamber of the system has a cylindrical regime of the discharge
form with diameter 30 cm and height 30 cm. From the top
an electrically insulated water cooled nozzle made of pure The operation of the plasma jet discharge in the pulse mode
Ti protrudes into the vacuum chamber. The nozzle serves enables measurement of the plasma flow velocity inside the
simultaneously as the inlet of the working gas that flows reactor chamber. The measurement was based on the following
continuously through the system and as a hollow cathode consideration. The ionization of the working gas takes place
for discharge generationsee figure 1. It is connected to almost entirely in the plasma jet nozzle during the pulse-
the negative pole of the dc power supply; the positive pole on time. The ions created leave the nozzle outlet and they
is grounded as is the vacuum chamber itself. The nozzle propagate through the reactor chamber together with the
works as a hollow cathode, i.e. in addition to the intensive neutral gas particles. If we measure the positive ion current at
125
Plasma Sources Sci. Technol. 22 (2013) 015020 J Kluson et al
Figure 2. Typical course of the ion current to the negatively biased probe for different distances between the probe and the nozzle outlet and
pressure p = 20 Pa on the left, and the evaluated dependence of the distance travelled by ions on time for several different pressures on the
right.
the plasma jet axis by a negatively biased Langmuir probe that of collisions between ions and neutral gas can be inferred.
is positioned downstream of the nozzle, every power impulse The ion velocity in the plasma jet axis was then calculated
applied to the cathode causesafter a certain delay time by differentiation of the parabolic fits that approximated the
an ion current impulse on the probe. The probe ion current experimental dependences of the travel distance on the travel
first rises fairly sharply, then, at a specific time, it attains its time. The ion velocity obtained in this manner was depicted
maximum value and subsequently gradually vanishes. With as a function of the distance from the nozzle. This distance is
increasing distance from the nozzle outlet the maximum of denoted as a z-coordinate; the z-axis coincides with the vertical
the current impulse becomes more and more blurred and its plasma jet axis, and z = 0 corresponds to the nozzle outlet.
position shifts in time. In order to measure the plasma flow
velocity we anticipate that the position of the ion current 2.3. Measurements of the neutral gas velocity using a Pitot
maximum corresponds to the most probable time of flight of tube
the ions between the nozzle and the probe.
In our case the ion current was detected by means of a As a second method independent of the measurements of
cylindrical Langmuir probe comprised of a tungsten wire of the positive ion flux to a Langmuir probe, the well-known
diameter 50 m and length 3 mm. A negative voltage of 90 V Pitot tube was used (Klopfenstein 1998). The applicability
via a 10 k resistor was applied to it. The probe was positioned of the Pitot tube has already been proved in a broad scale of
at the nozzle axis perpendicular to it. The distance between experiments, so it can be considered a standard tool for flow
the nozzle outlet and the probe could be adjusted since the velocity determination (Zhang et al 2006). The Pitot tube was
jet itself was movable in the vertical direction. The range constructed according to our own design to best suit the given
of achievable distances in our system spanned from 27 up to experimental conditions. In fact, it was realized by a set of two
57 mm. The duration of the voltage pulse on the cathode and stainless steel tubes welded side by side. One of the tubes had
the repetition frequency were adjusted in such a way to attain the inlet oriented in the flow direction, whereas the second one
sufficiently distinct maxima of the ion current and, at the same was at the closed end and the inlet was drilled in the wall of the
time, to keep the discharge stable. We managed to operate tube 1 mm below the end perpendicular to the flow direction
the discharge under these parameters: voltage on the cathode (as depicted in figure 1). The outer diameter of each tube was
U = 500 V, repetition frequency f = 400 Hz, duration of 1.2 mm. We treated the pressure at the inlet of the former tube
the voltage pulse Tp = 25 s (duty cycle 1%) and gas flow as the total pressure pT . As there is no outlet to allow the flow
rate Ar = 30 sccm. The average discharge current over the to continue from the tube the flow stagnates. We treated the
repetition period was 78 mA (discharge current in the power pressure at the inlet of the latter tube, where the gas flows along
impulse 7.8 A). The time dependence of the positive ion current the tube, as the static pressure pS . Under the assumption of the
to the probe was recorded synchronously with the negative laminar flow regime, the Bernoulli law in the following form
voltage pulse on the cathode at different distances between is valid:
the probe tip and the nozzle outlet. A typical result of such pT = pS + 21 v 2 , (1)
a measurement is depicted in the left panel of figure 2. The where is the gas density and v is the gas flow velocity. By
maxima of the smoothed curves were located and dependences means of a differential pressure transducer we measured the
of the distance travelled by ions on the time of flight were difference in the pressures p = pT pS . The accuracy of
constructed. The obtained distance versus time dependences this instrument is 0.5% of the full scale when the full scale
were fitted by a polynomial of second order, as shown in the range reaches 0.2 Torr.
right panel of figure 2. From the rather good parabolic fit The Pitot tube measuresby nature of its principle
a uniformly decelerated movement of ions as a consequence the velocity of the neutral gas. Therefore, most of the
126
Plasma Sources Sci. Technol. 22 (2013) 015020 J Kluson et al
127
Plasma Sources Sci. Technol. 22 (2013) 015020 J Kluson et al
Figure 4. Plasma flow velocity measured by the ion flux (left) and gas flow velocity measured by the Pitot tube (right) depending on the
distance from the nozzle outlet for different pressures in the reactor chamber. The respective points are connected in the sequence in which
the corresponding measurements were performed.
Figure 5. Dependences of the plasma/gas flow velocity in the plasma jet axis on the distance from the nozzle outlet. A comparison of results
from both methods at argon pressures p = 10, 30 and 100 Pa.
apparent velocity of the neutral gas obtained from the Pitot is lower than the critical pressure pc :
tube measurements compared with the maximum in the neutral 1
gas radial velocity profile. In our opinion, the difference in 2
pc = pIN , (2)
velocities observed in the right panel in figure 5 is therefore due +1
to the difference in the radial profile of ion density compared
with that of the neutral gas. The apparently greater plasma where pIN is the pressure inside the nozzle and is the
velocity compared with the neutral gas bulk velocity due to adiabatic constant. For monoatomic gases ( = 5/3) the
a similar effect has also been observed in flowing afterglow critical pressure is roughly one half of the pressure inside the
systems (see, e.g., Hierl et al 1996). nozzle. Using the approximation of ideal gas with friction and
heat absorption pIN can be expressed as (Hubicka et al 2007b)
The somewhat smaller ion velocity compared with that
of the neutral gas at 10 Pa pressure, see the left panel in
Q TN
figure 5, may be probably attributed to the larger mean free pIN = , (3)
T300 Ap RM
path and, consequently, to the less effective momentum transfer
from the neutral gas particles to the ions by collisions. At where Q is the volumetric flow rate measured by the flow
the intermediate pressure 30 Pa (the centre panel in figure 5) controller, TN is the thermodynamic temperature of the neutral
none of the above described effects applies and we get a fair gas at the nozzle outlet, T300 is the temperature of the gas
correspondence of the results of both scenarios within the immediately behind the flow controller, Ap is the cross section
measurement error. of the nozzle and RM is the universal gas constant divided
In all cases, the velocity of the particles decreases by the molar weight of argon. Under our specific conditions
downstream along the plasma jet axis and with increasing for the limit case, when all the delivered power is absorbed
pressure in the reactor chamber. Apparently, this is an effect by the cooling medium, the above formula gives the value
of the mutual collisions between the particles and the diffusion of the pressure at the nozzle outlet as approximately 20 Pa.
process. For sufficiently low pressures, supersonic velocities This means that for pressures in the reactor chamber of
are observed at smaller distances from the nozzle outlet. 10 Pa and less the supersonic flow should be established.
According to Jousten (2008), the velocity of the flowing gas The value of sonic speed for argon at room temperature is
can reach a supersonic value if the pressure inside the reactor 322 m s1 . In order to compare the above theoretical results
128
Plasma Sources Sci. Technol. 22 (2013) 015020 J Kluson et al
Figure 6. Neutral gas flow velocity for various pressures in the Acknowledgments
reactor chamber at a position 7 mm from the nozzle outlet.
Partial financial support by the Czech Science Foundation,
Grant Nos 202/03/H162 and P205/2011/0386, and by the
Charles University Grant Agency, Grant Nos 120510 and
604612, is gratefully acknowledged. The work pertains to
the studies performed in the frame of the CEEPUS III project
AT-0063.
References
de Araujo F O, de Almeida E O, Alves C, da Costa J A P and
Dumelow T 2006 Deposition of TiO2 on silicon by
sputtering in hollow cathode Surf. Coat. Technol.
201 29903
Barankova H and Bardos L 1999 Effect of gas and cathode material
on the rf hollow cathode reactive PVD Surf. Coat. Technol.
120121 7048
Barankova H, Bardos L and Gustavsson L E 2004 High-rate hot
Figure 7. Dependence of the neutral gas flow velocity on the flow hollow cathode arc deposition of chromium and chromium
rate at two different distances from the nozzle outlet; p = 10 Pa. nitride films Surf. Coat. Technol. 188 7037
Bardos L and Barankova H 2001 Hollow cathode PVD of nitride
and oxide films at low substrate temperature Surf. Coat.
with the experiment we measured the dependence of the argon
Technol. 146147 4638
flow velocity on the working pressure, see figure 6. The Cada M, Hubicka Z, Adamek P, Ptacek P, Schova H, Scha M and
supersonic velocity is reached at a pressure of about 17 Pa Jastrabk L 2003 Investigation of RF and dc plasma jet system
according to figure 6. This result is therefore in concordance during deposition of highly orientated ZnO thin films Surf.
with the stated theoretical conception. In particular, in the Coat. Technol. 174175 62731
closest surroundings of the plasma jet (the curve in figure 6 Hierl P M et al 1996 Flowing afterglow apparatus for the study of
ionmolecule reactions at high temperatures Rev. Sci. Instrum.
corresponds to a distance of 7 mm on the axis), the value 67 21428
of pressure has a considerable impact on the flow velocity. Hubicka Z, Cada M, Jakubec I, Bludska J, Malkova Z, Trunda B,
Nevertheless, the results obtained in the pressure range of Ptacek P, Pridal J and Jastrabk L 2003 Investigation of the RF
several Pa have to be interpreted with greater care because, plasma jet system for deposition of LiCoOx thin films Surf.
under these conditions, the mean free path of the particles is Coat. Technol. 174 6327
Hubicka Z et al 2007a Low pressure plasma-jet systems and their
greater than 1 mm and such a flow cannot be treated as a flow applications for deposition of ceramic thin films
of continuum. J. Optoelectron. Adv. Mater. 9 87580
In contrast, the dependence of the neutral gas flow velocity Hubicka Z et al 2007b Hollow cathodes and plasma jets for thin film
on the input gas flow rate is not very pronounced. It is apparent deposition Low Temperature Plasma Physics Fundamental,
from figure 7 that for commonly used values of several tens of Technologies, and Techniques vol 1 ed R Hippler et al (Berlin:
Wiley-VCH) pp 75984
sccm the neutral gas flow velocity increases only slightly with
Jousten K 2008 Handbook of Vacuum Technology (Weinheim:
the carrier gas flow rate. Wiley-VCH) pp 79162
Kluson J, Kudrna P, Kolpakova A, Hubicka Z and Tichy M 2013
4. Conclusion Experimental study of the discharge in the low pressure
plasma jet sputtering system Contrib. Plasma Phys.
in press
We conducted measurements of the plasma and neutral gas Klopfenstein R 1998 Air velocity and flow measurements using a
flow velocities streaming from a plasma jet nozzle by means Pitot tube ISA Trans. 37 25763
129
Plasma Sources Sci. Technol. 22 (2013) 015020 J Kluson et al
Kudrna P, Kluson J, Leshkov S, Chichina M, Pickova I, Hubicka Z Soukup L et al 1999 Copper nitride thin films prepared by the RF
and Tichy M 2010 A Study of plasma parameters in hollow plasma chemical reactor with low pressure supersonic single
cathode plasma jet in pulse regime Contrib. Plasma Phys. and multi-plasma jet system Surf. Coat. Technol.
50 88691 116119 3216
Selezneva S E and Boulos M I 2001 Supersonic induction plasma jet Tichy M, Scha M, Bardos L, Soukup L, Jastrabk L, Kapoun K,
modeling Nucl. Instrum. Methods Phys. Res. B 180 30611 Tous J, Mazanec Z and Soukup R 1994 A Study of the gas flow
Schrittwieser R et al 2010 Cavityhollow cathode-sputtering source in the RF low-pressure supersonic jet plasma chemical system
for titanium films J. Plasma Phys. 76 65564 Contrib. Plasma Phys. 34 76572
Smith D et al 1970 Ion- and electron-density decay rates in Zhang J, Sun J, Wang D and Wang X 2006 A novel cold plasma jet
afterglow plasmas of argon and argonoxygen mixtures generated by atmospheric dielectric barrier capillary discharge
J. Phys. B: At. Mol. Phys. 3 3444 Thin Solid Films 506507 4048
130