Novel Mo–Si3N4 based selective coating for high temperature concentrating solar power applications

No el Mo–Si 3 N 4 based selec i e coa ing o high empe a u e concen a ing sola powe applica ions E a Céspedes a,1 , Men Wi z b , J.A. Sánchez-Ga cía a,2 , L. Al a ez-F aga a , R. Escoba -Galindo a , C. P ie o a, n a Ins i u o de Ciencia de Ma e iales de Mad id, Consejo Supe io de In es igaciones Cien íficas, Can oblanco, 28049-Mad id, Spain b Depa men o Mechanical and P ocess Enginee ing, ETH Zu ich, 8092 Zu ich, Swi ze land a icle in o A icle his o y: Recei ed 3 July 2013 Recei ed in e ised o m 15 No embe 2013 Accep ed 9 Decembe 2013 A ailable online 1 Janua y 2014 Keywo ds: Selec i e coa ing ma e ial Sola abso be ce me s Pa abolic ough concen a o o CSP echnology abs ac A no el selec i e coa ing ma e ial based on Mo–Si 3 N 4 has been in es iga ed o he fi s ime. Accu a e combina ion o he indi idual componen laye s based on op ical simula ions and p ecise con ol o laye s hickness and composi ion leads o high sola abso p i i y (α Sol ¼0.926) and low he mal emissi i y (ε 25 1C ¼0.017) on a sil e film, esul ing in a high sola he mal ene gy con e sion e ficiency. The andem abso be is s able in mode a e acuum a empe a u es abo e 600 1C showing low emissi i y (es ima ed alue a 600 1Cisε 600 1C ¼0.109) e idencing i s ele ance o high empe a u e sola selec i e applica ions. A ull op ical and he mal analysis o a pa abolic ough concen a o (PTC) sys em confi med he po en ial o he no el coa ing ma e ial o be used o CSP echnology. C own Copy igh &2013 Published by Else ie B.V. All igh s ese ed. 1. In oduc ion Nowadays, he mal con e sion o sola ene gy is one o he mo e simple me hods o ha es ing enewable ene gy. In his di ec hea ing echnology, sola ene gy e ficiency is essen ially de e mined by he sola selec i e coa ings, equi ing high sola abso bance and low he mal emi ance, wi h he sola abso be ma e ial being o undamen al impo ance. Since e ficiency inc eases wi h wo king empe a u e in concen a ed sola powe (CSP) sys ems, esea ch has been aimed a designing selec i e sola coa ings wi h op imum op ical p ope ies a high empe a- u es (Z400 1C) [1]. Fo his pu pose, he ab ica ion o coa ings by physical apo deposi ion (PVD) echniques has become one o he mos popula ones in e ms o en i onmen al issues and also because hose PVD-deposi ed coa ings usually p esen highe he mal s abili y han when p epa ed by we chemis y me hods. Recen ly, an exhaus i e e iew o selec i e coa ings ma e ials p epa ed by PVD echniques has been epo ed [2]. Among hese PVD echniques, spu e ing appea s o be he mos sui able o indus ial p oduc ion o la ge a ea coa ings wi h e y p ecise con ol o laye hickness, which is a undamen al pa ame e o achie e he equi ed op ical abso bance. Among he di e en me hods o op imize he equi ed spec al selec i i y, mul ilaye coa ings a e o special in e es because o he e y high sola e ficiency ha may be ob ained a medium and high empe a u es. In his sense, he app oach which has p o en o gi e high sola pe o mances [3–5], by using a simple ab ica ion p oce- du e is he s uc u e o med by ou laye s: (i) an in a ed eflec i e me allic laye (IR-mi o ), (ii) a high me al olume ac ion (HMVF) ce me , (iii) a low me al olume ac ion (LMVF) ce me laye , and (i ) an an i- eflec i e (AR) laye . In a coa ing wi h his s uc u e, due o he g adual a ia ion o e ac i e index, sola adia ion is e ficien ly abso bed in e nally, and by phase in e e ence be ween hedoublece me and heARlaye s. Rega ding sola abso be ma e ials, a la ge lis o componen s has been epo ed. Mos o hem a e o med by ce me s composed o me als such as Mo, Ni, Fe, C , W o P and ce amics compiled wi h oxides, oxyni ides o ni ides. Oxide-based ce me s ha e been widely s udied, p o iding some o he ew comme cially a ailable high- empe a u e sola selec i e coa ings: o ins ance, Angelan oni ENEA (I aly) comme cializes Mo–SiO 2 [5,6], and W–Al 2 O 3 , while Siemens (Ge many) comme cializes Mo–Al 2 O 3 and W–Al 2 O 3 coa ings [7–9], which a e s able in he 350–500 1C empe a u e ange. No many examples can be ound o ni ides, whe e AlN has been s udied [10,11], o o m ni ide-based ce me abso be s, which ha e been comme cially used by Tu bosun (China), being he mally s able a 350–500 1C in acuum. Addi- ionally, a la ge lis o ce me s has been p oposed based on se e al Con en s lis s a ailable a ScienceDi ec jou nal homepage: www.else ie .com/loca e/solma Sola Ene gy Ma e ials & Sola Cells 0927-0248/$-see on ma e C own Copy igh &2013 Published by Else ie B.V. All igh s ese ed. h p://dx.doi.o g/10.1016/j.solma .2013.12.005 n Co esponding au ho . E-mail add ess: [email p o ec ed] (C. P ie o). 1 P esen add ess: Ins i u e o Science and Technology in Medicine, Keele Uni e si y, Guy Hil on Resea ch Cen e, Tho nbu ow D i e, Ha shill, S oke-on-T en ST4-7QB, UK. 2 P esen add ess: TECNALIA, Pa que Tecnológico de San Sebas ián, Mikele egi Pasealekua, 2. E-20009 San Sebas ián, Gipuzkoa, Spain. Sola Ene gy Ma e ials & Sola Cells 122 (2014) 217–225 mixed oxyni ide compounds, such as aluminum oxyni ide [12] o oxyni ides used in andem wi h a simila ni ide, such as TiAlN/ TiAlON [13], and mo e ecen ly NbAlN/NbAlON [14,15], and H MoN/H ON [16] o high empe a u e applica ions. Fo p ac ical high empe a u e applica ions, low emissi i y is a c i ical pa ame e , because he mal adia i e losses om he abso - be s inc ease wi h T 4 .Someo hesema e ialsha ebeen ound obe s able in acuum o ai up o 400–600 1C, showing excellen alues o abso p i i y in he ange o 0.93–0.96. Howe e , in mos cases, he he mal emissi i y ε h is no uned o e en e alua ed a wo king empe a u es. ε h is equen ly es ima ed a ound 100 1C ( ypical ε 100 1C ¼ε h (100 1C) a e be ween 0.04 and 0.1) while o some ma e ials, ε 350 1C has been epo ed (ε 350 1C ¼0.08–0.1 o W–AlN and Mo–AlN on o Cu o Mo–Al 2 O 3 ). Jus a ew e alua ions o ε h a high empe a u eha ebeendone,suchasW–Al 2 O 3 (ε 550 1C ¼0.14– 0.25) o Mo–SiO 2 (ε 580 1C ¼0.12) [2]. Fo highe empe a u es, he sola selec i e s acks gene ally deg ade, causing a dec ease in he sola selec i i y. The sea ch o new p omising ma e ials showing he mal s abili y bu also uned op ical p ope ies a high empe a u es is o g ea in e es in CSP applica ions. To ou bes knowledge, silicon ni ide has no been in es iga ed o da e as a candida e o he ce amic componen o he ce me in selec i e coa ings. I has been shown ha silicon ni ide is a p omising ma e ial o be used as insula o ma ix in me al/ce amic g anula sys ems and mul ilaye s, wi h po en ial applica ions in elec onic de ices due o i s op ical anspa ency, anspo p op- e ies [17] and chemical ine ness a high empe a u es [18].In addi ion, supe io mechano-elas ic p ope ies ha e been epo ed o Si 3 N 4 when p epa ed by eac i e spu e ing om pu e silicon [19] wi h espec o common silicon oxyni ide. Due o his mechanical beha io , Si 3 N 4 is used as coa ing o magne ic de ices [20] o o o m supe ha d TiN–Si 3 N 4 nanocomposi e coa ings [21]. Fu he mo e, om a ce me componen poin o iew, since me al ni ida ion is less ene ge ically a o able han oxida ion, silicon ni ide ce me s a e expec ed o ha e an enhanced chemical s abili y a high empe a u es compa ed o silicon oxyni ide. In his pape , we epo on a new p omising sola selec i e coa ing based on Si 3 N 4 /Mo–Si 3 N 4 . The Mo–Si 3 N 4 abso be consis s o a double laye o high and low me al olume ac ions, and on op a Si 3 N 4 laye ac s as a dielec ic AR coa ing. Op ical p ope ies o Mo–Si 3 N 4 ce me s and simula ions o he whole s ack ha e been in es iga ed by spec oscopic ellipsome y and by simula ion o achie e he op imum op ical selec i i y a high empe a u es. A p ecise expe imen al con ol o composi ion and hickness o he indi idual componen laye s by means o X- ay Reflec i i y (XRR) and Ru he o d Backsca e ing Spec oscopy (RBS) cha ac e iza- ions has led o op imized sola abso p i i y and he mal emissi - i y abo e 600 1C. Addi ionally, he mal s abili y unde mode a e acuum condi ions (110 2 mba ) has been s udied. Finally, a ull op ical and he mal analysis o he de eloped coa ings is conduc ed, de e mining he pe o mance o each s ack in a pa abolic ough concen a o (PTC) sys em and compa ing he esul s o exis ing selec i e coa ings. 2. Expe imen al The andem abso be s we e deposi ed a oom empe a u e on Si (100) and s ainless s eel AISI-321 subs a es (p e iously annealed a 600 1C in ai , labeled as SS h), using magne on spu e ing. The s ack ma e ials we e: sil e as me allic IR eflec o , molybdenum–silicon ni ide composi es as ce me laye s, and silicon ni ide as AR laye . The base p essu e was a ound 1 10 6 mba . The Ag laye was deposi ed by di ec cu en (DC) spu e ing o Ag wi h A a 910 3 mba a 15 W (deposi ion a e 5 nm/min). A e Ag deposi- ion, he Ag laye su ace was passi a ed a he chambe by flowing oxygen o achie e a p essu e o 510 2 mba du ing 5 min. Mo– Si 3 N 4 ce me s we e p epa ed by simul aneous co-spu e ing o Si and Mo a ge s using N 2 as eac i e gas a 710 3 mba . Radio- equency (RF) and DC spu e ing we e used o Si and Mo, espec i ely. The me al olume ac ion, o filling ac o s (FF), we e modified by uning he powe supplied o each a ge . Pa ame e s o ace me wi h37%filling ac o ( ypically used o he HMVF laye ) a e 100 W o bo h Si and Mo a ge s o ge 7.1 nm/min deposi ion a e and o 20% filling ac o ( ypically used o he LMVF laye ) a e 100 and 17 W o Si and o Mo, espec i ely, o ge 2.7 nm/min deposi ion a e. Si 3 N 4 AR laye s we e p epa ed by RF spu e ing o Si wi h N 2 a 710 3 mba and 100 W (deposi ion a e 1.7 nm/min). By a ying he Mo con en inside he ce me , as well as he hickness o he ce me and AR laye s, mul ilaye s uc u es we e op imized o high empe a u e pho o- he mal pe o mance. I should be no ed he impo ance o he Ag laye passi a ion be o e deposi ing he double ce me laye [22]. The eason o his Nomencla u e Abb e ia ions AR an i- eflec i e CSP concen a ed sola powe DNI di ec no mal i adiance FF filling ac o HMVF high me al olume ac ion HTF hea ans e fluid LMVF low me al olume ac ion MG Maxwell-Game PTC pa abolic ough concen a o PVD physical apo deposi ion RBS Ru he o d backsca e ing spec oscopy XRR X- ay eflec i i y Symbols A[W/m 2 μm], ASTM AM1.5D sola spec al i adiance A ap [m 2 ], To al ape u e a ea o he concen a o mi o s I DNI [W/m 2 ], Di ec no mal i adiance L[W/m 2 μm], Black body emission Planck unc ion P pump [W], Requi ed powe o pump he HTF Q gain [W], Ne hea gain o he HTF Q loss [W], To al hea loss om he abso be ube R[–], Reflec i i y G eek symbols α sol [–], Sola abso p i i y ε h [–], The mal emissi i y η el [–], E ficiency o powe block, gene a o , and HTF pump η op [–], Op ical e ficiency o he sola field loop η h [–], The mal e ficiency o he sola field loop λ[μm], Wa eleng h θ Sol [1], Incidence angle a he concen a o ape u e E. Céspedes e al. / Sola Ene gy Ma e ials & Sola Cells 122 (2014) 217–225218 passi a ion is o p e en he eac i i y be ween he IR-mi o and he abso be ce me , which esul s in wo sening he ab up ness o ansi ion om low o high alues in he eflec ance spec a. The explana ion should be ound in he oughness enhancemen allowed by he Ag–Si 3 N 4 eac i i y ha locally so en he e ac i e index change om he ce me o he IR-mi o and, consequen ly, a ec s he in e e ence pa e n o he mul ilaye . XRR was measu ed in a D8-B ucke AXS di ac ome e wi h X- ay ene gy o he Cu K α emission. Mic os uc u al ea u es, such as laye hickness and a e age oughness we e ob ained by fi ing he expe imen al XRR da a wi h he XREAL so wa e [23]. In o de o de e mine each laye composi ion, RBS expe imen s we e pe o med wi h he 5 MV HVEE Tande on accele a o loca ed a he Cen o de Mic o-Análisis de Ma e iales o he Uni e sidad Au ónoma de Mad id [24]. The RBS expe imen s we e pe o med bo h using 3.7 MeV He þ ions o make use o he esonance 14 N(α, α) 14 N, in o de o imp o e he sensi i i y o ni ogen, and 2.0 MeV He þ ions wi h he sample il ed (θ¼451) o imp o e he dep h esolu ion o he RBS analysis. The expe imen al spec a we e fi ed wi h he RBX so wa e p og am [25]. No app eciable oxygen amoun was de ec ed along he whole film when pe o ming esonan backsca e ing spec oscopy expe imen s. Deposi ed ce me s we e op ically cha ac e ized by ellipsome y o enable op ical simula ions o he eflec ance spec um by using he CODE (Coa ing Designe e sion 3.75) so wa e o op ical spec oscopy, de eloped by Theiss [26].UV– is–IR eflec ance measu emen s we e pe o med using bo h a Shimadzu SolidSpec-3700 spec opho ome e in he ange o 190–2600 nm, and a Va ian 660-IR FTIR spec ome e in he 1.5–25 mm wa e- leng h ange. Using he calcula ed o expe imen ally ob ained spec al eflec- ance da a R(λ), he sola -weigh ed abso p i i y α Sol is de e mined by α Sol ¼R λ 2 λ 1 ½1RðλÞAðλÞdλ R λ 2 λ 1 AðλÞdλ ð1Þ whe e A(λ) is he ASTM AM1.5D sola spec al i adiance, which is he e e ence a e passing h ough he a mosphe e 1.5 imes. α Sol is calcula ed wi h in eg a ion limi s λ 1 ¼0.25 μmandλ 2 ¼4μm. This angeiswide han heone ypically epo edinli e a u e(i.e.in Re . [6] he au ho s p esen α Sol in a 0.3–1.65 μm ange) oaccoun o he whole sola spec um conside ed in he AM1.5D da a. The he mal emissi i y ε h a a specific empe a u e Tis de e mined using ε h ðTÞ¼R 1 0 LðT;λÞ½1RðλÞdλ R 1 0 LðT;λÞdλð2Þ whe e L(T,λ) is he black body emission Planck unc ion. The spec al eflec ance da a R(λ) is assumed o con inue wi h cons an eflec i i y alues equal o he fi s and las measu emen poin o wa eleng hs below and abo e he measu emen limi s, espec i ely. E en hough he spec al eflec ance, and consequen ly he abso p i i y and emissi i y, is also di ec ionally dependen (R(λ,θ)), a good app ox- ima ion is ob ained om he nea no mal spec al eflec ance (R(λ)) p o ided by he measu emen s. Addi ionally, di ec emissi i y measu emen s o he samples we e ob ained a 82 1C, using a comme cial emissome e om De ices & Se ice Company (model AE1). 3. Resul s and discussion 3.1. Simula ion o op ical p ope ies In o de o use simula ion o s udying op ical p ope ies o selec i e coa ing mul ilaye s, he e ac i e index and ex inc ion coe ficien o each laye ha e o be known in he whole wa eleng h ange (n(λ) and κ(λ), espec i ely). Fo his pu pose, he knowledge o hese ce me p ope ies as a unc ion o i s filling ac o is also needed. Once n(λ) and κ(λ) ha e been ob ained, he comme cial CODE p o ides he needed eflec ance spec a o e alua e he sola abso p i i y and he he mal emissi i y. In his sec ion, he op ical p ope ies o a ce me o med by me allic Mo pa icles embedded in ce amic Si 3 N 4 a e s udied. A de ailed desc ip ion o he pe o med simula ion p ocedu e is epo ed because, as seen below, a gene al desc ip ion o he dielec ic cons an is needed o simula e he op ical beha io due o he high me al ac ion in some laye s o he coa ing. Fig. 1(a) shows he expe imen al an(ψ) (ampli ude a io o he pa allel and pe pendicula componen s o he eflec ed ligh ) and cos(Δ) ( ela i e phase change) spec a ob ained by spec oscopic phase modula ed ellipsome y o e a wa eleng h ange o 200– 2000 nm o Mo–Si 3 N 4 ce me s wi h wo di e en Mo filling ac o s ( ), ob ained by co-spu e ing on a Si(100) subs a e. Fig. 1. (a) Expe imen al ellipsome ic measu emen s o Mo–Si 3 N 4 single ce me s laye s ( o wo di e en Mo-con en s) deposi ed on Si subs a es (poin s) and simula ed spec a (dashed lines) ob ained by he CODE so wa e, whe e he Be gman ep esen a ion was used o simula ion. (b) Op ical cons an s ( e ac i e index, n, and ex inc ion coe ficien , κ) ob ained o ce me s and Si 3 N 4 . (c) Spec al densi y unc ion ob ained o wo ce me s. E. Céspedes e al. / Sola Ene gy Ma e ials & Sola Cells 122 (2014) 217–225 219 Expe imen al spec a ha e been fi ed wi h a double laye model, whe e he op ical cons an s o he Si subs a e laye we e aken om li e a u e [27].Si 3 N 4 and Mo con ibu ions we e simula ed by using Cauchy 0 s o mula and he D ude model, espec i ely, o gene a ing bo h e ac i e index (n) and ex inc- ion coe ficien (κ) spec a as a unc ion o wa eleng h, being in good ag eemen wi h alues epo ed elsewhe e [28]. The com- plex e ac i e index o ce me s ma e ials can be calcula ed om he combina ion o he e ac i e indices o he me al and dielec ic componen s using a physical model (Fig. 1(b)). The op ical esponse o hese ce me s is usually modeled by using he so-called e ec i e medium heo ies, whe e he me al pa icles ha e a much smalle size han he wa eleng h o he inciden ligh . The mos widely used physical models o he dielec ic unc ion o a composi e ha e been p oposed by Maxwell-Ga ne [29], B uggeman [30], and Be gman and Mil on [31,32]. The Maxwell-Ga ne (MG) o mula is applicable o sys ems o low olume ac ions (o filling ac o s) o he embedded me al sphe es pa icles, which mus also be a away om each o he . The B uggeman model is p obably he mos o en used e ec i e medium concep . Fo e y low olume ac ions, i is equi alen o he MG heo y, bu o inc easing olume ac ion i s esul s a e qui e di e en . Below ¼1/3, he e is no pe cola ion, bu abo e his h eshold he embedded pa icles a e assumed o be pa ially connec ed [29]. Finally, he so-called Be gman spec al ep esen- a ion heo y is he mos gene al o m o e ec i e medium app oxima ion. I s a es ha he e ec i e dielec ic unc ion (ε e ) o a he e ogeneous medium made up o wo phases (labeled m and c o me al and ce amic, espec i ely) wi h dielec ic unc ions ε m and ε c , and wi h a filling ac o o m ¼ is gi en by ε e ¼ε c 1þ C m ε m ε c 1  þZ 1 0 gðpÞ þpdp "# ! wi h ¼ε c ε c ε m ð3Þ whe e, o he m-phase, g(p) is he spec al densi y unc ion, C m is he so-called pe cola ion s eng h, and p(which uns om 0 o 1) can be iden ified as he depola iza ion ac o o he pa icles making up he composi e. In gene al, g(p) is a complica ed unc ion o he geome ic s uc u e o he composi e. This unc ion is de e mined by he shape and dis ibu ion o he in e aces be ween he di e en phases p esen in he composi e, being cons ained by se e al sum ules [33,34]. Mo eo e , some heo e- ical and empi ical models o g(p) ha e been desc ibed in he li e a u e [35,36], ofi expe imen al da a. Fo he sys em s udied he e, using he CODE so wa e [25], he fi o expe imen al eflec ance spec a and ψand Δ alues wi h he Be gman ep e- sen a ion has allowed us o ob ain he spec al densi y unc ions gi en in Fig. 1(c). The low olume ac ion ( ¼0.1) exhibi s a maximum a ound p¼0.3, which indica es no pe cola ion and would allow he use o any e ec i e medium heo y, because i does no o e pass hei co esponding filling ac o limi a ions [28,29]. On he o he hand, unc ions o highe alues p esen a defini e maximum shi ed o lowe p alues, which is p oo o m- phase pe cola ion and he e o e, he Be gman heo y is equi ed. The fi o ¼0.4 assumes ha Mo is pe cola ed, ob aining a pe cola ion s eng h o C m ¼0.14. In addi ion o he e ac i e indices o me al and ce amic o ming he ce me ma e ial, IR-mi o and subs a e da a a e necessa y o simula e he eflec ance o he whole selec i e s ack, om which sola abso p i i y and he mal emissi i y can be ob ained. Ag films and SS h subs a es, no shown he e, we e also cha ac e ized by spec oscopic ellipsome y, o ha e all inpu s o he op imiza ion o he whole s ack by means o CODE so wa e simula ions. Fig. 2 exhibi s he esul s using he CODE so wa e o Mo– Si 3 N 4 ce me s on he mally ea ed AISI-321 s ainless s eel subs a es wi h (a) di e en Mo olume ac ions ( anging om 10 o 50%) o a selec ed hickness o 50 nm, as well as (b) o di e en laye hicknesses (be ween 20 and 100 nm) o a selec ed Mo con en o 10%. Reflec ance cu es o Si 3 N 4 and Mo films, as well as o he annealed SS subs a e ha e been included o compa ison in Fig. 2(a). Simula ion shows how c i ical he Mo olume ac ion and laye hickness a e o achie e a o able spec al eflec i i y cha ac e is ics o he coa ing. Due o such a c i ical beha io , accu a e con ol o he composi ion and hickness o he deposi ed films is essen ial. In o de o achie e an op imum spec al selec i i y, i is necessa y o op imize he laye pa ame e s o a mul ilaye s ack (i.e., laye hickness, me al olume ac ion o he ce me laye s, and numbe o laye s), o minimize o maximize he eflec ance o specific wa eleng h egions. Fo his pu pose, Downhill-simplex and Gene alized Reduced G adien nonlinea op imiza ion codes we e used. Plainly speaking, o ha e a nea ideal selec i e abso be coa ing, eflec ance o he mul ilaye should be minimized in he UV– is–NIR op ical spec um and maximized in he IR ange, equi ing he op imiza ion o he eflec ance in bo h he sola spec um and he spec um a ec ed by he mal emission. The e o e, op ical simula ions o he SS h/Ag/Mo– Si 3 N 4 (HMVF)/Mo–Si 3 N 4 (LMVF)/Si 3 N 4 s ack allow in e ing he accu- a e hickness and olume ac ion alues o op imized selec i e coa ings o high empe a u e CSP applica ions. Op imiza ion o hicknesses and filling ac o s alues ha e been made ollowing a ial and e o p ocedu e. Table 1 summa izes ou selec ed s acks wi h di e en me al olume ac ions and Fig. 2. Spec al eflec i i y (R(λ)) simula ions using CODE o Mo–Si 3 N 4 single ce me s on o he mal annealed s ainless s eel (SS h) subs a es. (a) R(λ) spec a o ce me s wi h di e en Mo olume ac ions, a ying om 10 o 50% (wi h a cons an hickness o 50 nm). (b) R(λ) o di e en laye hicknesses, anging om 20 o 100 nm (wi h a Mo olume ac ion o 10%). E. Céspedes e al. / Sola Ene gy Ma e ials & Sola Cells 122 (2014) 217–225220 laye hicknesses, in which sola abso p i i y and he mal emis- si i y migh be conside ed as accep able magni udes. Simula ed eflec i i y spec a o hese ou selec i e coa ing s acks a e shown in Fig. 3. I can be obse ed how such small changes in hickness and composi ion o he cons i uen laye s s ongly a ec he eflec ance spec um and, consequen ly, he op ical abso p i i y and he mal emissi i y pa ame e s. Fo he simula ed spec a, sub le changes in he laye hicknesses o same (see Sim-1 s Sim-2) lead o some di e ences in ε 25 1C , om 0.02 o 0.03, and significan a ia ion o ε 600 1C , inc easing om 0.09 o 0.15. Fu he modifica ions o hickness and alues imply no ably di e ences in α sol and ε h pa ame e s (see Sim-3 and Sim-4), enhancing α sol o Sim-4 (up o 0.95) wi h an impo an wo sening o ε h (ε 25 1C ¼0.04 and ε 600 1C ¼0.28). Rega ding he op ical simula ions using CODE, sola abso p i i y abo e 0.92 wi h low emissi i y alues o high empe a u e applica ions (ε 600 1C be ween 0.1 and 0.2) a e expec ed o his no el sys em. 3.2. S uc u al cha ac e is ics o deposi ed selec i e coa ings Based on he op ical simula ion esul s o ce me hickness and filling ac o s, deposi ion o he selec i e mul ilaye s acks was pe o med o (i) assess he expe imen al beha io and (ii) o e alua e and op imize he ac ual sola e ficiency a high empe a u e. To ensu e a p ecise hickness and composi ion expe imen al con ol o he selec i e s acks, single film laye s we e in es iga ed o calib a e he deposi ion a e o di e en Mo olume ac ions. In his way, calib a ion may be used o se he app op ia e hickness and Mo–Si 3 N 4 ela i e composi ion. Mo–Si 3 N 4 ce me laye s o simila hickness as p edic ed o op imal op ical p ope - ies we e measu ed by XRR o an accu a e e alua ion o he deposi ion a e. On he o he hand, he me al olume ac ion o he p epa ed ce me s was de e mined by RBS. To acili a e RBS analysis, ce me laye s we e deposi ed on o Si(100). Fig. 4 epo s on se e al examples o illus a e he accu acy ob ained by he ollowed p ocedu e o con ol hickness and filling ac o o deposi ed laye s. Fig. 4(a) shows XRR spec a o selec ed Mo–Si 3 N 4 single laye s on Si, wi h di e en Mo olume ac ions and a hickness o a ound 50 nm. These single ce me laye s on silicon a e labeled as “MoSN–Fxx” o be dis inguished om he “Sxx”no a ion o he selec i e s acks. I can be obse ed ha XRR cu es exhibi in e e ence inges, e idencing smoo h laye s. Thickness alues we e ob ained by fi ing he expe imen al da a Table 1 Summa y o selec i e coa ing s acks cha ac e is ics and op ical pa ame e s o he simula ed eflec i i y spec a included in Fig. 3 (FF is he filling ac o ). Sola abso p i i y (α Sol ) has been calcula ed om R(λ) in he 0.25–4μm ange and he mal emissi i ies (ε 25 1C and ε 600 1C ) in he 0–1 ange by using he blackbody emission cu e a each empe a u e. S ack Subs . IR-mi o Ag Thickness (nm) HMVF ce me Mo–Si 3 N 4 LMVF ce me Mo–Si 3 N 4 AR laye Si 3 N 4 Thickness (nm) α Sol ε 25 1C ε 600 1C FF (%) Thickness (nm) FF (%) Thickness (nm) Sim-1 SS h 150 40 40 20 40 45 0.905 0.023 0.085 Sim-2 SS h 150 40 50 20 55 60 0.816 0.029 0.146 Sim-3 SS h 150 35 70 25 60 50 0.937 0.036 0.195 Sim-4 SS h 150 45 80 15 80 50 0.942 0.046 0.283 Fig. 3. Simula ed eflec i i y spec a o selec i e coa ing s acks based on Ag/ Mo– Si 3 N 4 (HMVF)/ Mo–Si 3 N 4 (LMVF)/ Si 3 N 4 wi h di e en me al filling ac o s and laye hicknesses. S acks cha ac e is ics and calcula ed α Sol and ε h alues a e summa ized in Table 1. The e e ence di ec AM1.5 sola spec um and he black-body emission a T¼600 1C a e also included o compa ison. Fig. 4. (a) XRR measu emen s (do s) and simula ions (lines) o Mo–Si 3 N 4 single ce me s on Si. (b) RBS da a o “S10”coa ing on s ainless s eel subs a e and wo Mo–Si 3 N 4 single ce me s on Si (do s) and simula ions (con ibu ions o “S10”s ack ha e been plo ed sepa a ely). Expe imen s we e made wi h 3.7 MeV He þ inciden ions. E. Céspedes e al. / Sola Ene gy Ma e ials & Sola Cells 122 (2014) 217–225 221 wi h he XREAL so wa e. The excellen ag eemen o simula ed spec a wi h he expe imen al XRR da a shown in Fig. 4(a) p o ides he laye hickness alues epo ed in Table 2 wi h a ypical accu acy o 0.5 nm. Fig. 4(b) shows selec ed expe imen al RBS da a (and hei co esponding fi s) o single laye ce me samples, including MoSN–F10 and MoSN–F13, as well as “S10”, a whole selec i e s ack. Accu a e Mo filling ac o alues ha e been achie ed by RBS simula- ions. Ac ual composi ion alues ha e been summa ized in Table 2. A e his calib a ion p ocedu e, he accu acy o designed selec i e s acks may be es ima ed in 71 nm o each p epa ed laye hickness and 71% o filling ac o alues. 3.3. Op ical p ope ies o deposi ed selec i e coa ings E alua ion o he op ical cha ac e is ics o he selec i e coa ings has been ca ied ou by measu ing he eflec ance o e a wide op ical wa eleng h ange ( om 200 nm in he UV o 25 μm in he IR), o accoun o he abso p ion o he whole sola spec um and o he he mal emission a wo king empe a u es. Fig. 5 shows he eflec i i y spec a measu ed a nea no mal incidence o se e al samples, also epo ed in Table 3, ha a e ep esen a i e o he expe imen ally pe o med op imiza ion. This op imiza ion has been guided by he ial and e o p ocedu e using he CODE simula ion. A fi s glance, Fig. 5(a) shows ha e y small changes in he mul ilaye s uc u e p oduce appa en di e ences in he op ical esponse. Typically, o quan i y hose changes, sola abso p i i y (α Sol ) and he mal emissi i y (ε h ) magni udes a e e y use ul o know he ac ion o abso bed sola ene gy and he ene gy los h ough he mal emission. Table 3 gi es he ob ained α Sol and ε 25 1C alues o each p epa ed s ack, as well as o spec al da a o selec i e coa ings p esen ed in he li e a u e ha app oxima e ecei e s de eloped by ENEA and Scho [5,37].“S10”and “S14”s acks ha e excellen op ical cha ac e is ics, i should be no ed ha “S10”has a highe sola abso p i i y and may be conside ed as an op imized coa ing o low empe a u e applica ions and “S14”p esen s ou - s anding alues o bo h sola abso p i i y and he mal emissi i y, which sugges i s use o high empe a u es applica ions. Values o α Sol and ε 25 1C , ob ained o he “S14”selec i e coa ing, gi e a high selec i i y a io defined by ξ 25 1C ¼α Sol /ε 25 1C ¼54.1. Mo eo e , a a empe a u e o 600 1C, he a io ξ 600 1C ¼8.5 also emains highe han he alues ob ained o he ENEA (ξ 600 1C ¼7.7) and Scho (ξ 600 1C ¼5.7) selec i e coa ings, al hough he la e is designed o lowe empe a u es. I should be commen ed ha his high selec i i y a io is based on wo op ical cha ac e is ics o he mul ilaye coa ing: he ab up ansi ion be ween he low and high eflec ance egimes and he wa eleng h unabili y o his ansi ion. Addi ionally, he lowe achie able alue o emissi i y is de e mined by he IR-mi o laye na u e. Fo example, in he “S14”sample, he abso be ce me sys em does no app eciably inc ease he sil e emissi i y allowing maximum alues o selec i i y a io when he low-high ansi ion is placed a he app op ia e wa eleng h. 3.4. Op ical and he mal analysis o a sola field In addi ion o calcula ing sola -weigh ed abso p i i y and he mal emissi i y, a ull op ical and he mal analysis is pe o med using he de ailed 3D hea - ans e model desc ibed elsewhe e [38]. This model allows a comp ehensi e e alua ion o he po en ial o he selec i e coa ings in a sola field o a PTC sys em because he conflic ing e ec s o sola abso p i i y and he mal emissi i y a e bo h included. Va ious pe o mance pa ame e s, such as hea loss and hea gain, as well as op ical and he mal e ficiencies o he sola field a e assessed o a loca ion in Se ille, Spain. Fo he ope a ing condi ions, weigh ed yea ly a e age alues Table 2 Summa y o he RBS esul s o Mo–Si 3 N 4 ce me laye s deposi ed on Si subs a es and “S10”selec i e coa ing s acks on ai -annealed s ainless s eel (SS h) subs a e. Sample Subs . Ag IR-mi o Thickness (nm) Mo–Si 3 N 4 (HVMF) Mo–Si 3 N 4 (LVMF) Si 3 N 4 (AR-laye ) FF (%) Thickness (nm) RBS (%) FF (%) Thickness (nm) RBS (%) Thickness (nm) RBS (%) MoSN–F10 Si –10 50 Mo: 9.570.5 Si: 36.073.0 N: 54.573.0 MoSN–F13 Si –13 54 Mo: 13.070.5 Si: 35.073.0 N: 5273.0 MoSN–F15 Si –15 49 S10 SS- h 175 37 50 Mo: 37.4 20 60 Mo: 19.8 80 Si: 37 Si: 20.9 Si: 26.9 N: 53 N: 41.7 N: 53.3 Fig. 5. (a) Reflec ance spec a o se e al selec i e coa ing s acks epo ed in Table 3. (b) Reflec ance spec a o “S14”coa ing a e annealing a se e al empe a u es. E. Céspedes e al. / Sola Ene gy Ma e ials & Sola Cells 122 (2014) 217–225222 a e used o he di ec no mal i adiance (DNI) I DNI ¼619 W/m 2 and o he sola incidence angle a he concen a o ape u e θ Sol ¼30.11. These alues ep esen a design poin ha p oduces he mal e fi- ciencies simila o he expec ed yea ly a e age pe o mance. Mo e de ailed in o ma ion on he used hea ans e model has been epo ed elsewhe e [39]. A single loop o he sola field consis s o 48 concen a o modules in se ies ha hea he hea - ans e fluid (HTF) om 290 1C a he inle o a maximum empe a u e a he ou le . Two sola field configu a ions wi h di e en HTFs a e analyzed: a s a e-o - he-a sys em using a comme cial mol en sal [40] and an ou le empe a u e o 550 1C, and a cu en gene a ion sys em using syn he ic oil [41] wi h a empe a u e limi o 390 1C. The he mal e ficiency o he sola -field loop is defined as he ne useable he mal powe (ne hea gainQ gain o he HTF minus he powe equi ed o pump he HTF) di ided by he DNI o e he o al ape u e a ea o he concen a o mi o s [39], η h ¼Q gain P pump =η el I DNI A ap ð4Þ whe e P pump is he equi ed pumping powe o ully-de eloped flow [42] (de e mined using a co ela ion o he ic ion ac o [43]), and η el ¼32.7% is he p oduc o ypical e ficiencies o he powe block, he elec ical gene a o , and he HTF pump [44,45]. The op ical e ficiency is defined as he o al sola powe abso bed a he abso be ube di ided by he DNI o e he ape u e a ea, η op ¼Q gain þQ loss I DNI A ap ð5Þ whe e Q loss is he o al hea loss om he abso be ube, including adia i e and con ec i e hea ans e o he glass en elope, as well as conduc ion h ough he suppo ing s uc u e [38].In hep esen analysis, he selec i e coa ings a e also compa ed o exis ing coa ings used in comme cial PTC plan s. Spec al eflec ance da a is used o app oxima e he beha io o a comme cial Scho ecei e [36] and a high- empe a u e ecei e de eloped by ENEA [5].Allo he sys em pa ame e s o he sola field a e iden ical o hose lis ed in Re . [39]. Mo e in o ma ion on hese pa ame e s and he hea - ans e model can be ound in Re . [38]. Using he spec al eflec ance cu es o he selec i e coa ings and he spec al da a ob ained om li e a u e, he sola -weigh ed abso p i i y and he he mal emissi i y a e calcula ed a 25 1C, 400 1C, and 600 1C. In Table 3, he calcula ed alues o α Sol and ε h a e shown o all coa ings. I is seen ha he Scho coa ing displays he highes sola abso p i i y o 0.957. Coa ing candida e “S14”has an abso p i i y ha is sligh ly below ha o he ENEA coa ing, bo h designed o highe empe a u es han he coa ing by Scho . The high- empe a u e coa ings ENEA and “S14”also display mo e a o able emissi i y alues a highe empe a u es compa ed o he Scho ecei e , wi h “S14”ha ing a lowe emissi i y han ENEA. In Table 4, esul s a e shown o he analysis pe o med assuming a cu en gene a ion PTC sys em, using syn he ic oil as he HTF and an ou le empe a u e o 390 1C. The hea gain and he e o e he o e all he mal e ficiency o he di e en coa ings a y conside ably since Q loss and η op depend s ongly on he spec al beha io . Fo his configu a ion, he Scho ecei e dis- plays he bes he mal e ficiency wi h η h ¼63.3%. The lowe empe a u es educe he hea losses om he abso be and hence inc ease he impo ance o he sola abso p i i y o he o e all he mal e ficiency. The high abso p i i y o he Scho ecei e hus enables a high he mal pe o mance. This e ec also equalizes he pe o mance o he ENEA and “S14”coa ings, bo h eaching i ually iden ical he mal e ficiencies. In Table 5, esul s a e shown o he he mal analysis o he s a e-o - he-a PTC sys em, using mol en sal and a high HTF ou le empe a u e o 550 1C. O all he coa ings, “S14”displays he bes pe o mance wi h a he mal e ficiency o η h ¼59.1%, which is 0.5% highe han ha o he ENEA coa ing. The lowe sola abso p i i y o “S14”is compensa ed by he educed hea losses due o he o e all lowe he mal emissi i y compa ed o he ENEA coa ing. The wo s pe o mance is obse ed wi h sample S2, as he low op ical e ficiency leads o a e y low hea gain ha canno be compensa ed e en by he lowes he mal emissi i y o all coa ings. 3.5. High empe a u e s abili y o deposi ed selec i e coa ings The p ac ical use o a selec i e coa ing a high empe a u es is de e mined by wo impo an ac o s. The fi s one is based on he chemical s abili y o he compounds and he mul ilaye s uc- u e o ming he s ack, and he second one is based on he he mal e ficiency, as he mal emissi i y is s ongly dependen on empe a u e. In o de o es he s abili y o he selec i e coa ings epo ed he e, we ha e ca ied ou he mal annealing unde mode a e acuum (10 2 mba ) o simula e wo king condi ions o coa ed ubes. The subsequen op ical cha ac e iza ion has allowed es ab- lishing any possible de e io a ion o he s acks, since he mal Table 3 Sola abso p i i y (α sol ) and he mal emissi i y (ε h )a 251C, 400 1C, and 600 1C, calcula ed om expe imen ally measu ed eflec i i y R(λ) and he blackbody emission cu e in he 0.25–4μm and 0–1 anges, espec i ely, o se e al coa ing s acks on ai -annealed s ainless s eel and spec al da a om li e a u e. Coa ing IR-mi o Ag Thickness (nm) Mo–Si 3 N 4 (HMVF) Mo–Si 3 N 4 (LMVF) Si 3 N 4 AR laye Thickness (nm) α Sol ε 25 1C ε 400 1C ε 600 1C FF (%) Thickness (nm) FF (%) Thickness (nm) S2 175 20 55 14 60 55 0.687 0.017 0.022 0.037 S5 175 37 50 18 55 80 0.903 0.024 0.089 0.170 S10 175 37 50 20 60 80 0.941 0.028 0.14 0.250 S13 175 32 50 14 55 75 0.866 0.011 0.039 0.075 S14 175 37 50 20 55 60 0.926 0.017 0.055 0.109 S19 175 37 52 18 57 60 0.905 0.010 0.049 0.108 ENEA 0.939 0.022 0.071 0.122 Scho 0.957 0.013 0.087 0.169 Table 4 Resul s o he hea ans e analysis o a s a e-o - he-a sola field, using mol en sal wi h a maximum HTF empe a u e o 550 1C. S ack Q gain (MW) Q loss (MW) η op (%) η h (%) S2 1.018 0.024 48.7 47.4 S5 1.280 0.079 63.5 59.6 S10 1.291 0.122 66.0 60.1 S13 1.267 0.038 61.0 59.0 S14 1.342 0.051 65.1 62.4 S19 1.318 0.044 63.7 61.3 ENEA 1.345 0.067 66.0 62.6 Scho 1.361 0.077 67.2 63.3 E. Céspedes e al. / Sola Ene gy Ma e ials & Sola Cells 122 (2014) 217–225 223 deg ada ion due o uns able mic os uc u e causes a dec ease in sola selec i i y. Fig. 5(b) shows he e olu ion o he spec al eflec ance measu ed a oom empe a u e a e se e al he mal annealings. In o de o elease possible s esses in he films, an ini ial wa m up o 450 1C du ing 15 h was pe o med wi h a slow hea ing a e (2 1C/min). A e his fi s annealing, subsequen he mal cycles we e pe o med up o di e en empe a u es du ing 15 h wi h hea ing a es o 4 1C/min (no dependence on hea ing a es has been obse ed). Reflec ance spec a depic a e y sligh change, simila o he p e iously obse ed [46,47], and calcula ions o he op ical cha - ac e is ics show an excellen s abili y pe o mance. A e all hose annealing p ocesses pe o med unde mode a e acuum, a max- imum a ia ion o 0.003 is ob ained o bo h sola abso p i i y (α Sol ¼0.92670.003) and o he mal emissi i y a oom empe a- u e (ε 25 1C ¼0.01770.003) o coa ing “S14”. A e annealing he “S14”s acks a 640 1C, he a ia ion o emissi i y a oom em- pe a u e migh be conside ed as negligible, bu emissi i y ob ained o 600 1C becomes ε 600 1C ¼0.154, which ep esen s an impo an a ia ion wi h espec o he emissi i y e alua ed o he as-p epa ed sample (ε 600 1C ¼0.109). The sligh change o eflec ance a e acuum annealing may be due o a e y small a ia ion o laye hickness o ming he abso be s ack. In o de o in es iga e his assump ion, we ha e pe o med RBS on “S14”s acks samples a e annealing a di e en empe a u es. Fig. 6 shows he signal co esponding o Mo and Ag componen s ob ained in RBS expe imen s pe o med wi h low ene gy He þ inciden ions o maximize he esolu ion. No a ia ion is obse ed up o annealing empe a u es highe han 500 1C wi h espec o he as-deposi ed samples. Since wid h and in ensi y o RBS-signal a e ela ed o laye hickness and a omic densi y, espec i ely, he sligh a ia ion ob ained a e annealing a 600 1C has o be explained by a sligh sh inkage o laye s. In his way, he mic o-s uc u al s abili y o ce me and IR-mi o is confi med, and i is e idenced ha eflec i i y a ia ions ob ained a e annealing a e due o small modifica ion o he op ical in e e ence be ween laye s. Much mo e expe imen al wo k has o be done o explain he o igin o he obse ed sh inkage a 600 1C. Ne e heless, i is clea ha annealing usually b ings abou deso p ion o he ypically occluded gas du ing spu e ing deposi ion. Acco ding o he measu ed RBS-signals, he obse ed dec ease is abou 6%. “S14” and “S19”as-p epa ed samples a e 4% di e en in ce me hickness; hei eflec ance spec a epo ed in Fig. 5(a) show he change due o ha di e ence, which is in good ag eemen wi h he obse ed a ia ion om he as-p epa ed o he 600 1C annealed “S14”samples spec a epo ed in Fig. 5(b). This is a e y impo an expe imen al ac ha should be aken in o accoun o u he op imiza ion o he selec i e s acks a high empe a u e. By ini ially op imizing he e ec i e laye hickness acco ding o he sh inking ac o o he annealed coa ings, enhanced he mal e ficiency a 600 1C, oge he wi h impo an beneficial e ec s, can be ob ained. As s a ed be o e, he small a ia ions obse ed in he eflec ance spec a a e he annealing p ocesses do no ha e any app eciable influence on he op ical pa ame e s e alua ed a oom empe a u e. Ne e heless, such eflec ance a ia ion has o modi y he high empe a u e coa ing beha io . By using Eq. (2) and in he assump ion o a empe a u e independen R(λ) spec a, he he mal emissi i y e olu ion may be easily calcula ed a di e en empe a u es, o ha e a comple e pic u e o his Mo–Si 3 N 4 based selec i e coa ing. Fig. 7 shows he mal emissi i y alues as a unc ion o he wo king empe a u e o he “S14”coa ing. Calcula ions ha e been made om oom empe a u e up o he annealing empe a u e by con- side ing he spec a o Fig. 5(b). I should be s a ed ha his commonly used me hod o e alua e he mal emissi i y a high empe a u es has been app op ia ely alida ed by measu ing he spec al emissi i y o simila selec i e coa ings, o de e mine he he mal emissi i y dependence a high empe a u es [48]. Addi ion- ally, he emissi i y alue ob ained a 82 1C wi h a comme cial emissome e is also gi en, which does no show a ia ion a e he acuum annealing pe o med a hose empe a u es. A high empe a u es, a small inc ease o he mal emissi i y may be obse ed, bu ne e heless, i should be commen ed ha esul s a e no modified a e some epe i i e annealing a he same empe a u e. 4. Conclusions A new sola abso be ce me , based on Mo–Si 3 N 4 , has been in es iga ed. The p ocedu e ca ied ou o de elop a sui able selec i e sola coa ing and o e alua e he pe o mance po en ial in a sola sys em is p esen ed. Selec i e coa ing op imiza ion has Table 5 Resul s o he hea ans e analysis o a cu en gene a ion sola field, using syn he ic oil wi h a maximum HTF empe a u e o 390 1C. S ack Q gain (MW) Q loss (MW) η op (%) η h (%) S2 0.992 0.050 48.7 46.3 S5 1.151 0.208 63.5 53.8 S10 1.087 0.325 66.0 50.8 S13 1.213 0.091 61.0 56.7 S14 1.265 0.129 65.1 59.1 S19 1.241 0.121 63.7 58.0 ENEA 1.255 0.157 66.0 58.6 Scho 1.232 0.206 67.2 57.4 Fig. 6. RBS da a o “S14”coa ing on s ainless s eel subs a e a e annealing a se e al empe a u es. Expe imen s we e made wi h 1.0 MeV He þ inciden ions. Fig. 7. The mal emissi i y dependence o he “S14”coa ing a e se e al annealing p ocesses a di e en empe a u es (calcula ions a e shown up o he empe a u e alue a which he annealing was pe o med). E. Céspedes e al. / Sola Ene gy Ma e ials & Sola Cells 122 (2014) 217–225224 included: (i) he ce me in insic op ical p ope ies (nand κ) ellipsome ic cha ac e iza ion, (ii) simula ion o he eflec ance spec um o s acks o med by sil e IR-mi o , double ce me abso be and on op a Si 3 N 4 AR laye , (iii) sola abso p i i y and he mal emissi i y op imiza ion by changing me al filling ac o and laye hickness o simula ed s acks, (i ) deposi ion o mul ilaye s acks wi hin a na ow ange o me al filling ac o and laye hicknesses, ( ) expe imen al cha ac e iza ion o eflec ance o es final esul s and co ec om a possible de ia ion wi h espec o he simula ion op imiza ion. Addi ionally, UV– is and FTIR abso p ion spec oscopy in es iga ions o he s acks be o e and a e consecu i e annealings indica e he mal s abili y abo e 600 1C. A e his p ocedu e, p omising high sola abso p i i y (α Sol ¼ 0.92670.003) and low he mal emissi i y (ε 25 1C ¼0.017 and es i- ma ed alue a 600 1Cisε 600 1C ¼0.109) ha e been ob ained, leading o a selec i i y a io a high empe a u e ha is among he highes epo ed ones. Fu he mo e, he analysis o he op ical and he mal pe o mance o he new ce me h ough simula ions o a high empe a u e PTC sys em yielded o e all he mal e ficiencies ha a e highe han he esul s achie ed wi h compa able, exis ing selec i e coa ings. Acknowledgmen s This wo k has been suppo ed by Eu opean Commission unde HITECO Collabo a i e p ojec con ac no. 256830 om he FP7- ENERGY p og am. Re e ences [1] C.E. Kennedy, Re iew o Mid- o High-Tempe a u e Sola Selec i e Abso be Ma e ials, NREL/TP-520-31267, Na ional Renewable Ene gy Labo a o y, Colo - ado, July 2002. [2] N. Sel akuma , H.C. Ba shilia, Re iew o physical apo deposi ed (PVD) spec ally selec i e coa ings o mid- o high- empe a u e sola he mal applica ions, Sol. Ene gy Ma e . Sol. Cells 98 (2012) 1–23. [3] Q.-C. Zhang, D.R. Mills, High sola pe o mance selec i e su ace using bi- sublaye ce me film s uc u esSol. Ene gy Ma e . Sol. Cells 27 (1992) 273–290. [4] Y. Yin, R.E. Collins, Op imiza ion and analysis o sola selec i e su aces wi h con inuous and mul ilaye p ofiles, J. Appl. Phys. 77 (1995) 6485–6491. [5] S. Esposi o, A. An onaia, M.L. Addonizio, S. Ap ea, Fab ica ion and op imiza ion o highly e ficien ce me -based spec ally selec i e coa ings o high em- pe a u e, Thin Solid Films 517 (2009) 6000–6006. [6] A. An onaia, S. Esposi o, Me hod o P oducing a Ce me -Based Spec ally Selec i e Coa ing o he Recei e Tube o a Sola Concen a o , and Ma e ial hus Ob ained, Wo ld In ellec ual P ope y O ganiza ion, WO2009/107157 A2, 2009. [7] Q.-C. Zhang, D. Mills, A. Monge , Thin Film Sola Selec i e Su ace Coa ing, US Pa en No. 5523132A, 1996. [8] Q.-C. Zhang, Y. Yin, D.R. Mills, High e ficiency Mo–Al 2 O 3 ce me selec i e su aces o high- empe a u e applica ion, Sol. Ene gy Ma e . Sol. Cells 40 (1996) 43–53. [9] T. Kuckelko n, W. G a , C. Hildeb and , A. Geo g, Radia ion Selec i e Abso be Pipe, Abso be Pipe wi h Said Coa ing, and Me hod o Making Same, US Pa en No. 0121225 A1, 2008. [10] Q.-C. Zhang, Me al–AlN ce me sola selec i e coa ings deposi ed by di ec cu en magne on spu e ing echnology, J. Phys. D:Appl. Phys. 31 (1998) 355–362. [11] Q.-C. Zhang, Recen p og ess in high- empe a u e sola selec i e coa ings, Sol. Ene gy Ma e . Sol. Cells 62 (2000) 63–74. [12] S. Yue, S. Yueyan, W. Fengchun, High- empe a u e op ical p ope ies and s abili y o Al x O y –AlN x –Al sola selec i e abso bing su ace p epa ed by DC magne on eac i e spu e ing, Sol. Ene gy Ma e . Sol. Cells 77 (2003) 393–403. [13] H.C. Ba shilia, N. Sel akuma , K.S. Rajam, D.V. S idha a Rao, K. Mu aleedha an, A. Biswas, TiAlN/TiAlON/Si 3 N 4 andem abso be o high empe a u e sola selec i e applica ions, Appl. Phys. Le . 89 (2006) 191909-1–191909-3. [14] H.C. Ba shilia, Spec ally selec i e NbAlN/NbAlON/Si 3 N 4 andem abso be o high- empe a u e sola applica ions, Sol. Ene gy Ma e . Sol. Cells 92 (2008) 495–504. [15] H.C. Ba shilia, V.K. William-G ips, K.S. Rajam, Sola Selec i e Coa ing Ha ing Highe The mal S abili y Use ul o Ha nessing Sola Ene gy and a P ocess o he P epa a ion The eo , US Pa en No. 7585568B2, 2009. [16] N. Sel akuma , N.T. Manikandana h, A. Biswas, H.C. Ba shilia, Design and ab ica ion o highly he mally s able H MoN/H ON/Al 2 O 3 andem abso be o sola he mal powe gene a ion applica ions, Sol. Ene gy Ma e . Sol. Cells 102 (2012) 86–92. [17] M. Vila, E. Román, C. P ie o, Elec ical conduc ion mechanism in silicon ni ide and oxy-ni ide-spu e ed hin films, J. Appl. Phys. 97 (2005) 113710- 1–113710-6. [18] N.S. Jacobson, Co osion o silicon-based ce amics in combus ion en i on- men s, J. Am. Ce am. Soc. 76 (1993) 3–28. [19] M. Vila, D. Cáce es, C. P ie o, Mechanical p ope ies o spu e ed silicon ni ide hin films, J. Appl. Phys. 94 (2003) 7868–7873. [20] M. Gau in, E.E. Fulle on, F.E. Talke, In es iga ion o Fe–Si–Nfilms as magne ic o e coa o high densi y eco ding disk d i es, J. Appl. Phys. 108 (2010) 063925-1–063925-5. [21] H.C. Ba shilia, B. Deep hi, A.S. A un P abhu, K.S. Rajam, Supe ha d nanocom- posi e coa ings o TiN/Si 3 N 4 p epa ed by eac i e di ec cu en unbalanced magne on spu e ing, Su . Coa . Technol. 201 (2006) 329–337. [22] E. Céspedes, C. P ie o, R. Escoba , J.A. Sánchez Ga cía, Co e ing ha Selec i ely Abso bs Visible and In a ed Radia ion, and Me hod o he P oduc ion The eo , Spanish Pa en No. PCT/ES2012/070444, Wo ld In ellec ual P ope y O ganiza ion, WO2012172148 A1, 2012. [23] N.D. Telling, S.J. Guil oyle, D.R. Lo e , C.C. Tang, M.D. C appe , M. Pe y, E idence o oughness dis ibu ions and in e ace smoo hing in Co/Cu mul i- laye s deposi ed unde ene ge ic pa icle bomba dmen , J. Phys. D: Appl. Phys. 31 (1998) 472–481. [24] A. Climen -Fon , F. Pász i, G. Ga cía, M.T. Fe nández-Jiménez, F. Agulló, Fi s measu emen s wi h he Mad id 5 MV andem accele a o , Nucl. Ins um. Me hods Phys. Res. B 219–220 (2004) 400–404. [25] E. Ko ai, Compu e me hods o analysis and simula ion o RBS and ERDA spec a, Nucl. Ins um. Me hods Phys. Res. B 85 (1994) 588–596. [26] W. Theiss, Ha d- and So wa e o Op ical Spec oscopy, 〈www.m heiss.com〉, 2013 (accessed: May 2013). [27] G.E. Jellison J ., Op ical unc ions o silicon de e mined by wo-channel pola iza ion modula ion ellipsome y, Op . Ma e . 1 (1992) 41–47. [28] E.D. Palik, Handbook o Op ical Cons an s o Solids, Academic P ess, New Yo k, 1985. [29] J.C. Maxwell-Ga ne , Colou s in me al glasses and in me allic films, Phil. T ans. R. Soc. Lond. A 203 (1904) 385–420. [30] D.A.G. B uggeman, Be echnung e schiedene physikalische Kons an en on he e ogenen Subs anzen, Ann. Phys., Leipzig 24 (1935) 636–679. [31] D.J. Be gman, The dielec ic cons an o a composi e ma e ial –a p oblem in classical physics, Phys. Rep. 44 (1978) 377–407. [32] G.W. Mil on, Bounds on he complex dielec ic cons an o a composi e ma e ial, Appl. Phys. Le . 37 (1980) 300–302. [33] R. Fuchs, S.H. Liu, Sum ule o he pola izabili y o small pa icles, Phys. Re . B 14 (1976) 5521–5522. [34] D. Langbein, No mal elec ic modes a small polyhed al pa icles, J. Phys. A: Ma h. Gen. 10 (1977) 1031–1041. [35] C. Pecha omán, F.J. Go dillo-Vázquez, Expansion o he spec al ep esen a- ion unc ion o a composi e ma e ial in a basis o Legend e polynomials: expe imen al de e mina ion and analy ic app oxima ions, Phys. Re . B 74 (2006) 035120-1–035120-10. [36] C. Pecha omán, F. Es ebán-Be egón, J.F. Ba olome, S. Lopez-Es eban, J.S. Moya, New pe cola i e BaTiO3–Ni composi es wi h a high and equency-independen dielec ic cons an (ε E80000),Ad .Ma e .13(2001)1541–1544. [37] N. Benz, W. G a , Z. Hacke , C. Hildeb and , M. Möllenho , J. Schul e-Fischedick, K. Silmy, Ad ances in ecei e echnology o pa abolic oughs, in: P oceedings o he 14 h In e na ional Sola PACES Symposium, Las Vegas, 2008. [38] M. Wi z, M. Roesle, A. S ein eld, Th ee-dimensional op ical and he mal nume ical model o sola ubula ecei e s in pa abolic ough concen a o s, J. Sol. Ene gy Eng. 134 (2012) 041012-1–041012-9. [39] M. Wi z, M. Roesle, A. S ein eld, Design poin o p edic ing yea - ound pe o mance o sola pa abolic ough concen a o sys ems, J. Sol. Ene gy Eng. 136 (2013) 021019-1–021019-9, h p://dx.doi.o g/10.1115/1.4025709. [40] Coas al Chemical Company, LLC, HITEC s Hea T ans e Sal , 〈h p://www. coal2nuclea .com/MSR%20-%20HITEC%20Hea %20T ans e %20Sal .pd 〉,2012 (accessed: Oc obe 2012). [41] Solu ia Inc., The minol VP-1 –Vapo Phase/Liquid Phase Hea T ans e Fluid, Technical Bulle in 7239115C, S . Louis, 2008. [42] F.P. Inc ope a, D.P. DeWi , T.L. Be gman, A.S. La ine, Fundamen als o Hea and Mass T ans e , 6 h Edi ion, John Wiley & Sons, New Yo k, NY, 2006. [43] T.K. Se ghides, Es ima e ic ion ac o accu a ely, Chem. Eng. 91 (1984) 63–64. [44] A.M. Pa node, Simula ion and pe o mance e alua ion o pa abolic ough sola powe plan s (Mas e hesis), Uni e si y o Wisconsin-Madison, Madi- son, WI, 2006. [45] G. Manzolini, A. Gios i, C. Saccilo o, P. Sil a, E. Macchi, De elopmen o an inno a i e code o he design o he modynamic sola powe plan s pa A: code desc ip ion and es case, Renew. Ene gy 36 (2011) 1993–2003. [46] H.G. C aighead, R.E. Howa d, J.E. Sweeney, R.A Buh man, G aded-index P – Al 2 O 3 composi e sola abso be s, Appl. Phys. Le . 39 (1981) 39–41. [47] Y. Yin, Y. Pan, L.X. Hang, D.R. McKenzie, M.M.M. Bilek, Di ec cu en eac i e spu e ing C –C 2 O 3 ce me sola selec i e su aces o sola ho wa e applica ions, Thin Solid Films 517 (2009) 1601–1606. [48] I. Se ién-Fe nández, T. Echániz, L. González-Fe nández, R.B. Pé ez-Sáez, E. Céspedes, J.A. Sánchez-Ga cía, L. Ál a ez-F aga, R. Escoba -Galindo, J. M. Albella, C. P ie o, M.J. Tello, Fi s spec al emissi i y s udy o a sola selec i e coa ing in he 150–600 1C empe a u e ange, Sol. Ene gy Ma e . Sol. Cells 117 (2013) 390–395. E. Céspedes e al. / Sola Ene gy Ma e ials & Sola Cells 122 (2014) 217–225 225