Chapter 7

PHOTONIC AND ELECTRONIC SPECTROSCOPIES FOR THECHARACTERIZATION OF ORGANIC SURFACES AND ORGANICMOLECULES ADSORBED ON SURFACES

Ana Maria Botelho do Rego, Luis Filipe Vieira FerreiraCentro de Química-Física Moleculal; Complexo Interdisciplinal; Instituto Superior Técnico,1049-001 Lisboa. Portugal

Contents

275276276277278280280280285286286296301301303305307310310310

I. lntroduction 2. Diffuse Reflectance Techniques for Surface Photochemistry Studies . . . . . . . . . . .

2.1. Oround-State Diffuse Reflectance Absorption Spectra (UV- Vis-NIR) . . . . . . .2.2. Time Resolved Laser lnduced Luminescence . . . . . . . . . . . . . . . . . . . .2.3. Diffuse Reflectance La!ier Flash-Photolysis . . . . . . . . . . . . . . . . . . . . .2.4. Other Techniques for Surface Studies . . . . . . . . . . . . . . . . . . . . . . . .

3. .ElectronicSpectroscopies 3.1. X-Ray Photoelectron Spectroscopy 3.2. High Resolution Electron Energy Loss Spectroscopy . . . . . . . . . . . . . . . .

4. Examples of Systems Studied by Electronic and Photonic Spectroscopies . . . . . . . .4.1. Microporous Solid Surfaces and Finely Divided Powders . . . . . . . . . . . . .4.2. Flat Surfaces. HREELS and XPS Studies on Organic Films . . . . . . . . . . . .

5. Combined Studies lnvolving Photonic and Electronic Spectroscopies . . . . . . . . . .5.1. Rhodamine Dye Covalently Bound to Microcrystalline Cellulose . . . . . . . . .

5.2. 2.2'-Cyanines Adsorbed onto Microcrystalline Cellulose 5.3. lnclusion Complexes of /3-Cyclodextrin and Cyanine Dyes . . . . . . . . . . . .

5.4. Sulforhodamine 101 and Rhodamine 60 Adsorbed on Different Pore Size Silicas

6. Conclusions Acknowledgment., References ,

. INTRODUCTION Microporous solids play a special role due to their largearea/volume ratio providing lhe opportunity for adsorption oflarge amounts of guest molecules on very small amounts ofsubstrate, due to the very high specific area of lhe adsorbent.In particular, powdered solids with controlled pore size solidsmay have specific uses such as molecular sieves, microreac-tors of controlled microsize, or catalysts. The surface-probeinteraction may simply be of electrostatic nature, of hydrogenbonding, or in terms of an acid or basic behavior (Brõnsted orLewis). Substrates may also interact with lhe adsorbed probe aselectronically active supports and have an important role in re-dox processes (as in advanced oxidative processes for persistent

pollutant destruction) [1-5].

Solid surfaces are a field of growing interest from both the ap-plied and fundamental points of view. Their quality of beingthe "face" of any material, their entrance doar, gives them amajor importance in applied areas as different as compositemateriaIs, corrosion, adsorption, biocompatibility, dyeing andlightfastness of dyes on fabrics, heterogeneous catalysis, amongmany other applications.

AIso infundamental areas, surfaces are privileged media toserve as substrates for monolayers (or a few monolayers) as inself-assembled on molecular systems or in Langmuir-Blodgettfilms (LB), for instance.

Handbook of Suifaces and lnteifaces of Materiais, edited by H.S. NalwaVolume 2: Suiface and lnteiface Analysis and Properties

Copyright @ 2001 by Academic PressAli rights of reproduction in any form reserved.

275

276 BOTELHO DO REGO AND VIEIRA FERREIRA

a black reference. Altematively, it is possible to use calibratedstandards either for white or black, both of which are commer-cially available.

Ground-state absorption spectrum for powdered microcrys-talline cellulose is also shown in Figure 1. It exhibits R valuesquite close to unity for visible (Vis) and near infrared (NIR)spectral regions and shows a significative absorption in lhe ul-traviolet (UV). Silica and silicalite also have a reflectance closeto unity in the Vis and NIR regions and differ from celluloseby the fact that UV absorption is smaller. In lhe case of silicaswith different porosity, but with the same particle size, signifi-cant variations of lhe reflectance were detected [7].

Let us consider now that I and J are the fluxes of incidentand disperse light traveling in the forward and reverse direc-tions. The incident flux, I, decreases as it penetrates the solidpowdered sample because not only is radiation absorbed, butbecause particles also disperse the incident light. At the sametime I increases with J dispersion. The flux of disperse light,J, has an analogous variation, although in the opposite direc-tion, as Figure 2 [8] shows.

The interest in studying surfaces is intimately related to thedevelopment and spreading of suitable techniques. In fact, thereare a number of conditions to be fulfilled in order to makea technique surface sensitive and surface specific. Apart fromthe cases of highly porous materiaIs and/or highly divided me-dia, surface usually represents a negligible amount of matter insolids. Therefore, and given the usual proportionality betweenthe number of species in a medium and the signal obtained in atechnique, some cace has to be taken to be sure that the obtainedinformation comes exclusively (surface specific) or mainly (sur-face sensitive) from the surface and not from the bulk.

At this point we will distinguish between surface-specificand surface-sensitive techniques: according to Oesimoni et aI.[6] the Corroer "are defined as those capable of collecting infor-mation relevant to the first few atomic layers of the surface of asolid specimen," i.e., a maximum sampling depth of 5-10 nm;the latter probe depths up to 100 nm.

For flat surfaces, the selectivity is achieved using lowpenetrating excitation probes and/or analyzing low penetrat-ing probes. As examples of low penetrating probes, we canpoint out heavy ions and electrons. X-ray photoelectron spec-troscopy (XPS) and high-resolution electron energy spec-troscopy (HREELS) are just two spectroscopies based on thelow penetration of electrons. XPS uses highly penetrating

probes-X-ray photons-but analyzes electrons, and HREELSuses very low energy electrons both as incident and analyzedprobes. These two electronic spectroscopies are surface spe-cific, the specificity being larger for HREELS when the impactmechanism is favored-a sampling depth of a few Angstrõm.In XPS, the sampling depth is of. the order of 3 to 10 nm.

The goal of this chapter is to emphasize the synergeticeffect obtained on putting together the larger number ofavailable techniques to the service of characterizing surfaces.Particularly, this chapter will deal with optical spectroscopies inthe UV- Vis-IR range, used in diffuse reflectance arrangement,combined with electronic spectroscopies, specifically XPS andHREELS.

200 300 4ÓO 5ÓO 600 700 800 9ÓO

Wavelength, nm

Reflectance spectra of white and black standards and microcrystallineFig. I.cellulose.

2. DIFFUSEREFLECTANCE TECHNIQUES FORSURFACE PHOTOCHEMISTRY STUDlES

2.1. Ground-State DitTuse Reflectance Absorption Spectra

(UV-Vis-NIR)

Ground-state absorption spectra of solid opaque samples canexperimentally be obtained using a procedure similar to the oneused to perform absorption spectra of transparent samples. Inthe latter case one has to use the Beer-Lambert law to deter-mine absorbances as a function of wavelength. In the case ofsolid samples, it is possible to determine the refiectance (R) asa function of the wavelength using an integrating sphere, fol-lowing an initial calibration of the apparatus as Figure I shows.An ideal diffuser has a unitary reflectance (in prac~ pure bar-ium sulphate or magnesium oxide can be used since for bothcompounds R "" 0.98 :!: 0.02 in the 200 to 900 nm wavelengthran~e). Finelv divided carbon oarticles (R "" m can he \J~ed a~

- - .- -x~

Fig. 2. lncident (I) and disperse (J) fluxes of radiation penetrating a disper.,,;v..m..rli,lm

PHOTONIC AND ELECTRONIC SPECTROSCOPIES 277

So we can write

dl(x) = -(K + S)l(x)dx + SJ(x)dx (1)dJ(x) = +(K + S)J(x)dx - SI (x)dx (2)

where K is the absorption coefficient, and S the dispersion co-efficient. Qne can very easily obtain the Beer-Lambert law fortransparent and homogeneous media, where no dispersion ex-ists, so S = O. Then,

dl(x) = -K I (x)dx (3)

In the case of more than one absorbing chromophore and foroptically thick samples we can write

,--K()") = KB + 2 L e; (),,)C; (10)

;

where K B concems the background substrate. From this, it foI-lows that a difference spectrum can be used to (;valuate lhetemporal evolution of a photochemicaI reaction (by comparingspectra before and after irradiation):

~K = S[ F(R)irradiated - F(R)nonirradiated] (11)

and therefore 1 = 10 exp( - K x), where K = eC, where e is lhenaperian coefficient for absorption and C is lhe concentration oflhe absorbing species.

Kubelka and Munk [8] established that, for an ideal diffuserand optically thick samples (alI those where a further increasein thickness does not change lhe reflectance ofthe sample), lhereflectance is given by

JR = - (4)10

where R is related to K and S by the remission function, F(R)

(I - R)2 KF(R) = ':(2R) = S (5)

and

K(À) = 2~(À)C (6)

2.2. Time Resolved Laser lnduced Luminescence

Time gated and intensified charge-coupled device [9] detec-tors coupled to pulsed lasers as excitation sources, which aremonochromatic and may have short pulses and high ftuences,are a very attractive and rigorous way of performing time re-solved luminescence studies. Opaque solid samples imply theuse of a reftection geometry, as shown in Figure 3. This figurealso shows in a schematic form the setup used to obtain eitherftuorescence or phosphorescence spectra. Specular reftectionshould be avoided in order not to damage the detector, whichis the maio device of the system.

Since 1997, we have been using in our laboratory an inten-sified charge-coupled device (ICCD, Oriel model Instaspec V,with a minimum temporal gate of 2.2 os) in a daily basis fortime resolved luminescence studies. The detectorhas 512 x 128pixels in a maximum spectral range of 200 to 900 um. With asingle laser pulse, a ftuorescence or a phosphorescence spec-trum can be instantaneously obtained, since the combined useof the delay unit and time gate enables one to separate promptfrom delayed emissions.

The fast photodiode detects the zero time, corresponding tothe laser pulse. The laser pulse can be optically delayed by theuse of an optical fiber to take into account the ICCD and de-lay unit intemal delays. Depending on the signal intensity, theICCD readings can be intensified or noto The delay unit (pi-

In Eq. (6) the factor 2 takes into account the average in-crease of distance travelled by the excitation light inside theideal diffuser. In the Kubelka-Munk theory the radiation isonly regarded as scattered when it is backward reflected intoa hemisphere whose boundary plane lies perpendicular to thex-direction [8c].

The Temission function varies linearly with the number ofabsorbing chromophores in the solid sample, which are consid-ered as being uniforrnly distributed. K and Sare independentof the light penetration into the sample. For optically thick sam-pIes we have

Oplica!Opaq~ sample Delay/'\ .,--,

~--SP~ -~:

: '... ~flection

~'~

Fast photodiode("Trigger")

4-0

1 (x) = 10 exp( -bSx)

J(x) = Rloexp(-bSx)(7)

(8)

where b = [1/(2R) - R/2], so that the depth of penetration ofthe exciting light into the sample (xo) can be given by

Analyzingmonochromator

(9)Xo =bS

Intensified CCOwith time pieand I (xo) = 10e-1 (the incident radiation with the intensity

10 is reduced in this case to e-I of the initial intensity, i.e.,63.1 % of the incident radiation). Another cri teria for definingthe penetration depth can be 5xo, which corresponds to 99.3%of absorption of the incident radiation.

Later (Section 4) we will give some specific examples ofpenetration depth in different substrates, for photon excitationin powdered solid samples.

<1--- +-Laser optical pedi Lumines= optical path Electrical signals

Fig. 3. Schematic diagram of lhe laser induced luminescence system.

278 BOTELHO DO REGO AND VIEIRA FERREIRA

OpaqueSample

" 250000tU

6 200000.~c2. 150000.5

100000

50000

o300

MonitoringLigbt

Fig. 5. Laser excitation pulse and monitoring light in lhe diffuse-reflectancemode.

~ 400 500 --- .--

Wavelength, nrn

Fig. 4. Room temperature time resolved phosphorescence spectra of ben-

zophenone crystals and benzophenone included into silicalite channels

(200 jlmol g-1 ).

fi()() 700

cosecond to second range) allows us to perform successive de-layed readings according to an initially programmed data ac-quisition sequence. 50 both luminescence time resolved spectraand emission decay curves can be obtained with this system.Therefore prompt and delayed luminescence can be obtainedwithout lhe use of lhe classical time consuming methods [2].

A good example of room temperature laser induced phos-phorescence is presented in Figure 4.

The same time scale was used in both spectra. The decay isslower in lhe case of inclusion of lhe ketone into lhe channelstructure of the powdered substrate. Clearly, bimolecular pro-cesses as well as nonradiative decay processes were reduced inthis case.

The technique is based on the study of the temporal evolu-tion of the absorption of excited species, which were createdby a laser pulse, through an analyzing light in the diffuse re-flectance mode. (See Fig. 5.)

The setup is identical to the one used for transmission stud-ies, but in this case, a reflection geometry is used, where themonitoring and the detected beams are in the same sample side.As in conventional flash-photolysis, the absorption spectra aredifference spectra. This means that the experimentally deter-mined absorption reflects the difference, in terms of absorption,between the excited state (created by laser excitation) and theground state (detected by the monitoring lamp), for a specificspecies at a specific wavelength. So the transient absorption(measured as a percentage of absorption, as a hypothesis) de-pends on both ground-state and excited-state extinction coeffi-cients. In many cases, it is important to consider the emissioncorrection (which is obtained by firing the laser alone) and it issubtractive relative to the transient absorption. (See Fig. 6.)

Apart from being the group responsible for the developmentof the diffuse-reflectance laser flash-photolysis technique in thetemporal range from nanosecond up to seconds [10], Wilkinsonet aI. were also the first authors to publish transient absorptionspectraof opaque materiais in the picosecond time domain [11].

In a diffuse-reflectance laser flash-photolysis experiment,the excited chromophores created by the laser excitation mayexhibit a nonhomogeneous spatial distribution. Theoreticaltreatments (Kessler et aI. [IOb] and Oelkrug et aI. [12]) haveshown that two extreme types of concentration profiles can beproduced. In the first type, the concentration of excited species

.-dccreases exponentially as a function of the penetration depth ofthe excitation radiation whereas, in the second type, the excitedspecies are distributed in a homogeneous manner within a spe-cific width, and then decrease ("plug" profile) as in [IOf, 10g].

2.3. DitTuse Reflectance Laser Flash-Photolysis

In lhe beginning of lhe 1980s, Frank Wilkinson and co-workerswere able to apply lhe laser flash-photolysis technique toopaque samples [10], thus allowing lhe study of a consider-able number of photochemical processes in heterogeneous sys-tems. Good examples of those studies are, among many others,lhe photochemical reactions in confined spaces, lhe study oforganic molecules adsorbed or included on substrates with cat-alytic activity, and dye~ adsorbed or covalently bound to naturalor synthetic fibres.

Everything points to lhe laser diffuse reflectance flash-photolysis technique playing in lhe near future, in regard totransient absorption studies in heterogeneous media, a role atleast as important as flash photolysis in lhe transmission modehas played, until now, for photochemical reactions in homoge-neous media.

PHOTONIC ANO ELECTRONIC SPECTROSCOPIES 279

çrultip/ier

+romator

Signal

~,./ /' Specular

/ ,/ Renection

Delay Unit

~Laser Trigger.oul

Pre.pulse

/ ,

Nd: Y AG Laserr+vf"Sample ~..- - -- - - -- - !'!!~P._- - - - - ~

-'-,. X..~robe

""h~" .. 450

-, XeLam~

450WXe

Lamp

Fig. 6. Schematic diagram of lhe laser ftash-photolysis setup in transmission and diffuse reftectanctmodes.

~d

The latter case can be found for large laser fluences (in termsof moles of photons per square centimeter) and low concen-trations of the ground-state absorbers. In this case there is atotal conversion of ground-state species into transient excitedspecies. The remission function F(R) can be used for opticallythick samples and is a linear function of concentration.

For a small percentage of conversion (high concentration ofground-state absorbers and low laser fluences) the concentra-tion of transients decreases exponentially and a representationof fj. RT as a function of time is a good measure of the transientconcentration for values of fj.RT < 0.1.

~;..-.-ri)QQ)

~-[RB - R(t)] ó 2E:OS -IE:OS 6E:OS 8E:OS

Time, s1E.:04dRr(t) = [1 - Rr(t)] = (12)

RB

3S~ ..'. (b)1115.~

..

~" 25-O

8'O 15-fi)

.o<

.'. 2.5 .,. ..I I- -.. ' -.. ,.' .. .10', -.. I .. . , , , ., ..' - .. " ., " - --7.5 .. ',-.." --- - ""'" I,.. -- --"..,. 75 "..._.' :~. ~. ~ *. .:_;M~;..

400 500 600 700

Wavelength, mnFig. 7. (a) Data traces for benzophenone microcrystals exciting at 355 nmand (b) time resolved triplet-triplet absorption spectra of microcrystalline ben-zophenone aI room temperalure.

, ,-- S (14)

and this equation justifies the occurrence of isosbestic points intransient absorption spectra.

Figure 7a presents the four traces needed to obtain a cor-rected transient absorption decay at a specific wavelength: base-line, top-line, emission, and absorption. By takin~ into account

-Si!:

.'ipe..lro-gr"ph

lIa';j

where RB is the substrate reflectance before firing the laser, andR(t) is the reflectance at time t after excitation.

For more concentrated samples, the variation of the remis-sion function before and after the laser firing is given by

~F(R(r)) = F(R(t)) - F(RB) ("3)

and K(r) = F(R(t))S = KB + 28GCG + 28*C*, K(O) =F(RB)S = KB + 28GCo and Co = CG + C*. G representsground state, K B the absorption coefficient, and * means, of

course, the excited state. Thus, we can write2(8* - 8 )C*~F(R(t)) = G

280 BOTELHO 00 REGO ANO VIEIRA FERREIRA

opaque surfaces that disperse the incident radiation. The exten-sion of this technique to obtain time resolved transient absorp-tion spectra in the IR wavelength range (laser flash-photolysiswith IR detection) will certainly play in the near future an im-portant role in terms of clarifying different reaction mecha-nisms in the surface photochemistry field [17c, 18].

We also want to refer lhe use of nuclear magnetic resonance,electron paramagnetic resonance, and Raman techniques forsolid surface studies [19].

In the next section, lhe basic principIes of XPS will be out-lined.

3. ELECTRONIC SPECTROSCOPIES

The generic designation of electronic spectroscopy is attributedto every technique that uses electrons as incident (ingoing)probes and/or as analyzed (outgoing) probes. Among the mostpopular ones, for both the qualitative and quantitative character-ization of surfaces, we can find XPS-formerly called electronspectroscopy for chemical analysis-and HREELS.

the entire wavelength range where the excited species absorbs,and algo by a suitable choice of the time scale, triplet-triplettime resolved absorption spectra can be obtained, such as theone presented in Figure 7b for benzophenone crystals. Thesespectra provide both spectroscopic and kinetic infonnation re-garding lhe powdered opaque sample. It is obvious that thesespectra will enable one to study the occurrence of a chemicalreaction on surfaces, provided the time scale is adequately cho-seu. Figure 7b algo shows the triplet-triplet absorption spec-tra of microcrystalline benzophenone at room temperature, firstpublished by Wilkinson andco-workers in 1984 [IOa].

Figure 7b clearly shows that lhe initial decay is faster thanthe long time decay, and lhe excited triplet disappears in lhe mi-crosecond time scale. The decay is a mixture of first and secondarder processes, with a maximum absorption at about 540 nm.The phosphorescence emission at room temperature algo has anidentical kinetic decay, within experimental error.

After lhe laser pulse, the transient decays. For a unimolecu-lar decay process, lhe concentration of lhe excited species ~e-cays exponentially, according to

C*(t) = C*(t = O)exp(-k)t) (15)

3.1. X-Ray Photoelectron Spectroscopy

3.1.1. Basic Principies

The photoelectric effect is lhe basis of this technique. Whena photon having an energy Ephoton impinges on a surface, asschematically displayed in Figure 8, an electron bound to lhenucleus with a binding energy Eb is ejected with a kinetic en-ergy Ek related by

Ephoton = hv = Eb + Ek + f/J (18)

where f/J is lhe work function of lhe spectrometer and needs tobe adjusted every time lhe equipment is vented [20].

Equation (18) can be rewritten in lhe following ways:

Eb = hv - Ek - f/J and Ek = hv - Eb - f/J (19)

And for low ftuences of the laser [I - RT (t = O)] ~ 0.10 andthe transient concentration is proportional to (I - RT). So

In[1 - RT(t)] = In[1 - RT(t = O)] - klt (16)

For high laser ftuences, optic;llly thick layers ofhomogeneousexcited species are produced and one has to use the Kubelka-Munk function to perform the ~ay analysis; Eq. (16) becomesnow

In{F[R(t)]-F(RB)} = In{F[R(t = O)]-F(RB)}-kit (17)

Equation (17) is generally used for data analysis and it predictsa linearrelation between In{F[R(t)] - F(RB)} with time. fromwhich the raie constant k. can be determined. For more com-plex cases one has to use eidrcr [I - RT(t)] or F[R(t)] de-pending on the specific conce.-Jration profile and on the decaykinetics [12-14], Eb and Ek vary, then, in a symmetrical way. In a photoelec-

tron spectrum, lhe intensity of ejected electrons as a functionof their kinetic energy is registered. With X-ray radiation, hv ishigh enough to eject inner shell electrons. These electrons, con-trarily to lhe valence ones, have their binding energy to lhe nu-cleus almost unchanged by lhe atomic environment. They are,then, a fingerprint of lhe element where they originate. Since

2.4. Other Techniques for s..face Studies

Apart from the techniques d~ed previously for surface pho-tochemistry studies, other aJ.-lllaches can be used.

We would like to stress ~tthe importance of X-rayphoto-electron spectroscopy (XPS) bsurface characterization, sinceit analyzes the first 10 or 20 3rlBrnic monolayers. It gives infor-mation regarding both co~ition and elemental concentra-tion, as well as the probe-SIEfRce illteractions. Quite recently[15,16], this technique allowcdllthe authors to study rhodamineand cyanine dyes physically alldr/or chemically bound to micro-crystalline cellulose.

AIso, Fourier transform infuared absorption spectroscopyprovides relevant informaliooJ lregarding the specific interac-tions of different probes widlinl$ubstrates [17], especially in thediffuse-reflectance mode wlK:nl:qpplied to the study of powdered

Sample

Fig.8. Schematic representation of an XPS experil

PHOTONIC AND ELECTRONIC SPECTROSCOPIES 281

C Is photoelectrons I--

~os=

-é=

,;.

.~c..

c-

CKLL Auge r

.-!Iectroffi

E

el.ction.

ri)00 O

~rgy lAIsses

I

1000 800 600 400 ---Binding Energy, eV

Fig. 10. Survey XPS spectrum of a sample of highly oriented pyrolithicgraphite showing several features, other Ihan lhe photoelectron peak.

(higher binding energies) larger than the background at higherkinetic energies (Iower binding energies). In Figure 10, we cansee a real survey spectrum of a sample containing a single ele-ment from the second period of the periodictable-highly ori-ented pyrolitic graphite-and exhibiting several of the abovementioned features. The spectrum was obtained with a freshlypealed sample with a Kratos XSAM800 spectrometer operatingin a fixed analyzer transmission mode (see Section 3.1.4) underapressure of the order of 10-9 mbar and using a pass energy of10 e V. The magnesium nonmonochromatic radiation was used(main comporient at hv = 1253.6 eV).

Fig. 9. Schematic representation of an "expected" XPS spectrum.

3.1.2. Chemical Shifts

One of the most useful characteristics of the XPS is the inner-shell binding energy shifts, usually named chemical shifts.A photoelectron ejected from an inner shell feels an attrac-tive action from the nucleus but also a repulsive action fromthe neighboring electrons, namely the valence electrons. There-Core, a given atom, like carbon, for instance, when bound to amore electronegative atom (O, N, F, Cl, ...) will have its va-lence electronic density rarefied and so lhe repulsion felt by itsinner electrons will be lower. This will increase lhe inner elec-tron binding energy. The inverse wilI occur when it is bound toa less electronegative atom. Also the hybridization state playsa role on lhe binding energy: lhe C ls photoelectron bindingenergy is around 284.7 eV in aromatic compounds and 285 eVin aliphatic chains. Theoretical calculations of chemical shiftsappearfrequently in literature [23, 24]. Besides general bindingenergy databases [25], a good database for organic polymers(and a few inorganic ones), was published in lhe 1990s [26].In Table I, the chemical shifts for the most common functionalgroups in organic and inorganic-:organic compounds are pre-sented.

the inner-shell electron binding energy is the variable allowingfor the identification of a given element, spectra are usually dis-played in the form of photoelectron intensity as a function ofthe decreasing binding energy rather than the increasing kinetic

energy.In a first approximation, we can consider the electron struc-

ture as frozen under the photoelectron emission process andidentify Eb with the Hartree-Fock energy eigenvalues of theorbitals (Koopman's theorem). A schematic representation ofan expected photoelectron spectrum could then be the one inFigure 9.

However, in an accurate calculation of a binding energy, therelaxation energy of the remaining electronic structure to a newbole state has to be included [21]. Besides, many other fac-tors contribute to render the schematic picture displayed in Fig-ure 9 very different in real situations and great care needs tobe taken in the interpretation of features appearing in a XPSspectrum. Some of the examples of features other than themaio photoelectron peak, which can appear in a XPS spec-trum, are shake-up and shake-off peaks, X-ray source satel-lites (for nonmonochromatic X-rays), "cross-talk" peaks, andAuger peaks [22]. Moreover, the ejection of an electron fromthe inner shell of a given element does not usually give rise toa single peak. Reasons for this are chemical shift, orbit-spincoupling, and spin coupling (when nonpaired electrons exist inthe element). Finally, photoelectrons suffering single or mul-tiple inelastic collisions in the medium lose energy and leavethe surface with a lower kinetic energy. This implies that everyohotoelectron oeak has a background at lower kinetic energies

2R2 BOTELHO DO REGO AND VIEIRA FERREIRA

Table Chemical Shift, in eV, for C Is Phoelectron for lhe Most CommonFunctional Groups in Organic and Organo-Inorganic Polymers I'

==

.e(Q

:i-.~2!..:~

Functional group Mean chemical shift (eV)

-C-O-H

-C-O-R

>C=O

-COOH

-COaR

-OCO-

-C-NO2-C-N<-C=N

-CF3(CFVn

-CFH-CH2--CSi--CS-C=C

1.55

1.45

2.90

4.26

3.99

2.93

0.76

0.94

1.74

7.69

7.48

2.91

-0.67

0.37

-0.27

~~~:~~~~~-'....-. -

83 78 73 68 63

Binding Energy, eV

Fig. 11. XPS AI 2p region: bold liDe: Aluminum covered with its native ox-ide; normal liDe: lhe same sample chemically etched to remove lhe oxide and

immediately introduced into lhe chamber.

The spectrum drawn in normalline displays two peaks cor-responding to the metallic aluminum (Iower binding energy)and to the oxidized aluminum (AI3+) due to the oxide film thatrapidly grows at the surface when it is exposed to the air. Thesetwo peaks are a good illustration of the chemical shift citedabove. The spectrum drawn with a bold line displays a singlepeak corresponding to the species A13+ in the oxide filmo But,this time, the oxide film is so thick that the metallic aluminumis hidden. Due to its extreme thickness, the film is more insulat-ing than the one existing after the etching and the A13+ peak ischarge shifted toward larger binding energies. Simultaneously,the peak is broadened. This example is, therefore, a good exam-pIe to illustrate the effect of the insulating character of a sampleon an XPS spectrum: charge shift and peak broadening. TheAI 2p peak, as any np peak, has two components - the 2P3/Zand the 2PI/z. However, the energy splitting is very small-"-'0.4 eV [25]-and they are only separable by curve fitting.

The correction of the binding energy for this charging effectis made using one of the peaks in the spectrum; the most used isthe C I s contamination peak, to which a binding energy equalto 285 eV is assigned [26]. For spectrometers equipped withmonochromatic sources, a flood gun for charge neutralizationis used [28].

Reference is C Is in saturated hydrocarbons, Eb(Cls) = 285 eV [26].

3.1.3. Charge Shifts

Most of the samples referred to here-nonconjugated organicfilms and porous materiaIs like cellulose, silica, zeolites-areelectrically insulating. Since the X-ray induces the escape ofelectrons, the surface of the sample becomes positively charged.A recent paper by Cazaux [27.] makes a detailed analysis ofthe role of various parameters involved in the chargingmecha-nisms of insulating materiaIs, mainly in XPS and Auger spec-troscopies. An escaping electron having a kinetic energy, Ek,will, then, "feel" an attractive force toward the surface and itsmovement toward the analyzer is retarded. It is similar to a sit-uation where the electron is ejected with a lower kinetic en-ergy, Ek - ~E (~E > O), i.e., an apparent binding energy

Eb + ~E. Therefore, if no charge compensation exists, thecharge may increase enormously during the spectrum acquisi-tion. In the limit, no defined peaks are obtained: the apparentbinding energy changes continuously. For rapid acquisitions,well-defined peaks are obtained but much broader than for notcharged samples. Hopefully, when nonmonochromatic X-raysources are used, the sample is very close to the window ofthe source and the number of electrons emitted by the sourcefilaments, crossing the window and impinging on the sample,is enough to rapidly reach a stationary charge on the surface.The entire spectrum is displaced from its real binding energiesbut peaks are well defined and their full width at half-maximum(fwhm) has a value very close to the fwhm obtained for a con-ducting material. In Figure 11, an XPS spectrum of the regionAI 2p taken on a sample of aluminum covered with a thickfilm of its native oxide and the same sample freshly chemicallyetched are shown. This spectrum was acquired in the same con-ditions as the ones used for the spectrum in Figure 10.

3.1.4. Equipment

A large number of books and book chapters were publishedcontaining a description of the X-ray photoelectron spectrom-eters. The "bible" in this domain is the book edited by Briggsand Seah [29, 30]. The main components in a spectrometer arethe X-ray gun, the sample holder, an electrostatic gun, the en-ergy selector, and the electron multiplier. In addition, of course,there is alI the electronics needed to apply the voltages ir) alIthe lens and energy selector electrodes as weII as the softwaresuitable to acquire and treat spectra. It is out of the scope ofthis chapter to provide a thorilllgh description of the instru-mentation details. We just would like to stress the utility ofhaving a multianode X-ray gun, at least a dual-anode one. Infact, in the task of assigning alI the features appearing in a

PHOTONIC AND ELECTRONIC SPECTROSCOPIES

trance by a constant factor and analyzing their energy by sweep.ing lhe voltage difference between lhe analyzer electrodes.

3.1.5. Electron Escape Mechanisms and Surface Sensitivity

ane important parameter in the XPS technique is the electronefTective atenuation length (EAL), À. It is defined as the dis-tance normal to the surface at which an electron should beejected from the atom in order for the probability of escapingthe surface without losing energy due to inelastic scattering pro-cesses to have lhe value I/e for an angle () = O (see Figure 8).For a good discussion about the concept and some confusion inlhe literature between this parameter and the inelastic free meanpath (IMFP), see, for instance, [31]. If the number of photo-electrons produced at depth xis denoted no(x), the number ofelectrons, n(x), escaping lhe surface is, therefore, given by

n(x) = no(x)e

The effective attenuation length is a function of electron ki-netic energy and of the medium density. However, when werepresent À in number of monolayers, an approximately univer-sal curve is obtained [32] as displayed in Figure 12.

~..

~u

'so-=

~c~c.9

§c!!.

~.<

spectrum, it helps a lot to have at least two different anodes(the most common combJ!!-3;tion is magnesium and aluminum).For instance, photoelectron binding energies are independentof lhe photon energy whereas Auger electrons have constantkinetic energies. Therefore, for spectra of a same sample ac-quired with a Mg anode (h v = 1253.6 eV) and with an AIanode (hv = 1486.6 eV = (1253.6 + 233) eV), in a bindingenergy scale, photoelectron peaks will remain in lhe same po-sitiaDo On lhe contrary, Auger peaks will appear displaced by233 eV toward larger binding energies in lhe AI-anode spec-trum rei ative to lhe Mg-anode afie. AIso lhe satellite peaks areat different distances from lhe maio photoelectron peak whendifferent anodes are used, as can be seen in Table li, and lhecomparison between two spectra allows lhe complete assign-ment of satellite features in nonmonochromatic sources.

The last generation of spectrometers is monochromatic, al-lowing a best resolution in energy (an intrinsic resolution of0.28 eV is estimated [26]). AIso lhe lateral resolution has in-creased: in conventional equipment lhe lateral resolution wasof lhe arder of magnitude of I mm (in fact, lhe size of lhe area"seen" by lhe monochromator defined as lhe fwhm of lhe spa-tial intensity distribution is about 3 mm x I fim) whereas nowit is about 10 Jlm.

Ali lhe spectrometers where electrons are analyzed need towork under conditions of high vacuum. Moreover, to keep sur-faces clean (at lhe atomic levei), ultrahigh vacuum conditionsare generally used in XPS spectrometers.

The electron energy selection is usually made through twodifferent modes: fixed analyzer transmission and fixed retar-dation fatia. The first one is lhe one used when a quantitativeanalysis is required. It consists of keeping lhe analyzer voltagesconstant, lhe kinetic energy being swept by a variable retarda-tion voltage between lhe sample and lhe analyzer entrance. Inthis mode lhe energy resolution, defined as lhe fwhm of lheelectron energy distribution leaving lhe energy analyzer, is keptconstant. The second mo de is used when lhe sensitivity at lowkinetic energy needs to be increased for identification purposesand is used exclusively in survey spectra. It consists of retard-ing lhe photoelectrons between lhe sample and lhe analyzer en-

1 10 100 1000 10000

Electron Energy, eV

Fig. 12. Schematic representation of electron effective attenuation length as

function of electron energy.

Magnesium and Aluminum X-Ray Spectrum for Magnesium and Aluminum K a LinesTable 11

X-ray

1253.7

1253.4

1258.2

1262.1

1263.7

1271.0

1274.2

1302.0

67

33

1486.7

1486.3

1492.3

1496.3

1498.2

1506.5

1510.1

1557.0

67

33Kal

Ka2

Ka'

Ka3

Ka4

Ka5

Ka6

K/3

100.0

1.0

7.8

3.3

0.42

0.28

2.0

100.0

1.0

9.2

5.1

0.8

0.5

2.0

-

4.6

8.5

10.1

17.4

20.6

48.4

5.7

9.7

11.6

19.9

23.5

70.4

a dE is computed relative to lhe Kal.2 lines taken respectively at 1253.6 and 1486.6 eV.

1000

100

10

1

284 BarELHO DO REGO AND VIEIRA FERREIRA

Table 111 Table IV. Inelastic Mean Free Palh, Ài. Relative to lhe One for C I s. forSome of lhe Most Common Photoelectrons Ejected from Organic Polymers

and Some Inorganic Compounds Using lhe Magnesium Kal.2 Line

Parameters k and m to Compute Ài in Organic Polymers [35] andSome Inorganic Compounds [36]

Compound k m

Photoelectron Àj/Àj(C ls)0.145

0.107

0.139

0.137

0.149

0.138

0.138

0.112

0.103

0.0936

0.0933

0.794

0.790

0.782

0.790

0.786

0.790

0.791

0.760

0.777

0.770

0.766

oN

F

CI

Si

S

AI

0.794 % 0.005

0.906 % 0.005

0.655 % 0.005

1.068 % 0.005

1.146%0.005

1.095 % 0.005

1.166%0.005

26-n-Paraffin

Polyacetylene

Poly(butene-l-sulfone)

Polyethylene

PMMA

Polystyrene

Poly(2-vinylpyridine)

AI203

SiO2SiC

Si3N4

Given lhe low value of à for lhe energy range correspond-ing to photoelectrons generated by Mg or AI sources, this tech-nique analyzes depths of lhe order of a few atomic monolayers:for an homogeneous material, 63.2% ofthe signal comes from adepth equal to Ã, 86.5% from 2Ã,95% from 3Ã, 98.1 % from 4Ã,and 99.3% from SÃ. Given these values, it is generally consid-ered that lhe information depth is 3à but some authors take SÃas lhe information depth. EAL are generalÍy measured by lhetechnique of overlayer film [31] which presents many practicalproblems. However, many semiempirical and empirical formu-las exist to compute another related quantity, lhe IMFP Ã;. EALis different from IMFP by about 15-30% depending on lhe rel-ative importance of the elastic interaction processes [31, 33].The greatest differences arise for high atomic numbers and lowelectron energies. One of the most useful relations to obtainIMFP (for kinetic energies, Ek, larger than 500 eV), given itssimplicity, is expressed by [34]:

ing to Shirley's method [38]: lhe background intensity withina peak is assumed proportional to lhe peak integrated inten-sity (area between lhe peak and lhe background) at higher ki-netic energy. The spectrum is constrained to be zero at ei-ther end and lhe background to be subtracted is computed byan iterative method. Another very popular method is due toTougaard who has developed much work on this subject dur-ing lhe last 20 years [39]. Also Monte Carlo methods can beused to simulate lhe inelastic background to be subtracted [40,41]. The problem with these two methods is that they require avery detailed knowledge ofall lhe electron-medium interactionmechanisms and respective cross-sections to electron transportthrough tfle medium, a task still in progress [42]. That is whylhe old Shirley's method cited above keeps its popularity.

Once lhe background is subtracted, lhe area of a peak, due toaw photoelectron ejected from an element B, is given by [43].

IBw«(J) = K«(J) fi CB(X) exp(- X) dX (22)

10 ÀBw cos(J

where I Bw «(j) is the intensity of photoelectrons coming from el-ement B and having an effective att~nuation length ÀBw, CB(X)is the density of the element B as a function of the depth x(concentration profile), (j is the angle between the normal tothe specimen surface and the mean direction of analysis (seeFigure 8) I, and t' is the sample thickness. K «(j) contains thew photoelectron ejection cross section from the element B, theX-ray intensity,2 and a response function which depends on thew photoelectron kinetic energy, Ek (usually, <xEi: I). For a sam-pIe homogeneous in depth (CB(X) = constant = cB)andhavinga depth much larger than À (t' = 00), Eq. (22) becomes

(23)/wB =K'QwBCB

where QwB is called the sensitivity factor and K' is a constantindependent of the analyzed photoelectron. If two elements, B

Àj = kEí:' (21)

where k and m are empirical parameters. In Table III values ofthese parameters proposed by Tanuma et aI. [35, 36] for someorganic polymers and inorganic compounds are presented.

The order of magnitude of IMFP for a given photoelectrondepends on the crossed medium. For instance, Àj (C I s) com-puted using Eq. (21) and parameters in Table III varies between30 and 34 A in organic materiais and between i8 and 22 A in in-organic materiais. In many appiications, the reiative IMFPs aresufficient and, contrary to the absoiute ones, they are very insen-sitive to the medi um. In Tabie IV IMFPs reI ative to C Is IMFPfor several photoelectrons ejected from eiements contained inmateriaIs here studied are presented.

A NIST database is available in electronic form [37].

.' -

lThis angle (J is lhe complementary to lhe take-off angle [29] defined as lhe

angle between lhe specimen surface and lhe mean direction of analysis.

2Since lhe altenuation length for X-rays is much larger than for electrons,X-ray intensity is considered constant Ihrough lhe entire analyzed deplh.

3.1.6. Quantitative Aspects and ARXPS

As in any other spectroscopy, the major source of error in thecomputation of the area of a given peak is the background sub-traction. In XPS, background is generally curve fitted accord-

ls

ls

ls

2p

2p

2p

2p

PHOTONIC ANO ELECTRONIC SPECTROSCOPIES 285

and C, exist in lhe same sample, we can write

Fig. 13. Schematic representation of:lm 18tEELS experiment: Ei is lhe pri-mary energy, ai is lhe angle of incidemDe. 8Id au is lhe analysis angle, bothrelative to lhe surface normal.

about excited states from IR to.. UJV. more precisely from vibra-tions to electronic excitations aailionizations, without changingthe probe source or the detecti~ system.

3.2.2. Electron-surface lnteradilln Mechanisms

Depending on lhe mechanism (dipole, impact, or resonance)[47,49], electron interactions C3D be produced at different dis-lances. In particular, impact inter.:tions (short range) producedat typical interaction distances. of a few Angstr6m and observedin off-specular conditions becOllK: mainly sensitive to molec-ular groups exposed at lhe film-vacuum interface [50, 51]. Itis also very important to empilasize that HREELS selectionmies enable optical forbidden transitions to be observed. Infact, during lhe interaction, electron exchange between incidentandmolecular electrons can OCcDr. allowing optically spin for-bidden excitations (singlet-tripIet transitions, for instance). Onlhe other hand, symmetry modifications of lhe molecular or-bitals induced by lhe electromagnetic field associated with lheincident electrons can also break symmetry selection rules ofoptical excitations. Furthermore. in HREELS, electrons can ex-change lhe momentum as well as energy with lhe medium. Thisfact is important in lhe case of crystaIline materiais, since lhemomentum and energy transfers may lead to lhe observation ofnonvertical electronic excitations. which cannot be detected by

optical spectroscopy.

- ~=~~ (24)

lyc QyC Cc

Therefore, only rei ative values of Q are needed to makequantitative analysis. Usually, sensitivity factors are normalizedto fluorine F Is (Q(F Is) = I) or to C Is (Q(C Is) = I) and

are included in libraries fumished with spectrometers.Equation (22) is particularly useful when a concentration

gradient in depth exists. In this case, several spectra at differ-ent values of () are taken and lhe analysis is called angle re-solved X-ray photoelectron spectroscopy. However, for a max-imum efficiency, a flat surface (at an atomic levei) is needed toavoid shade effects as shown by Fadley in his early works in lhe1970s [44]. An additional problem exists: lhe extraction of con-centration profiles, CB(X), from Eq. (22) is an inverse problem:lhe intensity as a function of lhe analysis angle is lhe Laplacetransform of lhe composition depth profile of lhe sample [45]and does not have a unique solution. Several algorithms to solvelhe inversion problem were developed and tested [46]. They areali very unstable and sensitive to small statistical fluctuationsin lhe photoelectron intensity and to small uncertainties in lheanalysis angle (smaller than lhe experimentally available ones).The most usual and practical way of applying Eq. (22) is, there-roce, to test concentration profiles, for two or more elements,needing a few parameters (for instance, linear or exponentialprofiles needingjust lhe atomic densities at lhe extreme surfaceand at lhe bulk and lhe rale of variation with depth) and to com-pute ruem by fitting to lhe experimental data IB«(})/ Ic«(}). Themost simple concentration profile, requiring a single parameter,is enough for example to test if a given sample is made of amaterial covered by a overlaying film of homogeneous compo-sition and coostant thickness. In this case, admitting that ele-ment B onlyexists in lhe overlaying film and element C onlyexists in lhe material undemeath, we will have:

1mB QwB CB 1- exp(-t'/ÀwB)(2~)-=-x-x -

lyc: QyB Cc exp(-i/Àyc) ,--,

A single point is enough to obtain i. However, an angulardistribution is needed to confirm the assumption that the thick-ness film is amstant.

3.2.3. Equipment

3.2. High ~ution Electron Energy Loss Spectroscopy

3.2.1. Basicrrinciples

In a HREEIS experiment, a monokinetic electron beam witha primary e8:1"gY, E p, interacts with lhe surface region indifferent ways exciting vibrational modes [47] and electronicstates [48].l.-:nsity of the backscattered electrons is measuredversus prima:y energy for a given direction as shown in Fig-ure 13.

Primary e8:1"gy can be varied in a continuous way from Oto a few tens of eV (or even a few hundreds). Energy lossescan be meaS8al CromO to typically 1 eV ar; in some modifiedequipment, rn.n O to 15 e V, allowing oneto obtain information

The pioneering work in HREELS equipment development wasmade by several groups in parallel in Canada [52], France [53].Germany [54], and in United States [55]. A good reviewabout this item can be found in Chapter 2 of lbach andMills' book [56]. ln Figure 14, a schematic representation ismadeof the spectrometer LK 2000-R [57]-an old generationspectrometer-showing the main components: a filament whichemits an electron beam broad in energy and two electrostatic1270 cylindrical energy selectors to obtain a beam with a fwhmof about 5 me V. This beam impinges on the sample (which canbe rotated).

Backscattered electrons are analyzed by a third energy se-lector (in other models, two selectors are also used to analyzebackscattered electrons) and multiplied and counted by a chan-neltron. Between the filament and the first ener~y selector and

BOTELHO 00 REGO ANO VIEIRA FERRElRA286

4T10

9an. 8inc.

..~ 5

. .'::' ...' ".

A.,;'lyzer -', 2nd Mf)DOChromator. "

\, \

I SamnleLo

10o 5

En~ (b)(a)Channeltro~6T~",I.6T

10 T-10

. . . . . . . ., . . . . . . . . . .~ Electron Gun \ "

\ "," .-"--~~

1st Monochromator

~ 14. Schematic representation ofLK2000-R spectrometer. Sample can be

i3ted. The analyzer can also be rotated from 8"/1 = 13° to 8"/1 = 68°.

~~ 5 r,.js

./Ep= 1.020IE.-0.0804

1R = 0.9998. . . I . . . ,5

E.

o o

(c) (d)

Fig. 15. Extent of HREELS spectra. E p, as a functionof nominal energy, En.for: (a) 4-thiophene; (b) 5-thiophene; (c) 6-thiophene films on gold with chainsperpendicular to lhe substrate; (d) 6-thiophene films on graphite with chainsparallel to lhe substrate. Straight lines fitted by a least square method are also

displayed having lhe following equations: (a) (0.93:t:0.06)+(0.910:t:0.015)En;(b) (0.53 :t:0.13) + (0.952:t:0.024)En; (c) (0.81 :t:0.08) + (1.010:t:0.016)En;(d) (0.00:t: 0.09) + (1.02:t: 0.02)En. Energies are in eV.

between lhe two energy selectors injection lenses existo Be-

t\\"een the sample and the neighboring energy selectors there

are also focusing lenses. Ali magnetic components must be ex-

cluded from the instrument and shielding (assured by cylindersof mu-metal) is required to prevent magnetic deflection from

extemally based fields (the magnetic earth field, for instance).

The new generation models, providing energy resolutions bet-

ter than I me V, have essentiaI1y lhe same components but the

injection lenses have differentdesigns [58].Primary energy can be varied from O to 250 eV, in the par-

ticular model presented in Figure 14, by varying the voltage

applied between the filament midpoint and the sample. For the

same model, energy losses cao 00 analyzed from O to 15 e V, en-

abling the acquisition of complete spectra for incident energies

lower than 15 eV. However, for lhe ultimate resolution spec-

trometer, the range of measurable energy losses is much more

limited, ranging from O to 1 or 2 eV.

3.2.4. Sample Charging

only consequence is that the primary energy depends on the fi-nal charge. This fact can be used to obtain qualitative informa-tion about the reI ative conductivity of samples. For instance for

three different a-oligothiophenes films-quaterthiophene (4T),quinquethiophene (5T), and sexithiophene (6T)-deposited ongold samplesby a procedure described elsewhere [61] therela-tion between E p and En becomes increasingly linear and theslope of the variation approaches unity as the chain size in-

creases (see Fig. 15).However, for the two 6T samples, no difference exists be-

tween slopes. The only difference comes from the interceptvalues. This means that their contact potentials with the spec-trometer are different since graphite was replaced by gold.

Other sample requirements are low vapor pressure sinceHREELS spectrometers work under ultra-high-vacuum condi-

tions.

4. EXAMPLES OF SYSTEMS STUDIED BYELECTRONIC AND PHOTONIC SPECTROSCOPIES

When a low energy electron illJl)inges on an insulating solidsurface, a large probability f(W ~netrating the solid and be-ing trapped exists. Hencefortlt. ak surface charges negativelyif lhe electron primary energy às k!>wer than the ionization en-ergy and mar charge negative1y w positively for larger primaryenergies depending on the sea.-llary electron yield. Since theelectron energy is determined by tne voltage difference, ~ VII'between the emitting filame" aIRCiI the sample surface, if thesurt.ace charges, the primary ~y changes continuously dur-ing 11leacquisition of a spectnDml.. The primary energy, Ep, isrelated to the nominal energy. F.".,,!by

Ep = En + e<l>cl + e<l>c = #!i(~ Vn + <l>cl + <l>c) (26)

where <l>cl is the contact poteoti:lil, <l>c is the charge potential,and e is the electron char",e. P~ymeric films always chargeunder electron irradiation. F.." very thick films, a ftood gun isfrequently used to stabilize dE accumulated charge [59], even-tually combined with spectral re$toration algorithms [60]. Forvery thin films, with thickness arl1llind a few hundreds of A, thissurface charge attains an equihorimm value very rapidly and the

4.1. Microporous Solid Surfaces and Finely Divided

Powders

4.1.1. Surface Photochemistry Studies

There is a growing interest in studies of photoprocesses regard.ing molecules on surfaces, either physically adsorbed or chefioically bound to solid surfaces, grains, powders. or gels [1-5].

PHOTONIC ANO ELECTRONIC SPEC 'ROSCOPIES

A*.s.. B,

~

A' B-(a)

(b)

Common adsorbents are oxides such as silica, alumina,alumino-silicates, and clays, among others. Rarer studies werealso presented for cellulose and cellulose derivatives or starch[2a]. Cyclodextrins and calixarenes are good examples of or-ganic host molecules, which can form inclusion complexes withmany guest molecules.

In a great variety of solid supports, the number and natureof the surface reactive groups drastically affect the distributionand local organization of the adsorbates. The adsorbent pre-treatment and handling, or the solvent used for probe deposi-tion, strongly influence the surface and, as a consequence, themolecule's adsorption. Therefore, their photochemistry and/orphotophysics vary according to the surface pretreatment.

Two goals usually exist in these studies: to observe the wayinteractions with the surface affect the probe's behavior in theexcited state, and also how to use photochemistry as a tool toprobe the surface of an unknown substrate.

Apart from the interest in the above.;mentioned substrates,which we can call "electronically inert substrates" (Fig. 16a),we should also refer to the intense research activity in the fieldof "electronically active substrates" (Fig. 16b), namely semi-conductors.

In these systems, as the solid substrate absorbs the excitationradiation, electrons are promoted to the conduction band andholes are formed in the valence band, which may react with theadsorbate at the solid surface. The excited states of the adsorbedmolecule may also be quenched as a consequence of an electrontransfer from the probe to the surface. In this text, we shall notconsider these cases.

4.1.2. Pioneer Work in This Field

The initial studies on the photochemistry and photophysics ofmolecules adsorbed DotO solid surfaces are relatively recentand were done in the 1960s (apart the work of Boer et aI. inthe 1930s [62]). Leermakers [63] presented an excellent reviewof the work produced up to 1970, where he also included bisown work concerning ketones adsorbed on silica gel, the cis-trans isomerization of stilbene on silica, the spiropyranes pho-tochromism, and the photoc1eavage of cyclohexadienones.

These studies clearly established that in many cases, after ad-sorption, energetic changes have occurred, as well as changes inthe nature of the electronic excited state. As a consequence, theefficiency of the various photophysical and photochemical pro-cesses also changes. The adsorbate-adsorbent interaction may

be simultaneously nonspecific and specific, namely, the inter-actions of the surface-active groups (as an example, the -OHgroups on the oxide surfaces). The interaction forces are re-sponsible for adsorption and may have an electrostatic or dis-persive nature or hydrogen bondlng formation.

Several spectroscopic and nonspectroscopic techniques maybe used to study the bonding nature ofthe adsorbate to the sur-face [2a, 4]. In the first case we want to emphasize the im-portance of diffuse reftectance techniques for absorption andemission studies in the ultraviolet (UV), Visible (Vis), andnear infrared (NIR) spectral ranges, X-ray photoelectron spec-troscopy, and Fourier transform infrared spectroscopy. In thesecond group, we refer the heat adsorption and the isotherm ad-sorption techniques, among others.

It is obviously crucial to have a detailed knowledge of thesurface structure and its modification under different experi-mental conditions.

We will refer here to some of the initial studies on silica gel,Vycor porous glass, alumina, and some zeolites. In the 1970sand in the 1980s several groups produced important work in thefield, namely de Mayo and Ware et ai. for silica surfaces [64]:Oelkrug et aI. [65], and Thomas et ai. [66] who presented sev-eral studies of different probes on alumina surfaces, amongothers; Turro and Scaiano published very interesting work forsilicas, zeolites, and other surfaces [67, 68]. AlI these studiesprovided a solid base for this new discipline.

Wilkinson et ai. developed, at the end of the 1980s, the dif-fuse reftectance laser ftash-photolysis technique [I, 10], whichproved to be crucial for transient absorption and emission stud-ies on surfaces, providing both spectroscopic and kinetical in-formation. This technique for studying solid and opaque me-dia became so important for surface studies as lhe conventionalftash-photolysis was and still is for transparent media, after itsdiscovery by G. Porter in lhe 1950s.

Pyrene was one of lhe widely used probes in the initialsurface photochemistry studies due to the long lifetime of itsmonomer, lhe capacity of excimer formation, and also its spec-traI sensitivity. The 111/1 (370 nm/390 nm) vibronic band fatiawas successfully used to monitor lhe microscopic polarity ofthe adsorbent, either onto silica or alumina [66]. Peak I, lhe 0-0band of lhe So -+ S) absorption, is symmetry forbidden andgrows in polar media. In lhe case of alumina, this surface ex-hibits a surface polarity similar to lhe one presented by polarsolvents such as methanol [66a].

On lhe silica surface, pyrene shows a 111/1 peak fatia alsocharacteristic of a strongly polar and hydrophilic surface. Thederivatized chlorotrimethylsilane surface revealed higher val-ues for lhe 111/1 fatia, showing an increase in surface hydropho-bicy due to this treatment. Solvent co-adsorption also changesthe 111/1 peak fatia, therefore showing the variation of lhe hy-drophobic and hydrophilic surface characteristics [66b].

de Mayo et ai. [64] and other authors [69, 70] used severalpolycyclic aromatic hydrocarbons (PAHs) such as naphthalene.pyrene, and anthracene as probes to study lhe surface mobil-ity and also aggregation effects on different surfaces. The per-centage of lhe surface coverage (usually expressed in terms of

288 BOTELHO DO REGO AND VIEIRA FERREIRA

I~CX:)::)...H/O-~

H,

'ImJirr1/;

Fig. 17. Schematic representation of naphthalene and acridine adsorption onsilica surface.

emission are similar to thos~_obtained in solvem such as wa-ter or ethanol, lhos suggesting that acridine is simply boundto the substrate by hydrogen bonds and is the ooly emissivespecies. Al203 pretreatment at 600 °C results in Au+ complexformation between the lone electron pair of the nitrogen atomof acridine and the Lewis acidic sites of alumina. whereas AH+species are unfavored by dehydroxilation.

The acidity of surfaces such as the Vycor ~s glass wasalgo studied using appropriate probes: Lin et aI. [nJ used 9,10-diazofenanthrene as proton acceptor from the BlÕnsted acidicsites of the surface, which bind to the nitrogeo atoms of the

probe.Suzuki and Fujii [73] used acridone as a probe for differ-

ent pretreatments of silica. The results obtained for acridoneadsorbed on moderately pretreated silica (simply warmed at200 °C under a reduced pressure of """ I 0-6 mbar) were com-pared with the emission in benzene, ethanol. and H2S0418 N. The conclusion was that the main emissive species isthe acridone molecule, which forms hydrogen bonds with thesilanol surface groups of silica (the carbonyl group of acridoneor N -methylacridone interacts with the surface hydroxil groups,whereas the amino group hardly interacts). The protonatedspecies algo emits on lhe silica surface, but this component isless important.

It is worthwhile to reter here to the use of rhodamine B andrhodamine 6G and other xanthene dyes as probes for the studyof organic crystals, calcium fluoride, and quartz plates [74].

ln alI the examples quoted until now, molecules with singletexcited states were used as probes, with relatively short life-times, usually in the nanosecond time range.

The use of triplet excited states as probes to study the sur-face properties of many solids is particulary interesting, doe tothe fact that they usually exhibit long lifetimes which in mostcases come closer to those obtained for rigid matrices. Theselong lifetimes increase in many cases the efficiency of severalphotochemicalprocesses. Therefore spectroscopic and kineticstudies can be performed in a wide and interesting variety ofsituations. As we said before, the development of the diffusereflectance laser flash photolysis technique [I, 10] by Wilkin-son et aI. was crucial for the development of these studies onsurfaces.

Later we will present some examples of studies of moleculeswhich exhibit high intersystem crossing yields used as probesfor surface studies, but before doing that, it is important to de-scribe the methods for sample preparation. We algo describesome of the substrates used for surface photochemistry studies.

% of the monolayer) is crucial with respect to the two above-mentioned aspects. PAHs molecules have delocalized 7r elec-trons, which strongly interact with the adsorbent surface byforming hydrogen bonds, provided neither steric hindrance norphysically adsorbed water (which may prevent the probe-hostdirect interaction) acts as a barrier.

This specific host-guest interaction may also occur withnonbonding electron pairs of heteroatoms, as Figure 17 shows.

The fluorescence emission of adsorbed PAHs (either frommonomers or excimers) is usualIy multiexponential, in accor-dance with the adsorbent heterogeneity, thus showing differentadsorption sites. UsualIy those multiexponential decays are an-alyzed with two or three exponential decays, according to

I (t) = ale-I/TI + a2e-I/T2 + aJe-I/T3 (27)

where i is the average lifetime decay

i = La;r;2/ La;r; (28)

and fi = ai r; / L a; ri is the fraction of the excited molecules,which have a lifetime of r;. Several other kinetic models werepresented for data analysis in heterogeneous media [65, 66, 71].

Oelkrug et ai. used several silicas and alumina as adsor-bents, and different PAHs, diphenylpolyenes, and acridine andderivatives, among others, as adsorbates [65]. They have shownthat both spectroscopic and kinetic data were strongly depen-dent on the surface's pretreatment and also on the adsorptionprotocol used for sample preparalian. In the case of acridine(A) and for moderate pretreatment at low temperatures of thealumina surface (activation temperature Ta "-' 100°C), the in-

teraction with the surface is essentialIy composed of hydrogenbonds and protonation of the adsorbate (AH+). For silica acti-vated at 300 °C, both ground-state absorption and fluorescence

4.1.3. Some Solid PowderedSubstrates: Cellulose, SilicaGel, and Silicalite

4.1.3.1. Cellulose

Cellulose has been used as a solid powdered substrate for lhestudy of photophysical and photochemical studies of severa!organic probes, mostly dyes. Some of lhe properties of this sub-strate, namely lhe capacity of adsorbing molecules both by en-trapment and on lhe surface of lhe natural polymer (forming

PHOTONIC AND ELECTRONIC SPECTROSCOPIES 289

"II

011 i3 lHO,.~ .0 I"

lk)k" I<9 CH2OH I

@

)n

~

1ig. 18. Cellulose

the removal of the solvent used for sample preparation, and for aswelling solvent, a chain-guest-<:hain interaction is promoted,replacing the previous chain-solvent-<:hain interaction.

Df particular relevance is the study of dye photodegradation,either on wet or dry cellulose, due to the importance of thiseffect in the textile industry [89]. Today we know that an in-crease of the humidity content in the fiber promotes a del:reasein the lightfastness of many dyes. This effect exists both forcotton and other polymers where the dye may be adsorbed orcovalently bound [89-91]. It also occurs for wool, although toa smaller extent, since in this substrate the mechanism is essen-tially reductive, while in cellulose it is an oxidative one [89].

4.1.3.2. Silica Gel

in many cases hydrogen bonds) and also the absence (or ex-tremely reduced) of diffusion of oxygen, make this substrate aparticularly attractive one for room temperature luminescencestudies [75-83]. We recently published some ftuorescence andphosphorescence studies of rhodamine dyes [15, 81, 82], au-ramine O [81b], 2,3-naphthalimides [84], oxazine [85], acridineorange [86], and cyanine dyes [16,87,88] adsorbed on cellu-lose.

In ali these studies, microcrystalline cellulose was used asthe solid powdered substrate. The structural formula of cellu-lose is presented in Figure 18.

From a structural point of view, cellulose is a polymerof D-glucose in which the individual units are connected by,B-glucoside bonds between the anomeric carbon of one unitand the hydroxyl group in the C4 position of the neighbor unit.Cellulose is probably the most abundant organic compound thatexists on earth. It is the chief structural component of vegetalcells. Wood strength is derived from the hydrogen bonds be-tween the hydroxyl groups of neighbor polymer chains. Thesehydrogen bonds are favored by the linear structure of the natu-ral polymer, which presents conformations that favor those in-teractions. In the case of starch, although it may also presentlinear structures, it has no linear conformations; thus the hydro-gen bonds are unable to become the main interaction betweenchains.

X-ray diffraction studies have shown that native celluloseis a two phase system: one is amorphous, with lower order andcompactness, and is localized at the elemental fibril surface; theother one is highly ordered and compact (crystallites), wherethe polymer chains are well organized (crystalline structure)and strongly bound by hydrogen bonds.

Microcrystalline cellulose is simply apure form of celluloseobtained by ao acid treatment of native cellulose. The amor-phous regions are preferentially attacked and transformed, andthe final residue is highly crystalline.

The swelling of cellulose in the presence of moisture is awell known property of this material. Other protic and nonpro-tic solvents such as methanol, ethanol, acetonitrile, and acetonealso have the capacity to swell microcrystalline cellulose. How-ever, solvents such as benzene, toluene, or dichloromethane donot promote this effect. Therefore it is possible to control ad-sorption of probes on microcrystalline cellulose, either on thesurrare or entrapped within the natural polymer chains. After

Silica gels are porous and granular forrns of amorphous silicas,formed by a complex net of microscopic pores which attract andretain water or organic solvents by means of physical adsorp-tion [92-98]. Porous silica has a sponge structure, from whichresults a very high specific surface area, that varies greatly withpore size (from 20 to 750 m2fg). This surface area is essentiallylhe internal area of lhe pore walls. The average pore size canbe obtained by surface area measurements. Due to lhe fact thateven lhe smaller pores are larger than most molecules, althoughof lhe same arder of magnitude, it is not a surprise that restric-tions of mobility do occur for adsorbed molecules [7].

Porous silica surface contains both silanol (Si-OH) andsiloxane groups (Si-O-Si). Silanols are considered to bestrong sites for adsorption, while siloxanes are hydrophobicsites [92, 96]. Silanols may be isolated, vicinal, or geminateand may be linked by hydrogen bonds to lhe surface water. Fig-ure 19 shows lhe silica surface in a schematic way.

In lhe pioneering study of Snyder and Ward (1966) [92]H-bonded silanols forrn pairs and these pairs were considered tobe lhe more active sites on a surface (due to an enhanced acidityof lhe proton not engaged in lhe H-bond), responsible for probeadsorption at lhe silica surface. We now know (by temperature-programmed desorption studies) that isolated silanols are lhemore reactive sites on lhe silica surface [98]. H-bonded silanolshave a desorption energy of 50-60 kcal moI-I and lhe isolatedsilanols 90 kcalmol-1 [98].

Moderate heating of silica in vacuum (100-120 °C) ensuresa quasi-complete removal of lhe physically adsorbed water.

BOTELHO DO REGO AND VIEIRA FERREIRA290

Vício aI (I)

.aSubstratewilhadsorbedprobe

~ Solution (a)H

IH -+-

", . H Vicinal (fi)O O /

O- / HGeminate

0-7HO

OH Si-OHIsolated Internal

Fig. 19. Schemalic represenlalion of silica surface. ~SolveDt~

evapomtion

(b)~ with

adsorbcd

probe

,--,. ~~.Solution

Substrate '" " L--' 'I. -Dccantation Solution

Fig. 20. Sample preparation methods: (a) slurries. (b) solvent evaporation,

(c) equilibrium.

withadsorl>ed (C)

probe

t

However, the final water monolayer is removed at activationtemperatures of 200 oCo The use of higher temperatures (200-1000 °C) promotes removal of chemisorbed water [96].

The importance of the hydroxyl groups from the point ofview of adsorption comes from the fact that the higher lhepercentage of active silanols per surface area unit, the largerthe efficiency of the adsorption processo SmaIl pore size silicaspresent a higher percentage of active silanols relative to largerpore size silicas [92, 93, 96]. In the former case we detectedconformer formation for several dyes, depending both on thepore size and on the dye structure itself [7].

4. J .3.3. Si/ica/ite

Silicalites are very specific forros of pentasilic zeolites withsmall pores and, opposite to alumina-silicate zeolites, theypresent a strong hydrophobic and organophylic character. Inter-nally they have linear channels of e\liptic cross-section (5.7 x5.1 A), which intercross with zig-zag channels of an almost cir-cular cross-section (5.4 ::I:: 0.2 A) in the case of silicalite I orlinear ones in lhe case of silicalite 11 [99, 100].

Silicalites are characterized by an almost complete ab-sence of aluminum in the structure, while zeolites ZSM-5 andZSM-II, although with a similar structure, present sma\lerSi/AI ratios [94, 100]. The hydrophobic characteristic of thesemateriaIs arises from the absence of AIO2 units of the crys-talline structure.

Silicalites are used for removal of organic compounds fromwater or from industrial exhaust smokes. Inclusion of organicmolecules into the silicalite channels may impose significantconformational restrictions [83]. The reduced dimension of thechannels enables the selective chromatographic use of this ma-terial [99b].

Generally speaking, one of the simplest and widely usedmethods is the solvent evaporation method. Solvent evapora-tion methodology means the addition of a solution containingthe probe to the previously dried or thermally activated pow-dered solid substrate, followed by solvent evaporation from theslurry. This can be made in a fume cupboard or by the use of arotating evaporator. The final removal of solvent can be madeunder moderate vacuum, '"'"" 10-3 mbar, and the evaluation ofthe

existence of final traces of solvent can be monitored by the useof IR spectra. (See Fig. 20.)

The simplicity of this procedure and the possibility of asimple calculation of the adsorbate concentration are impor-tant advantages of this method. Therefore, ground-state diffuse-reflectance absorption spectra of the above-mentioned pow-dered samples (and also of a blank sample) enable a calculationof the molar extinction coefficients of the probe. These can becompared with the one obtained for transparent samples (ho-mogeneous solutions, films, or solid matrices) by the use oftheBeer-Lambert Law.

An alternative procedure for sample preparation is theequilibrium method: the solution still containing some probeis removed (by centrifugation, for instance) after prolongedcontact with the sample (so that an equilibrium situation canbe reached), followed by the residual solvent evaporation. Theprobe concentration calculation can be made after determina-tion of the amount of probe left in the decanted liquido The bigdisadvantage of this method is that large error~ can occur in theevaluation of the probe concentration, namely by solvent evapo-ration. The obvious advantage is that a thermodynamic equilib-rium is reached where the probe is shared between the solvent(as solute) and surface (as adsorbate), according to the inter-active forces present in each specific case. This is a dynamicequilibrium and the probe molecules keep on going from thesolution to the solid surface or from the surface to the solvent.

4.1.4. Sample Preparation: Slurries, Solvent EvaporatedSamples and "Equilibrium" Samples

The adsorption of probes on microcrystalline cellulose or nativecellulose, on the surface of different pore size silicas ,:,r alumina,and on silicalite surfaces (or other zeolites) has to be made in adifferentiated manner, according to each's adsorbent character-istics. Adsorption of probes onto these powdered solids can beperformed from a solution containing the probe or from a gasphase.

PHOTONIC ANO ELECTRONIC SPECTROSCOPIES 291

Therefore no "forced" probe aggregation occurs due to lhe fast

solvent evaporation.In lhe samples prepared by both methods and after solvent

removal, ground-state absorption spectra may exhibit devia-tions either to lhe red or to lhe blue, depending on lhe sur-face characteristics, for surface coverages less than lhe mono-layer. Some broadening of lhe absorption bands can also bedetected, or even new bands in lhe absorption spectra, or sig-nificant changes in lhe extinction coefficients for lhe differentvibronic bands. The formation of "new" species may be relatedwith lhe "pool" effect which occurs during evaporation (probemolecules may aggregate or form small crystals as lhe solventevaporates). Due to lhe limited number of available adsorptionsites at lhe adsorbent surface, as soon as each site is occupiedby a first molecule, lhe forthcoming molecules have a smallerinteraction with lhe surface. As long as lhe solvent is being re-moved, lhe amount of these weakly interacting molecules in-creases, and lhe probability of forming larger ground-state ag-gregates or small crystalsalso grows. Some examples of studieson surfaces where absorption spectra exhibit hypsochromic andbathochromic shifts and also data regarding ground state aggre-gate formation will be given in lhe next section.

Slurries provide mixed information, refiecting both adsorbedmolecules and solution molecules. In this sense, this type ofsample should always be avoided.

In Section 4.1.5 we will describe some of ourstudies on sur-faces using lhe techniques and substrates described so faro

4.1.5. Some Examples of Surface Photochemistry Studies

sition) in samples prepared by both solvent evaporation and theequilibrium method (see [14] and other data not yet published).

The use of polar protic or nonprotic solvents for samplepreparation (methanol, ethanol, acetonitrile, acetone) and evena nonpolar solvent as dioxane promotes hypsochromic shifts inthe n ~ n* transition of benzophenone, while with solventssuch as benzene, isooctane, or even dichloromethane, thesebands are shifted to the red and exhibit some vibrational struc-turco These facts are a consequence of lhe different swellingcapacities of these solvents toward lhe cellulose polymeric ma-trix [14, 79b, 80]. The chain-<:hain interaction is replaced bylhe chain-solvent-<:hain interaction, allowing different probemolecules to penetrare within lhe matrix in different ways, oreven lhe probe simply remains in lhe external cellulose surfacefor nonswelling solvents. In this case, increasing concentrationsof the probe easily give rise to lhe formation of ketone micro-crystals. Therefore, benzophenone could be used as a probe toevaluate lhe swelling capacity of several solvents regarding lhecellulose matrix.

The consequences of this larger or smaller intimare contactof cellulose with benzophenone are also remarkable from lhepoint of view of lhe probe photochemical behavior. In fact, lhetime resolved absorption spectra for benzophenone/cellulose/dichloromethane samples, where lhe probe is deposited on lhesurface, present spectroscopic and kinetic characteristics verysimilar to the microcrystal case. On the contrary, in lhe case ofethanol or other swelling agents, lhe ketyl radical of benzophe-none is formed with a lifetime tanger than lhe triplet state ofbenzophenone [14, 87], and lhe transient absorption peaks atabOlir 550 nm.

Another interesting example of lhe photochemistry of ben-zophenone within microcrystalline cellulose is lhe geminateradical pair formed following laser excitation of samplesof coadsorbed benzophenone and 2,4,6-trimethylphenol. Thediffuse-reftectance absorption spectrum immediately after laserfire shows both lhe ketyl radical of benzophenc"e (BZPH8)and lhe phenoxyl radical, thus formed (PO8) [13a]. The ki-netic study of lhe geminate recombination of these radicais hasshown short ('"" I O J1.s) and long components ('"" 100 J1.s) anda multiexponentialdecay pattern which reftects heterogeneousadsorption sites for lhe radical pairo Cellulose provides very lowmobility for both radicais, which form a contact pairo No exter-nal magnetic effect was detected in this system.

4.1.5.1. Ketones Adsorbed on Microcrystalline Cellulose.Photochemistry and Geminate Radical PairsFormation FollowinR Laser Excitation

Benzophenone. The use of the diffuse-reflectance ground-state spectra methodology presented in Section 2 has lead usto several interesting results in the study of benzophenoneadsorbed on microcrystalline cellulose. Figure 21 shows theground-state absorption of that aromatic ketone (n -+ n* tran-

\ ..":7"'\~:1 Mixture1

{J-Phenylpropiophenone. Another interesting example of theadsorbent's inftuence on a specitic probe is the case of{J-phenylpropiophenone included in the channels of silicalite.It is well known that this molecule presents different con-formations in a solvent [68] as Figure 22 shows. In one ofthem (I), the phenyl group may approach the excited carbonylgroup; therefore a fast de-excitation of the triplet excited stateof this molecule occurs. The triplet lifetime of a benzene so-lution of {J-phenylpropiophenone is about one nanosecond.{J-Phenylpropiophenone inclusion into the silicalite channelsincreases the lifetime of this molecule tive orders of magni-tude [83b], as Figure 22 shows.

~.- I§.ri:ü I

ti>. Iog~ I

Me!

v

Fig.21.celIulose.

o . 750 JI.:ol g-l . . ~300 320 340 360 380 400

Wavelength, nm

Remission function for benzophenone adsorbed on microcrystalline

0.8

0.6

0.4

0.2

BOTELHO DO REGO AND VIEIRA FERREIRA292

IJ - PP conformations

~)°r§ft

::iai. 50000

P'0;.:

E'8 30000 2 ' f.:

~ ~;~;;~- 8 3 ~ .2 10000 4 ~ 5 O

.Q O~

300 400 500 600 700

Wavelengili, nm

Fig. 22. Time resolved room lemperalure phosphorescence emission speclra

of f3-phenylpropiophenone included inlo silicalile channels.

1-0115

2-1001153 - 200 ~s

4 -300~s5 - 400 us

A

\

ever, by expressing concentrations in tenns of moles of probeper surface area unit (the surface area was ~tennined by N2-BET measurements), the cate constants for dIe quenching pro-cess become similar [67].

We perfonned similar experiments for tIM: systems benzo-

phenone/l-methylnaphthalene [101], benzophenone/oxazine725 [85], and acetonaphthone/acridine orange [86], as donor/acceptor pairs, ali systems co-adsorbed on microcrystalline cel-lulose. In some of these cases the energy transfer process wasalgo studied in solution (where the quenching process is diffu-sion controlled) for comparison purposes.

The main conclusion was that on cellulose the quenchingprocess has a static nature. Both direct absorption of the excita-tion radiation by the acceptor, as well as radiative transfer, haveto be quantified for a correct evaluation of lhe energy transfer

efficiency.A simple model for static quenching was presented, where

two types of benzophenone molecules exist: those that have anacceptor as the nearest neighbor and are immediately quenchedafter laser excitation, and the others that have no such neigh-bor and are not quenched. This simple model was enough forthe interpretation of the results on the cellulose surface. A ki-netic analysis showed that the lifetime of the donor remainedunchanged within experimental errar, only the emission inten-sity at time zero decreased, again in accordance with the staticnature of the quenching processo

In the acridine orange and in the oxazine case, delayed fluo-rescence was detected, by direct excitation in the first case andsensitization in the second one [85, 86].

Another very important conclusion that emerged from thesestudies (where microcrystalline cellulose was used as powderedsubstrate) was that molecular oxygen did not quench the tripletstate of molecules entrapped within the polymer chains of thisnatural polymer. AIl these studies could be done with air equi-librated samples, and nitrogen or argon purged samples providesimilar results for the efficiencies of the quenching process,within experimental errar. This property of cellulose makes ita very special substrate to be used for room temperature fluo-rescence and phosphorescence studies.

This phosphorescence emission was obtained at room tem-perature and with air equilibrated samples. Oxygen diffu-sion inside lhe silicalite channels is reduced into lhe channelswhich already have f3-phenylpropiophenone molecules andmany triplet excited molecules are not quenched and are there-fore emitting. In argon purged samples lhe triplet lifetime in-creases about ten times [83c]. f3-Phenylpropiophenone includedwithin microcrystalline cellulose chains enabled us to detect lhekety I radical of this species and therefore contrast with its solu-tion photochemical inertia [83a].

Several other diaryl and alkylarylketones also exhibit roomtemperature phosphorescence in air equilibrated samples whenincluded in silicalite [83c] or forming inclusion complexes withcyclodextrins [83c, 102], depending on lhe probe and cavitysize. Both substrates provide some degree of protection fromoxygen quenching, as well as imposed conformational restric-tions that decrease lhe nonradiative mechanisms of deactiva-tion.

4.1.5.2. Energy Transfer on Suifaces: Some Examples

The first study of a triplet-triplet energy transfer process onsurfaces (silicas with different porosity) was reported by Turroand co-workers [67] regarding the benzophenone/naphthalenesystem. Following the selective excitation of benzophenone(355 nm), the triplet absorption of naphthalene appeared, peak-ing at about 400 nm. Increasing amounts of added naphthaleneresulted in an increase of the rate constant for lhe benzophe-none decay, presenting evidence for lhe dynamic nature of lhequenching process on the silica surface. The authors algo showkinetic data for 255 and 95 A pore silicas, which seem to evi-dence larger quenching for the silica with lhe larger pore. How-

4.1.5.3. Fluorescence Quantum Yield Detennination onSurfaces: The lnfiuence 01 Aggregation

Luminescence quantum yield determination of dyes and otherorganic molecules adsorbed on solid substrate surfaces is animportant problem, although with a difficult approach. One hasto determine the amount of light absorbed by the sample at theexcitation wavelength and also use the appropriate standards.A careful study of the probe aggregation has to be made üueto the fact that aggregation may deeply affect emission. Self-absorption also may exist and has to be taken into account. The

same applies to concentration quenching effects.We have studied this problem using first rhodamines 101

and 6G and later sulforhodamine 101, adsorbed onto microcrys-talline cellulose. After detailed concentration studies, we con-cluded that these two compounds could be used as reference

293PHOTONIC AND ELECTRONIC SPECTROSCOPIES

;:j~

008.-ci'o

°p§ 0.6

~d 0.4o

o~'"

Os 0.2

~O

6

5

450 500 550 600 650 700

Wavelength, nm

Fig. 24. Ground-state diffuse reftectance spectra of sulforhodamine 101adsorbed on microcrystalline cellulose. Curve 1-{}.01 /lmoVg-l; 1-0.05 /Lmol/g-l; 2-{}.10 /Lmol/g-l; 3-{}.25 /Lmol/g-1 4-ü.50 /lmoVg-15-{}.75 /Lmol/g-l; 6-1.0 /Lmol/g-l.

::j=

i-.~]8~~

~I.:

0.2 0.4 0.6

(l-Re) . f ~0.8o

~-Rhodamine 101 , -.. '-~ ~ Sulforbodamine 101

COOH lo,D"~.. ;" ~ 1 .;;s?H """ O N CIO~ 'H)l?,- H. . l..-, .

Fig. 23. Luminescence quantum yield deterrnination of molecules adsorbedon surfaces.

This figure also shows the analogies in the main equationsused for determining the fluorescence quantum yield in solutionandon solids(foroptically thick samples) [7,14,15.81-88]. Inthis figure u stands for unknown sample, s for standard, G forgeometrical factor, and f for fraction of the incident light ab-sorbed by the emitting probe under study. AlI other parametershave the usual meaning.

Several molecules with low fluorescence emission quantumyields in solution, such as several cyanine dye and auramineO, may exhibit a three, falir, and sometimes five orders ofmagnitude increase in the quantum yield of emission at roamtemperature, simply because the probe was entrapped withinmicrocrystalIine celIulose polymer chains. The host imposessevere restrictions to the mobility of the guest molecule, there-Core reducing the nonradiative mechanisms of deactivation

[81, 83b].Residual amounts of moisture may have a significant quench-

ing effect in the fluorescence quantum yield of many entrappedmolecules [81 b, 82].

For most dyes, but not alI, aggregation is particularly rele-vant. A good example is sulforhodamine 101 adsorbed on cel-lulose, as we show in Figure 24 [81 b, 82], where aggregation isonly present for the two higher loadings. For auramine O, a veryflexible dye, almost no aggregation was detected on celIulose,using ethanol as solvent for sample preparation [81b].

compounds for fluorescence quantum yield determination ofprobes on solid surfaces [81]. Later this method was extendedto studies of other surfaces and other molecules [15, 16, 82, 88,

102-104].The method is based on the comparison of the slopes of the

linear part of the curves IF versus (I - R)fprobe for an un-

known sample and for the standard one. These curves may beobtained by plotting the fluorescence emission intensity (I F )as a function of the absorbed light at the excitation wavelengthwhich is proportional to (I - R)fprobe for each sample with

a specific dye concentration [81]. The reflectance R has to bedetermined at the excitation wavelength, with the use of an in-tegrating sphere, as described in Section 2.1.

Figure 23 shows the calibration curves we currently use inour laboratory for luminescence quantum yield determinationof probes adsorbed on powdered solids, for optically thick sam-pIes. We would like to stress the importance of knowing withaccuracy the excitation energy profile for each specific appara-tus being used, as we have shown before [104].

4.1.5.4. Photochemistry o/ Dyes Adsorbed on MicrocrystallineCellulose and Different Pore Size Silicas

Cyanines. Photochemical and photophysical studies of cya-nine dyes are an important and up-to-date research domain, dueto their use in several relevant applications, such as black andwhite and colar photography, laser dyes, potential sensitizers incancer photodynamic therapy, and also devices for optical stor-

age of data [88].Cyanine aggregation in an electronically inert substrate is

exemplified in Figure 25, where several diffuse-reflectancespectra of thiacyanines are shown, namely 3,3'-diethylthiacar-bocyanine (TCC) and 3,3'-diethyl-9-methylthiacarbocyanine

BOTELHO DO REGO AND VIEIRA FERREIRA294

TCCrgr::~"""",~"1QJ

bzH5 b2Hs.

3;3' -.li.lhylthiacarbocyaau.. iodid.

TCC,8

::3d

f:!o

.~~~ocncn

"§~

~

c~

~1--""",l~<~b2Hs ~2Hs I

3,3' ~1by-"'mdhyU-"rbocy'D;" iodid.

9-MeTCC

TCC

~

IJJ

~(.)Q)

~

~

80;~::$~~~~~~"I' 0.02~ pmolg-'

0.15

60-

40. " \'.

20'' 3.7

01 400 . 500 . 600 . 700

Wavelength, nm

~ o.s]

i- I°Ê 0.6.QJ

:s I

8 0.41~ IQJt)ê 0.2

1gf.:

500 550 600 650 700 .-- 800Wavelengili, nm

Fluorescence emission of TCC for low and high dye concentrations.

750

(9-MeTCC) entrapped within the polymer chains of microcrys-talline cellulose.

Kubelka-Munk remission function curves are also shownin Figure 26, clearly showing for the second dye that themethyl group in the ninth position of the carbocyanine fa-vors the formation of sandwich aggregates (H aggregation),which absorb at energies higher than the monomer absorp-tion. Head to tail aggregates (J aggregation) are also formed,in the 5 to 15 JLmole of dye per gram of cellulose concentra-tion range, with an absorption peaking at about 610 nm. 3,3/-Diethylthiacarbocyanine no longer exhibits a clear J absorptionband, showing in this way that small structural differences mayplay a major role-t-oward aggregation.

The inftuence of concentration, and as a consequence, of ag-gregation in the ftuorescence emission intensity is presented inFigures 27 and 28. Fig.27.

PHOTONIC ANO ELECTRONIC SPECTROSCOPIES 295

:::s

cd~

:t;-~d~=-

~Qd~Qrn~

o=.-.

~

~~ 100 9-MeTCC

.

i- 80~~

5 60d-d) 40-u~8 20-~d)~o l-

.E[L.

1 ? 9-MeTCC, II

/-/

:::S~ --

~ 10.~= 8~

..s 6-~u= 4.~u

~ 2-

o~ 500 550 600 650 700

W avelen~ nmFig. 29. Laser induced fluorescence for TCC and 9.MeTCC as a funclion oflhe sarnDle concenlralion.

10pmol g-1"~ . -.....

.F=O.95:f;O.O3/"=",' !

()~ , . . . .

o 0.2 0.4 0.6 0.8 1~ (I-Re) . .f dye

Fig. 28. Fluorescence quantum yields for TCC and 9-MeTCC adsorbed onmicrocrystalline cellulose.

0.1

,\ 4.7"-

It is a well documented fact that lhe main path\\"ay for lheSI state of polymethinic dyes at room temperature is lhe trans-cis isomerization (see [16, 88] and references quoted therein).For monomeric cyanines and in lhe absence of steric hindrance,both fluorescence quantum yields and intersystem crossingquantum yields are usually very low. For TCC and 9-MeTCCentrapped within lhe polymer chains of microcrystalline cellu-Jose cP F becomes close to unity, evidence of lhe decrease of lhenonradiative pathways of deactivation.

In sharp contrast with lhe steady-state behavior described sofar, pulsed laser excitation of lhe same samples produces quitedifferent results: for concentrated samples and for high laser flu-ences (above about 10 m] per pulse and per square centimeter)a second emission appears in lhe case of TCC, as Figures 29and 30 show. This emission is sharp and appears for lower en-ergies rei ative to lhe monomer.

The cyanine that forms J aggregates does not exhibit thisnew emission or, at least, it is very much reduced. The newemission occurs in lhe nanosecond time range follo\ving laserexcitation and has its origin in lhe cyanine monomers only.

Figure 27 shows that, for low loadings ofthe dye (up to about0.5 ILmole of dye per gram of cellulose) the fluorescence emis-sion intensity increases with the increase of the light absorbedby the sample, <:ir better, with the fraction of the light which isabsorbed by the dye alone (the substrates also absorb at the ex-citation wavelength). In the 0.5 to 5 ILmole/g-1 concentrationrange a clear decrease in the fluorescence emission intensity isdetected, due to the fact that a fraction of the photons goes tothe aggregated forros of the dye from which no emission wasdetected.

Due to the overlap of the monomer fluorescence emis-sion and absorption from the aggregates, a quenching pro-cess may occur by a resonance interaction. However, in othercases [85, 86] where we determine the extinction coefficientsfor ground-state monomers and dimers, by simply taking intoaccount the fraction of the excitation radiation absorbed by themonomers (which are emissive species) and aggregates (fromwhich we could not detect any fluorescence emission), a goodsuperposition between the calculated and experimental fluores-cence intensity curve was obtained.

Both TCC and 9-MeTCC in ethanolic solution present a flu-orescence emission quantum yield (I/>F) of about 0.05. Whenthese molecules are adsorbed on microcrystalline, I/> F increases

to about 0.95 as Figure 28 shows.

296 BOTELHO DO REGO AND VIEIRA FERREIRA

. silical50 A

csilica 100 A. silica60 A

x silica40 A. silica25 A. cellulose

140

120

100

80

60 .D D.X401

I

k 08 ',p I 8 ~8 ..

i'; ,.n n.,

;:;~

i-

.~d-u

~uuri)

~ri:: ju, ...

- v.~ 0.4 0.6 0.8(l-R) fd~

Fig. 31. Fluorescence intensity (measured as lhe total area under lhe correctedemission spectra) as a function (I - R)fdve (see text) for sulforhodamine 101.

9-MeTCC::i~ 8000

100%:i-.- 6000cn~Q)

:s 4000Q)()5 2000()cn~g ~OO --- 000650 ---

f:i:: Wavelength, nmFig. 30. Laser induced fluorescence for TCC and 9-MeTCC as a funclion oflhe laser energy.

I.O,.mol g-1

550 700

only sulforhodamine 101 has a unitary cPF, and also that mois-ture has a strong quenching effect in alI cases [7, 82].

Sulforhodamine 101 and rhodamine 6G were further usedfor fluorescence quantum yield determination on different silicasurfaces [7]. Silicas with controlIed pore (22, 25, 40, 60, 100,150 A), and particle sizes were used. A systematic study of theinfluence of pore size and silica pretreatment on the emissionproperties of the two adsorbed probes has shown that sulforho-damine 101 fluorescence emission is severely affected by thosefactors, as Figure 31 shows, while rhodamine 6G is rather in-sensitive.

The combined information from photonic and electronictechniques was used in this particular study, which wilI be de-scribed in Section 5. These techniques proved to be comple-mentary in the study of the specific interactions of the tworhodamines with this substrate.

4.2. Flat Surfaces. HREELS aod XPS Studies 00 OrgaoicFilms

This new emission was also detected in other cyanines[16,88], such as 2,2'-cyanine [88b] and oxacyanine dyes [88a].Energy studies revealed a supralinear dependence on laser en-ergy.

From the above-mentioned studies with several cyaninedyes, it has been possible to establish the origin of this newemission. It occurs as a consequence of a two photon absorp-tion process, lhe first photon creating a photoisomer which marbe excited by the absorption of a second photon if the laser flu-ence is enough and the energy appropriated for the photoisomerto absorb. This second excited species then emits its own fluo-rescence [88a].

Picosecond laser excitation ("pump and probe" experiments)allowed us to determine lifetimes of the singlet excited states ofthese dyes in some cases [88c, 88e].

Rhodamines. The luminescence quantum yield study for mol-ecules adsorbed on surfaces, which started with rhodamine 6Gand 101 [8Ia], was later extended to zwitterionic rhodamines(sulforhodamine 101 and B) and to other nonrigid rhodamines(rhodamines B and 3B [7, 82]. Important conclusions arisingfrom those studies are that apart from rhodamines 6G and 101,

Organic films surfaces, and particularly polymeric ones, can bestudied bya great variety of analytical techniques and micro-scopies.

The association of several techniques to study the same sys-tem is always advantageous. For the specific case of surfaces,techniques like XPS, contact angle, diffuse reflectance infraredwith Fourier transform, surface arca and pore size measure-ments, scanning electron microscopy arrear frequently asso-ciated. This is the case for the study of carbon or glass fibers[105-109], activated carbons [110], and polymeric films ob-tained by plasma polymerization [111]. In Section 5, the spe-cific combination of diffuse reflectance spectroscopies in theregion of UV Nis and XPS will be emphasized.

In this section the application of HREELS to the study ofpolymeric surfaces will be reviewed with relevance for some ofits specific aspects: quantitative application of the vibrationalstudies and electronic studies.

PHOTONIC ANO ELECTRONIC SPECTROSCOPIES 297

4.2.1. Sample Requirements

Both HREELS and XPS work under ultra-high-vacuum con-ditions [29]. Therefore, organic films need to have low vaporpressures. Among organic films, polymeric ones always fulfillthis condition. However, these techniques induce a charge onthe surface (see Section 3) if the sample is insulating, whichis the case for a large part of organic polymers. With films ofmolecules of a long oligomer, the hexatriacontane (C36H74), ithas been possiblc. to record HREELS spectra without chargecompensation for up to thicknesses around 1000 A [112].Thicknesses lower than this value are obtainable by spin-coating, dipping, or deposition, followed by evaporation, froma polymer dilute solution. Provided that a conducting or a semi-conducting substrate is used, in principie, the film will naturallydischarge to the earth the charges trapped on its surface [49].The same occurs with Langmuir-Blodgett (LB) films [113].Experiments with perdeuterated polystyrene films on siliconwafers showed that a single dipping is not enough to get asubstrate entirely covered by the filn-. independent of the con-centration of the solution or the contact time between the sub-strate and the solution. A second dipping is necessary to assurea complete substrate covering [114]. This result is compatiblewith a mechanism with a first step consisting of a fast nucle-ation followed by slow growth of the small domains formedduring nucleation-as verified for the deposition of films ofpoly(o-methoxyaniline) (POMA) on glass substrates [115].During the second dipping, another fast nucleation seems tooccur on the uncovered substrate, the dissolution of the firstdomains being slower than the nucleation. With Langmuir-Blodgett films of perdeuterated stearic acid CD3(CD2) 16COOHon silanized surfaces (hydrophobic) studied by vibrationalHREELS, it was found that the film evaporated under ultra-high-vacuum conditions. However, the hydrogen bonding to thesubstrate, present when the substrate surface is polar, is enoughto keep the film on the substrate. As a result of this, many stud-ies were carried on LB films using XPS [116] and HREELS(one of the first works of HREELS on LB films can be foundin [113]).

c=

.eou

i-.~~-=-

1800 2800 3800

Energy Loss, crn-1

Fig.32. C-O and C-H stretching vibrations in polystyrene (symmetric andasymmetric modes are not resolved). Comparisoll of HREELS spectra fromtwo deuterated polystyrenes which differ only in lhe terminal groups: (I) oneend group is CH3CH(C6H5) lhe other one is C-H (normalline); (11) both end

groups are C-H (bold line).

Aliphatic ~ Aromatic

==

.eou

i-.~=..

E \~""" A~ !"\~::::::::::,:;;;:..,.;.;

, . . . . I . . . . I . . . , .

2500 3000 3500 4000

Energy Loss, crno\

Fig.33. C-H stretching vibrations in polystyrene (symmetric and asymmet-ric modes are nol resolved). 70% isolaclic film (normal line) and alaclic film

(bold line) show Ihal lhe firsl one is much richer in lhe aromalic componenlIhan lhe second one. Isolaclic film exposes preterenlially lhe phenyl groups aI

lhe surface whereas lhe alaclic one expo~es phenyl groups and chain segmenlsin comparable amounls. Speclra were normalized 10 lhe aromalic componenl.

4.2.2. Polystyrene

HREELS is usually applied in a qualitative way for lhe studyof adsorbed molecules. The theoretical frame for these studieswas established some 20 years ago [47]. However, its applíca-tíon to polymer surfaces is more recent [117, 118]. Polystyreneís a very common polymer and has been extensively studíedin optical absorption spectroscopy and ís often taken as a stan-dard sample. It was extensively used as a standard sample toevídence lhe ability of HREELS to study polymeríc surfaces,namely íts quantitatíve capabílítíes.

vibrational modes. End group segregation was also evidencedthrough the intensity fatia C-D/C-H for the stretching modesin completely deuterated polystyrenes, where just the chainending was hydrogenated (see Fig. 32).

One of the most striking pieces of evidence of the greatsensitivity of this technique to the extreme surface was ob-tained by the comparison of spectra obtained with two normalpolystyrene films made of isotactic and atactic polystyrenes,which is displayed in Figure 33.

Figure 32 shows that, even for a polystyrene with two hy-drogen atoms per chain, there is an intense peak in the re-gion around 3000 cm-l corresponding to the C-H stretchingmodes, aliphatic and aromatic. This is due to multiple losses,which are much more important than in infrared spectra. Blendsof completely deuterated and normal polystyrene, having thesame molecular weight, in several compositions were used toshow how quantitative the technique was and also to try to findrelative excitation cross sections for the vibrational modes cor-

4.2.2.1. Vibrational HREELS Studies

Conceming the vibrational region of the spectrum, the use ofselectively deuterated polystyrenes [119] allows a better sep-aration of the main-chain contributions and of pendant-group

298 BOTELHO DO REGO AND VIEIRA FERREIRA

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

O

CQ"'5ou.!:00~

responding to c-o and C-H bonds [120]. Films were pre-pared by dipping lhe substrate-films of gold 1000 A thickevaporated on glass plates-in CCl4 solutions.

Oenoting by x D and x H lhe molar fractions of deuterated andhydrogenated monomers, lhe following relation is expected,similar to lhe observed for solutions in CCI4:

AH ~ ~ (?Q)

Results clearly show that polystyrene segregates at the sur-face. The driving force for this segregation is the lower solidstate surface tension of PS (36 mJ/m2) compared to the PEO(44 mJ/m2) value [126]. This was confirmed by Thomas andO'Malley for several PS-PEO diblock [127] and PEO-PS-PEOtriblock [128] copolymers films cast from different solventsusing XPS. However, as already pointed out, techniques havedifferent surface sensitivities: with HREELS we measure thecomposition of the very first atomic layers (about 5 A) whereaswith XPS the extreme surface composition is always obtainedby indirect modes from angle resolved measurements and re-sults are therefore less reliable. The danger of confusing con-taminant C- H vibrations with those intrinsic to the system isavoided by using the whole spectrum from both the PEO and PSto curve fit the copolymer spectrum, since the vibrational spec-tra from both polymer~ are completely different from saturatedhydrocarbon, the usual contaminant. Figure 34 also shows thatthe amount of segregation at the surface, before annealing, ishighly sensitive to substrate, preparation method, and solvent.However, after annealing, only slight differences are noticed.

- '-/1

AD UD XD

UD and UH are the excitation crosssections and AD and AH

are the peak areas for the C-H and C- D stretching vibra-

tions, respectively. Surprising1y, instead of a straight line, a

curve with a plateau was obtained, suggesting the existence of

deuterated chain segregation at the surface. A thorough study

of the surface segregation in b1ends of perdeuterated and per-

hydrogenated polystyrenes as a function of relative molecular

weights, annealing time, and fi1m thickness was later carried

on using surface-enhanced Raman scattering [121]. The seg-

regation of deuterated chains at the surface when chains have

comparable length was a1so verified for a different substrate-

silicon wafer-and a different film preparation method-spin-

coating.Later, randomly deuterated polymers were used with the

same purpose and a linear relation between A D/A H and x D/X H

was finally obtained as expected [122]. From Eq. (33) the rel-

ative excitation cross-sections UH /UD were obtained as a func-

tion of the primary energy and were used to deduce the ratio

XD/XH for block copolymers of hydrogenated polystyrene ter-

minated by deuterated blocks at both ends as a function of the

electron primary energy [50].. Results were compatible with a

segregation of the end groups at the extreme surface (first 5-

10 A) followed by a layer with end group depletion and were

among the first experimental evidence of this surface segre-

gation predicted a few years before by de Gennes [123]. Fur-

ther experimental evidence of this end group segregation at

the surface carne from neutron reflectometry on the same sys-

tem [124].

Still with polystyrene (PS) but, this time, integrated in a di-

block copolymer with polyethyleneoxide (PEO), it has been

possible to quantitatively estimate the reI ative amounts of PS

monomers and PEO monomers at the extreme surface through a

method consisting of a previous normalization of spectra to the

background [125]. This norma1ization, previously suggested by

lbach and Mills [47], aIlows the use of HREELS spectra ac-

quired in different occasions with different films (having, for

instance, different roughnesses), one from pure PS film and the

other one from PEO film, to be combined, after being multi-

plied by factors whose sum is unitary, to yield the normalized

HREELS copolymer film [[25]. Annealing and substrate ef-

fects were studied by this method and results were obtained

for fi1ms spread on two type of substrates: si1icon wafers (keep-

i-ng;íheir native oxide film and with their surface si1anized) and

spin-coating and casting from di1uted solutions (0.1 gn) in two

different solvents (carbon tetrach10ride and tetrahydrofuran).

These results are presented in Figure 34.

4.2.2.2. Electronic HREELS Study

HREELS is mostly used as a vibrational spectr°s.copy and theextremely high resolution requirements are a consequence ofthat use. However, low energy electrons can be very useful tostudy the energy position and relative excitation probabilitiesof electronic states optically forbidden for spin or symmetryreasons.

Polystyrene is also, in this respect, a good material to test thecapabilities of this technique. Its molecular structure is made ofan aliphatic chain and phenyl pendant groups. lt has been ex-tensively studied in optical absorption spectroscopy and theirsinglet excited electronic states are well known: they are essen-tiallv the same detected in a mono~llh~tittltect henzene rI29].

Fig. 34. Surface PS fractions (symbols) for PS-PEO copolymer films pre-

pared by dipping (dip) and by spin coating (spinrfrom solutions in CCI4 andtetrahydrofuran (THF) on two different substrates: silicon wafer covered by itsnative oxide (denoted SiO2) and silanized silicon wafer (denoted Silane) fornonannealed (empty squares) and annealed films (fuI! squares). The gray line

represents copolymer stoichiometric composition. Stoichiometric compositioncorresponds to a PS fracti".. of 0.09. Horizontallines represent average valuesfor annealed (-) and non-annealed (-" - -) films.

PHOTONIC ANO ELECTRONIC SPECTROSCOPIES 299

Table V. Assignrnent of lhe Principal Features Present in HREELS Spectraof Polystyrene Filrns in Figure 35

Energy loss (eV) Assignrnent

TI (3 Blu)

SI (I B2u) + T2(3 Elu)

S2(1 Blu) + T3(3 B2u) + T4(3 E2g]

S3(IElu)

3.9

4.7

5.9

6.7

Energy Loss (eV)

Fig.35. HREELS s~tra for polystyren.: films prepared by dipping a siliconwafer in a polymer solution in CCI4 1.0 gL -I. Spectra were recorded at anincidence angle of 6()0 and an analysis direction of 50° (both directions arerelative to lhe normal to lhe film surface) in a loss region between 3 and 11 eV.

S~tra correspondto primary energies ranging from 4.7 (bottom) to 11.2 (top)in steps ofO.5 eV, which are here ser off for clarity of presentation.

For lhe assignment of lhe electronic energy losses, exper-imental and theoretical data for lhe electronic excitation ofsimple molecules having electronic structure comparable topolystyrene were used: benzene and toluene for lhe side groupsand polyethylene or one of its oligomers, such as hexatriacon-tane, for lhe chain. For incident energies lower than 7 eV, noelectronic excitations are induced by electrons in polyethylene-like compounds. Henceforth, in the range of incident energyfrom 3 to 7 eV, ali lhe comparisons were done with benzene andtoluene. Singlet excitations are well known from UV absorp-tion spectra of benzene and toluene liquid solutions [129, 131].Triplet and singlet states of benzene were also known by elec-tron impact experiments in lhe gas phase [132-135], adsorbedon metais [136], or included in xenon matrices or multilay-ers [137]. In Table V, lhe assignment ofelectronic excitations,clearly visible in spectra of Figure 35, is presented.

Comparison between HREELS spectra and optical absorp-tion spectra of polystyrene was revealed to be very useful inimproving lhe assignment of lhe peak centered at 4.7 e V. Infact, two hypotheses were plausible: to assign that peak to lheSI+- So or to lhe Tz +- So transitions. The comparison al-lowed us to establish that it should be a mixture of both with amajor triplet character. The comparison was also helpful for lheassignment of lhe peak at 5.9 eV. Ali lhe electronic transitionsare meant as occurring from lhe highest occupied molecular or-bital to different localized molecular electronic states coexistingin lhe gap ofthe insulator.

Features corresponding to fixed kinetic energies, located atfixed positions from lhe end of lhe spectrum, are assigned to lheaccumulation of electrons relaxed in high density of state levelsin lhe conduction band, above lhe vacuum levei following lhemultiple relaxation processes suffered by incident electrons.

AIso lhe analysis of relative intensities (as a function of pri-mary energy and geometric conditions) is important to assesslhe type of mechanism acting on lhe excitation. In polystyreneand for electronic excitations, that analysis showed that lhe be-havior of relative intensities as a function of primary energy istypical of resonant interactions. This is compatible with lhe factthat triplet excitation (optically forbidden for spin reasons) in-tensity is favored reIative to singlet excitation.

4.2.3. Oligothiophenes

Thiophene oligomers are compounds of rue form

-f-()t

In one of the first applications of low energy backscatteredelectron spectroscopy to polymers [118], polystyrene electroniclosses corresponding to electronic excited states, singlets andtriplets were studied with a low resolution spectrometer (fwhm,..,., 80 meV). Higher quality spectra and in a wider range of pri-mary energies and geometries were later published confirmingthat ali the excitations detected for energy losses lower than7 eV correspond to the electronic excited states in the side

groups [130].Polystyrene films were prepared by doubly dipping the

substrate-a silicon wafer previously cleaned by pure solventsand etched by hydrogen fluoride-into solutions of I gL -1 of

pure polystyrelK: in carbon tetrachloride (spectroscopic grade)and allowing tIK: solvent to evaporate. Thickness of the filmsthus obtained wz estimated to be of the arder of 100 A by elas-tic recoil diffusioo analysis (ERDA) measurements [122].

Figure 35 slMJWs HREELS spectra recorded at an incidenceangle of 60° and an analysis directions of 50° (both directionsare reI ative to tIK: normal to the film surface) in a loss regionbetween 3 and I1 eV. Spectra correspond to primary energiesranging from 4_7 (bottom) to 11.2 e V (top) by steps of 0.5 e V,which are here Kt off for clarity of presentation.

As in any ~plete spectrum (Ioss energy, ~E, from Oto E p) two kilKl5 of features are observed:

. those remai1i8g at a constant energy loss in the differentspectm, loC8d at fixed positions from the elastic peak,correspondi-:to the excitation of electronic individualstates and i-*ated in Figure 35 by verticallines;

. those corr~ing to fixed kinetic energies, located atfixed positicBS from the end of the spectrum (vacuum leveI)which are ~ associated to density of states maxima inthe conductil8band [112].

300 BOTELHO DO REGO AND VIEIRA FERREIRA

. ~where n is lhe number of thiophene rings, and lhe compoundname is usually abbreviated to nT.

Apart from being good models for lhe polythiophene, olig-othiophenes occupy, by themselves, a special place among ma-teriaIs suitable for opto-electronics, namely in lhe fabricationof light emitting devices and field-effect transistors [138]. Infact, these molecules constitute interesting materiaIs for elec-tronic devices as they can be easily used as blocks in lhe forma-tion of good quality, well-organized, thin organic films obtainedby evaporation. For this application, an accurate knowledgeof their electronic structures, excitation, and relaxation mecha-nisms is needed. Many theoretical and experimental studies onisolated molecules (or in dilute solution) [139] and also on LBfilms [140] were published. However, a very limited knowledgeexists on their solid-state electronic structure [141]. In fact, in-termolecular interactions mar piar a crucial role in lhe opticaland electronic properties of oligothiophenes in particular andof conjugated systems in general. For instance, until recently,lhe picture for lhe electronic structure of sexithiophene (6T) insolid phase was mainly based upon optical data [61, 141] andcould be summarized as follows:

'Bu '

agau~QJ=~

IACrystal gI

IAgMolecule

Fig. 36. Energy leveI diagram of lhe 11 Bu exciton band structure of a-sexithiophene (auapted from [142a]). Copyright 1998, American lnstitute ofPhvsics.

_/~~::=:::~~=~~:::;:~~;~~~~~':;J:::" \ ~

~'--~ -

'"

-02=.c..OQ

i-o~=~c- E

~\

\--

, I

o 2 4 6 8

Energy Loss, eV

Fig. 37. Complete spectra (from ô. E = O to ô. E = E p) for surfaces of quin-quethiophene films for primary energies ranging from Ep = 2.5 eV (bottom)

to E p = 7 eV (top) by steps of 0.5 eV. Spectra were recorded at an incidenceangle of 60° and an analysis directions of 30° (both directions are relative to

lhe normal to lhe fi 1m surface) and were seI off for clarity of presentation.

(i) Oligothiophene films present a large u excitonicbando

(ii) Transitions to the bottom of the band at 2.275 e Vareoptically forbidden by symmetry reasons.

(iii) The top of this band corresponds to the intense peakat 3.5 eV in the absorption spectrum obtained withUV radiation polarized in a direction parallel to themain molecular axis.

(iv) The 0-0 transition determines the position of theexciton band bottom. This was observed at18335 cm-1 (2.288 e V) in the single crystalabsorption spectrum recorded at low temperature[61,141].

Table VI. Main Features Positions (in eV and a Standard Deviation of:J::O.O2 eV) Assignment in HREELS Spectra of 4T. 5T. and 6T Films Surfaces

2BuBu crCompound Au

3.65

4.05

4.55

2.27

2.38

2.64

2.55

2.65

2.85

2.97

3.24

3.60

6T

5T

4T

For low energy losses, below 1 eV, vibrational structuresinduced by lhe incident electrons in interaction with lhe sur-face of lhe film arrear. For higher energy losses above 1 eV,a structured region, corresponding to electronic tosses Witl1 aí1edge above 2 eV, emerges. Different thresholds assigned to ex-citon and electronic gaps were extrapolated from this wide band[144-146] and are presented in Tab.le VI.

AlI lhe thresholds present in HREELS spectra are alsopresent in UV- Vis optical absorption spectra except lhe sec-ond threshold, named CT in lhe table; it is therefore assignedto electronic gaps, via charge transfer levels [145]. Similarthresholds were detected in spectra of similar samples usinga single and higher primary energy (15 eV) [147]. However,different assignments were proposed which were not compat-

More recently, the polarized absorption and fluorescencespectra obtained at 4.2 K from an oriented 6T single crystalallowed the assignment of the lowest singlet electronic transi-tion (Iocated at 18360 cm-l) to the au 4i- 1I A g transition (142]and a qualitative scheme of the 1I Bu molecular leveI Davydovsplitting could be drawn as represented in Figure 36 (142a].

Theoretical calculations using a quantum chemical model,which considers the total molecular wavefunctions for eachtransition, shows good agreement with the experimental find-ings for the energy and polarization of the optically allowedcrystallevels (142b].

In Figure 37 a series of complete HREELS spectra (from~E = O to ~E = Ep) is shown for a quinquethiophene filmsurface for primary energies rar:ging frum Ep = 2.5 eV to

Ep=7eV.For energy losses near zero, a very intense elastic peak ap-

pears, preselrtTng a very sharp angle distribution contained in alobe of fwhm around 12°. Such a high intensity and direction-ality associated with the elastic peak are characteristic of verywell organized and atomically fIat surfuces (143].

301PHOTONIC ANO ELECTRONIC SPECTROSCOPIES

advantageous. For a completely planar molecule, an equiva-lent electronic distribution around the two nitrogen atoms ex-ists, therefore a unique XPS N Is phófõefectron binding energyshould be detected. In contrast, in a slightly distorted conformerthe two nitrogen atoms are not equivalent since there is a breakin lhe conjugation and they should have different binding ener-gies. As a consequence, a broadening of the N Is peak shouldoccur. The XPS N Is peak is, therefore, expected to be a goodprobe for the planarity of this kind of dye molecules. More-over, it can also be used to probe the interaction between thedye molecules and the substrate. In fact, if a nitrogen atom isinvolved in strong hydrogen bonding, as an electron donor, theXPS N Is binding energy increases [151]. In the cases wherethe molecule also contained sulphur, the XPS S 2p region wasalso revealed to be interesting as we will show later.

ible with observations in optical absorption and fluorescence

spectra.The observation of strong secondary emissions led to the

evaluation of ionization potentials of 4.7 :!:: 0.5 eV for 5Tand 6T films. A large difference between optical absorptionand HREELS spectra was noticed [145]. The large differencebetween the interaction of electrons with the surface and ofphotons with the surface was related to crystalline effects onspectra: they strongly affect the optical absorption spectra shapewhereas they are completely absent in HREELS spectra. This iscertainly due to the fact that, for equal energies, the wavelengthassociated with electrons is much smaller than for photons and,therefore, electrons do not probe large delocalization effects.

In addition, HREELS vibrational spectra have shown thatmolecular orientation is not the same when graphite or goldsubstrates are used for the deposition of a-6T evaporatedfilms [148]. HREELS studies did corroborate reflection-absorp-tion infrared spectroscopy (RAIRS) analysis [149], confirm-ing that graphite substrates induce films where molecules larflat on the surface, whereas gold substrates generate a molec-ular orientation close to perpendicularity. Complete HREELSspectra recorded with incident electron energy between 2 and7 eV demonstrated the existence of two relaxation channels inthe electron surface interaction--electronic excitation and ion-ization. In fact, secondary electron emission originating fromionizations is always more efficient for escaping directions per-pendicular to the plane of the molecule. This result is associatedwith the 7r-orbital origin of this emission. Oppositely, the effi-ciency of the electronic excitation is higher for electrons emerg-in,g in directions parallel to the plane of the molecule.

5.1. Rhodamine Dye Covalently Bound to MicrocrystallineCellulose

5. COMBINEO STUDIES INVOLVING PHOTONICANO ELECTRONIC SPECTROSCOPIES

The combination "f several techniques to study a specificsystem, as already pointed, may provide a better insightand understanding of its constitution. In particular, photonicspectroscopies associated to X-ray photoelectron spectroscopyare being used more and more to study the adsorption of dyeson the most varied substrates [150].

In this section, special emphasis will be given to the study ofthe adsorption of dyes belonging to the cyanine and to the rho-damine families onto microporous substrates (cellulose,. silicas,and cyclodextrins).

AlI the dyes presented here have a common feature: thepresence of nitrogen atoms in the conjugated system. On theother hand. me substrates do not contain nitrogen. The organicsubstrates (cellulose and cyclodextrins) contain carbon, alsopresent in me dye, and oxygen. For the X-ray photoelectronspectroscopy studies, the XPS N ls region is therefore expectedto be particularly useful. On one hand, concerning the quantita-tive evaluation of the amount of dye adsorbed on the substratesby XPS, me N/O (more suitable than the N/C afie) and N/Si(for the adsorption on silicas) fatias should be considered. Onthe other band, for qualitative studies, the N ls region is again

The most important application of reactive dyes is lhe dyeingand printing of cellulose fibers and cellulose fiber-based mate-riais. These dyes form covalent bonds with lhe substrate thatis to be colored in lhe dyeing process [15]. The dye moleculecontains specific functional groups that can undergo additionor substitution reactions with lhe -OH, -SH, or -NH2 groupsthat are present in lhe textile fibers. According to lhe numberof reactive systems they contain, reactive dyes can be classi-fied as mono-, double-, and multiple-anchor dyes. RhodamineB isothiocyanate is a reactive monoanchor dye which can easilybe chemically bound to cellulose in lhe presence of a base [15].

The information on lhe nature of lhe interactions of rho-damine B with lhe natural polymer chains, namely to establishclear differences between physical and chemical adsorbed rho-damine B molecules within cellulose, is therefore needed.

Rhodamine B isothiocyanate was adsorbed onto microcrys-talline cellulose by two different methods: deposition fromethanolic and aqueous solutions followed by solvent evapo-ration (Type I) and also from aqueous solutions in equilib-rium with lhe powdered solid and following a dyeing protocol(Type 11). Figure 38 displays lhe scheme of lhe dyeing proce-dure used to bind rhodamine molecules to microcrystalline cel-lulose [15].

After carefully washing lhe above mentioned samples, theirground state absorption spectra presented clear differences asFigure 39 shows.

Also lhe fluorescence quantum yields (4>F) determined wereabout 0.40:!: 0.03 and 0.28:!: 0.03 for ethanol and water respec-tively (solvents which efficiently swell cellulose), when thesetwo solvents are used for Type I sample preparation. For dyedsamples, 4>F is only 0.10:!: 0.05. These values for 4>F can becompared with O. 70:!:0.03 obtained for rhodamine B entrappedinto lhe polymer chains of microcrystalline cellulose as Fig-ure 40 shows.

X-ray photoelectron spectroscopic studies of lhe same set ofsamples were essentially centered in lhe nitrogen as explainedabove.

BOTELHO DO REGO AND VIEIRA FERREIRA302

Rhodamine Bisothiocyanate

.coou

H,c"

~

.ã=

.Q...

-

.~a..

E

~:fj'c=.c..~

..'O

g~

'::\Y2

~\~

50

40

30

20 i ~;;~~:~. . . . .. A

. ..O. - I ~

3'

J--r;.'"10

,~1,,\~

'~420 470 520 ~.- 620

Wavelength I nm

Fig. 39. Remission function values of rhodamine B isothiocyanate onto mi-crocrystalline cellulose, normalized to lhe maximum of lhe absorption of lhedye. Curve (I) is a Type II sample with 0.035 /lmol/g-1 . Curve (2) is a Type Isample 0.030 Itmol/g-l. Curve (3) is a 0.035 /lmol/g-1 rhodamine B physi-cally adsorbed sample. The dashed curve is a diluted aqueous solution of lhe

dye containing only monomers.

~7n o 0.2 0.4 0.6 0.8 J

(l-R)!.,.

Fig. 40. Variation of lhe fjuorescence intensity of rhodamine B and rho-damine B isothiocyanate adsorbed onto microcrysta\line ce\lulose measuredas lhe total area under lhe corrected emission spectrum, I F. as a function of(1- R)jctye. Curve (I)-Type I samples ofrhodamine B prepared from ethano-lic solutions. Curve (2)-Type I samples of lhe reactive dye prepared fromethanol. Curve (3)-Type I samples of lhe reactive dye prepared from water.Curve (4)- Type 11 or dyed samples. Samples from curves 3 and 4 were repeat-edly washed after lhe initial solvent evaporation.

Figure 41 shows the XPS spectra for the N 1 s region ofadsorbed (Type I) and chemically bound (Type 11) samples ofrhodamine isothiocyanate as well as rhodamine B. Clear dif-ferences between them are observed: curve a is from a type 11sample and is centered around 399.5 eV; curves b and c showincreasing shifts toward larger binding energies. They corre-spond respectively to a type I sample (physically adsorbed dyefrom aqueous solutions), subsequently washed with water., andto rhodamine B type I sample from an ethanolic solution.

These results show that the positive charge density on the ni-trogen atom is smaller for dy~~amples when compared withthe adsorbed ones. This is compatible with nitrogen atoms thatdo not participate in the conjugated system. On the other hand,this fact is indicative of the nonplanarity of at least a fraction

of the dye molecules. This is due to the fact that the chemi-cal reaction of the isothiocyanate group occurs with cellulosedeep sites and since that molecular end group stays constrainedby the chemical bond, geometrical hindrance to the planarityof the whole molecule exists, putting the positive charge in

the xanthene moiety of the molecule instead of the nitrogenatoms. The sample's behavior in which the dye is physicallyadsorbed from an aqueous solution and subsequently washed is

also very interesting: it presents a shift toward higher binding

energies (BE).These data are consistent with the UV- Vis diffuse reflectance

absorption and fluorescence data: in aqueous solution and in

I

0,8

0,6

0,4

0,2

303PHOTONIC AND ELECTRONIC SPECTROSCOPIES

6000

d 4000cd

--

By studying simple cyanines adsorbed onto microcrystallinecellulose it was possibl,tiQ get some insight into the natureof the interactions of these cyanines with the polymer chains,namely in regard to the importance of hydrogen bonding of theprobe to the substrate, which is of major importance in the im-mobilization process, and in minimizing the nonradiative tran-sitions of the excited states.

2,2'-Cyanines are dyes with a general formula as follows:2000

.ro.J NI ICzHs CzHs

2,2 '-Cyanine

o405 400 395

Bíndíng Energy / eV

Fig.41. XPS spectra forthe N Is region ofsamples with loadings of3.5 JLmolof rhodamine B isothiocyanate per gram of cellulose. (a) Type 1i sample repeat-edly washed with water. (b) Type I sample from aqueous solution after washingwith water. (c) Rhodamine B physically adsorbed from ethanolic solutions and

subsequent removal of solvent by evaporation.

( Q""",,~X)I ~ 7" ~

~ +b 1#N

I I

C2H~ C2H~

2,2 '-Carbocyaninethe absence of dyeing conditions, dye molecules adsorb on themost favorable sites, being free to acquire a planar configura-tion. Nitrogen atoms (or, at least, one of them) become, in thisway, an active part of the conjugated system. With time evolu-tion, some of the dye molecules react with the substrate (oth-erwise, they would be removed by washing). However, in theabsence ofthe dyeing procedure, the reaction occurs slowly af-ter the equilibrium adsorption is established and has very littlear no effect on the final conformation of the dye molecule. InFigure 41 (curve c) the spectrum of a sample where pure rho-damine B is physically adsorbed from an ethanolic solution isalgo included for comparison purposes. It displays a slight shifttowards higher BE. This is assigned to the fact that, for this highloading of the dye, some aggregation on the external surfaceexists and, as a consequence, the amount of planar conformersincreases relatively to curve b case.

These findings indicate that rhodamine B has different con-formers in dyed samples as compared to samples where it isadsorbed. In the Corroer case the chemical bond, anchoring thedye to microcrystalline cellulose, leads to nonplanar conform-ers with smaller tPF and iF values [15]. In the latter case planarconformers predominate, with the consequent increase of both

lifetime and fluorescent quantum yield.

I, 1'-Diethyl-2,2'-cyanine iodide and 1, 1'-diethyl-2,2'-carbo-cyanine iodide were adsorbed onto microcrystalline celluloseby two different methods: by deposition from ethanolic solu-tions followed by solvent evaporation (Type I) and also fromethanolic solutions in equilibrium with the powdered solid(Type 11). Both methods provided the same fluorescencequan-tum yield ofthe adsorbed dyes in the 0.01 to 5.0 JLmoles of dye

per gram of cellulose concentration range.Ethanol swells cellulose and some dye molecules stay en-

trapped into the natural polymer chains and in close contactwith the substrate. The use of dichloromethane, a solvent thatdoes not swellmicrocrystalline cellulose, provides samples thatexhibit a smaller fluorescence quantum yield. This is consistentwith a larger degree of mobility (and also formation of non-planar and less emissive conformers) of the cyanines adsorbedon the surface of the solid substrate, while entrapment providesmore rigid, planar, and emissive fluorophors.

For 2,2'-cyanine the fluorescence quantum yields (cPF) de-

termined were about 0.08 whenever dichloromethane (solventwhich does not swell cellulose) was used for sample prepara-tion, while, with ethanol, cPF was approximately 0.30. Similarvalues were also obtained for 1,l'-diethyl-2,2'-carbocyanine asFigure 42 shows. These values are about three orders of mag-nitude higher than in solution, showing the importance of therigid dry matrix in reducing the nonradiative pathways of deac-tivation of the (n, n*) first excited singlet state of this cyanine.

The adsorption isotherms of 2,2'-cyanine on cellulose fromalcoholic and dichloromethane solutions using a Langmuir ad-sorption model [16] showed that cellulose surface area acces-sible to dye adsorption is about twice larger when ethanol isused for sample preparation instead of dichloromethane, con-firming the larger swelling power of cellulose by ethanol than

by dichloromethane [14].

5.2. 2,2'-Cyaoioes Adsorbed ooto MicrocrystallioeCellulose

The photochemistry and photophysical studies of cyanine dyesare an important research field owing to the economical impor-tance of these substances in color and black and white photogra-phy, in dye lasers, and as potential sensitizers for photodynamic

therapy.Polymethine cyanine photochernistry in fluid solution is

dominated by trans-cis isomerization as the maio SI state decayroute.

304 BOTELHO DO REGO AND VIEIRA FERREIRA

10

"a"-e'"

~"õnc.,

5

~J\;\ c

;~5

o

- J .j~: IVJ'Vv"Vt~Y ri V \1' \MIV". I/'f b

, '~I\~r"'ltV v a, ..

415 405 395 385

Binding Energy/eV

Fig.44. X-ray photoelectron spectra of lhe N Is region for three samples of2,2'-cyanine adsorbed on cellulose in three different conditions: (a) mechanicalmixture; (b) from dichloromethane solutions and allowing for complete solventevaporation; (c) from alcoholic solutions and allowing for complete solventevaporation. The concentration in ali samples is 5.0 /lmol of dye per gram ofcellulose.

o 0.5 1 1.5 2 2.5(Conc.) li] / (J.lmolg-11i]

Fig. 42. (a) Varialion of lhe inlensilY of fluorescence of 1.I'-dielhyl-2.2'-cyanine adsorbed onlo microcryslalline cellulose (sleady slale) mea.~ured as lhetotal area under lhe correcled emission speclrum. I F. as a funclion of lhe squarerool of lhe concenlralion of dye adsorbed onlo microcryslalline cellulose. Thesolvenls used for sample preparalion were: D-elhanol. 8-<1ichloromelhane.Full squares and open squares are used for Iype I and Iype II samples, respec-

lively. (b) Same data a.~ in (a) bul now for 1.I'-dielhyl-2.2'-carbocyanine.

15

those with other dye molecules, as occurs in H aggregates.Further support for this assignment was provided by lhe XPSanalysis of thick films ( 10 monolayers) of 2,2'-cyanine de-posited onto silicon wafers as well as mechanical mixtures of2,2'-cyanine and cellulose containing dye loadings equivalentto 1.0 and 5.0 ,umol/g. The peak at higher binding energies( 405 eV), usually associated with nitrogen bound to highlyelectronegative atoms [25], is here assigned to strongly hydro-gen bonded nitrogen. Its intensity decreases with increasing dyeloading. This fact suggests that at low loadings a high percent-age of molecules is intimately surrounded by substrate hydroxylgroups (entrapped) and algo that with increasing dye loadingthat percentage decreases.

At lower binding energies, another new component devel-ops as dye concentration increases, indicating that an increas-ing fraction of molecules has a larger electron density near lhenitrogen atoms (on both or at least on one of them). The assign-ment of this component is rather difficult. However, by com-parison with results from optical absorption in lhe visible regionwe can say that it exists whenever J aggregates are present; onepossibility therefore is to assign that component as arising fromthis kind of aggregate. An alternative explanation could be lheexistence of nonplanar conformers in which lhe loss of symme-try of lhe molecule causes lhe appearance of different electrondensities around lhe nitrogen atom: one strongly bound to lhesubstrate and lhe other dangling free above lhe surface.

In ali samples where lhe cyanine was adsorbed to cellulosefrom dichloromethane solution, lhe same effects as for ethano-lic samples were observed (Fig. 44b). However, some quanti-tative differences exist: (i) lhe low binding energy componentappears for lower dye loadings. This fact algo supports lhe as-signment of this component to lhe existence of J aggregates,which begins at lower concentrations in lhe dichloromethanecase. (ii) The high binding energy component is less intense andcentered at energies slightly lower than in lhe ethanolic case.

A third set of samples prepared by mt:chanically mixing cya-nine with cellulose was studied. In this case, no intimate con-

-'2 10='.cia"E--~ 5c.,

.5

,.~c

o415 405 395 385

Binding Energy/eV

Fig. 43. X-ray photoelectron spectra of lhe N 15 region for three samplesof 2,2'-cyanine adsorbed on cellulose from alcoholic solutions and allowingfor lhe complete solvent evaporation. Curve (a) 0.1, curve (b) 1.0, curve (c)10.0 jlmol of dye per gram of cellulose.

Figure 43 shows lhe XPS spectra for lhe N Is region ofseveral samples of 2,2'-cyanine adsorbed onto microcrystaUinecelIulose from an ethanolic solution after total solvent evapo-ration. We can clearly see that lhe N Is peak has a main com-ponent centered at a binding energy õH99.9:t 0.1 eV for alIconcentrations under study.

This component is assigned to lhe nitrogen in a moleculewith a planar conformation free from interactions aDart from

PHOTONIC ANO ELECTRONIC SPECTROSCOPIES 305

5.3. lnclusion Complexes or fJ-Cyclodextrin and Cyanine

Dyes

Cyclodextrins are compounds with a torus-shaped hydropho-bic cavity, and with a limited number of D-( +)- glucopyra-nose units joined by a-(1,4) linkages. fJ-Cyclodextrin (fJ-CD)in particular is composed of seven of these units while a-CDand y-CD have six and eight, respectively. Figure 45 shows lheshape and internal dimensions of these olygosaccharides.

They constitute therefore a substrate chemically similar tocellulose but provide a better characterized environment fromlhe point of view of possible sites for adsorption. It should be agood medium to check for lhe correctness of lhe interpretationabout lhe dye/cellulose interaction.

5.3.1. 1,1'-Cyanine and 1,1'-Carbocyanine

XPS studies on lhe N 1 s region show that lhe behavior ofl.l'-cyanine dyes adsorbed onto cellulose in low concentration(0.1 JLg/g of celIulose) is similar to lhe behavior of lhe samedyes included in {J-cyclodextrin. Figure 46 displays lhe first ex-

ample.

~~'2"

.etO

-à'0;c~.5

~

tact between the dye and the substrate exists: dye is mainly inthe form of tiny crystals. We can see (Fig. 44, curve a) thatthe N Is peak becomes narrower, but is centered at a higherbinding energy (~400.4 e V), than the central component of thesamples prepared from the dissolved dye. This fact is attributedto a change of the electron density around the nitrogen atomswhen the dye molecule is included within a microcrystal.

XPS N ls spectra for 2,2'-carbocyanine also for increas-ing loadings and for samples prepared by three differentmethods (adsorption from ethanolic solution, adsorption fromdichloromethane, and mechanical mixture) are qualitatively notvery different from the ones for 2,2'-cyanine in Figures 43and 44. However, a few differences were noticed. They canbe summarized as follows: (a) with increasing dye loading,the high energy component relative intensity increases; (b) thecomparison of samples where the dye was adsorbed from twodifferent solvents and also mechanically mixed with celluloseshowed that only "ethanolic" and "dichloromethane" samplesexhibited a strong interaction of the nitrogen atom with the sub-strate which obviously could not exist in the mechanical mix-ture; (c) 2,2'-carbocyanine nitrogen atoms are not equivalent asthey are in 2,2'-cyanine.

Quantitative XPS studies performed through the evalu-ation of the atomic nitrogen/oxygen (N/O) ratio also al-lowed us to establish some important differences between 2,2'-carbocyanine and 2,2'-cyanine: (i) for 2,2'-cyanine moleculesadsorbed onto cellulose from both solvents, lhe N/O ratio wasinvariant, within the experimental errar. However, it increasedwith dye loading for lhe mechanical mixture. These resultsstrongly suggest that 2,2'-cyanine molecules occupy deep sitesin the cellulose, most probably pores, when they are adsorbedfrom solution. In lhe case of the mechanical mixture, the ab-sence of a close dye-substrate contact precludes that type ofoccupancy. (ii) In lhe case of 2,2'-carbocyanine there is a dif-ferent behavior for ethanolic and dichloromethane samples:in the ethanolic case the invariance of the N/O ratio breaksdown for dte highest dye loading (10 J,l,mol/g) whereas fordichloromethane samples the invariance disappears or it breaksdown for much lower dye loadings. This points to a generallygreater difficulty for this molecule in occupying deep sites in thesubstrate as compared with lhe smaller 2,2' -cyanine molecule.The difference between the two kinds of samples is consis-tent with ~ fact that ethanol is a better swelling solvent thandichloroI1K:lbane.

X-ray pbotoelectron spectroscopic studies present evidencefor hydrogen bonding of 2,2' -cyanine to cellulose for low load-ings and agpegates formation for higher loadings in the ethanoland dichkxumethane cases. This hydrogen bonding is assignedto the situ.-.ion where the dye molecule is entrapped into cel-lulose cb3im. On the other hand, for 2,2'-carbocyanine, evi-dence exim for an increase of hydrogen bonding with dye load-ing. This result together with ground-state diffuse reflectanceabsorption and luminescence results is compatible with dyemolecules king firmly bound to lhe substrate by one of thenitrogen -.os, lhe other one dangling over the substrate's freesurface.

\~v. ' . , , I . , . , I ' , , , I I

412 407 402 397 392

Binding Energy I eV

Fig. 46. N Is region X-ray photoelectron spectrum: comparison of 2,2'-cyanine adsorbed on cellulose from an alcoholic solution, with a concentra-tion ofO.1 /lg per gram of cellulose (thin line), with 2,2'-cyanine included in

f3-cyclodextrin (I : 50, moi: moi) (bold line).

306 BOTELHO DO REGO AND VIEIRA FERREIRA

weaker than it is with nitrogen in 2,2'-cyanine, as had been al-ready evidenced with the substrate cellulose.

A common point with the 2,2'-cyanine case is the presenceof a low binding energy component ('"'"'397 e V) when celluloseis the substrate and its absence when the substrate is the ..8-

cyclodextrin.

5.3.2. 3,3'-Ethylthiacarbocyanine

AIso lhe inclusion of 3,3'-ethylthiacarbocyanine (TCC) in ,8-cyclodextrin shows a greater similitude with lhe cellulose lowloading sample, as displayed in Figure 48.

However, in lhe case of TCC, similar to lhe case of lhe car-bocyanine, lhe N Is region does not show evidence for a largeinteraction between lhe nitrogen atoms and lhe substrate (com-parable to lhe one detected in lhe case of lhe 2,2'-cyanine). Inthis case, lhe most interesting XPS region is not lhe nitrogenregion but lhe XPS sulphur region. Figure 49 displays lhe XPSS 2p region for falir different samples: TCC adsorbed on cellu-Jose with two different concentrations (0.1 and 1 ,ug ver gram

tJ.2~

-fJ

?;;-.~c:P..s

1

Figure 46 shows that the N Is region of the XPS spectrumof 2,2'-cyanine included in {J-cyclodextrin (1 : 50) is very sim-ilar to the same spectrum for 2,2'-cyanine adsorbed on cellu-Jose from an alcoholic solution with a very low concentration-0.1 JLg per gram of cellulose. As seen in Section 5.2, one of thepeaks, located at 399.9:J:: 0.1 eV, is assigned to the nitrogen ina molecule with a planar configuration free from interactions,apart from those with other dye molecules, as occurs in H ag-gregates. Theother one, at higher binding energies (---405 e V),is assigned to strongly hydrogen bonded nitrogen and its inten-sity decreases for lower loadings. When compared to the low-est 10aded sample, the 2,2'-cyanine included in {J-cyclodextrin(I : 50) presents also these two components but their relative in-tensity is different: alI the spectrum is displaced toward higher

binding energies and the component at higher binding energyhas a larger intensity reI ative to the component at lower BE.These two facts denote a larger interaction between nitrogenand the substrate's hydroxyl groups when the substrate is {J-cyclodextrin than when it is cellulose.

Another striking difference is the existence of a small com-ponent at lower binding energy in the case of the adsorption oncellulose which is absent when the substrate is {J-cyclodextrin.This reinforces the assignment of this component to nitrogenbelonging to molecules in J aggregates. Figure 47 displays thesame comparison as Figure 46 but the adsorbed dye is the 2,2'-

carbocyanine.Figure 47 shows that the N Is region of the XPS spec-

trum of 2,2'-carbocyanine included in {J-cyclodextrin (I : 50)is very similar to the same spectrum for 2,2'-carbocyanine ad-sorbed on cellulose from an alcoholic solution with a very lowconcentration-O.1 JLg per gram of cellulose. Also in this casethe greater similarity between the two cases arises for the leastconcentrated sample adsorbed on cellulose. (We remember that,contrary to the case of 2,2'-cyanine, the high binding energycomponent increases with dye loading on cellulose: see Sec-tion 5.2). Thi~ means that the interaction of the substrate's

hydroxyl groups with nitrogen in 2,2'-carbocyanine is much

rt.I

-p<I

v

. I . , . . I . . , . .410 405 400 395

Binding Energy / eV

Fig. 48. N Is region X-ray photoelectron spectrum: comparison of TCC ad-sorbed on cellulose from an alcoholic solution, with a concentration of O. I JLg

per gram of cellulose (thin line) and TCC included in f3-cyclodextrin (I: 50,moi: moI) (bold line).

~~.2::I

~

à.~~Q)

..s

~'2;:I

-fJ--

::-';ncG)C-

-f'\f"'\

~

\;~'V-i ' , , , I . . . . I . . . . ,

410 405 400 395

Binding Energy / eV

Fig. 47. N Is region X-ray photoelectron spectrum: comparisoli"Uí'2.2'-carbocyanine adsorbed on cellulose from an alcoholic solution. with a con-centration of 0.1 Jl..g per gram of cellulose (thin líne), with 2.2'-carbocyanineincluded in IJ-cyclodextrín (I : 50, moI: moi) (bold tine).

,--'--' . . I . , , . I ' , , , I . -, , . I180 175 170 165 160

BindingEnergy/eV

Fig. 49. XPS S 2p region for TCC. From bottom to top: included in ,B-CDI: 100, moi: moI (8), adsorbed on ceIlulose (0.1 J1.g'g of ceIlulose) (o), ad-sorbed on ceIlulose (I J1.g'g of ceIlulose) (--), and mechanicaIly mixed withceIlulose (+).

PHOTONIC ANO ELECTRONIC SPECTROSCOPIES 307

of cellulose), a mechanical mixture of TCC with cellulose, andTCC included in ,B-CD with lhe concentration 1 : 100.

The XPS S 2 p region for TCC mechanically mixed with cel-lulose shows a single component, as expected, since no intimatecontact exists between lhe dye molecule and lhe substrate. Thepeak is asymmetric since it is a doublel (S 2P3/2 and S 2Pl/2)with an energy splitting of about 1 eV [26]. For TCC adsorbedon cellulose at least two components are exhibited: one at aboutlhe same energy as lhe one detected for lhe mechanical mixtureand another one at higher binding energies usually assigned tosulphur bound to highly electronegative atoms [25]. Decreasinglhe dye loading, lhe low BE component almost disappears andonly lhe high BE component is left. Similar to lhe observationmade on lhe XPS nitrogen region, these results point to a stronginteraction between lhe sulphur atom and lhe substrate, lhe sul-phur atom acting as a strong electron donor. An obvious can-didate is lhe hydrogen bonding interaction between hydroxylgroups in lhe substrate and sulphur lone pairs. The same kindof interaction exists also in lhe inclusion complex ,B-CD:TCC.Thisconfirms that this interaction occurs only when moleculesare deeply entrapped into lhe cellulose chains.

In parallel, <PF for TCC entrapped within lhe cellulose poly-mer chains is about 0.95 (see Figure 28). For TCC included in,B-CD, <PF is much smaller, around 0.20. This shows that molec-ular mobility is larger in lhe latter case.

ronments or lhe study of interactions of probes with specificsurfaces.

-. Sulforhodamine 101 and Rhodamine 6G were chosen as

probes for lhe latter study [7]. They were adsorbed auto sili-cas with different pore sizes ranging from 22 to 150 A. Groundstate diffuse reflectance absorption spectra revealed lhe forrna-tion of different forms of adsorbed sulforhod:lmine 101 depend-ing on concentration and on lhe pore size of lhe silica. For lowloadings (0.00 I to about 0.025 j1mol of dye per gram of silica)lhe absorption spectra are broad, hypsochromically shifted inrelation to lhe monomer spectra, and quite different from lheethanolic solution spectra. For high loadings (0.050 to about0.20 j1mol/g) they are similar to lhe solution spectra with asmall shift of about 7 fim. For rhodamine 6G spectra are muchmore of lhe "solution type" in lhe entire range of concentrationsunder study (0.001 to about 0.20 j1mol/g). (See Figure 50.)

The "weighed" fluorescence quantum yields (Li fi4>F;) de-terrnined for sulforhodamine 101 were 0.10:i:0.03, 0.35:i:0.05,and 0.50 :i: 0.10 for low loadings and for 25, 60, and 150 A sil-icas, respectively. For high loadings Li /;4>F; = 0.70:i: 0.10,as Figure 31 in Section 4.1 shows. These values for Li /;4>Fican be compared with a value of about 0.70 obtained for rho-damine 6G in ali silicas (Iow loadings) and a unitary 4>F for

high loadings.XPS experiments were also performed with sulforhodamine

101 and rhodamine 6G on silicas. The dye loadings under studywere 0.002, 0.025, 0.10, and 1.0 j1mol/g auto silica gel with25 A pores, and 0.002, 0.04, 0.10, and 1.0 j1moVg auto silicagel having pores of 150 A diameter. Within lhe êrror associ-ated with XPS quantifications, it was shown that lhe atomic ra-tio N/Si was constant (of lhe arder of 10-:-3) and independentof dye loading (although a slight variation is detected for sul-

5.4. Sulforhodamine 101 and Rhodamine 6G Adsorbed onDifferent Pore Size Silicas

Powdered solids with controlled porosity and particle size areprivileged media for the study of reactions in restricted envi-

Rhodamine 6G

0.5

:J 1"c=.d; 0.5

~~ O I , .

400 450 500 550 600 650o

400 450 500 550 600 650

Sulforhodamine 101

10R BOTELHO 00 REGO ANO VIEIRA FERREIRA

o ~.l ~

150

g 5Q

~ 4

~Z ~ A .

to

6

.

2nMI

-v.vv. 0.01 0.1 I 10

Dye Loading I ~mol g-1

Fig.51. Atomic fatia !':l/Si for sulforhodamine 101 (full symbols) and rho-damine 6G (open symbols) adsorbed on 25 or 150 A pore silicas as a function

of dye concentration. Straight lines in lhe plot represent average values.

forhodamine adsorbed on 25 A pore silica). That fatia was alsolarger in lhe silica with lhe 25 A pore than in lhe silica with lhe150 A pore (by a factor around 1.5) as displayed in Figure 51.

These findings are compatible with lhe following picture forthese systems: lhe dye must be mainly adsorbed near lhe ex-treme surface of lhe pores. Moreover lhe adsorption must occuressentially in depth (filling lhe pores) and not with an increasein lhe number of lhe occupied pores with increasing loadings.Dye molecules penetrate deeper into lhe pores in lhe larger poresize silica, lhe constraints concerning penetration being moreimportant in lhe 25 A pore silica than in lhe 150 A pore silica.

To try to understand why lhe adsorption process occurs fol-lowing this pattern, it is important to make an estimation of lhespecific number of pores, n p, which allows computation of lheavailable number of pores per adsorbed molecule.

The experimental available data for powdered microporousmateriaIs are lhe average particle size L, obtained by sieving,lhe total specific area, Asp, and lhe average pore radius, rp, orpore volume, vp, obtained by gaseous adsorption isotherms orspecific porosimetry methods [152]. These three parameters arenot independent and are usually related by

dv2 IJ (30)

Fig. 52. Schematic representation of cubic particles having diameter L andnonintersecting cylindrical capillaries of radius rp.

Table VII. Typical Experimental ?arameters for Two of lhe Silicas Used inThi!' Work Together with the Specific Number of Pores Computed from

Modell

Pore size

(A)

Particle size Specific area

(jlm) (m2/g)

Pore volume

(cm3/g)

Specific number of

pores (J1moVg)

3.38 x

1.69 x

2.82 x

1.41 x

25 75

150

75

150

600 0.38

0.38

1.13

1.13

150 300

is given by equation

2npp~ (31)np =mp mp

where n pp is the number of nonended cylindrical pores per par-ticle, n ps is the number of pores emerging at the extreme sur-face, and m p is the mass of a particle. On the other hand, thespecific total surface, Asp, neglecting the contribution of the ex-treme surface, is

=

dA = rksp

derived with lhe aid of lhe conventional Kelvin formula,where rk is lhe so-called "Kelvin radius" which coincides withlhe pore radius, for nonintersecting pores of circular cross-section [152], provided L » r p.

For a rough estimation of np we will use two very simplemodels. In lhe first one, here called model I and schematicallyrepresented in Figure 52, pores are considered to be a systemof nonintersecting cylindrical capillaries and lhe particles arecubes, with an edg~ L. Given lhe absence of shared volumes,it will give lhe minimum number of pores for a given porevolume. Moreover, this model is compatible with lhe fact thathighly porous silica present self-similar surfaces with a fractaldimensionality approaching 3 [152, 153]. --

For a powdered solid containing particles with that shape,and assuming that ali lhe pores are available even in the regionsof contact between particles, lhe specific number of pores, n n.

10-3

10-3

10-4

10-4

~OQPHOTONIC ANO ELECTRONIC SPECTROSCOPIES

Table VIII. Estimated Particle Mass and Values for lhe Specific Number of

Pores for Two of lhe Silicas Used in Ihis Work. Assuming Model 28000

Particle mass (/lg) 6000Assuming Specific number of

dexp pores (Ilmol/g)

Pore size Particle size

(Â.) (/Lm)

Assuming

dSiO2 = 2.2 g/cm'4000

34.1 x 10-3

17.0 x 10-3

1.79 x 10-3

0.90 x 10-3

75

150

75

150

0.51

4.1

0.27

2.1

0.30

2.4

0.16

1.3

ri}

.~==

..c~=-c.~ri}=~

=~

25

2000150

o

there is an upper unreachable limit: lhe entire extemal surfaceoccupied by tangent pores of circular cross-section in a close-compact arrangement, here called model 2. In this case,

6L2 6L2 nA.)

6000

4000"p.' - ~ and np = ;~2~ ,_.,

In this model, n p depends on the inverse of the particle mass.The mass of the particle, m p, was estimated by two methods:

(I) Assume that the particle-a cube of edge L-is madeof pores and massive amorphous silica with a massdensity dSilica = 2.2 g/cm3 [154]. In this case, thespecific volume is given by I /dSilica + vp. Therefore,the particle mass is given by

Particle volume L 3

"

2000

o390410 405 400 395

Binding Energy I e V

mp= -Specific volume l/dsilica + vp

(11) Use "experimental" mass density, dexp, valuesdetermined by weighting a given volume of drypowder (dexp = 0.72 g/cm3 for the 25 A silica anddexp = 0.38 g/cm3 for the 150 A silica). In this case,the particle mass is simpIy

-

mp = dexpL3

These two values are clearly upper and lower limits: it is un-expected that lhe walls between pores do have lhe same densityas lhe massive silica or larger; on lhe other hand, lhe experi-mental value considers lhe volume between particles as particlevolume, thus underestimating lhe particle mass. The real valueshould, then, lie between these two extreme values. In orderto be sure to compute an upper limit for np, displayed in Ta-ble VIII, lhe lower mass value was taken into account.

As a conclusion, we can say that lhe real number of poresshould be between lhe values presented in Tables VII and VIII.

This estimation of lhe boundaries for lhe number of poresemerging at lhe extreme surface shows that alI lhe dye loadingsused in this study are larger than lhe number of pores except,perhaps, for lhe lowest loading in lhe case of lhe silica havinga 25 A pore diameter. This means that conclusions about poreoccupancy by dye molecules based on XPS results are reliable.

The qualitative analysis of lhe N ls peak (see Figure 53)does not reveal any measurable change in shape and position,

suggesting that the interaction between dye molecules and thesubstrate leading to the dye malecule deformatian, and the con-sequent Jr delocalization extent decrease, does nat involve thenitrogen atom as much as it did in the case of the rhodaminedyes covalently described in Section 5.1.

However, and since optical studies of the same system un-doubtedly point to the existence of distorted malecules, the in-variance of the XPS N I s peak may be due to the fact that XPSonly "sees" surface layers of about 100 A width. Therefore, wemainly see probe molecules poorly interacting with the silicasurface, and the strongest interactions occur deeply inside the

pore.These studies indicate that sulforhodamine 101 forms non-

planar conformers in small pore size silicas as compared talarge pore silica samples where the amount of conformers beingformed is reduced. Rhodamine 60 samples exhibit very littleconformer formation but their <P F are still slightly dependent onpore size. Both rhodamines exhibit smaller ftuorescence quan-tum yields when compared to thé case ofadsorption onto mi-crocrystalline cellulose, this effect being more relevant in the

sulforhodamine 101 case.

Fig. 53. XPS spectra for lhe N I.~ region of samples of lhe two rhodaminesadsorbed on 25 A silica. (a) Sulforhodamine 101 (--) 0.002 ILmol/g. (-)0.025 ILmol/g. and (---) 0.1 ILmol/g. (b) Rhodamine 6G (--) 0.002 J1mol/g.(-) 0.020 ILmol/g. and (- - -) 0.1 ILmol/g. (Ali samples are Type I.)

310 BOTELHO DO REGO AND VIEIRA FERREIRA

6. CONCLUSIONS

X -ray photoelectron spectroscopy and UV Nis absorption andluminescence studies have proved to be complementary tech-niques in the study of dyes in several environments: when ad-sorbed onto or bound to a natural polymer, microcrystalline cel-lulose, when forming inclusion complexes with cyclodextrins,and also when interacting with silicas with different pore sizes.

UV Nis ground-state diffuse-reflectance absorption and lu-minescence studies showed in most cases a huge increase offluorescence emission quantum yield when compared to solu-tion studies for several dyes. when they are entrapped into mi-crocrystalline cellulose, due to the reduction of mobility andformation af planar and emissive conformers and to the amountand type of aggregates which are formed. The emission dependson the solvent used to adsorb the probes. In some cases a pho-toisomer fluorescence emission was detected following a twophoton absorption process, lhe first one used for lhe formationof the photoisomer and the second one creating the excited stateresponsible for this special second emission from the cyanineswhich coexists with lhe monomer emission.

A complementary view is given by X-ray photoelectronstudies which provide evidence for hydrogen bond formationas well as preferential attachment by one or two nitrogen atomsper cyanine molecule to the hydroxyl groups in the substrates.In the thiacarbocyanine case, hydrogen bond formation was de-tected for microcrystalline cellulose and f3-cyclodextrin caseswhich involved the sulphur atom of the cyanine.

The combined use of diffuse reflectance techniques and XPSallowed us to establish of a clear picture of lhe specific inter-actions of dyes and substrates in several cases and enabled amuch deeper insight than the sum of information provided byeach techniQue alone.

Acknowledgments

We thank Drs. M. Rei Vilar, A. S. Oliveira, and O. Pellegrinoand Mrs. M. J. Lemos for their collaboration in some of lhework described here. We also thank Dr. A. S. Oliveira for a crit-ical reading of lhe manuscript and Dr. O. Pellegrino and Mrs.M. J. Lemos for helping us with some of lhe drawings.

REFERENCES

4. C. Bohne, R. W. Redmond, and J. C. Scaiano. ia ~IJJistry in Or-

ganized and constrained Media" (V. RamamlBlby.Y.).Chap. 3, p. 79.VCH Publishers, New York, 1991.

5. L. J. Johnston, in "Photochemistry in Micl";~us Systems"(K. Kalyanasundaram, Ed.), Chap. 8, p. 359. A&:.--: Press, afiando,

1987; P. V. Kamat, Chem. Rev. 93,267 (1993).6. E. Desimoni and P. G. Zambonin, in "Surface ~zation of Ad-

vanced Polymers" (L. Sabbalini and P. G. ~ &Is.), p. 5. VCH,Weinheim, 1993.

7. L. F. Vieira Ferreira, M. J. Lemos, M. J. Reis, - A- M. Botelho doRego, Langmuir 16,5673 (2000).

8a. P. Kubelka and F. Munk, Z. Tech. Phys. 12,593 (1931).8b. W. Wendlandt and H. G. Hecht, in "Reftecta/K% ~opy," Wiley,

New York, 1966.

8c. G. Kortum, "Refleclance Speclroscopy. Principk:5. Ioktods, Applica-tions," Springer-Verlag, Berlin, 1969.

9. P. M. Epperson, J. V. Sweedler, R. B. Bilhom. G. R- Sims, and M. B.Denton, Anal. Chem. 60, 327A (1988); J. V. S~. R. B. Bilhom,P. M. Epperson, G. R. Sims. and M. Bonner ~ Anal. Chem. 60,

283A (1988).

lOa. F. Wilkinson and C. J. Willsher, Chem. Phys.lLII- 101.272 (1984).10b. R. W. Kessler and F. Wilkinson, J. Chem. SO(.. FanDI,- Trans. 177,309

(1981).10c. F. Wilkinson, J. Chem. Soco Faraday Trans. 1182. ~ (1986).IOdo R. W. Kessler, G. Krabichler, S. Schaller, S. UhI. D. <kIkrug, W. P. Ha-

gan, J. Hyslop, and F. Wilkinson, Opto Acta 30.1099(1983).10e. F. Winkinson and G. P. Kelly, in "Photochemistry 00 Solid Surfaces"

(M. Anpo and T. Matsuara, Eds.), p. 31. Elsevier. AmsIerdam, 1989.

10f F. Wilkinson, Tetrahedron 43, 1197 (1987).lOgo F. Wilkinson and D. R. Worrall, Proc. Indian Acud. Sci. (Chem. Sci.)

104,287 (1992).11. G. P. Kelly, P. A. Leicester, F. Wilkinson, D. R. Wonall, L. F. Vieira

Ferreira, R. Chittock, and W. Toner, Spectrochim- Acla A 46, 975 (1990)./2. D. Oelkrug, W. Honnen, C. J. Willsher, and F. Wilkinson. J. Chem. SOCo

Faraday Trans. 1 81, 2081 (1987).13a. P. P. Levin, L. F. Vieira Ferreira, and S. M. B. CcNa. Chem. Phys. Lerr.

173,277 (1990).13b. P. P. Levin, L. F. Vieira Ferreira, S. M. B. CcNa. anil I. V. Katalnikov,

Chem. Phys. Lerr. 193,461 (1992).13c. P. P. Levin, L. F. Vieira Ferreira, and S. M. B. CcNa. Langmuir 9, 1001

(1993).13d. P. P. Levin, S. M. B. CosIa, and L. F. Vieira Ferreira,J. Phys. Chem. 100,

15171 (1996).14. L. F. Vieira Ferreira, J. C. Nello-Ferreira, I. V. Khmelinskii, A. R. Gar-

cia, and S. M. B. Costa, Langmuir 11,231 (1995).15. L. F. Vieira Ferreira, P. V. Cabral, P. Almeida, A. S. Oliveira, M. J. Reis,

and A. M. Botelho do Rego, Macromolecules 30. 3936 (1998).

16. A. M. Bolelho do Rego. L. Penedo Pereira, M. J. Reis, A. S. Oliveira,

and L. F. Vieira Ferreira, Langmuir 13,6787 (1997).17a. L. M. IIharco, A. R. Garcia, J. Lopes da Silva. and L. F. Vieira Ferreira,

Langmuir 13,4126 (1997).17b. L. M. Ilharco, A. R. Garcia, J. Lopes da Silva. M. J. Lemos, and L. F.

Vieira Ferreira, Langmuir 13,3787 (1997).

17c. L. M. Ilharco, Química 69, 34 (1998).18. N. Câmara de Lucas, J. C. Netto-Ferreira, J. Andros, J. Lusztyk, B. D.

D. Wagner, and J. C. Scaiano, Tetrahedron Len. 38,5147 (1997).

19. "Perspectives in Modero Chemical Spectroscopy" (D. L. Andrew, Ed.),

Springer- Verlag, Berlin/Heideiberg, 19~O.20. See, for instance, XSAM 800, "Operators Handbook," Vol. 2. Kratos

Analytical, Manchesler, 1990.21. K. Siegt.ahn, "Electron Spectroscopy for Atoms, Molecules and Con-

densed Matter," Nobel Lecture, 1981; N. Martensson and A. Nils-

son, J. Electron Spectrosc. Relat. Phenom. 75,209 (1995); T. L. Barr,E. Hoppe, T. Dugall, P. Shah, and S. Seal, J. Electron Spectrosc. Relat.

Phenom. 99, 95 (1999).

J. F. Wilkinson and G. P. Kelly, in "Handbook of Organic Pholochemislry"(J. C. Scaiano, Ed.), Vol. I, p. 293. CRC Press, Boca Ralon, 1989, andreferences quoled Ihere.

2a. R. J. Hurlubise, "Phosphorimelry: Theory, Inslrumenlalion and Appli-calions," VCH Pub!ishers, New York, 1990.

2b. R. J. HUrlubise, "Solid Surface Luminescence Analysis. Theory, Inslru-

menlalion and Applicalions," Dekker, New York, 1981.2c. R. J. HUrlubise, Ana/. Chem. 61, 889A (1989).

3. "Surface Pholochemislry" (M. Anpo, Ed.), Wiley, Baffins Lane, 1996;"Pholochemislry on Solid Surt'aces" (M. Anpo and T. Malsuara, Eds,),Elsevier. Amslerdam. 1989,

11PHOTONIC AND ELECTRONIC SPECTROSCOPIES

22. D. Briggs and J. C. Riviere, in "Pratical Surface Analysis by Auger andX-Ray Photoelectron Spectroscopy" (D. Briggs and M. P. Seah, Eds.),Chap. 3. Wiley, New York, 1983. ,-'

23. See, for instance, M. Sastry and P. Ganguly, J. Phys. Chem. 102,697

(1998) and references therein.24. Z. B. Maksic, T::. Kovacek, K. Kovacevik, and Z. Medven, J. Mal. Struc-

fure 304,151 (1994); D. Kovacek, K. Kovacevik, D. Korenic, and Z. B.Maksic. J. Mal. Structure 304, 163 (1994); D. Kovacek, Z. B. Maksic,

1. Petanjek. and 1. Basic, J. Mal. Structure 304, 261 (1994).25. See, f(X' instance, C. D. Wagner, "N1ST X-Ray Photoelectron Spec-

tro~ Database," version I, V.S. Dept. of Commerce, NIST, 1989,avai~ on diskette. Version 2 is now available on lhe Web site hup://

srdata.lD5l.gov/xps/ .26. G. Be:8Dson and D. Briggs, "High Resolution XPS of Organic Polymers.

The Sc:KIMa ESCA300 Database," Wiley, New York, 1992.

27. J. Cazax, J. ElectlVll Spectrosc. Relat. Phenom. 105,155 (1999).28. D. A-lkbital and R. T. Mckeon, Appl. Ph.,'s. Leu. 20, 158 (1972).29. D. 8Ifts and M. P. Seah (Eds.), "Practical Surface Analysis," Wiley,

Newv..k. 1983.30. D. 8Ifts and M. P. Seah (Eds.), "Practical Surface Analysis," 2nd ed.,

Voll- Wiley, New Ym, 1996.31. C. J_I\JWeIl, J. EleclllNl Spectrosc. Relat. Phenom. 47,197 (1988).32. L. c. Fàhnan and J. W. Mayer, "Fundamentais of Surface and Thin

Fiba~sis," p. 129. North-Holland, New York, 1986.33. c. 1. ft.-dl. A. J~ki, I. S. Tilinin, S. Tanuma, and D. R. Penn,

1. a.- SpectmR: Relot. Phenom. 98-99, 1 (1999).34. c. D.~, L E.lmis, and W. M. Riggs, Surf. Interface Anal. 2, 53

(l~35. 5. 1:.-. C. J. "-dI. and D. R. Penn, Sutj. lnterface Ana/. 21, 165

(l~S. C. J. I'-dJ. and D. R. Penn, Su1f lnterface Ana/. 20, 7736.(I~

37. L J. .-di, "NIST Inelastic Electron-Mean-Free-Path Database,"~_disldk.

38. D-A.!i8ky. Ph)1"" B 5, 4709 (1972).39. Str;.k8§1allCe. P.C. J. Graat, M. A. J. Somers, and A. Bõttger, Suif.

1"""'~'-ll."'(1995).40. Scr..~~x:c;., E.~in and E. Ragaini, J. Electron Spectrosc. Relat.

~73.53(lm).41. R.~"'Z--1Ding, Rep. Prog. Phys. 55,487 (1992); J.-Ch. Kuhr

-"~.l.E1ft:tron Spectrosc. Relat. Phenom. 105,257 (1999).42. A.~A.~. and S. Tougaard, Prog. Suif. Sci. 63, 135 (2000).43. c. S.~. R.. J. Baird, W. Siekhaus, T. Novakov, and S. A. L.

~.L ~Spectrosc. Relat. Phenom. 4,93 (1977).44. c. S :,. ~'i in Surface Science" (S. G. Davison, Ed.),

'Wi m.~, Oxford, 1984, and references therein.45. J_~"" s-j:.,rface Anal. 20,161 (1993).46. p'.r~.La-m1 Spectrosc. Relat. Phenom. 73,25 (1995).47. H_~_D-L~, "Electron Energy Loss Spectroscopy and Sur-

r..:~--;"~ 3. Academic Press, New York, 1982.48. )L"~)L ~an, and M. Schott, Chem. Ph)'s. Lett. 94, 552(-49. 1~~ _"'s.&ce Characterization of Advanced Polymers"

(L.~ - r. fi Zambonin, Eds.), Chap. 2. VCH, Weinheim,~-

50. ~A.--~lho do Rego, J. Lopes da Silva, F. Abel, V. Quil-b,-~ s..~, and R. Jérôme, Macromolecules 26, 5900

51. u_- M.~, W. Gopel, and H. Schier, Suif Sci. Leu. 237,

56. H. Ibach and D. L. Mills, "Electron Energ [i)Ss Spectroscopy and Sur-

face Vibrations," Chap. 2. Academic Pre~Hew York, 1982.

57. "Owner's Manual for lhe Model LK2Km.~ LK Technologies, Inc.,

Bloomington, IN.

58. H. Ibach, "Electron Energy Loss Spe~ters: The Technology ofHigh Performance," Springer Series iiD ~tical Sciences, Vol. 63.

Springer, Berlin, 1991

59. M. Liehr, P. A. Thiry, J.- J. Pireaux, anil! R. Caudano, Phys. Rev. B 33,

5682 (1986).60. L. L. Kesmodel, S. Wild, and G. Apai. ~ Sci. 429, L475 (1999).

61. F. Deloffre, F. Garnier, P. Sriva.'itava, A\.. Y~ar, and J.-L. Fave, SYI//h.

Mer. 67, 223 (1994).62. J. H. de Boer, Z. Phys. Chem. Ab/. B 1 ~., 163 (1931).

63. C. H. Nikolls and P. A. Leemakers, Ad\t Plwrl}chem. 8,315 (1971) and

references therein; L. D. Weis, B. W. ~en, and P. A. Leemakers,

J. AmeI: Chem. Soco 88,3176 (1966).

64, P. de Mayo, L. V. Natarajan, and W. R. ~. in "Organic Phototransfor-mations in Nonhomogeneous Media" (M A. Fox, Ed.), ACS Symp. Se-

fies, Vol. 278, p. I. Am. Chefio Soc., Washington, DC, 1984, and ref-erences therein; P. de Mayo, Pure Appu.. Chem. 54 1623 (1982); P. deMayo, L. V. Natarajan, and W. R. WalTe. Chem. Phys. Ler/. 107, 187

(1984); P. de Mayo, L. V. Natarajan, andJW. R. Ware, J. Phys. Chem. 89,

3526 (1985).

65. D. Oelkrug, M. Plauschinat, and R. W~ Kessler, J. Lumil/. 18/19,434(1979); R. W. Kessler, S. Uhl, W. HornM:ll, and D. Oelkrug, J. Lumil/.24/25, 551 (1981); W. Honnen, G. Kraft*.-hler, S. Uhl, and D. Oelkrug,

J. Phy,t. Chem. 87,4872 (1983); K. Kempfer, S. Uhl, and D. Oelkrug,J. Moi. S/ucrllre 114,225 (1984); D. Oelbug, W. Flemming, R. Fuller-man, R. Gunter. W. Honnen, G. Krabicl1ler, S. Schatler, and S. Uhl, Pure

Appl. Chem. 58,1207 (1986) and references therein.

66a. K. Kalyanasumdaram and J. K. Thoma.'i.J. AmeI: Chem. Soco 99,2039(1977); G. Beck and J. K. Thomas, Chl'nL Phys. Ler/. 94,553 (l983);J. K. Thomas, Chem. Rev. 93, 301 (1993)and references therein; J. K.

Thomas, Cheln. Rev. 80, 283 (1980).

66b. L. Francis, J. Lin, and L. A. Singer, ChenL Phys. Letl. 94, 162 (1983).

67. N. J. Turro, M. B. Zimmit, and I. R. Gold, J. AmeI: Chem. Soco 107,5826 (1985); N. J. Turro, I. R. Gold, and M. B. Zimmit, Chenl. Phys.Lerr. 119,484 (1985); N. J. Turro, C. C. Cheng, and W. Mahler, J. AmeI:Chem. Soco 106,5022 (1984); N. J. Turro, Pure Appl. Chem. 58, 1219

(1986).68. J. C. Scaiano, H. L. Casal, and J. C. Netto-Ferreira, AmeI: Chem. Soco

S."lnpos. SeI: 13,211 (f985); H. L. Casal and J. C. Scaiano, CaI/aJo J.

Chem. 62,628 (1984); H. L. Casal and J. C. Scaiano, CaI/aJo J. Chem.63, 1308 (1985); H. L. Casal, J. C. Netto-Ferreira, and J. C. Scaiano,J. Il/clusion Pilel/om. 3,395 (1985); R. Boch, C. Bohne, and J. C. Sca-

iano, J. Org. Chem. 61,1423 (1996).69. H. Ishida, H Takahashi, and H. Tsubomura,J. AmeI: Chem. Soco 92, 275

(1970); H. Ishida and H. Tsubomura. J. Pil%chem. 2, 285 (1973).

70. R. Dabestani, lI/ler-AmeI: Ph%chem. Soco Newslerr. 20,24 (1997).

71. W. J. Albery, P. N. Bartlett, C. P. Wilde. and J. R. Darwent, J. AmeI:Chem. Soco 107, 1854 (1985); K. F. Scott, J. Chem. Soco Farada.,' Trans. I76,2065 (1980); A. Habti, D. Keravis, P. Levitz, and H. Van Damme,

J. Chem. Soco Faraday Trans. li 80, 67 (1984).

72. C. T. Lin and W. C. Hsu, J. Phys. Chem. 92, 1889 (1998); C. T. Lin,

W. L. Hsu, and M. A. EI-Sayed, J. Phys. Chem. 91,4556 (1987).

73. S. Suzuki and T. Fujii, in "Photochemistry on Solid Surfaces" (M. Anpoand T. Matsuara, Eds.), p. 77. Elsevier, Amsterdam, 1989.

74. K. Kemnitz, T. Murao, I. Yamazaki, N. Nakashima, and K. Yoshihara,Chem. Phys. Lerr. 101,337 (1983); K. Kemnitz, N. Tamai, I. Yamazaki,N. Nakashima, and K. Yoshihara, J. Phys. Chem. 91, 1423 (1987);K Kemnitz, N. Tamai, I. Yamazaki, N. Nakashima, and K. Yoshihara,

J. Phys. Chem. 91,5094 (1986).75. A. D. French, N. R. Bertoniere, O. A. Battista, J. A. Cuculo, and D. G.

Gray, in "Ullmanns Encyclopedia of Industrial Chemistry," 4th ed.,

Vol. 3, p. 476. VCH, New York, 1993.

52. D~I. o.caa, in "Electron Spectroscopy for Surface Analy-~ (:.:J. ~cs in Current Physics, Vol. 4. Springer- Verlag,~ 1911.

53. Y.~-lI!1a._'. 3,46 (1968).54. 8-~-. 8 and S. Lewhald, Rev. Sci. Instrum. 46, 1325

, -55. . . - - ".L'" Sci. Technol. A 1, 1456 (1983).

BOTELHO DO REGO AND VIEIRA FERRElRA312

76. O. A. Battista, in "Encyclopedia of Polymer Science and Technology,"(H. F. Mark, N. G. Gaylord, and N. M. Bikales, Eds.), Vol. 3, p. 285.

Wiley, New York, 1965.77. T. Vo-Dinh, "Room- Temperature Phosphorimetry for Chemical Analy-

sis," Wiley, New York. 1984.78. E. M. Schulman and C. Walling, Science 178, 53 (1972); E. M. Schul-

man and C. Walling, J. Phys. Chem. 77, 902 (1973); E. M. Schulman

and R. T, Parker, J. Ph.,'s. Chem. 81, 1932 (1977).79a. T. Vo-Dinh, E. L. Yen. and J. D. Winefordner, Anal. Chem.48, 1186

(1976).79b. D. L. McAlease and R. B. Dunlap, Anal. Chem. 56,2244 (1984).80. J. Murtagh and J. K. Thomas, Chem. Phys. [ett. 148,445 (1988).

8/a. L. F. Vieira Ferreira, M. R. Freixo, A. R. Garcia, and F. Wilkinson,

J. Chem. Soco Farada.,' Trans. 88,15 (199:!).8/b. L. F. Vieira Ferreira. A. R. Garcia, M. R. Freixo, and S. M. B. Costa,

J. Chem Soco Farada.,' Trans. 89,1937 (1993).82. M. J. Lemos and L. F. Vieira Ferreira, in "Eurolights ll-Light on Or-

ganized Molecular Systems," Hengelhoef. Belgium, 1995, p. 90; L. F.

Vieira Ferreira, M. J. Lemos, M. J. Reis. and A. M. Botehlo do Rego,

Langmuir 16,5673 (2000).83a. J. C. Netto-Ferreira, L. F. Vieira Ferreira. and S. M. B. Costa, Quim.

Nova 19.230(1996).83b. L. F. Vieira Ferreira, J. C. Netto-Ferreira. and S. M. B. Costa, Spec-

trochim.ActaA5I,1385(1995).83c. L. F. Vieira Ferreira, A. S. Oliveira, and J. C. Netto-Ferreira, in "Pro-

ceedings of lhe Third Conference on Fluorescence Microscopy and Flu-

orescent Probes," Prague, June 1999, pp. 199-208.84. L. F. Vieira Ferreira. M. J. Lemos, V. Wintgens, and J. C. Netto-Ferreira,

Quim. Nova 22, 522 (1999); L. F. Vieira Ferreira, M. J. Lemos, V. Wint-

gens, and J. C. Netto-Ferreira, Spectrochim. Acta A 55,1219 (1999).85. F. Wilkinson, P. A. Leicester, L. F. Vieira Ferreira, and V. M. M. Freire,

Photochem. Photobiol. 54,599 (1991).86. L. F. Vieira Ferreira, A. S. Oliveira, 1. V. Khmelinskii, and S. M. B.

Costa. J. Lumin. 60&61,485 (1994); F. Wilkinson, D. R. Worrall, and

L. F. Vieira Ferreira, Spectrochim. Acta A '+8, 135 (1992).87. L. F. Vieira Ferreira, J. C. Netto-Ferreira. A. S. Oliveira, and S. M. B.

Costa, Química 60, 50 (1996).88a. A. S. Oliveira, L. F. Vieira Ferreira, F. Wilkinson, and D. R. Worrall,

J. Chem. Soco Farada.,' Truns. 92,4809 (1996).88b. L. F. Vieira Ferreira, A. S. Oliveira, F. Wilkinson, and D. R. Worrall,

J. Chem. Soco Farada.,' Trans. 92,1217 (1996).88c. L. F. Vieira Ferreira, A. S. Oliveira, K. H. Henbest, F. Wilkinson, and

D. R. Worrall, "RAL CLF Annual Repon;' p. 143, 1997.88d. A. S. Oliveira, P. Almeida, and L. F. Vieira Ferreira, Col/. Czech. Chem.

Comm. 64,459 (1999).88e. L. F. Vieira Ferreira, A. S. Oliveira, P. Matousec, M. Towrie, A. W.

Parker, F. Wilkinson, and D. R. Worrall, "RAL CLF Annual Repon,"

p. 81,1998.89. H. E. A. Kramer, Chimia 40, 160 (1986).

90. N. S. Allen, Rev. Prog. Colo 17,61 (1987).9/. L. M. G. Jansen, t. P. Wilkes, D. C. Greenhill, and F. Wilkinson, J. Soco

Dyes Cal. 114,327 (1998).92. L. R. Snyder and J. W. Ward, J. Phys. Chem. 70,3941 (1966).

93. N. J. Turro, Tetrahedron 43,1589 (1987).94. W. J. Leigh and L. Johnson, "Handbook of Organic Photochemistry"

(J. C. Scaiano, Ed.), Vol. 2, Chap. 22, p. 401. CRC Press, Boca Raton,

1989, and references therein.95. Aldrich Technicallnfomlation Bulletin, AL-I44.

96. J. Nawrocki, J. Chromatogl: 779,29 (1997).97. K. K. Unger, "Porous Silica," EIsevier, Amsterdam, 1979; R. lIer, "The

Chemislry of Silica;' Wiley, New York, 1979.98. P. Van de Voon, 1. Gillis-D'Hamers, and E. F. Vansant, J. Chem. Soco

Faraday Trans. 86,3751 (1992); 1. Gillis-D'Hamers, 1. Comelissens,K. C. Vrancken, P. Van de Voon, and E. F. Vansant, J. Chem. Soco Fara-

day Trans. 88,723 (1992).99a. E. M. Ranigen, J. M. B.:nnett, R. W. Grose. R. L. Patton, R. M. Kirchner,

and J. V. Smith. Nalure :!71. 512 (1978).

99b. G. M. W.Shultz-Sibbel, D. T. Gjerde, C. D. Chriswell, J. S. Fritz, andW. E. Coleman, Talanta 29, 447 (1982).

99c. G. T. Kokotailo, S. L. Lawton, and D. H. Olson, Nature 272, 437 (1978).100. D. M. Bilby, N. B. Milestone, and L. P. Aldridge, Nature 280, 664

(1979).101. F. Wilkinson and L. F. Vieira Ferreira, J. Lumin. 40&41,704 (1988).102. J.C. Netto-Ferreira, L. M. Ilharco, A. R. Garcia, and L. F. Vieira Ferreira,

Langmuir 16, 10392 (2000).103. M. Gabriela Lagorio, L. E. Dicerio, M. I. Litter, and H. San Roman,

J. Chem. Soco Faraday Trans. 94,419 (1998).104. L. F. Vieira Ferreira, S. M. B. Costa, and E. J. Pereira, J. Photochem.

Photobiol. A 55, 361 (1991); L. F. Vieira Ferreira and S. M. B. Costa,

J. Lumin. 48&49, 135 (1991).105. M. Desaeger, M. J. Reis, A. M. Botelho do Rego, J. D. Lopes da Silva,

and I. Verpoest, J. MateI: Sci. 31,6305 (1996).106. U. Zielke, K. J. Huttinger, and W. P. Hoffman, Carbon 34, 983 (1996).

/07. H. Zhuang and J. P.Wightman, J. Adhes. 62, 213 (1997).108. A. E. E. Jokinen, P. J. Mikkola, J. G. Matisons, and J. B. Rosenholm,

J. Colloid lntefj: Sci. 196, 207 (1997)./09. A. Provatas, J. G. Matisons, and R. S. Smart, Langmuir 14, 1656 (1998)./ /0. S. H. Park, S. Mcclain, Z. R. Tian, S. L. Suib, and C. Karwacki, Chem.

MateI: 9, 176 (1997)./ / /. G. A. Hishmeh, T. L. Barr, A. Sklyarov, and S. Hardcastle, J. Vacuum

Sci. Technol. A 14, 1330 (1996)./ /2. M. Rei Vilar, M. Schott, and P. Pfluger, J. Chem. Phys. 92,5722 (1990)./13. See, for instance, M. Schreck, M. Abraham, A. Lehmann, H. Schier, and

W. Gopel, Sufj: Sci. 262,128 (1992)./ /4. A. M. Botelho do Rego, unpublished results./ /5. M. Raposo, R. S. Pontes, L. H. C. Mattoso, and O. N. Oliveira, Jr.,

Macromolecules 30,6095 (1997)./ /6. See, for instance, A. Maroufi, L. J. Mao, and A. M. Ritcey, Macromol.

Sympos. 87,25 (1994)./ /7. J.-J. Pireaux, P. A. Thiry, R. Caudano, and P. Pfluger, J. Chem. Phys. 84,

6452 (1986)./18. A. M. Botelho do Rego, M. Rei Vilar, J. Lopes da Silva, M. Heyman,

and M. Schott, Sufj: Sci. 178,367 (1986)./ /9. M. Rei Vilar, M. Schott, J.-J. Pireaux, C. Grégoire, P. A. Thiry, R. Cau-

dano, A. Lapp, A. M. Botelho do Rego, and J. Lopes da Silva, Sufj: Sci.

189/190,927 (1987)./20. M. Rei Vilar, M. Schott, J.-J. Pireaux, C. Grégoire, R. Caudano, A. Lapp,

A. M. Botelho do Rego, and J. Lopes da Silva, Sufj: Sci. 211/212,782

(1989).12/. P. P. Hong, F. J. Boerio, and S. D. Smith, Macromolecules 27, 596 (1994)122. M. Rei Vilar, A. M. Botelho do Rego, J. Lopes da Silva, F. Abel, V. Quil-

leI, M. Schott, S. Petitjean, and R. Jérôme, Macromolecules 27, 5900

(1994)./23. P.-G. de Gennes, C. R. Acad. Sci. B 307,1841 (1988).124. W. Zhao, X. Zhao, M. H. Rafailovich, J. Sokolov, R. J. Composto, S. D.

Smith, M. Satkovski, T. P. Russell, W. D. Dozier, and T. Mansfield,

Macromolecules 26, 561 (1993)./25. A. M. Botelho do Rego, O. Pellegrino, J. M. G. Martinho, and J. Lopes

da Silva, Langmuir 16, 2385 (2000); ibid, Surface Science (2001), in

press./26. "Polymer Handbook" (J. Brandrup and E. H. Immergut, Ed.), Wiley,

New York, 1989.127. H. R. Thomas and J. J. O'Malley, Macromolecules 12,323 (1979)./28. J. J. O'Malley, H. R. Thomas, and G. M. Lee, Macromolecules 12,996

(1979)./29. "UV Atlas of Organic Compounds," Vol. I. Butterworths, Lon-

don/Verlag Chemie, Weinheim, 1966./30. A. M. Botelho do Rego, M. Rei Vilar, and J. Lopes da Silva, J. Electron

Spectrosc. Relat. Phenom. 85, 81 (1997).13/. H. H. Jaffe and M. Orchin, "Theory and Applications of Ultraviolet

Spectroscopy," V;';~-New York, 1962.132. E. Yamamoto, T. Yoshidome, T. Ogawa. and H. Kawazumi, J. Electron

Spectrosc. Relat. Phenom. 63,341 (1993)./33. A. Skerbele and E. N. Lassettre. J. Chem. Phys. 42, 395 (1965).

PHOTONIC AND ELECTRONIC SPECTROSCOPIES 313

/44. O. Pellegrino. M. Rei Vilar, G. Horowitz. F. Kouki. F. Garnier. J. D.

Lopes da Silva, and A. M. Botelho do Rego. Thin Solid Films 327-329.252 (1998).

/45. O. Pellegrino, M. Rei Vilar. G. Horowitz, F. Kouki. F. Garnier. J. D.

Lopes da Silva. and A. M. Botelho do Rego, Thin Solid Films 327-329.291 (1998).

/46. A. M. Botelho do Rego, O. Pellegrino. J. D. Lopes da Silva, G. Hora-

witz, F. Kouki. F. Gamier, and M. Rei Vilar, S.\'nth. Met. 101.606(1999)./47. D. Oeter, H.-J. Egelhaaf. Ch. Ziegler. D. Oelkrug, and W. Gtipel.

J. Chem. Phys. 101,6344 (1994)./48. M. Rei Vilar, G. Horowitz, P. Lang. O. Pellegrino, and A. M. Botelho do

Rego, Adv. MateI: Opto Electron. 9, 211 (1999)./49. F. Kouki, These d'Université, Univ. de Pierre et Marie Curie. Thiais.

1998./50. See. for instance, M. Kawakasi and H. Inokuma, J. Phys. Chem. 103.

1233(1999)./5/. S. J. Kerber, J. J. Bruckner, K. Wozniak. S. Seal. S. Hardcastle. and T.l.

Barr.J. Vac.Sci. Technol.A 14, 1314(1996)./52. S. J. Gregg and K. S. W. Sing. "Adsorption, Surface Area and Porosity:'

p. 140. Academic Press, New 'fork. 1978./53. A. W. Adamson. "Physical Chemistry of Surfaces," 5th ed.. p. 564. \\1-

ley, New York. 1990./54. R.C. Weast. "CRC Handbook of Chemistry and Physics," 78th ed.. p. B-

155. CRC. Press. Cle'"eland. 1997.

134. E. N. Lasseltre, A. Skerbele, M. A. Dillon, and K. J. Ross, J. Chem.

Phys. 48, 5066 (1968).

135. J. P. Doering, J. Chem. Phys. 512866 (1969).

136. Ph. Avouris and J. E. Demuth, J. Chem. Ph.vs. 75,4783 (1981).

1i7. L. Saliche and M. Michaud, Chem. Ph.,.s. Utt. 80, 184 (1981).

138. A. Dodabalapur, L. Torsi, and H. E. Katz, Science 268, 270 (1995);B. Servet, G. Horowitz, S. Ries, O. Lagorse, P. Alnot, A. Yassar, F. De-

loffre, P. Srivastava, R. Hajlaoui, P. Lang. and F. Gamier, Chem. Mattet:

6, 1089 (1994); P. Ostoja, S. Guerri, S. Rossini, M. Servidori, C. Taliani,and R. Zamboni, S.vnth. Met. 54,447 (1993).

139. D. Bimbaum and B. E. Kohler, J. Chem. Phys. 96,2492 (1992); D. Bim-baum, D. Fichou, and B. E. Kohler, J. Chem. Ph.vs. 96, 165 (1992); R. S.

Becker, J. Seixas de Meio, A. L. Maçanita. and F. Elisei, J. Phys. Chem.

100,18683 (1996), and references therein.

140. See, for instance, M. Schmelzer, S. Roth. P. Bãuerle, and R. Li, ThinSolid Films 229, 255 (1993).

141. M. Muccini, E. Lunedei, C. Taliani, F. Gamier, and H. Baessler, Synth.Met. 84,863 (1997).

142a. M. Muccini, E. Lunedei, A. Bree, G. Horowitz, F. Gamier, and C. Tal-iani, J. Chem. Phy.ç. 108,7327 (1998);

142b. M. Muccini, E. Lunedei, C. Taliani. D. Beljonne, J. Comil, and J.-L.

Brédas, J. Chem. Phys. 109, 10513 (1998).

143. P. Dannetun, M. Rei Vilar, and M. Scholt, Thin Solid Films 286, 321

(1996).