Towards nanometric resolution in multilayer depth profiling: a comparative study of RBS, SIMS, XPS and GDOES

Towa ds nanome ic esolu ion in mul ilaye dep h p o iling: a compa a i e s udy o RBS, SIMS, XPS and GDOES Ramón Escoba Galindo & Raul Gago & Da id Duday & Ca los Palacio Abs ac An inc easing amoun o e o is cu en ly being di ec ed owa ds he de elopmen o new unc ionalized nanos uc u ed ma e ials (i.e., mul ilaye s and nanocompo- si es). Using an app op ia e combina ion o composi ion and mic os uc u e, i is possible o op imize and ailo he inal p ope ies o he ma e ial o i s inal applica ion. The analy ical cha ac e iza ion o hese new complex nano- s uc u es equi es high- esolu ion analy ical echniques ha a e able o p o ide in o ma ion abou su ace and dep h composi ion a he nanome ic le el. In his wo k, we compa a i ely e iew he s a e o he a in ou di e en dep h-p o iling cha ac e iza ion echniques: Ru he o d backsca e ing spec oscopy (RBS), seconda y ion mass spec ome y (SIMS), X- ay pho oelec on spec oscopy (XPS) and glow discha ge op ical emission spec oscopy (GDOES). In addi ion, we p edic u u e ends in hese echniques ega ding imp o emen s in hei dep h esolu- ions. Subnanome ic esolu ion can now be achie ed in RBS using magne ic spec ome y sys ems. In SIMS, he use o o a ing sample holde s and oxygen looding du ing analysis as well as he op imiza ion o loa ing low-ene gy ion guns o lowe he impac ene gy o he p ima y ions imp o es he dep h esolu ion o he echnique. Angle- esol ed XPS p o ides a e y powe ul and nondes uc i e echnique o ob aining dep h p o iling and chemical in o ma ion wi hin he ange o a ew monolaye s. Finally, he applica ion o ma hema ical ools (decon olu ion algo i hms and a dep h-p o iling model), pulsed sou ces and su ace plasma cleaning p ocedu es is expec ed o g ea ly imp o e GDOES dep h esolu ion. Keywo ds Dep h p o iling .XPS .SIMS .GDOES .RBS . Mul ilaye .High esolu ion In oduc ion The cu en oad o minia u iza ion (i.e., sub-32 nm echnologies) demands no el o ad anced analy ical ech- niques wi h ex emely high dep h esolu ions ha a e capable o cha ac e izing sample ea u es on he nanoscale [1–3]. In pa icula , ul a hin laye s and mul ilaye sys ems a e gaining inc easing in e es because hey can exhibi no el s uc u al, physical and chemical p ope ies ha di e om hose o he co esponding bulk ma e ials. Wi h he dec easing hicknesses o such laye s, ilm cha ac e iza ion wi h dep h esolu ions a he scale o a ew in e a omic dis ances mus be achie ed. Fo a single in e ace be ween wo laye s A and B, he mos common de ini ion o he dep h esolu ion as ecommended by IUPAC and ASTM E-42 is Δz, which co esponds o he in e al o he dep h R. Escoba Galindo Cen o de Mic o-Análisis de Ma e iales, Uni e sidad Au ónoma de Mad id, 28049 Mad id, Spain R. Escoba Galindo (*):R. Gago Ins i u o de Ciencia de Ma e iales de Mad id, Consejo Supe io de In es igaciones Cien í icas, 28049 Mad id, Spain e-mail: [email p o ec ed] D. Duday Dépa emen de Science e Analyse des Ma é iaux, Cen e de Reche che Public Gab iel Lippmann, 41 ue du B ill, Bel aux 4422, Luxembou g C. Palacio Depa amen o de Física Aplicada, Uni e sidad Au ónoma de Mad id, Can oblanco 28049 Mad id, Spain coo dina e za an in e ace o e which he e is a 84% o 16% change in in ensi y [3]. Nowadays, he chemical analysis o hese e y hin ilms is ca ied ou using ei he des uc i e o nondes uc i e me hods. Ac ually, o hin ilms abo e 10 nm in hickness, he me hod mos equen ly applied o ob ain he concen- a ion dep h p o ile (CDP)— ha is, he composi ion o he hin ilm as a unc ion o dep h—is ion spu e ing in combina ion wi h any o he well-known su ace analysis echniques [4,5]. Among hese, X- ay pho oelec on spec oscopy (XPS) [6,7], seconda y ion mass spec oscopy (SIMS) [8,9], and, mo e ecen ly, glow-discha ge op ical emission spec oscopy (GDOES) [10,11] a e becoming he mos popula . In addi ion, Ru he o d backsca e ing spec- oscopy (RBS) is a nondes uc i e echnique ha is commonly employed o in-dep h composi ional analysis [12,13]. The me hod ha should be used depends speci ically on he ac ual analy ical p oblem o be sol ed, al hough he gene al end is owa ds he simul aneous use o se e al echniques due o he complemen a y in o ma ion p o ided. The p incipal mul ilaye sys ems (i.e., ni ide, oxide, me al, e c.) ha e been ex ensi ely s udied by a combined analyses, such as XPS-RBS [14–16], RBS-GDOES [17,18], GDOES-SIMS [19,20], SIMS-RBS [21,22], XPS-GDOES [23], XPS-SIMS [24,25], XPS-RBS-GDOES [26]o RBS- SIMS-GDOES [27]. In his pape we will e iew he cu en s a us o each o hese ou cha ac e iza ion echniques, as well as he mos ecen de elopmen s ega ding imp o emen s in he dep h esolu ions o hese echniques. Rele an examples o hei applica ion o nanome e -scale mul ilaye analysis will be p esen ed. X- ay pho oelec on spec oscopy (XPS) Among he ins umen al echniques used o he chemical analysis o solids, X- ay pho oelec on spec oscopy (XPS) is one o he mos app op ia e o examining su aces. I is based on Eins ein’s quan um in e p e a ion o he pho o- elec ic e ec . This in e p e a ion assumes ha he ene gy o he emi ed pho oelec ons can be calcula ed as he di e ence be ween he inciden pho on ene gy and he binding ene gy (BE) o he elec ons in he a oms o he sample. The mode n XPS echnique was o igina ed by Siegbahn and cowo ke s a Uppsala Uni e si y in Sweden in he mid-1950s. They de eloped a high- esolu ion spec ome e o de e min- ing he binding ene gies o he a omic shells o he less igh ly bound o bi als. As a esul o his p og am i was disco e ed ha he BEs o he co e elec ons we e dependen on he chemical en i onmen ; ha is, he chemical shi s in XPS e lec changes in chemical bonding o oxida ion s a e and can he e o e be used o ollow such changes on su aces. XPS is sensi i e (0.01 monolaye s o mos ma e ials p oduce a de ec able signal), quan i a i e, and signals can be measu ed o all elemen s excep H and He. S ic ly speaking, XPS is a su ace echnique, bu i can also be used o analyze a laye o adso bed ma e ial in o de o see bo h he su ace and he adso ba e, since only he elec ons c ea ed in a zone cha ac e ized by a dep h lcosθ(called he escape dep h) om he su ace escape o con ibu e o he XPS spec um. He e lis he a enua ion leng h o he emi ed pho oelec on and θis he angle o emission wi h espec o he su ace no mal. Values o llie in he ange 0.5−2.5 nm, and his phenomenon makes XPS a su ace-sensi i e echnique, as s a ed be o e. Nowadays, when employed o dep h-p o iling analysis, XPS is mos equen ly used in combina ion wi h ion spu e ing [28,29]. An example o dep h p o iling wi h XPS is depic ed in Fig. 1. The ul ima e dep h esolu ion depends on ins umen al ac o s, he induced su ace oughness, a omic mixing, enhanced induced di usion and he in o ma ion dep h o he su ace analysis echnique used. Howe e , applying some e inemen s can lead o eliable esul s, wi h dep h esolu ions as good as 1–2nm [30–32]. Expe imen al CDPs a e usually ex ac ed om he in ensi y I i ( ) ha is cha ac e is ic o a sample componen i as a unc ion o he e osion ime . Howe e , he desi ed esul is he concen a ion c i (z) o each componen as a unc ion o he dep h z. The p oblem is hen he quan i ica ion o he expe imen al in o ma ion o ge he desi ed in o ma ion. This quan i a i e e alua ion o measu ed p o iles in ol es he calib a ion o he hickness scale ( ha is, he con e sion o he spu e ing ime in o he spu e ed dep h) and he calib a ion o he concen a ion scale ( ha is, he con e sion o he elemen al signal in ensi y in o he elemen al concen a ion, and he co ec ion/decon olu ion o he e ec s ha modi y he dep h esolu ion). The p ocedu e used o pe o m he quan i a i e analysis is shown schema ically in Fig. 2. In mos cases, he ime →dep h and in ensi y → concen a ion ela ions a e p ac ically linea and can easily be ob ained om app op ia e spu e ing a es and elemen al sensi i i y ac o s, espec i ely [30,31]. E ec s ha modi y he dep h esolu ion can be co ec ed o using he so-called esolu ion unc ion g(z′−z) along wi h Eq. 1(below) in o de o ob ain he eal concen a ion c i (z) using he app op ia e decon olu ion p ocedu es [33,34]: c0 iz0 ðÞ¼ Z 1 1 ciðzÞgz 0zðÞdz ð1Þ The esolu ion unc ion can be modeled ei he expe i- men ally o ia heo e ical models ha ake in o accoun all o he abo emen ioned ac o s ha modi y he dep h esolu ion. In ac , o he in o ma ion dep h o he su ace 2726 R. Escoba Galindo e al. analysis echnique used, he esolu ion unc ion is p ope ly ep esen ed by Eq. 2: gzz0 ðÞ¼exp zz0 licos q  ð2Þ whe e l i cosθis he a enua ion leng h o he emi ed pho oelec ons, and θis he ake-o angle wi h espec o he su ace no mal [33]. Fo hin ilms wi h hicknesses below 10 nm, he changes ha occu du ing ion bomba dmen be o e a s eady s a e is es ablished du ing he spu e ing p ocess make his app oach comple ely inaccu a e, and al e na i e me hods mus be used. Such me hods, o example angle- esol ed X- ay pho oelec on spec oscopy (ARXPS), a e nonde- s uc i e [32,35,36]. Since chemical in o ma ion is e y apidly des oyed by dep h-p o iling me hods ha in ol e spu e ing, angle- esol ed me hods a e also ad an ageous o p ese ing chemical s a e in o ma ion. Figu e 3shows a schema ic diag am o he ARXPS expe imen al se -up. In his igu e, ls ands o he a enua ion leng h o he emi ed pho oelec ons. I spec a a e eco ded a di e en emission angles θ, he pho oelec on escape dep h, lcosθ, can be a ied in such a way ha he dep h p obed is a ied, so he se o angula measu emen s will con ain in o ma ion on he CDP o each species. The in ensi y o he emi ed pho oelec ons co esponding o a pa icula XPS band o a speci ic elemen is gi en by Eq. 3: IiqðÞ¼ I0 i licos q Z 1 0 ciðxÞexp x licos q  dx ð3Þ whe e c i (x) is he concen a ion o elemen ia a dep h x, assumed o be uni o m a he su ace, l i is he a enua ion leng h o he pho oelec ons om his elemen , θis he ake-o angle wi h espec o he su ace no mal, and I i 0 is a cons an ha depends on ins umen al ac o s, angula asymme y and pho oioniza ion c oss-sec ion. Equa ion 3 is an in eg al equa ion o he i s kind and ep esen s a p o o ype o ill-posed p oblems [35]. Two di e en app oaches ha e been used o ex ac he concen a ion dep h p o iles, c(x): simple pa ame ic models ha make some assump ion abou he o m o he dep h p o ile (see he bo om panel o Fig. 3), and gene al algo i hms ha use egula iza ion in addi ion o singula alue decomposi ion (SVD) echniques [35] wi hou any p io hypo hesis o he shape o he concen a ion dep h p o ile. In gene al, hese me hods display common limi a ions imposed by he low I i( ) i c' (z') calib a ion o he hickness scale calib a ion o he concen a ion scale ci(z) decon olu ion o bomba ding e ec s Fig. 2 Scheme o he dep h-p o ile quan i ica ion p ocedu e Fig. 1 XPS concen a ion– dep h p o iling da a om a (le ) C (N)/C(DLC) nano- mul ilaye ed coa ing ( aken om [28] and ep oduced wi h he pe mission o Else ie ) and ( igh ) a 15 nm Si/15 nm Al mul ilaye ed sys em ( aken om [29] and ep oduced wi h he pe mission o Else ie ) A S 020406080 IA/ I S(a.u.) ake-o an g le (0) c(x)= 1 i 0 x/λ1 0 o he wise Sample No mal o he su ace Analyze λ θ cos θ λ ≥≥ Fig. 3 Top: schema ic diag am o he ARXPS expe imen al se -up. Bo om: schema ic diag am o a s ep-like dep h p o ile o a pu e subs a e S, co e ed by a laye A, o hickness d=land cons an composi ion. The simula ed in ensi y a io I a /I s as a unc ion o he ake-o angle is also shown ( igh -hand side). Taken om [35] and ep oduced wi h he pe mission o Else ie Towa ds nanome ic esolu ion in mul ilaye dep h p o iling: a compa a i e s udy o RBS, SIMS, XPS and GDOES 2727 in o ma ion con en o ARXPS measu emen s, wi h he dep h esolu ion Δzlimi ed o abou 0.8z[32]. Al hough his is a poo e ac ional dep h esolu ion han spu e dep h p o iling allows in he dep h ange o ~10–500 nm, i is be e han he spu e dep h p o iling in he nea -su ace (0–5 nm) egion accessible o ARXPS. Anyway, he wo s alue o he dep h esolu ion in his hickness in e al is abou 4 nm, as can be obse ed in he le -hand panel o Fig. 4. This igu e shows he dep h p o iles calcula ed using he expe imen al da a collec ed o an ul a hin laye p epa ed by he unde po en ial deposi ion (upd) o sil e on polyc ys alline pla inum using he in e sion p ocedu e desc ibed abo e [35]. The egula iza ion algo i hm p oduces a easonable es ima e o he concen a ion dep h p o ile whe e wo laye s can be obse ed: a su ace ca bon laye ~1.6 nm hick, ollowed by a complex laye con aining S, O, and Ag wi h a hickness o ~1 nm. Figu e 4( igh -hand panel) shows ARXPS esul s o a na u ally oxidized Al laye [36]. Acco ding o he esul s shown in Fig. 3,an inc ease in he signal in ensi y (o , equi alen ly, he concen a ion) wi h inc easing ake-o angle indica es ha he co esponding species is nea e o he su ace han species ha display he opposi e beha io . The e o e, ARXPS o Fig. 4( igh -hand panel) indica es ha C con amina ion co e s he na u ally oxidized Al laye [36]. I is impo an o poin ou ha he dep h esolu ion in ARXPS is limi ed by he signal- o-noise a io, no he numbe o emission angles o which da a a e acqui ed. Table 1compa es he main ea u es o ARXPS o hose o s anda d XPS dep h p o iling. Finally, i should be no ed ha dep h p obed by XPS can be also a ied using a unable-ene gy exci a ion sou ce, such as synch o on adia ion (SR), since he pho oelec on escape dep h depends on he pho oelec on kine ic ene gy [37]. XPS spec a om di e en dep hs can be ob ained by a ying exci a ion ene gy. This echnique can p o ide he dep h p o ile nondes uc i ely, jus like he ake-o angle a ia ion echnique. Howe e , only a ew s udies ha e used his app oach, p obably due o he need o a la ge-scale ins alla ion o ca y ou he expe imen s and he ma hema ical p oblems in ol ed in eco e ing he CDP; also, he ac ional dep h esolu ion ha can be ob ained is simila o ha achie ed wi h ARXPS measu emen s. Seconda y ion mass spec ome y (SIMS) Seconda y ion mass spec ome y is based upon spu e ing a ew a omic laye s om he su ace o he sample h ough “p ima y ion”bomba dmen . A p ima y ion impac igge s a cascade o a omic collisions ha induce he emission o a oms and a omic clus e s. Al hough mos o he a oms and molecules emo ed om he sample by he in e ac ion o he p ima y beam and he sample su ace a e neu al, some pe cen age o hem a e ionized. These “seconda y ions”a e cha ac e is ic o he composi ion o he analyzed a ea. I he p ima y beam is composed o posi i ely cha ged ions, he esul ing ioniza ion a o s he p oduc ion o nega i e ions, and p ima y beams o nega i e ions a o he gene a ion o posi i e ions. These ions a e hen accele a ed, sepa a ed acco ding o hei mass, and analyzed by a mass spec ome e . Finally, an image con aining quan i a i e in o ma ion is ob ained o a selec ed mass. In [8], Benningho en e al. de ail he basic concep s, ins umen al aspec s and applica ions o SIMS. Con en ional SIMS sys ems equipped wi h ion beam dep h p o iling can achie e nanome e esolu ion a la ge dep hs, oge he wi h a high le el o de ec ion and quan i ica ion (below 1 ppm) [38,39]. Ne e heless, such nanome ic dep h esolu ion is only ob ained when using p ima y ion ene gies in he sub-keV ange [40]. Howe e , he ad an ages o lowe ing he impac ene gy a e comp o- mised by no only signi ican su ace oughening bu also a dec ease in spu e ing yield (p opo ional o he numbe o emi ed/de ec ed pa icles). The dep h esolu ion is limi ed o a high ion lux, since ene ge ic in e ac ions o he 0 2 4 6 8 10 12 14 0 10 20 30 40 50 60 70 80 Ag S O C c(x)[%] dep h (nm) Fig. 4 Le : ARXPS dep h p o iles calcula ed using he expe imen al da a collec ed o all species de ec ed in an ul a hin laye o Ag deposi ed elec ochemically on P using he gene al in e sion algo i hm desc ibed in [35]. Righ : esul s o ARXPS in es iga ions o a na i e Al oxide sample using a model o a C con amina ion/Al oxide/Al subs a e mul ilaye . Taken om [36] and ep oduced wi h he pe mission o Else ie 2728 R. Escoba Galindo e al. p ima y ion wi h he sample lead o p o ile dis o ions due o he p ima y inco po a ion o he inciden p ojec iles and he damage induced by he collision cascades. Di e en app oaches a e now being de eloped wi h he aim o achie ing he ul ima e dep h esolu ion. Vande o s ecen ly e iewed hese di e en app oaches in o de o assess i he SIMS me hod will be able o ul il he needs o indus ial semiconduc o analysis [41]. A i s app oach (EXLE-SIMS) in ol es using ex emely low bomba dmen ene gies as low as 100 eV. New p ima y ion columns called loa ing low-ene gy ion guns (FLIG) ha e been de eloped, as shown in Fig. 5, as a eplacemen o classical ion columns, in o de o dec ease he ion beam ene gy down o 100 eV while keeping ch oma ic abe a ions a a low le el in he newes gene a ion o EXLE-SIMS echniques [42]. Se e al eams ha e al eady achie ed a dep h esolu ion o 1–1.2 nm by using p ima y ion ene gies o 100–550 eV [41,43,44]. This made i possible, o example, o be e analyze in e aces in o ganic–ino ganic mul ilaye s (as shown in he le -hand panel in Fig. 6) o polyme sola cells [43]. Mo eo e , in [45], nanome e dep h esolu ion was also ob ained o a 50×(16 monolaye s 28 Si/10 monolaye s 30 Si) iso opic supe la ice sys em (see he igh -hand panel in Fig. 6) using a 1 keV Cs p ima y beam. The na u e o he p ima y ions used has an impo an e ec on he dep h esolu ion. Fo example, he esul s ob ained wi h cesium (Cs) ions a e no as good as hose achie ed wi h oxygen (O) ions, which gi e a much la ge decay leng h, e en o Table 1 Compa ison o he main ea u es o he XPS and ARXPS con igu a ions XPS + ion spu e ing ARXPS P obing pa icle Pho ons Pho ons De ec ed pa icle Elec ons Elec ons Spu e ing Ion beam (noble gas) <10 keV No A omic mixing? Yes No Sampling dep h (nm) 1–10 0.3–3 Dep h esolu ion (nm) 1–2 nm (in he bes cases) (wi h o a ing samples) ~2–4nm (in he wo s cases) La e al esolu ion (μm) ~5 ~5 Max. dep h analysis (nm) 100–1000 10 De ec ion limi 0.01 a om% 0.01 a om% Elemen al ange All excep H, He All, excep H, He Reliable quan i ica ion (elemen al composi ion)? Yes bu des oyed du ing spu e ing Yes Des uc i e? Yes No Chemis y in o ma ion? No, des oyed du ing spu e ing Yes Vacuum (mba ) 10 −10 10 −10 Decon olu ion needed? Yes (di icul o de e mine he esolu ion unc ion) Yes (a enua ion leng h unc ion as esolu ion unc ion) Analysis ime Slow (hou s) Mode a e (minu es) Fig. 5 Schema ic d awing (le ) and pho og aph ( igh ) o he FLIG ins alled on he spu e deposi ion chambe o he p o o ype S o ing Ma e ins umen . Taken om [42] and ep oduced wi h he pe mission o Else ie Towa ds nanome ic esolu ion in mul ilaye dep h p o iling: a compa a i e s udy o RBS, SIMS, XPS and GDOES 2729 ene gies as low as 150 eV [41]. The bes dep h esolu ion ob ained o ge manium (Ge) p o iles in a hin (2.5 nm) SiGe laye g own in a Si subs a e we e achie ed wi h a 100 eV O + p ima y ion beam and an impac angle o 57°. In heo y, e en lowe p ima y ion ene gies will allow u he dec eases ha will imp o e he dep h esolu ion. Un o una ely, below 100 eV he spu e ing e iciency (e osion a e) dec eases qui e apidly, and bo h he ime need o de ec a su icien numbe o ions and he ime equi ed o analysis apidly inc ease. The de ec ion limi also dec eases as he p ima y ion ene gy is lowe ed. Fo example, in he case o bo on de ec ion in silicon, he de ec ion limi wo sens om 0.001 ppm a 10 keV o 0.1 ppm a 500 eV. Ins umen al de elopmen s ha enable imp o ed p ima y beam cu en s in he 50–200 eV ange a e p esen ly being a ge ed o imp o e he spu e ing pe o mance o he 1 nm/min le el [41], bu his is a huge challenge. In some cases, he use o oxygen (o some o he elemen s wi h a low hea o sublima ion) looding a he same ime as p ima y ion spu e ing can inc ease he nega i e seconda y ion yield, hus compensa ing o he lowe speed o spu e ing a low p ima y ion ene gies [46,47]. Sample holde o a ion can also be use ul in condi ions whe e high oughness is c ea ed in he bo om o he analysis c a e , and especially when analysis a la ge dep hs (mic ome e ange) is needed. In such cases, o a ing he sample holde dec eases o supp esses oughening in he analyzed a ea, allowing o be e dep h esolu ion. Finally, i appea s ha he dep h- esolu ion limi s o EXLE-SIMS a e close o being eached o he semiconduc o indus y, bu o o he applica ions—such as s udies o he oxida ion o he wea o hin ilms— his could become a e y in e es ing ool o s udying he i s s eps in deg ada ion. A second app oach uses clus e beams ins ead o ions beams in o de o imp o e he dep h esolu ion. I was shown ha he use o clus e beams (C60, I -based clus e s) yields no ad an age in e ms o dep h esolu ion, as he o al ene gy appea s o be mo e in luen ial han he ene gy pe cons i uen a om, and he educ ion in he o al ene gy is es ic ed by he onse o clus e deposi ion [48]. Howe e , u he explo a ion o hese clus e s may lead o mo e in e es ing esul s. Fig. 6 Le : SIMS dep h p o iles ob ained o a P3HT-P4VP:PCBM/ PEDOT:PSS ac i e laye wi h 17% PCBM (equi alen o C17) deposi ed on an oxidized silicon wa e , whe e 28 Si (c osses), 12 C (squa es), 16 O( iangles) and 12 C 14 N(ci cles) we e p obed. Inse : 12 C 14 N signal no malized by he 12 C signal (polyme ma ix) o a enua e possible ma ix e ec s. The expe imen s epo ed he e we e ca ied ou in ul ahigh acuum (UHV) using a Cs + p ima y ion sou ce wi h an impac ene gy o 500 eV and nega i e ion de ec ion o op imize he seconda y de ec ion. The 12 C SIMS signal was used o iden i y he polyme laye , while he 12 C 14 N SIMS signal was chosen o p obe P4VP. Taken om [43] and ep oduced wi h he pe mission o Wiley. Righ : SIMS dep h p o iles o a 28 Si and b 30 Si iso opes in a 50×(16 monolaye s 28 Si/10 monolaye s 30 Si) supe la ice sys em. The expe imen s we e pe o med using a 1 keV Cs + ion beam a 45°. Taken om [45] and ep oduced wi h he pe mission o Else ie 2730 R. Escoba Galindo e al. New app oaches ha aim o mee he highe dep h- esolu ion a ge s, such as ze o-ene gy SIMS o a om-p obe omog aphy, elimina e he use o a p ima y ion beam. In he ze o-ene gy SIMS concep he ion beam is comple ely absen [49,50]. Localized ma e ial emo al is ins ead based on he adso p ion o eac i e gases a he sample su ace and he c ea ion o a ola ile compound. Selec i e ma e ial emo al is achie ed using an elec on beam, which locally s imula es a chemical eac ion be ween he adso bed species and he ma ix elemen s. Theo e ically his p ocess has monolaye dep h esolu ion, as se by he monolaye adso p ion and deso p ion p ocesses, and a high spa ial esolu ion (se by he spo size o he elec on beam, which could be as small as a ew nanome e s), and i is quan i a i e, as he deso bed species a e non esonan ly ionized using a lase beam. Howe e , his me hod will end o be applied o s anda d o e y homogeneous samples, a leas o he nex ew yea s, because many de iciencies can occu , such as p e e en ial e ching and/o ou pla ing, nonplana e osion, incomple e ioniza ion, e c. A second me hod ha does no use a p ima y ion beam is a om-p obe omog aphy [51–53]. He e, a om emo al is induced by applying a e y high (pulsed) elec ical ield. The emi ed a oms a e ully ionized ( he e is no ma ix e ec o conside du ing quan i ica ion) and de ec ed in a spa ially esol ed ime-o - ligh de ec o . Due o he op ical ion magni ica ions in ol ed, spa ial esolu ion is on he o de o <0.5 nm. As he a oms a e emo ed one by one, he dep h esolu ion is on he same o de as his. Clea ly, he excellen h ee-dimensional esolu ion makes he me hod much mo e e sa ile han egula SIMS, in pa icula o he analysis o he e ogeneous samples such as nanocomposi e coa ings. Complica ions and limi a ions a ise om he need o special sample p epa a ion ( he o ma ion o a needle- shaped specimen wi h a op adius o abou 50 nm, which ypically equi es ex ensi e ocused ion beam milling), sample conduc i i y equi emen s ha enable he p opaga ion o he high- ol age pulse ( his equi emen is now pa ially elaxed due o he addi ion o lase -assis ed e osion), he small ield o iew (<100 nm), which limi s he o al numbe o a oms analyzed and hus he achie able sensi i i y (>10 18 a /cm 3 ), and he ex ensi e (3D) da a econs uc ion needed. In p ac ice, sample b eakage (due o he la ge mechanical s esses induced by he s ong elec ical ields) is an impo an mechanism ha limi s he ou ine applica ion o he a om p obe. Finally, a e y signi ican ad an age o he a om p obe ela i e o SIMS is i s 3D esolu ion a he subnanome ic scale, which enables he e ogeneous samples and/o ough in e aces o be analyzed wi hou any se ious loss in dep h esolu ion [54]. In conclusion, se e al app oaches wi h imp o ed dep h esolu ion a e being de eloped, wi h he mos p omising being ze o-ene gy SIMS and a om-p obe omog aphy, which a ge he sub-nm ange, bu hese wo app oaches a e much mo e complex o use han EXLE-SIMS. Table 2p esen s a compa ison be ween he igu es o me i o ul alow-ene gy SIMS and hose o a con en ional sys em. Glow-discha ge op ical emission spec oscopy (GDOES) GDOES is a ela i ely new su ace analysis echnique wi h he ad an ages o mode a e acuum equi emen s and e y high spu e ing a es ( ypically >1 µm min −1 ), which enable he CDP o be ob ained in a ew minu es wi h a dep h esolu ion compa able o mo e s anda d echniques, such as AES, XPS o SIMS [4,55–57]. Du ing GDOES expe imen s, he samples a e spu e ed by A + ions and accele a ed neu al species wi h e y low ene gies (<50 eV). The spu e ed a oms a e hen exci ed by he plasma and de-exci ed by emi ing pho ons wi h a cha ac e is ic wa eleng h, enabling elemen disc imina ion. These pho ons a e di ac ed by a di ac ion g a ing and collec ed by pho omul iplie (PM) de ec o s posi ioned in a Rowland ci cle [10,11]. Fu he mo e, he use o a adio equency ( ) sou ce o spu e ing ex ends he applica ion o GDOES o he s udy o insula o s used bo h as coa ings and as subs a es [58–60]. In he las ew yea s, GDOES has been applied ex ensi ely o he s udy o mul ilaye dep h p o iling [17,18,61–65], al hough a se ious loss o esolu ion wi h dep h du ing GDOES analysis has been epo ed. The majo ac o s con ibu ing o his loss o esolu ion a e ela ed o he high e osion a e, sample hea ing, ion-induced su ace oughening [66,67], and in pa icula , he c a e geome y du ing he spu e ing p ocess, which esul s in he mixing o he consecu i e laye s in a mul ilaye sys em when he indi idual laye hicknesses each down o he nanome e scale [68–72]. Among he di e en s a egies ha a e cu en ly being employed o imp o e he dep h esolu ion o GDOES dep h p o iling, we will ocus he e on: i) he use o pulsed sou ces; ii) he applica ion o decon olu ion algo i hms and he modeling o mul ilaye p o iles, and; iii) he u iliza ion o plasma cleaning p ocedu es p io o analysis. The use o pulsed sou ces in GDOES analysis (pulsed glow discha ge, o PGD) is based on he pionee ing wo k o Winches e and Ma cus [73]. Pulses o a sho du a ion (10 μs o 1 ms) wi h a ious possible epe i ion a es (du y cycles) and high ins an aneous powe s (up o hund eds o wa s) can he e o e be gene a ed. I has been epo ed in he li e a u e [74–76] ha such pulses p oduce an enhancemen in he emission yield, esul ing in inc eased sensi i i y and lowe de ec ion limi s oge he wi h simpli ied calib a ion cu es due o he educed sel -abso p ion o he emission lines. Mo eo e , pulsed glow discha ges cause a lowe a e age powe o be applied han in DC mode, and so hey Towa ds nanome ic esolu ion in mul ilaye dep h p o iling: a compa a i e s udy o RBS, SIMS, XPS and GDOES 2731 a e ideal o he analysis o hea -sensi i e ma e ials such as glasses, polyme s and low- empe a u e me als. I is wo h no ing he e he excellen e iew on pulsed glow discha ges ha was ecen ly published by Belengue and coau ho s [77]. The applica ion o pulsed GDOES o he dep h-p o iling analysis o ma e ials has a isen due o he abili y o con ol bo h he la ness o he c a e and he spu e ing a e using he pulsed sou ce pa ame e s (i.e., equency and wid h). The e o e, hinne su ace laye s can be p ope ly esol ed using PGD, as shown in he ini ial wo ks o Yang and Oxley [78,79] o 10 nm Cu coa ings on s eel. Mo e ecen ly he e ha e been published examples o he applica ion o PGD o he analysis o lead zi cona e i ana e (PZT) hin ilms (which ind signi ican use in mic oelec onic and mic oelec ome- chanical sys ems) [80] and hin- ilm sola cells based on Cu(In,Ga)Se 2 and Cu(In,Ga)S 2 mul ilaye s [81]. As exp essed abo e in Eq. 2, an expe imen ally ob ained CDP is a con olu ion o he ue dep h p o ile and he esponse unc ion o he expe imen al se -up. In he pa icula case o a GDOES dep h p o ile, he esponse unc ion is de e mined by he spu e ing c a e shape. The c a e shape can change ma kedly when he in e aces be ween di e en ma e ials a e a e sed in a mul ilaye dep h p o ile. The e o e, ma hema ically desc ibing he esponse unc ion in GDOES analysis is no a i ial ask. Oswald and coau ho s [82] we e he i s o a emp o apply decon olu ion p ocedu es o GDOES dep h p o iles (al hough he me hod was ini ially designed o SIMS analysis). Based on his ini ial wo k, P äßle and coau ho s [83] de eloped an i e a i e decon olu ion algo i hm whe e— using he calib a ed mass– ime p o ile, he pa ial densi ies o he sample cons i uen s and he measu ed inal shape o he spu e ing c a e as inpu da a— he dep h p o iles a e imp o ed by u ilizing in o ma ion on c a e o ma ion. In Fig. 7(le -hand panel), an example o he success ul applica ion o such a decon olu ion p ocedu e o a Cu/C Ni mul ilaye sys em is p esen ed. Yea s la e , in he excellen e iew by Winches e and Payling [10], i was w i en ha his success “makes decon olu ion a eal and exci ing possibili y o GDOES.”Ne e heless, mo e han a decade la e , decon olu ion p ocedu es a e ye o be in eg a ed in o any comme cially a ailable GDOES so wa e. In an al e na- i e app oach, Escoba and Albella [84] ecen ly epo ed on a simple model o in e p e ing and p edic ing he dep h p o iles o pe iodic mul ilaye s uc u es o wo elemen s, whe e he indi idual laye hicknesses a e in he ange o 10– 100 nm (i.e., wi hin he o de o magni ude o oughness induced when dep h p o iling using GDOES echniques). The model is based on he assump ion ha he su ace oughening p oduced by he ion bomba dmen gi es ise o he pa ial mixing o he laye s and hei in e aces, leading o a smoo hing o he o he wise ab up p o iles. Fi ing he model o he expe imen al p o iles o a se o samples made o al e na ing Ti and C laye s, wi h hicknesses a ying o e a wide ange, allows he deg ada ion cons an b(p e- exponen ial ac o ) o be ob ained o each ma e ial. The alues esul ing om he i ing p ocedu e co ela e wi h he e osion a es o Ti and C , and suppo he idea ha he laye b oadening o each ma e ial is de e mined by i s e osion a e. The deg ee o adjus men can be conside ed o be qui e accep able (see he igh -hand panel o Fig. 7) o mos o he samples in es iga ed, aking in o accoun he app oxima ions in ol ed. Table 2 Compa ison o he cha ac e is ics o SIMS and EXLE-SIMS SIMS EXLE-SIMS P obing pa icle Ions Ions De ec ed pa icle Ions (m/q) Ions (m/q) Spu e ing Ion beam <10 Ke Ion beam <100 eV A om mixing? Yes Yes Sampling dep h (nm) 5 10 Dep h esolu ion (nm) 5 1 La e al esolu ion (nm) 100 100 De ec ion limi (ppm) 10 −3 –10 10 −1 –10 Elemen al ange All All Calib a ion me hod Complex (need s anda ds) Complex (need s anda ds) Des uc i e? Yes Yes Iso ope in o ma ion? Yes (+) Yes (+) C a e e ec ? Yes (−) Yes (−) C a e diame e (mm) <1 < 0.5 Vacuum (mba ) 10 −10 10 −10 Analysis ime Slow (hou s) Slow (hou s) 2732 R. Escoba Galindo e al. Finally, i is clea ha in o de o be able o analyze ul a hin laye s nea he su ace using GDOES, apid s abiliza ion o he plasma a he s a o he spu e ing p ocess is equi ed. The echnique whe e a sac i icial sample ( ypically a monoc ys alline silicon wa e [85]) is employed o emo e he con aminan s (ca bon, oxygen, hyd ogen) om he inne walls o he anode p io o he analysis o he sample has been ex ensi ely applied by GDOES use s. Molchan e al. ha e ecen ly p oposed [86] ha his p ocedu e can be imp o ed by using a low-ene gy plasma (<5 W) o gen ly emo e con aminan om he su ace o he sample. In hei wo k hey demons a e how such a ea men is below he spu e ing h eshold and does no cause any damage o he analyzed specimen. Figu e 8 shows he la ge educ ion in su ace con aminan s ( en old o he ca bon in ensi y) achie ed when he plasma cleaning p ocedu e is applied o an elec opolished aluminum sample. Mo eo e , no e how he o al plasma esponse (Fi) s abilizes mo e apidly. This plasma cleaning me hod has been implemen ed o ou ine su ace analysis in he la es e sion o he QUANTUM so wa e om Ho iba Jobin Y on ins umen s. In conclusion, since GDOES is a ela i ely no el echnique o ul a hin dep h p o iling, he e is mo e oom o imp o e i s esolu ion. The use o so plasma cleaning p io o spu e ing using pulsed discha ges, and he applica ion o decon olu ion p ocedu es a e wa ds, will su ely u n GDOES in o an ideal echnique o subnanome e dep h p o iling in e ms o i s ease o use and he accu acy o he esul s i yields. Table 3compa es he expec ed bene i s o an op imized GDOES (OPT-GDOES) o he ypical ea u es o s anda d GDOES analysis. Fig. 7 Le : composi ion dep h p o ile o 100 nm hick Cu/C Ni mul ilaye s on Si: aa e decon olu ion wi h a cons an c a e shape ( he esul is e y simila o he uncon olu ed p o ile); ba e decon olu ion wi h i e a ion o de e mine he une en c a e shape. Taken om [83] and ep oduced wi h he pe mission o Else ie . Righ : a he GDOES expe imen al dep h p o ile o a 10×(70 nm Ti/150 nm C ) mul ilaye ; b he simula ed dep h p o ile ob ained using he nominal indi idual hicknesses om (a) and he deg ada ion cons an shown in Table 1o [84]. Rep oduced wi h he pe mission o Else ie Fig. 8 GD OES dep h p o ile o an elec opolished aluminum sample, ob ained a 750 Pa and 35 W, a e no plasma cleaning (le ) and a e spu e ing he silicon wa e o 1 min and hen cleaning he specimen a 750 Pa and 3 W o 1 min ( igh ). No e he educ ion in con aminan s ( en old o ca bon) in he igh -hand dep h p o ile. Taken om [86] and ep oduced wi h he pe mission o Else ie Towa ds nanome ic esolu ion in mul ilaye dep h p o iling: a compa a i e s udy o RBS, SIMS, XPS and GDOES 2733 and pulsed and dc glow discha ges. J Phys Con Se ies 133:012017 77. Belengue PH, Ganciu M, Guillo PH, Nelis TH (2009) Pulsed glow discha ges o analy ical applica ions. Spec ochim Ac a B 64(7):623–641 78. Yang C, Ingene i K, Mohill M, Ha ison WW (1999) Dep h p o iling o hin ilms wi h pulsed glow discha ge a omic emission spec ome y. Anal Chem 71(23):5328–5334 79. Oxley E, Yang C, Ha ison WW (2000) Quan i a i e dep h analysis using mic osecond pulsed glow discha ge a omic emission spec ome y. J Anal A Spec om 15(9):1241–1246 80. Schwalle P, Aebe ha d M, Nelis T, Fische A, Thapliyal R, Michle J (2006) Rapid dep h p o iling o lead zi cona e i ana e (PZT) hin ilms by pulsed glow-discha ge op ical emission spec oscopy. Su In e ace Anal 38:757–760 81. E imo a V, Ho mann V, Ecke J, Abou-Ras D, Die ich J (2009) Pulsed glow discha ge: om elec ical pa ame e s o applica ion in GDOES. In: Colloquium Spec oscopicum In XXXVI, 30 Aug—3 Sep 2009, Budapes , Hunga y 82. Oswald S, Ho mann V, Eh lich G (1994) Con ibu ions o compu e -aided in e p e a ion o ion spu e ing dep h p o iling. Spec ochim Ac a Pa B 49:1123–1145 83. P äßle F, Ho mann V, Schumann J, We zig K (1996) Quan i a i e dep h p o iling in glow discha ge spec oscopies: a new decon olu ion echnique o sepa a e e ec s o an une en e osion c a e shape. F esenius J Anal Chem 355:840–846 84. Escoba Galindo R, Albella JM (2008) Modelling o glow discha ge op ical emission spec oscopy dep h p o iles o me al (C ,Ti) mul ilaye coa ings. Spec ochim Ac a Pa B 63(3):422–430 85. Klemm D, S angl M, Pee a A, Ho mann V, We ziga K, Ecke J (2008) Analysis o in e ace impu i ies in elec opla ed Cu laye s by using GD-OES and TOF-SIMS. Su In e ace Anal 40:418–422 86. Molchan IS, Thompson GE, Skeldon P, T igoule N, Chapon P, Tempez A, Malhe be J, Lobo Re illa L, Bo del N, Belengue Ph, Nelis T, Zah i A, The ese L, Guillo Ph, Ganciu M, Michle i J, Hohl M (2009) The concep o plasma cleaning in glow discha ge spec ome y. J Anal A Spec om 24:734–741 87. G ö zschel R, Klein Ch, K use O (2001) The Rossendo b oad- ange magne ic spec ome e o high esolu ion RBS and NRA. Nucl Ins um Me h B 183:3–9 88. Kimu a K, Joumo i S, Oo a Y, Nakajima K, Suzuki M (2004) High- esolu ion RBS: a powe ul ool o a omic le el cha ac- e iza ion. Nucl Ins um Me h B 219–220:351–357 89. A noldbik WM, Wol swinkel W, Inia DK, Ve leun VCG, Lobne S, Reinde s JA, Labohm F, Boe ma DO (1996) A high esolu ion magne ic spec og aph o ion beam analysis. Nucl Ins um Me h B 118:566–572 90. And zejewski R, Lucas J, Gui ao A, Go dillo N, Boe ma DO (2006) A wide-angle magne ic spec og aph o a no el design. Nucl Ins um Me h B 249:939–942 91. G ö zschel R, Klein C, Mäde M (2004) RBS wi h high dep h esolu ion using small magne ic spec ome e s. Nucl Ins um Me h B 219–220:344–350 92. Nagel R, Alo C, Balogh AG, A noldbik WM, Boe ma DO (2001) S udy o magne ic mul ilaye s by RBS wi h nanome e esolu ion. Nucl Ins um Me h B 183:140–145 93. Kimu a K, Mannami M (1996) RBS wi h monolaye esolu ion. Nucl Ins um Me h B 113:270–274 94. Nakajima K, Joumo i S, Suzuki M, Kimu a K, Osipowicz T, Tok KL, Zheng JZ, See A, Zhang BC (2003) S ain p o iling o H O 2 /Si(001) in e ace wi h high- esolu ion Ru he o d back- sca e ing spec oscopy. Appl Phys Le 83:296 95. Sakai W, Nakajima K, Suzuki M, Kimu a K, B ijs B (2006) Obse a ion o he in e acial laye in H O 2 (10 nm)/Si by high- esolu ion RBS in combina ion wi h g azing angle spu e ing. Nucl Ins um Me h B 249:238–241 96. Suzuki M, Takashima A, Koyama M, Iijima R, Ino T, Takenaka M (2004) Cha ac e iza ion o Si(1 0 0)/H SiON in e ace. Nucl Ins um Me h B 219–220:851–855 97. Nakajima K, Suzuki M, Kimu a K, Yamamo o M, Te amo o A, Ohmi T, Ha o i T (2006) La ice dis o ion a SiO 2 /Si(001) in e ace s udied wi h high- esolu ion Ru he o d backsca e ing spec oscopy/channeling. Jpn J Appl Phys 45:2467 98. Hosoi S, Nakajima K, Suzuki M, Kimu a K, Shimizu Y, Fuka su S, I oh KM, Uema su M, Kageshima H, Shi aishi K (2006) Obse a ion o Si emission du ing he mal oxida ion o Si(0 0 1) wi h high- esolu ion RBS. Nucl Ins um Me h 249:390–393 99. Yamamo o T, Miyamo o T, Ka en A (2004) Quan i ica ion o ni ogen p o iles in H SiON ilms o ga e dielec ics. Appl Su Sci 231–232:561–564 100. Kimu a K, Nakajima K, Zha M, Nohi a H, Ha o i T, Koba a M, Ikenaga E, Jin Kim J, Kobayashi K, Cona de T, Vande o s W (2008) Combina ion o high- esolu ion RBS and angle- esol ed XPS: accu a e dep h p o iling o chemical s a es. Su In e ace Anal 40(3-4):423–426 2740 R. Escoba Galindo e al.