Materials Chemistry and Physics 94 (2005) 298–307

Highly stable amorphous silicon hydride froma helium plasma reaction

R. Mills ∗, J. He, B. DhandapaniBlackLight Power Inc., 493 Old Trenton Road, Cranbury, NJ 08512, USA

Received 10 November 2004; received in revised form 18 April 2005; accepted 7 May 2005

Abstract

A novel highly stable hydrogen terminated silicon coating was synthesized by microwave plasma reaction of mixture of silane, hydrogen,and helium wherein it is proposed that He+ served as a catalyst with atomic hydrogen to form highly stable silicon hydrides. In contrast,the replacement of helium (He+) with xenon (Xe+) yielded nothing extraordinary. Novel silicon hydride was identified by time of time-of-flight-secondary-ion-mass spectroscopy (ToF-SIMS) and X-ray photoelectron spectroscopy (XPS). The ToF-SIMS identified the coatings ashydride by a large SiH+ peak, extensive-length SinHm

− fragments, and an increased Si2+ in the positive spectrum as well as a dominant H− inthe negative spectrum. Since hydrogen is the only element with no primary element peaks, XPS identified the H content of the SiH coatingsa ate elementa hly stablea ssivation oft ncorporationm©

K

1

ttiacsoiocwte

,tionthe

aus-ethe

ionelec-

end,dingeV

sedtion. Anlue

or-elec-tors

0d

s comprising novel silicon hydrides due to new peaks at 11, 43, and 55 eV in the absence of corresponding peaks of any candidt higher binding energies. The silicon hydride surface was remarkably stable to air as shown by ToF-SIMS and XPS. The higmorphous silicon hydride coating may advance the production of integrated circuits and microdevices by resisting the oxygen pa

he surface and possibly altering the dielectric constant and band gap to increase device performance. The percentage hydride iay be used to increase or tune the band gap and optical properties in photovoltaics and electronics devices.2005 Elsevier B.V. All rights reserved.

eywords: Catalysis; Chemical vapor deposition; Hydrides; Plasma processing; Silane

. Introduction

Aqueous HF acid etching of silicon surfaces results inhe removal of the surface oxide and produces hydrogenerminated silicon surfaces, Si–H. HF etching is a key stepn producing silicon surfaces which are contamination-freend chemically stable for subsequent processing in the semi-onductor industry[1–3]. In fact, chemical oxidation andubsequent HF treatment of Si surfaces are used prior to gatexidation, where surface contamination (<10 ppm level) and

nterface control are crucial to device performance[4]. Flu-rine termination was initially considered the basis of thehemical stability of HF-treated surfaces. Subsequently, itas found that fluorine is a minor species on the surface and

hat the remarkable surface passivation achieved by HF isxplained by H termination of silicon dangling bonds pro-

∗ Corresponding author. Tel.: +1 609 490 1090; fax: +1 609 490 1066.E-mail address: rmills@blacklightpower.com (R. Mills).

tecting the surface from chemical attack[3,5,6]. Howeverthe replacement of the oxide layer with the H terminaof the silicon dangling bonds by HF can be attributed toincreased electronegativity of fluoride ion versus oxide cing an enhanced reactivity of H+ which attacks the oxidlayer. The electron affinity of halogens increases frombottom of the group VII elements to the top. Hydridemay be considered a halide since it possess the sametronic structure. And, according to the binding energy trit should have a high binding energy. However, the binenergy is only 0.75 eV which is much lower than the 3.4binding energy of a fluoride ion. And, once the HF is rinfrom the surface, the Si–H layer undergoes rapid oxidawhen exposed to oxygen or solvents containing oxygenSi–H layer with enhanced stability would be of great vato the semiconductor industry.

Amorphous Si–H films, the active component of imptant semiconductor devices such as photovoltaics, optotronics, liquid crystal displays, and field-effect transis

254-0584/$ – see front matter © 2005 Elsevier B.V. All rights reserved.oi:10.1016/j.matchemphys.2005.05.002

R. Mills et al. / Materials Chemistry and Physics 94 (2005) 298–307 299

are formed by plasma enhanced chemical vapor deposition(PECVD) techniques[7]. Hydrogen stability is also impor-tant for optoelectronic performance and durability to devicessuch as photovoltaics[8]. A review of the current state ofphotovoltaic technology is given by Green[9]. Typically thefilm is grown on a silicon wafer substrate exposed to a plasmaof silane, hydrogen, and often argon using a reactor with adiode configuration in which the plasma is confined betweentwo parallel electrodes. In this study, we find that the aque-ous HF acid etched surface undergoes rapid oxidation whenexposed to air and provides little protection from such expo-sure. Whereas, a novel highly air stable amorphous siliconhydride surface coating designated�-SiH was synthesizedby microwave plasma reaction of mixture of silane, hydro-gen, and helium wherein it is proposed that He+ served asa catalyst with atomic hydrogen. The novel�-SiH film mayadvance semiconductor fabrication and devices.

It was reported previously that a new plasma source hasbeen developed that operates by incandescently heating ahydrogen dissociator to provide atomic hydrogen and heatsa catalyst such that it becomes gaseous and reacts with theatomic hydrogen to produce a plasma called a resonant trans-fer or rt-plasma. It was extraordinary, that intense VUV emis-sion was observed[10] at low temperatures (e.g.,≈103 K)and an extraordinary low field strength of about 1–2 V cm−1

from atomic hydrogen and certain atomized elements or cer-t germ eV.It hichi ans-fi areh ableh eeno ningpu erm

ibil-i cesa ) andp ltra-v -t ucts[ ]clh wd so sg

conh times n. Int e in

a helium–hydrogen and control xenon–hydrogen microwavedischarge plasmas at the surface of a nickel foil. After theplasma reaction processing, the surface was characterized byToF-SIMS and XPS.

2. Experimental

2.1. Synthesis

Amorphous silicon hydride (�-SiH) films were grownon nickel substrates by their exposure to a low-pressuremicrowave discharge of SiH4 (2.5%)/He (96.6%)/H2 (0.9%).The experimental set up comprising a microwave dischargecell operated under flow conditions is shown inFig. 1.The SiH4 gas was introduced into a 1000 ml reservoirby a gas/vacuum line where it was mixed with premixedHe (99%)/H2 (1%) to obtain the reaction mixture SiH4(2.5%)/He (96.6%)/H2 (0.9%) by controlling the individ-ual gas pressures. In control experiments Xe replaced He.Nickel foil (5 mm× 5 mm and 0.05 mm thick, Alfa Aesar,99+%) substrates were used instead of silicon wafers toavoid charging during ToF-SIMS and XPS characterizationas well as background interference in the analysis. The sub-strates were placed inside of a quartz tube (1.3 cm in diame-ter× 15.5 cm long) with vacuum valves at both ends. Thet vity( . Theq tly tor SiH( eq witha con-t y ana wn inF ncy:2 witha in.Y wall

F l oper-a

ain gaseous ions which singly or multiply ionize at inteultiples of the potential energy of atomic hydrogen, 27.2

n this study, He+ was used as a catalyst to form�-SiH sincehe second ionization energy of helium is 54.417 eV, ws equivalent to 2.272 eV. In this case, 54.417 eV is trerred nonradiatively from atomic hydrogen to He+ whichs resonantly ionized. The catalysis reaction productsighly stable intermediates which react to form novel stydrides. Hydride ions with high binding energies have bbserved by XPS and by solid state magic-angle spinroton nuclear magnetic resonance (1H MAS NMR) havingpfield shifted peaks[11]. The unique role of fast H in othaterial processes was discussed previously[12].Additional prior related studies that support the poss

ty of a novel reaction of atomic hydrogen which produchemically generated or assisted plasma (rt-plasma

roduces novel hydride compounds include extreme uiolet (EUV) spectroscopy[10,13–25,27,29], characterisic emission from catalysts and the hydride ion prod17–20], lower-energy hydrogen emission[13–15,21–22,27,hemically formed plasmas[10,16–20,23–24], Balmer �ine broadening[10,12,14–15,17,22,25,27,29,30], invertedydrogen populations[29], anomalous plasma afterglouration[23–24], power generation[15,26–28], and analysif novel chemical compounds[12,31–34]. The theory waiven previously[13–15].

Even a partial coating of an extremely stable siliydride may stabilize a silicon surface to unprecedentedcales to increase the yield in integrated chip fabricatiohis paper, we report the results of the reaction of silan

ube was fitted with an Opthos coaxial microwave caEvenson cavity) and connected to the gas/vacuum lineuartz tube and vacuum line were evacuated sufficienemove any trace moisture or oxygen. The gas mixture42.5%)/He (Xe) (96.6%)/H2 (0.9%) was flowed through thuartz tube at a total pressure of 0.7 Torr maintainedgas flow rate of 40 sccm controlled by a mass flow

roller with a readout. The cell pressure was monitored bbsolute pressure gauge. The microwave generator shoig. 1was an Opthos model MPG-4M generator (freque450 MHz). The microwave plasma was maintained40 W (forward)/15 W (reflected) power for about 20 m

ellow-orange coatings formed on the substrates and the

ig. 1. The experimental set up comprising a microwave discharge celted under flow conditions.

300 R. Mills et al. / Materials Chemistry and Physics 94 (2005) 298–307

of the quartz tube. The quartz tube was removed and trans-ferred to a drybox with the samples inside by closing thevacuum valves at both ends and detaching the tube fromthe vacuum/gas line. The coated substrates were mounted onXPS and ToF-SIMS sample holders under an argon atmo-sphere in order to prepare samples for the correspondinganalyses. One set of samples was analyzed with air exposurelimited to 10 min and another for 20 min while transferringand mounting during the analyses. Separate samples wereremoved from the drybox and stored in air at room tem-perature for 48 h or 10 days before the analyses. Controlscomprised a commercial silicon wafer (Alfa Aesar, 99.99%)untreated, and HF cleaned silicon wafers exposed to air for10 min or 3 h. Further controls comprised nickel foils coatedby reaction of SiH4–Xe–H2 plasmas and exposed to air for10 min or 1 h.

2.2. ToF-SIMS characterization

The commercial silicon wafer, HF cleaned silicon wafer,and coated nickel foil samples from SiH4–He–H2 andSiH4–Xe–H2 plasmas were characterized using PhysicalElectronics TRIFT ToF-SIMS instrument. The primary ionsource was a pulsed69Ga+ liquid metal source operated at15 keV [35–36]. The secondary ions were extracted by a±3 keV (according to the mode) voltage. Three electrostatica ino ionso 5 kHzr was6 Thet -d lsede er tor ce fora ing a8 at ne si-t sen-t dataw tron-i nedr

2

usinga rans-m used.T was0 was0 , andt , thet 30. C1

The valance band edge of the�-SiH film formed by theSiH4–He–H2 plasma reaction was determined relative to thatof crystalline silicon using XPS of the valance region in orderto estimate the band gap. This required careful minimizationof the charging effect that causes a spectral shift. The chargingeffect was nullified by two procedures: (1) charge neutraliza-tion was applied during XPS data acquisition, and (2) thespectra of the�-SiH film and the crystalline silicon samplewere further calibrated by the Si 2p3/2 core level that was setat a binding energy of 99.8 eV.

3. Results and discussion

3.1. ToF-SIMS

The positive ToF-SIMS spectra (m/e = 0–100) of the com-mercial silicon wafer, HF cleaned silicon wafer, and coatednickel foil samples from SiH4–He–H2 and SiH4–Xe–H2 plas-mas are shown inFig. 2a–f. The positive ToF-SIMS spectra ofthe noncoated cleaned commercial silicon wafer and a nickelfoil coated with an�-SiH film by reaction of a SiH4–He–H2plasma and exposed to air for 10 min are shown inFig. 2a ande, respectively. The positive ion spectrum of the control wasdominated by Si+, oxides SixOy

+, and hydroxides Six(OH)y+;whereas, that of the�-SiH sample contained essentially noo Sia mS assba o-r 9.6.R as,fpi

S dtd le to

TS

N

22334444455

nalyzers (Triple-Focusing-Time-of-Flight) deflect themrder to compensate for the initial energy dispersion off the same mass. The 400 pA dc current was pulsed at aepetition rate with a 7 ns pulse width. The analyzed area0�m× 60�m and the mass range was 0–1000 amu.

otal ion dose was 7× 1011 ions cm−2, ensuring static conitions. Charge compensation was performed with a pulectron gun operated at 20 eV electron energy. In ordemove surface contaminants and expose a fresh surfanalysis, the samples were sputter-cleaned for 30 s us0�m× 80�m raster, with 600 pA current, resulting in

otal ion dose of 1015 ions cm−2. Three different regions oach sample of 60�m× 60�m were analyzed. The po

ive and negative SIMS spectra were acquired. Repreative post sputtering data is reported. The ToF-SIMSere treated using ‘Cadence’ software (Physical Elec

cs), which calculates the mass calibration from well-defieference peaks.

.3. XPS characterization

A series of XPS analyses were made on the samplesScienta 300 XPS Spectrometer. The fixed analyzer tission mode and the sweep acquisition mode werehe angle was 15◦. The step energy in the survey scan.5 eV, and the step energy in the high resolution scan.15 eV. In the survey scan, the time per step was 0.4 s

he number of sweeps was 4. In the high resolution scanime per step was 0.3 s, and the number of sweeps wass at 284.5 eV was used as the internal standard.

xide or hydroxide peaks. Rather, it was dominated by+

nd a peak atm/z = 29 which comprised a contribution froiH+ and29Si+ which were difficult to separate definitively,hown inFig. 3. However, the contribution due to SiH+ coulde determined by calculating the ratioR = 28Si/(28SiH +29Si)fter Fahey and Messenger[37]. For comparison, the theetical ratio of28Si/29Si based on isotopic abundance is 1for the clean noncoated silicon wafer was 8.1. WhereR

or the �-SiH sample was 1.15 indicating that them/z = 29eak was overwhelmingly due to SiH+. Additional positive

on peaks of minority species are identified inTable 1.The positive spectrum of a nickel foil coated with an�-

iH film by reaction of a SiH4–He–H2 plasma and exposeo air for 10 days is shown inFig. 2f. In this case,R was 1.75emonstrating that the sample was extraordinarily stab

able 1ummary of�-SiH-related peaks from the positive spectra ofFig. 2a–f

ominal mass,m/z Compound or fragment

1 H8 28Si9 29Si, 28SiH0 30Si, 29SiH, 28SiH2

1 28SiH3, 29SiH2,30SiH, etc.0 SiC2 SiN3 SiNH, C3H7

4 SiO5 SiOH2 SiC2

6 Si2

R. Mills et al. / Materials Chemistry and Physics 94 (2005) 298–307 301

Fig. 2. Positive ToF-SIMS characterization (m/e = 0–100) of the commercialsilicon wafer, HF cleaned silicon wafer, and coated nickel foil samples fromSiH4–He–H2 and SiH4–Xe–H2 plasmas. (a) The positive ion ToF-SIMSspectrum of a noncoated cleaned commercial silicon wafer (Alfa Aesar,99.9%). (b) The positive ion ToF-SIMS spectrum of the HF cleaned sil-icon wafer exposed to air for 10 min before ToF-SIMS analysis. (c) Thepositive ion ToF-SIMS spectrum of a nickel foil coated by reaction of aSiH4–Xe–H2 plasma and exposed to air for 10 min before ToF-SIMS anal-ysis. (d) The positive ion ToF-SIMS spectrum of a nickel foil coated byreaction of a SiH4–Xe–H2 plasma and exposed to air for 1 h before ToF-SIMS analysis that showed the loss of the large SiH+ peak. (e) The positiveion ToF-SIMS spectrum of a nickel foil coated with an�-SiH film by reac-tion of a SiH4–He–H2 plasma and exposed to air for 10 min that showed alarge SiH+ peak. (f) The positive ion ToF-SIMS spectrum of a nickel foilcoated with an�-SiH film by reaction of a SiH4–He–H2 plasma and exposedto atmosphere for 10 days before the ToF-SIMS analysis that retained a largeSiH+ peak.

air exposure. In contrast,R was 2.45 in the positive spectrumof the HF cleaned silicon wafer exposed to air for only 10 minas shown inFig. 2b. The exceptional air-stability of the�-SiH film formed by reaction of a SiH4–He–H2 plasma wasfurther demonstrated by comparison with air-exposed coat-ings from SiH4–Xe–H2 plasmas. The positive-ion ToF-SIMSof coatings from a SiH4–Xe–H2 plasma with 10 min and 1 hair exposure are shown inFig. 2c and d, respectively. TheSiH4–Xe–H2 plasma-film showed the loss of a large SiH+

peak with a relatively short duration air-exposure. At 1 hexposure,R of the SiH4–Xe–H2 plasma-film was 2.41 whichwas similar to that of the HF cleaned silicon wafer exposed toair for 10 min; whereas, the SiH4–He–H2 plasma-film shown

Fig. 3. High resolution positive ToF-SIMS characteristic (m/e = 28.95–29.02) of HF-cleaned Si wafer and a Ni foil-coated with an�-SiH film.

in Fig. 2f retained the large SiH+ peak at 10 days continu-ous air exposure. These results support the role of He+ as acatalyst to form a highly stable hydride.

ToF-SIMS is not directly quantitative for the stoichiome-try of silicon hydrides. However, based on the isotopic abun-dance ratio29Si/28Si of (0.051), the28SiH/28Si ratio can beestimated. The ToF-SIMS determined28SiH/28Si ratio for theHF-etched Si(1 1 1) surface and the�-SiH formed by reactionof a SiH4–He–H2 plasma was 0.35 and 0.82, respectively. Itis generally recognized that each Si atom of the top layer of anHF-etched Si(1 1 1) surface is terminated by one H atom, i.e.,the stoichiometry of the top layer is H–Si(1 1 1)[38]. SinceToF-SIMS measures the species mainly from the top layer,the maximum28SiH/28Si ratio of an HF-etched Si(1 1 1) sur-face is unity. But, fragmentation of SiH into Si and H occursduring ToF-SIMS analysis due to the ionization processescaused by the bombardment with a Ga+ source. Therefore,the 28SiH/28Si ratio from HF-etched Si(1 1 1) surfaces wasobserved to be far less than 1, but it was quantifiable andreproducible. Furthermore, the28SiH/28Si ratio of the�-SiHformed by reaction of a SiH4–He–H2 plasma was about 2.3times higher than that of the HF-etched Si(1 1 1). Given theH–Si(1 1 1) stoichiometry of HF-etched Si(1 1 1), it can beinferred that the average stoichiometry of the�-SiH investi-gated was SiH2.3.

Si2+ ion does not occur naturally, but it may be observedi rd-m ll.Hs thec mt ni feri t2 ch adfi is

n the positive ToF-SIMS spectrum formed by ion bombaent. Typically, the Si2+ is very weak if it is observed at aowever, the�-SiH film formed by the SiH4–He–H2 plasmahowed an Si2+ peak was abnormally intense relative toorresponding Si+ peak. SiH2+ was also observed only frohe�-SiH film formed by the SiH4–He–H2 plasma. As shown Table 2, control experiments with HF cleaned silicon wandicate that the normalized Si2+ of the�-SiH film is abou0–50 times higher than typical of the clean Si wafer. Sueviation implies that the Si2+ comes directly from the�-SiHlm rather than from primary (Ga+) ion bombardment. Th

302 R. Mills et al. / Materials Chemistry and Physics 94 (2005) 298–307

Table 2Summary of peaks from the positive spectra ofFig. 2a–f

Sample 28SiH+/28Si+ 28Si2+/28Si+ (×104)

HF-treated silicon wafer(5s ion sputter)

0.146 8.45

�-SiH film (nonsputtered) 0.972 403�-SiH film (5s ion sputter) 0.576 205

is further evidence of a highly stabilized hydride which per-mits silicon to be exceptionally electropositive. A potentialapplication is as an intrinsic silicon p-dopant.

The negative ion ToF-SIMS spectra (m/e = 0–100) ofthe commercial silicon wafer, HF cleaned silicon wafer,and coated nickel foil samples from SiH4–He–H2 andSiH4–Xe–H2 plasmas are shown inFig. 4a–g. The nega-tive ion spectra of the noncoated cleaned commercial siliconwafer and a nickel foil coated with an�-SiH film by reac-tion of a SiH4–He–H2 plasma and exposed to air for 10 minare shown inFig. 4a and f, respectively. The control spec-trum was dominated by oxide (O− m/z = 16) and hydroxide(OH− m/z = 17); whereas, spectrum of the�-SiH film wasdominated by hydride ion (H− m/z = 1). Very little oxide orhydroxide was observed. Silicon hydride clusters SinHm

−shown inFig. 5a–d were observed from the�-SiH film formedby the SiH4–He–H2 plasma, but not from the controls. Addi-tional negative ion peaks of minority species are identified inTable 3.

The negative spectrum (m/e = 0–100) of a nickel foilcoated with an�-SiH film formed by the SiH4–He–H2 plasma

Table 3Summary of some predominant peaks from the negative spectra ofFig. 4a–g

Nominal mass,m/z Compound or fragment

1 H11111122223333344444666777

Fig. 4. Negative ToF-SIMS characterization (m/e = 0–100) of the commer-cial silicon wafer, HF cleaned silicon wafer, and coated nickel foil samplesfrom SiH4–He–H2 and SiH4–Xe–H2 plasmas. (a) The negative ion ToF-SIMS spectrum of the noncoated cleaned commercial silicon wafer (AlfaAesar, 99.99%). (b) The negative ion ToF-SIMS spectrum of the HF cleanedsilicon wafer exposed to air for 10 min before ToF-SIMS analysis. (c) Thenegative ion ToF-SIMS spectrum of the HF cleaned silicon wafer exposedto air for 3 h before ToF-SIMS analysis showing a dominant oxide peak. (d)The negative ion ToF-SIMS spectrum of a nickel foil coated by reaction ofa SiH4–Xe–H2 plasma exposed to air for 10 min before ToF-SIMS analysisshowing a dominant hydride ion peak. (e) The negative ion ToF-SIMS spec-trum of a nickel foil coated by reaction of a SiH4–Xe–H2 plasma exposed toair for 1 h before ToF-SIMS analysis showing a dominant oxide peak and lossof the hydride peak. (f) The negative ion ToF-SIMS spectrum of a nickel foilcoated with an�-SiH film by reaction of a SiH4–He–H2 plasma and exposedto air for 10 min before ToF-SIMS analysis that was dominated by hydrideion. (g) The negative ion ToF-SIMS spectrum of a nickel foil coated with an�-SiH film by reaction of a SiH4–He–H2 plasma and exposed to air for 10days before the ToF-SIMS analysis that retained the dominant hydride ionpeak.

and exposed to air for 10 days is shown inFig. 4g. In this case,hydride ion also dominated the negative spectrum demon-strating extraordinary air stability of the�-SiH film. Thenegative spectrum (m/e = 0–100) of the HF cleaned siliconwafer exposed to air for only 10 min shown inFig. 4b alsoshows a dominant hydride as well as oxide, hydroxide, andsome fluoride (F− m/z = 19). However, the HF treated surfacewas not stable with prolonged air exposure. A dominant oxidepeak was observed in the negative spectrum (m/e = 0–100)

2 C3 CH4 CH2

6 O7 OH9 F5 C2H6 C2H2

8 28Si9 29Si1 30SiH, 29SiH2, 28SiH3

2 28SiH4

3 HO2

5 35Cl7 37Cl0 28SiC1 29SiC2 30SiC4 28SiO5 28SiOH0 29SiO2

1 29SiO2, 28SiO2H2 28SiO2H2, 29SiO2H5 SiO2NH6 28SiO

3

7 29SiO3, 28SiO3Hof the HF cleaned silicon wafer exposed to air for only 3 has shown inFig. 4c. Hydride was also observed in lesser

R. Mills et al. / Materials Chemistry and Physics 94 (2005) 298–307 303

Fig. 5. Negative ToF-SIMS characterization of an�-SiH film formed by the reaction of a SiH4–He–H2 plasma and exposed to air for 10 min before ToF-SIMSanalysis. In addition to a dominant hydride peak, silicon hydride clusters SinHm

− were observed that were not observed from the HF cleaned silicon wafer orthe film from the SiH4–Xe–H2 plasma. (a) The negative ion ToF-SIMS spectrum (m/e = 27.5–48) showing [HmSi]− fragments. (b) The negative ion ToF-SIMSspectrum (m/e = 55.5–78.5) showing [HmSi2]− fragments. (c) The negative ion ToF-SIMS spectrum (m/e = 80–108) showing [HmSi3]− fragments. (d) Thenegative ion ToF-SIMS spectrum (m/e = 110–145) showing [HmSi4]− fragments.

amounts and may have resulted as a fragment of the observedhydroxide. Fluoride (F− m/z = 19) was also observed. TheToF-SIMS results from the HF treated surface is consistentwith predominantly H termination of silicon dangling bondsas reported previously[3,5,6] that has undergone rapid oxi-dation to form mixed oxides such as SiOH.

These results indicate that the SiH4–He–H2 plasma reac-tion formed a highly stable hydrogenated silicon coating inthe absence of fluorine observed on the HF treated surface.Remarkably, the�-SiH film was stable even after 10 days;whereas, the HF treated surface showed signs of oxidationover a 1500 times shorter time scale—10 min. At 3 h theHF treated surface had similarities to the control untreatedsilicon wafer which comprised a full oxide coating. Theexceptional air-stability of the�-SiH film formed by reactionof a SiH4–He–H2 plasma was demonstrated by compari-son with air-exposed coatings from a SiH4–Xe–H2 plasma.The negative-ion ToF-SIMS of coatings from a SiH4–Xe–H2

plasma with 10 min and 1 h air exposure are shown inFig. 4dand e, respectively. The SiH4–Xe–H2 plasma-film showedthe loss of a large H− peak with a relatively short durationair-exposure. At 1 h exposure, the SiH4–Xe–H2 plasma-filmwas covered by an oxide with a spectrum which was sim-ilar to that of the HF cleaned silicon wafer exposed to airfor the same period of time of only 10 min; whereas, theSiH4–He–H2 plasma-film shown inFig. 4g retained the dom-inant hydride peak at 10 days continuous air exposure. Theseresults support the role of He+ as a catalyst to form a highlystable hydride.

The SiH4–He–H2 plasma-reaction-formed�-SiH is pro-posed to comprise a more stable hydride ion than the Hterminated silicon from HF treatment. Thus, the ion pro-duction efficiencies in ToF-SIMS analysis could be differentmaking a comparison only qualitative and indicative of rel-ative changes that occurred with timed air exposure. Sincethe Si 2p electron of all samples was equivalent except for

304 R. Mills et al. / Materials Chemistry and Physics 94 (2005) 298–307

Fig. 6. XPS survey scan characterizations of the commercial silicon waferand coated nickel foil samples from SiH4–He–H2 and SiH4–Xe–H2 plasmas.(a) The XPS survey scan of the noncoated cleaned commercial silicon wafershowing a large amount of oxide and carbon contamination of the surface.(b) The XPS survey scan of a nickel foil coated by reaction of a SiH4–Xe–H2

plasma and exposed to air for 10 min before XPS analysis showing a largeamount of oxide and carbon contamination of the surface. (c) The XPS surveyscan of a nickel foil coated with an�-SiH film by reaction of a SiH4–He–H2

plasma and exposed to air for 20 min before XPS analysis showing minimaloxide and carbon.

energy shifts due to the presence of ordinary or novel hydride,or oxide, qualitative analysis was possible as given in the XPSsection. As shown in this section, the ToF-SIMS results wereconfirmed by XPS.

3.2. XPS

The XPS survey spectra of the noncoated cleaned commer-cial silicon wafer (control), a nickel foil coated by reactionof a SiH4–Xe–H2 plasma and exposed to air for 10 min (con-trol), and a nickel foil coated with an�-SiH film by reactionof a SiH4–He–H2 plasma and exposed to air for 20 min wereobtained over the regionEb = 0 to 1200 eV and are shownin Fig. 6a–c, respectively. The survey spectra permitted thedetermination of all of the elements present and detectedshifts in the binding energies of the Si 2p peak, which alsoidentifies the presence or absence of SiO2. The major speciesidentified in the XPS spectra of the control samples weresilicon, oxygen, and carbon. The�-SiH sample containedessentially silicon with negligible oxygen and carbon.

XPS spectra (96–108 eV) in the region of the Si 2ppeak of the commercial silicon wafer, HF cleaned siliconwafer, and coated nickel foil samples from SiH4–He–H2and SiH4–Xe–H2 plasmas are shown inFig. 7a–e. The XPSspectra (96–108 eV) in the region of the Si 2p peak of the non-coated cleaned commercial silicon wafer (control), a nickelf dtb for

Fig. 7. XPS spectra (96–108 eV) in the region of the Si 2p peak of thecommercial silicon wafer, HF cleaned silicon wafer, and coated nickel foilsamples from SiH4–He–H2 and SiH4–Xe–H2 plasmas. (a) The XPS spec-trum of the noncoated cleaned commercial silicon wafer showing a largeSiO2 in the region of 104 eV. (b) The XPS spectrum of the HF cleaned sil-icon wafer exposed to air for 10 min before XPS analysis showing a verylarge SiOx peak in the region of 101.5–104 eV. (c) The XPS survey scan ofa nickel foil coated by reaction of a SiH4–Xe–H2 plasma and exposed toair for 10 min before XPS analysis showing a large SiO2 in the region of104 eV. (d) The XPS spectrum of a nickel foil with an�-SiH film by reactionof a SiH4–He–H2 plasma and exposed to air for 20 min before XPS analysisshowing no oxide in the region of 104 eV. (e) The XPS spectrum of a nickelfoil coated with an�-SiH film by reaction of a SiH4–He–H2 plasma andexposed to air for 48 h before the XPS analysis showing no oxide at 104 eVand possibly trace SiOH in the 102 eV region.

20 min are shown inFig. 7a, c and d, respectively. The XPSspectra of the controls showed a large SiO2 content at 104 eVas given by Wagner et al.[39]. In contrast, the SiH4–He–H2plasma-film had essentially no SiO2. In addition, spin-orbitalcoupling gives rise to a split Si 2p peak in pure silicon, butthis peak changed to a single broad peak upon reaction toform the�-SiH film indicative of amorphous silicon.

The XPS spectrum (96–108 eV) in the region of the Si2p peak of a nickel foil coated with an�-SiH film by reac-tion of a SiH4–He–H2 plasma and exposed to air for 48 h isshown inFig. 7e. Essentially no SiO2 was observed at 104 eVdemonstrating that the sample was extraordinarily stable toair exposure. Perhaps trace SiOH is present in the region of102 eV potentially due to less than 100% coverage of the sur-face with the�-SiH film; rather, some silicon deposition may

oil coated by reaction of a SiH4–Xe–H2 plasma and exposeo air for 10 min (control), and a nickel foil with an�-SiH filmy reaction of a SiH4–He–H2 plasma and exposed to air

R. Mills et al. / Materials Chemistry and Physics 94 (2005) 298–307 305

Fig. 8. XPS spectra (525–540 eV) in the region of the O 1s peak of theHF cleaned silicon wafer and coated nickel foil samples from SiH4–He–H2

and SiH4–Xe–H2 plasmas. (a) The XPS spectrum of the HF cleaned siliconwafer exposed to air for 10 min before XPS analysis showing a very largeoxide peak. (b) The XPS survey scan of a nickel foil coated by reaction ofa SiH4–Xe–H2 plasma and exposed to air for 10 min before XPS analysisshowing a very large oxide peak. (c) The XPS spectrum of a nickel foil coatedwith an�-SiH film by reaction of a SiH4–He–H2 plasma and exposed to airfor 48 h before XPS analysis showing a minimal amount of oxide.

have occurred. In contrast, the XPS spectrum (96–108 eV) inthe region of the Si 2p peak of the HF cleaned silicon waferexposed to air for 10 min was essentially fully covered by par-tial oxides SiOx such as SiOH. The mixed silicon oxide peakin the region of 101.5–104 eV shown inFig. 7b was essen-tially the same percentage of the Si 2p as that of the SiO2peak of the uncleaned wafer shown at 104 eV inFig. 7a. Inaddition, the O 1s peak of the HF cleaned wafer exposedto air for 10 min and the coating formed by reaction of aSiH4–Xe–H2 plasma were intense as shown inFig. 8a and b,respectively. Whereas, the�-SiH film formed by reaction ofa SiH4–He–H2 plasma and exposed to air for 48 h shown inFig. 8c was negligible.

The 0–70 eV binding energy region of high resolutionXPS spectra of controls, the commercial silicon wafer, a HFcleaned silicon wafer exposed to air for 10 min, and a nickelfoil coated by reaction of a SiH4–Xe–H2 plasma and exposedto air for 10 min, are shown inFig. 9a–c, respectively. Only alarge O 2s peak in the low binding energy region was observedin each case. InFig. 9a–c, the C, O, and Si valance bands areobserved in the region 0–10 eV. The 0–70 eV binding energyregion of a nickel foil coated with an�-SiH film by reaction

Fig. 9. The 0–70 eV binding energy region of high resolution XPS spectraof the commercial silicon wafer, HF cleaned silicon wafer, and a coatednickel foil sample from a SiH4–Xe–H2 plasma. (a) The XPS spectrum ofthe commercial silicon wafer showing only a large O 2s peak in the lowbinding energy region. (b) The XPS spectrum of the HF cleaned siliconwafer exposed to air for 10 min before XPS analysis showing only a largeO 2s peak in the low binding energy region. (c) The XPS survey scan of anickel foil coated by reaction of a SiH4–Xe–H2 plasma and exposed to airfor 10 min before XPS analysis showing only a large O 2s peak in the lowbinding energy region.

of a SiH4–He–H2 plasma and exposed to air for 20 min beforeXPS analysis is shown inFig. 10. By comparison of the�-SiH sample to the controls, novel XPS peaks were identifiedat 11, 43, and 55 eV. These peaks do not correspond to anyof the primary elements, silicon, carbon, or oxygen, shown

Fig. 10. The 0–70 eV binding energy region of a high resolution XPS spec-trum of a nickel foil coated with an�-SiH film by reaction of a SiH4–He–H2

p peakso mentsi highlysh

lasma and exposed to air for 20 min before XPS analysis. The novelbserved at 11, 43 and 55 eV which could not be assigned to the ele

dentified by their primary XPS peaks matched and were assigned totable silicon hydrides formed by the catalytic reaction of He+ and atomicydrogen.

306 R. Mills et al. / Materials Chemistry and Physics 94 (2005) 298–307

in the survey scan inFig. 6c, wherein the peaks of these ele-ments are given by Wagner et al.[39]. Hydrogen is the onlyelement which does not have primary element peaks; thus,it is the only candidate to produce the novel peaks and cor-respond to the H content of the SiH coatings. These peaksclosely matched hydrides formed by the catalytic reactionof He+ with atomic hydrogen and subsequent reactions toform highly stable silicon hydride products�-SiH that werediscussed previously[34].

These results indicate that the plasma reaction formed ahighly stable novel hydrogenated coating; whereas, the con-trol comprised an oxide coating or an unstable hydrogenterminated silicon surface which rapidly formed an oxidepassivation layer. The hydrogen content of the�-SiH coat-ing formed by reaction of a SiH4–He–H2 plasma appears tobe novel silicon hydrides with high binding energies whichaccount for the exceptional air stability.

The normalized XPS spectra of the valance band of the�-SiH film formed by the SiH4–He–H2 plasma reaction and thecrystalline silicon sample are shown inFig. 11. The energyof each valance band edge is determined as the intercept ofthe extrapolated linear part of the band edge just above thethreshold with the baseline. The valance band edge of the�-SiH film was shifted about 0.5 eV higher when comparedto that of the crystalline silicon sample. This result indicatesthat the valance band in the�-SiH film is 0.5 eV lower thant

n thes anyc n bea withs allert nb n-

Ff liness hichi .5 eV.

duction band in crystalline silicon. Based on the 1.1 eV bandgap of crystalline silicon, the band gap of the�-SiH film isestimated to be in the range of 1.6–2.1 eV.

4. Conclusions

Nickel substrates were coated by the reaction productof a low-pressure microwave discharge plasma of SiH4(2.5%)/He (96.6%)/H2 (0.9%). The ToF-SIMS identified thecoatings as hydride by the large SiH+ peak in the posi-tive spectrum and silicon hydride clusters SinHm

− and thedominant H− in the negative spectrum. XPS identified theH content of the SiH coatings as hydride corresponding topeaks at 11, 43, and 55 eV, respectively, in the absence of anyother identifiable primary element peaks. The novel hydridesare proposed to form by the catalytic reaction of He+ withatomic hydrogen with subsequent reactions to form highlystable silicon hydride products. The role of He+ as a cata-lyst was supported by the observation that the replacement ofhelium (He+) with xenon (Xe+) resulted in no notable char-acteristics from the films. The SiH coating was amorphous asindicated by the shape of the Si 2p peak and was remarkablystable to air exposure. After a 48 h exposure to air, essen-tially no oxygen was observed as evidence by the negligibleO 1s peak at 531 eV and absence of any SiOx Si 2p peakii sicp oat-i andm sur-f bandg dridei d gapa icesa niuma

R

.

ppl.

73.ht,

iver-

rgy

[ 002)

[ ro-

[ ter.

he valance band in crystalline silicon.An estimate of the band gap can be made based o

hift and the known band gap of crystalline silicon. In mases, the shift of the conduction band edge which canalyzed by a Si K-edge photoabsorption experimentynchrotron radiation is about the same or slightly smhan the valance band shift[40–42]. Thus, the conductioand in the�-SiH film is up to 0.5 eV higher than the co

ig. 11. The normalized XPS spectra of the valance band of the�-SiH filmormed by the SiH4–He–H2 plasma reaction (dotted line) and the crystalilicon sample (solid line). The valance band edge of the�-SiH film washifted about 0.5 eV higher than that of the crystalline silicon sample wndicates that the corresponding relative valence band is lowered by 0

n the region of 102–104 eV. The extraordinary Si2+ peakn the positive ToF-SIMS suggests a possibility for intrin-doping. The highly stable amorphous silicon hydride c

ng may advance the production of integrated circuitsicrodevices by resisting the oxygen passivation of the

ace and possibly altering the dielectric constant andap to increase device performance. The percentage hy

ncorporation may be used to increase or tune the bannd optical properties in photovoltaics and electronic devs attempted with hydrogenated amorphous silicon–selelloys and by chemical annealing processes[43,44].

eferences

[1] W. Kern, Semicond. Int. 7 (4) (1984) 94.[2] F.J. Grunthaner, P.J. Grunthaner, Mater. Sci. Rep. 1 (1986) 69[3] M. Grudner, H. Jacob, Appl. Phys. A 39 (1986) 73.[4] G.S. Higashi, Y.J. Chabal, G.W. Trucks, K. Raghavachari, A

Phys. Lett. 56 (1990) 656.[5] H. Ubara, T. Imura, A. Hiraki, Solid State Commun. 50 (1984) 6[6] E. Yablonovitch, D.L. Allara, C.C. Chang, T. Gmitter, T.B. Brig

Phys. Rev. Lett. 57 (1986) 249.[7] R.A. Street, Hydrogenated Amorphous Silicon, Cambridge Un

sity Press, Cambridge, 1991, pp. 18–61.[8] S. Sheng, X. Liao, Z. Ma, G. Yue, Y. Wang, G. Kong, Solar Ene

Mat. Solar Cells 68 (2001) 123.[9] M.A. Green, Energy Policy 28 (2000) 989.10] R. Mills, M. Nansteel, P. Ray, IEEE Trans. Plasma Sci. 30 (2

639.11] R. Mills, B. Dhandapani, M. Nansteel, J. He, A. Voigt, Int. J. Hyd

gen Energy 26 (2001) 965.12] R.L. Mills, J. Sankar, A. Voigt, J. He, B. Dhandapani, Chem. Ma

15 (2003) 1313.

R. Mills et al. / Materials Chemistry and Physics 94 (2005) 298–307 307

[13] R. Mills, P. Ray, Int. J. Hydrogen Energy 27 (2002) 301.[14] R.L. Mills, P. Ray, J. Phys. D: Appl. Phys. 36 (2003) 1535.[15] R.L. Mills, P. Ray, B. Dhandapani, M. Nansteel, X. Chen, J. He, J.

Mol. Struct. 643 (1–3) (2002) 43–54.[16] R. Mills, J. Dong, Y. Lu, Int. J. Hydrogen Energy 25 (2000) 919.[17] R.L. Mills, P. Ray, Int. J. Hydrogen Energy 28 (2003) 825.[18] R. Mills, Int. J. Hydrogen Energy 26 (2001) 1041.[19] R. Mills, P. Ray, Int. J. Hydrogen Energy 27 (2002) 183.[20] R.L. Mills, P. Ray, Int. J. Hydrogen Energy 27 (2002) 927.[21] R. Mills, P. Ray, Int. J. Hydrogen Energy 27 (2002) 533.[22] R.L. Mills, P. Ray, J. Dong, M. Nansteel, B. Dhandapani, J. He, Vib.

Spectrosc. 31 (2003) 195.[23] H. Conrads, R. Mills, Th. Wrubel, Plasma Sources Sci. Technol. 12

(2003) 389.[24] R. Mills, T. Onuma, Y. Lu, Int. J. Hydrogen Energy 26 (2001) 749.[25] R.L. Mills, P. Ray, B. Dhandapani, J. He, J. Appl. Phys. 92 (2002)

7008.[26] R.L. Mills, P. Ray, N. J. Phys.www.njp.org, 4 (2002) 22.1.[27] R.L. Mills, X. Chen, P. Ray, J. He, B. Dhandapani, Thermochim.

Acta 406 (1–2) (2003) 35.[28] J. Phillips, R.L. Mills, X. Chen, J. Appl. Phys. 96 (6) (2004)

3095–3102.[29] R. Mills, P. Ray, R.M. Mayo, Appl. Phys. Lett. 82 (2003) 1679.[30] R.L. Mills, P. Ray, E. Dayalan, B. Dhandapani, J. He, IEEE Trans.

Plasma Sci. 31 (2003) 338.[31] R. Mills, B. Dhandapani, M. Nansteel, J. He, T. Shannon, A.

Echezuria, Int. J. Hydrogen Energy 26 (2001) 339.

[32] R. Mills, B. Dhandapani, N. Greenig, J. He, Int. J. Hydrogen Energy25 (2000) 1185.

[33] R. Mills, E. Dayalan, P. Ray, B. Dhandapani, J. He, Electrochim.Acta 47 (2002) 3909.

[34] R.L. Mills, J. He, P. Ray, B. Dhandapani, X. Chen, Int. J. HydrogenEnergy 28 (2003) 1401.

[35] B. Sefsaf, B. Carriere, J.P. Deville, Microsc. Microanal. Microstruct.3 (1) (1992) 15.

[36] For recent specifications see PHI Trift II, ToF-SIMS TechnicalBrochure, Eden Prairie, MN, 1999.

[37] A.J. Fahey, S. Messenger, Int. J. Mass Spectrosc. 208 (2001)227.

[38] G.S. Higashi, Y.J. Chabal, G.W. Trucks, K. Raghavachari, Appl.Phys. Lett. 56 (1990) 656.

[39] C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder, G.E. Mulilen-berg (Eds.), Handbook of X-ray Photoelectron Spectroscopy, Perkin-Elmer Corp., Eden Prairie, MN, 1997.

[40] J. He, Y. Ba, C.I. Ratcliffe, J.A. Ripmeester, D.D. Klug, J.S. Tse,Appl. Phys. Lett. 74 (1999) 830.

[41] J. He, Y. Ba, C.I. Ratcliffe, J.A. Ripmeester, D.D. Klug, J.S. Tse,K.F. Preston, J. Am. Chem. Soc. 120 (1998) 10697.

[42] T. van Buuren, T. Tiedje, J.R. Dahn, B.W. Way, Appl. Phys. Lett.63 (1993) 2911.

[43] S. Al-Dallai, S. Aljishi, M. Hammam, S.M. Al-Alawi, M. Stutzmann,S. Jin, T. Muschik, R. Schwarz, J. Appl. Phys. 70 (1991) 4926.

[44] W. Futako, S. Takeoka, C.M. Fortmann, I. Shimizu, J. Appl. Phys.84 (1998) 1333.