Books - Non-fiction

7 pages
5 views

The origin of inverse absorption bands observed in the far-infrared RAIRS spectra of SnCl 4 and SnBr 4 adsorbed on thin-film SnO 2 surfaces

Please download to get full document.

View again

of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Share
Description
The origin of inverse absorption bands observed in the far-infrared RAIRS spectra of SnCl 4 and SnBr 4 adsorbed on thin-film SnO 2 surfaces
Transcript
  The srcin of inverse absorption bands observed in thefar-infrared RAIRS spectra of SnCl 4  and SnBr 4  adsorbedon thin-film SnO 2  surfaces A. Awaluddin  a , M.J. Pilling  b , P.L. Wincott  a , S. LeVent  b , M. Surman  c ,M.E. Pemble  a , P. Gardner  b,* a Division of Chemistry, School of Sciences, University of Salford, Salford M5 4WT, UK  b Department of Chemistry, UMIST, P.O. Box 88, Manchester M60 1QD, UK  c CCLRC Daresbury Laboratory, Daresbury, Warrington WA4 4AD, UK  Abstract The adsorption of SnCl 4  and SnBr 4  on polycrystalline SnO 2  has been studied using synchrotron radiation based far-infrared reflection absorption infrared spectroscopy FIR-RAIRS. In order to exploit the sensitivity advantages of theburied metal layer method, the SnO 2  is in the form of a thin film deposited on a tungsten foil substrate. Adsorption of SnCl 4  and SnBr 4  on an oxygen sputtered surface at 120 K results in spectra characteristic of condensed multilayers. Inaddition, both spectra exhibit an inverse absorption band centred at 355 cm  1 . Modified 4-layer, wavelength-dependent,Greenler calculations show that this inverse absorption band is induced by the presence of the adsorbate but ischaracteristic of the SnO 2  layer. The lack of any frequency shift upon changing the adsorbate from SnCl 4  to SnBr 4  rulesout the possibility that the inverse absorption band is due to a dipole-forbidden parallel mode of the molecule excitedvia the interaction with free electron oscillations in the metal, resulting from the radiation induced oscillating electricfield just below the surface.    2002 Elsevier Science B.V. All rights reserved. Keywords:  Semi-empirical models and model calculations; Infrared absorption spectroscopy; Reflection spectroscopy; Dielectricphenomena; Chemical vapor deposition; Glass surfaces; Tin oxides; Polycrystalline thin films 1. Introduction The chemical vapour deposition (CVD) of tin-(IV) oxide based films on glass surfaces is a keystep in the production of energy-efficient architec-tural glass and a number of other smart-glassproducts. However, in contrast to the semicon-ductor-growth industry where the adsorption andinteraction of precursors with the substrate is wellunderstood, very little is known about the funda-mental processes essential to the CVD of tin oxide.Recently, in our laboratories we have used anumber of techniques to study this process, both atatmospheric pressure and under ultra-high vacuum(UHV) conditions. In particular, we have used thesynchrotron radiation facility at Daresbury tocarry out far-infrared RAIRS studies of SnCl 4  andH 2 O adsorption on model glass substrates con-sisting of SiO 2  and SnO 2  films [1–3]. In order toenhance the sensitivity of these experiments, we Surface Science 502–503 (2002) 63–69www.elsevier.com/locate/susc * Corresponding author. Tel.: +44-161-200-4463; fax: +44-161-200-4430. E-mail address:  peter.gardner@umist.ac.uk (P. Gardner).0039-6028/02/$ - see front matter    2002 Elsevier Science B.V. All rights reserved.PII: S0039-6028(01)01899-4  have used the buried metal layer (BML) method, inwhich a thin film of the oxide of interest is depos-ited on top of a highly reflecting metallic surface.Provided that the oxide film is optically thin, i.e.thickness is considerably smaller than the wave-length of radiation used, then the reflection ispredominantly from the underlying metal surfacebut the surface chemistry probed is that of theoxide gas interface. Since the vibrations of interestare below 500 cm  1 , the oxide films can be up to100 nm thick and still only attenuate the sensitivityof the experiment compared to the pure metal by afactor of two.In our previous experiments, the results fromSnCl 4  adsorption on SiO 2  showed weak SnCl 4 stretching vibrations from a chemisorbed surfacespecies in the region 360–380 cm  1 , with multilayerbands appearing at 412–420 cm  1 at low temper-ature [1–3]. In all cases, including the situationwhere spectral features shifted to lower frequencydue to the interaction of adsorbed Na [3], thebands appeared as conventional absorption bands.The results obtained from SnCl 4  adsorption onSnO 2  surfaces were not quite so straight-forward[4]. On an oxygen-sputtered substrate (oxygensputtering, rather than argon, helps to reducepreferential loss of Sn which is known to occur [5– 9]), adsorption at 300 K results in a band at ca 355cm  1 . However, rather than being a conventionalabsorption band, it develops in the opposite di-rection, i.e. as an inverse absorption band. Con-ventional absorption is only observed at lowtemperature when a condensed multilayer of SnCl 4 starts to form. The observation of inverse ab-sorption features in far-infrared RAIRS spectra is,of course, not new. Almost exclusively however,they involve adsorbates, normally CO or NO butalso H, adsorbed on Cu or other single crystalmetal surfaces [10–13]. Persson and Volokitin (PV)have shown that these features for adsorption onmetal surfaces can be attributed to dipole-forbid-den parallel vibrational modes of the adsorbateinteracting with free electron oscillations in themetal, induced by the oscillating electric field justbelow the surface, this having been created by andof the same frequency as the radiation electric field[14,15]. However, the situation on a metal is verydifferent from that of the complex thin-film SnO 2 substrate. Although in the BML method, the re-flection is dominated by the metal, the free elec-tron oscillation important for the model is that just below the SnO 2  surface. In a separate paperthe PV model has been modified and applied to theSnO 2  BML system [4]. However, a lack of reliabledata on a number of parameters required for themodel make it impossible to say if this model isimportant for this system. In a separate approach,using modified (4-layer, wavenumber-dependent)Greenler-type calculations the inverse absorptionband could be reproduced. 2. The modified Greenler model The Greenler equations [16– 18], based on theFresnel continuity relationships (for normal com-ponents of electric displacement and magnetic in-duction, and tangential components of electricfield and magnetic field) at phase interfaces [19]were devised to analyse the optical characteristicsof multilayer systems. They have usually beenapplied to 3-layer systems (vacuum–adsorbate– substrate) for isotropic layers of fixed complexdielectric constants with a variety of incidenceangle at the vacuum–adsorbate interface. In fixingthe dielectric constants, the radiation is essentiallytreated as monochromatic. Reflectivity,  R 0 , de-duced by the equations for both s- and p-polari-sations of the incident light and for a particularthickness of the adsorbate layer, is then comparedwith the reflectivity,  R , in the absence of an ad-sorbate. D  R =  R  ¼ ð  R 0   R Þ =  R , when negative is thenconsidered as a measure of light absorption by theadsorbate.For the BML situation, it is necessary to con-sider the reflectivity of a 4-layer system (infinitevacuum, adsorbate of a particular thickness, thin-film substrate (tin oxide), infinite thickness sub-strate (tungsten)) and compare this as D  R =  R  with a3-layer system where the adsorbate is absent. Theequation for such a 4-layer system, from Knittl[20],  1 and the development from the two and three 1 Note that Knittl calls this a ‘‘two-layer’’ situation, but herethe ‘‘two’’ refers to the finite-thickness layers.64  A. Awaluddin et al. / Surface Science 502–503 (2002) 63–69  layer systems is given in Appendix A. The othermodification required is an extension from varia-tion of incidence angle only to variation of boththis and wavenumber, i.e. variation of dielectricconstants with wavenumber  ~ mm . This is essentialsince unlike on a metal where the dielectricconstants are essentially invariant over a smallwavenumber range, those of SnO 2  are not: theoxide exhibits a number of strong spectral fea-tures close to the region of interest.Calculations were performed using MATH-CAD Plus 6 for the following type of model 4-layer system: Layer 4.  A main (infinite) substrate with a fixedrefractive index of 46 : 5    93 : 7i,  2 characteristic of polycrystalline tungsten at 403 cm  1 [21]. Layer 3.  A thin, random polycrystalline, layer(usually 100 nm) of SnO 2 , for which variablewavenumber formulae given by Summitt [22] havebeen used for complex dielectric constants, withthe electric radiation vector either perpendicular orparallel to the unique tetragonal axis of a singlecrystal; these dielectric constants are respectivelydesignated  e ?  and  e k . Layer 2.  A very thin hypothetical adsorbatelayer (usually 1 nm), (i) for which Lorentzianpeaks, of various heights, and wavenumber centres( ~ mm 0 ) and widths can be created for the imaginaryrefractive index component ( k  ), and (ii) for whichthe real refractive index component ( n ) can then becalculated (within an arbitrary constant  C   whichcan be given an assigned value but was usuallychosen as 1.3) by application of the Kramers– Kronig relationship [23]; in the Lorentzian for-mula  k   ¼  A = ½  B  þ ð ~ mm    ~ mm 0 Þ , common choice for  A (the height parameter) was either 1 or 0 (the lattercorresponding to  k   ¼  0) and for  B   (the width pa-rameter) was 10. Layer 1.  Infinite vacuum.Using the above model, the calculated spectrado show an inverse absorption feature in the re-gion of interest. Fig. 1(a) shows a calculatedspectrum for a hypothetical adsorbate having anabsorption at 420 cm  1 . Clearly a conventionalabsorption band at 420 cm  1 is produced, as mightbe expected, but at higher angles of incidence, astrong inverse absorption band is observed at 320cm  1 . Most significantly, as indicated in Fig. 1(b), 2 The Nebraska convention with a negative sign between realand imaginary parts is used in this paper.Fig. 1. The result of modified 4-layer wavelength-dependentGreenler calculations at 82  , 84  , 86   and 88   angles of inci-dence (88   higher baseline) for a randomly oriented polycrys-talline SnO 2  substrate and a hypothetical adsorbate having (a) n  ¼  1 : 3 and  k   maximum  ¼  420 cm  1 and (b)  n  ¼  1 : 3 and  k   ¼  0.(c) is the same as for (b) except that the calculation has beenperformed for a preferentially oriented SnO 2  film and only the88   angle of incidence has been shown for clarity. A. Awaluddin et al. / Surface Science 502–503 (2002) 63–69  65  the inverse absorption band is still present even if the hypothetical adsorbate has no absorptionsin the range of interest, provided that the real partof the refractive index is different from that of theSnO 2  layer. The inverse absorption band in thesespectra appears at a lower wavenumber than thatobserved experimentally. However, the calcula-tions have been performed assuming a randomlyoriented polycrystalline SnO 2  layer. If there is apreferential growth orientation of the SnO 2  layer,as we believe there to be [6], then this can shift theposition of the inverse absorption band. Fig. 1(c)shows the calculation where 90% of   E   is parallel tothe  c -axis of the crystallites. As can be seen theinverse absorption band now appears at ca 350cm  1 , i.e. much closer to the experimentally ob-served position. According to these results, theinverse absorption band is induced by the presenceof an adsorbate having a different refractive indexfrom SnO 2  but its wavenumber position and in-tensity are characteristic of the oxide and thus areunrelated to the vibrational structure of the ad-sorbate. If this is the case, it should appear uponthe adsorption of adsorbates other than SnCl 4 . Inthis paper we report the results of such an exper-iment. We have used SnBr 4 , which may reasonablyassumed to have a similar refractive index to SnCl 4 but importantly has no fundamental modes of vi-bration above 300 cm  1 . The asymmetric  m 3  Sn–Brstretching mode of SnBr 4  is at 284 cm  1 in the gasphase [24]. If the inverse absorption band appearsat the same frequency as that observed for theSnCl 4  then the above model probably provides thecorrect explanation for the observed behaviour. If however the feature shifts close to the frequency of the SnBr stretching mode then the PV model maybe applicable to this system. 3. Experimental The far-infrared spectra were obtained usingthe RAIRS station on the infrared beamline 13.3at the Daresbury synchrotron radiation source(SRS). The complete experimental arrangementhas been described in detail elsewhere [25]. Briefly,the infrared radiation is extracted from the syn-chrotron source and passes via suitable optics,through an evacuable FTIR spectrometer (Nicolet20F). The radiation is then directed through theUHV chamber where it reflects off the sample be-fore being focussed onto the detector. In our pre-vious work the UHV windows have been CsI butfor these experiments presented here, diamondwindows were used in order to improve trans-mission at lower wavenumber. The backgroundpressure in the chamber during these experimentswas typically 2    10  10 mbar. Spectra were gener-ally recorded using 8 cm  1 resolution and con-sisted of either 256 or 512 scans. The sample,prepared in a separate vacuum chamber, consistedof a 50    50    0 : 05 mm tungsten foil coated with100 nm SnO 2  by magnetron sputtering of a tintarget in an oxygen atmosphere. The sample wascleaned in the RAIRS chamber prior to each ex-periment by oxygen ion sputtering, and the surfacecleanness checked with in situ XPS. 4. Results Since the inverse absorption band is believed tobe characteristic of the oxide film rather than theadsorbate, it is important to check that this par-ticular substrate behaves in the same way as ourother samples. This is necessary since the sampleswere prepared in a different vacuum chamberseveral months apart and the calculations showthat variations in film thickness, etc., all effect theappearance of the band. Fig. 2 shows the spectrumof SnCl 4  adsorbed on the SnO 2  thin film at 120 K.The approximate exposure was 20 nbars. Themain band in the spectrum is at 417 cm  1 indica-tive of a condensed SnCl 4  multilayer and, as aconsequence of the diamond windows, the equiv-alent T 2  bending mode  m 4  is observed at 133 cm  1 .In addition to these absorption bands, an inverseabsorption band is clearly observed at 355 cm  1 .This frequency is the same as that observed pre-viously but the band is slightly less intense.Fig. 3 shows the spectra for SnBr 4  adsorptionon the SnO 2  thin film after exposures of 20 and 50nbars. In each case multiplayer absorption is in-dicated by the strong  m 3  Sn–Br stretching mode of SnBr 4  at 291 cm  1 . In addition, a broad absorp-tion feature is observed at 386 cm  1 , which we 66  A. Awaluddin et al. / Surface Science 502–503 (2002) 63–69  believe is due to a small degree of unavoidableisotopic exchange with Cl atoms from residualSnCl 4  adsorbed on the chamber walls, etc. Im-portantly, an inverse absorption band is againobserved at 355 cm  1 . 5. Discussion From the spectra of SnCl 4  adsorption it is clearthat the SnO 2  thin films exhibit qualitatively re-producible behaviour. The intensity of the inverseabsorption band is reduced compared with previ-ous experiments but this can be explained by slightchanges in the optical properties of the polycrys-talline SnO 2 , which can vary with the quality of thefilm [4].The spectra of SnBr 4  adsorption clearly showan inverse absorption band. It appears at exactlythe same frequency as that observed using SnCl 4 .Although some unavoidable isotopic scramblinghas occurred with residual Cl, there is no intensityat 417 cm  1 indicative of molecular SnCl 4 . Weconclude therefore, that the appearance of the in-verse band is not a vibrational feature of the ad-sorbate but is an adsorbate-induced feature of thesubstrate. This is supported by the fact that thereis no inverse absorption intensity close to or justbelow the position of the  m 3  Sn–Br stretching modeof SnBr 4  at 291 cm  1 . These results thus enable usto confirm the validity of our modified 4-layerGreenler model for our BML system and, just asimportantly, rule out the PV model as an expla-nation for the inverse absorption bands observedin this system. 6. Conclusion Adsorption of SnCl 4  and SnBr 4  on to oxygensputtered SnO 2  thin-film surfaces both give rise toan inverse absorption band at ca 355 cm  1 . At lowtemperatures (120 K) this is accompanied byconventional absorption bands of a condensed Fig. 3. The FIR-RAIRS spectrum of SnBr 4  adsorbed on a thin-film polycrystalline SnO 2  surface, after an exposure of (a) 20nbars and (b) 50 nbars.Fig. 2. The FIR-RAIRS spectrum of SnCl 4  adsorbed on a thin-film polycrystalline SnO 2  surface, after an exposure of 20 nbars. A. Awaluddin et al. / Surface Science 502–503 (2002) 63–69  67
Related Documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x