Jackson S.D., Hargreaves J.S.J. Metal Oxide Catalysis

Подождите немного. Документ загружается.

In the case of semiconductor systems initially containing TMI in a d

state and

subsequently reduced, the presence of band - like absorptions assignable to d – d

transitions, instead of edge - like features, arises from the formation of localized

reduced centers and not the injection of electrons into the conduction band of the

support. This has been discussed in the case of the V

/WO

/TiO

EUROCAT

SCR catalyst [72] .

Focusing on the shape of d – d bands, a proﬁ le asymmetry due to the distortion

of local structures by the Jahn – Teller effect can be useful in monitoring the isola-

tion of TMI sites with respect to aggregated structures exhibiting a sharp band - gap

transition in a similar position. An example is provided by the dependence of the

spectral features of CuO/Al

catalysts on the copper content (Figure 2.14 ). It can

be observed that Cu

octahedra are strongly distorted by the Jahn – Teller effect,

changing from O

to D

symmetry.

Besides asymmetry, band width may be a source of information also. The width

of the d – d bands of TMI complexes adsorbed on an oxidic support can be compa-

rable to or even narrower than those observed for the aqueous precursor com-

plexes. This can be related to a high homogeneity of the molecular environment

of TMI on the surface, suggesting a possible molecular recognition character of

the adsorption process [73] .

2.3 UV-Vis-NIR Absorption Spectroscopy 73

Figure 2.13 DR UV - Vis spectra of (a) calcined Ti - MCM - 41, (b)

silylated Ti - MCM - 41, (c) calcined Ti - MCM - 48, silylated Ti -

MCM - 48. In all cases the Ti loading was of ca 2 wt%; the

samples were outgassed at 523 K for 2 h before spectroscopic

measurements. Reprinted with permission from ref. [71] .

74 2 The Application of UV-Visible-NIR Spectroscopy to Oxides

Of course, a relevant source of information on the features of a TMI center at

the surface of a catalyst, or on its evolution during the catalyst preparation, is the

position of d – d bands. These have a dependence upon the composition of the

ligand sphere, including oxygen atoms/hydroxy groups at the surface of an oxidic

support. If the system can be treated with the model of the cubic ligand ﬁ eld

(i.e. octahedral and tetrahedral stereochemistry) and, once known the nature of

the ligands, it is possible to analyze the band positions by applying the law of

average environment [15] . This states, taking a [MA

6 - n

] complex as an example,

that the ligand ﬁ eld will be

∆∆ ∆

00 0

TOT A B

=+−

()

[]

nn (provided that the partial

substitution of B with A ligands results in negligible splitting of the octahedral

terms). For instance, using this approach, analysis of the progressive shift of d – d

bands along a series of preparation steps of Ni/SiO

catalysts (Figure 2.15 ), has

allowed evaluation of the ∆

related to surface

≡

SiO

−

present in the ligand sphere

of supported Ni

ions [74] . Such an approach extended to a larger series of oxidic

supports gave rise to a “ spectrochemical series of supports ” (alumina, zeolite Y

and silica) [75] :

∆∆∆

000

AlO ZO SiO

()

consistent with what was found for square - planar Pd

complexes [76] .

Furthermore, the comparison of the values of the Racah parameter, B , obtained

from the d – d spectra of a TMI dispersed on a series of supports can give informa-

tion on the difference in the covalence level of the TMI – support interaction among

Figure 2.14 DR UV - Vis spectra of copper oxide supported or

not on Al

in the fresh state: (a) bulk copper oxide diluted

into Al

; (b) 2.1 wt% CuO/Al

; (c) 4.8 wt% CuO/Al

;

(d) 9.2 wt% CuO/Al

. Reprinted from ref. [116] , with

permission from the Royal Society of Chemistry.

the various cases. Again in the case of Ni

ions, the following “ nephelauxetic series

of supports ” was proposed [75] :

βββAlO ZO SiO

()

indicating that the weaker - ﬁ eld surface ligands have more covalent bonding to the

supported ions.

Finally, in deriving structural information from the features of d – d spectra of

TMIs, it must be considered that, because of the Laporte selection rule, ion sites

with octahedral symmetry can contribute to the spectra only to a very limited

extent, and so can escape spectroscopic detection. However, this behavior can be

turned into a tool to monitor the distribution of TMIs in sites with different struc-

ture as a function of loading, as in the case of CoAPO zeotype materials. In this

case, the attainment of a plateau level of the intensity of the d – d bands due to Co

ions with tetrahedral symmetry that became inserted in the framework indicated

the formation of extra - framework species, containing “ d – d silent ” octahedral Co

sites with increasing loading (Figure 2.16 ) [77] .

2.3 UV-Vis-NIR Absorption Spectroscopy 75

Figure 2.15 DR UV - Vis - NIR spectra of a 1.7 wt% Ni

(II)

/SiO

catalyst after different preparation steps: (a) wet after

centrifugation; (b) ﬁ ltered and dried at 293 K in air for 15 h; (c)

ﬁ ltered and dried at 353 K in an oven for 15 h; (d) calcined at

773 K in O

and rehydrated for 1 year; (e) unexchanged

ammoniated SiO

dried in an oven at 353 K for 15 h.

Reprinted from ref. [74] with permission from Elsevier.

76 2 The Application of UV-Visible-NIR Spectroscopy to Oxides

2.4

UV - V is - NIR Photoluminescence Spectroscopy

Photoluminescence spectroscopy studies the emission of electromagnetic radia-

tions by a system that returns to its ground electronic state from an excited one

previously populated by light absorption. This kind of electronic spectroscopy

is a powerful tool in determining the structure, coordination and reactivity of

metal centers, especially for metal loadings below 1 – 2 wt%. In fact, at higher metal

loading, a concentration quenching may occur and the photoluminescence

becomes less informative.

Photoluminescence ( PL ) is widely applied to investigate surfaces and surface

chemical phenomena with a high degree of sensitivity. The technique provides

extremely rich information when applied to the study of photoluminescence sites

on bulk oxides with a large surface to volume ratio; on sites located on the surface

of a support, for example oxide - supported catalysts; on sites that can be modiﬁ ed

by thermal treatments (calcination, reduction, etc.); and when the local environ-

ment of the emitting sites is altered by the adsorption of molecular probes. By way

of introduction, basic photophysical aspects essential for the rationalization of PL

data will be summarized.

2.4.1

Franck – Condon Principle

The Frank – Condon principle is based on the fact that the time of an electronic

transition (of the order of 10

− 16

s) is shorter than that of a vibration (of the order

of 10

− 14

s). This means that during an electronic transition the nuclei do not change

their positions. This phenomenon can be illustrated using the Morse potential

energy curves for diatomic molecules (Figure 2.17 ). The series of horizontal lines

Figure 2.16 Plot of the intensity of the KM

function [ F ( R

∞

)] against the Co - content in

CoAPO - 5 catalysts: (A) as synthesized;

(B) after activation in O

at 373 K. Curve 1:

NIR band; curves 2: Vis bands. Co content

is expressed as the value “ x ” in the formula

(Co

· n (N/T) · m H

O [ x + y = 0.5; z =

0.5; n = 0.38 (triethylamine, T), 0.25 (N,N -

diethylethanolamine, N); m = 8.5 (with T),

10.3 (with N). Reprinted from ref. [77] , with

permission from Elsevier.

represents the vibrational states of the anharmonic oscillator either in the ground

singlet state (S

) in which all electrons are paired or are in the electronic excited

singlet state (S

) in which the unpaired electrons remain with opposite spin. Con-

sequently the electronic transitions, in both absorption and emission processes,

occur at constant internuclear distance and should be drawn vertically, as shown

in Figure 2.17 .

Upon absorption of UV - Vis light, the system moves to an electronic excited state.

There are many ways by which the system can return to the ground state involving

radiative or non - radiative decays. In the case of radiative decay, the system loses

the excitation energy as a photon and two different mechanisms can be involved:

ﬂ uorescence and phosphorescence. Fluorescence is an emission process between

states with the same spin multiplicity (i.e. singlet – singlet transitions), while phos-

phorescence involves electronic states with different spin multiplicity (i.e. triplet –

singlet transitions). Fluorescence lifetimes are normally very short (10

− 9

s for

organic molecules). Phosphorescence lifetimes are longer (ranging from 10

− 3

to minutes) which is because transitions between states of different multiplicity

are forbidden by the selection rule ∆ S = 0. They thus have very low probability.

Frequently, non - radiative decay may also compete.

Figure 2.17 Potential energy curves for ground (S

) and

excited (S

) states. Reprinted from ref. [11] , with permission

from Elsevier.

2.4 UV-Vis-NIR Photoluminescence Spectroscopy 77

78 2 The Application of UV-Visible-NIR Spectroscopy to Oxides

In the case of solids, the excited ion pair is subject to the inﬂ uence of its lattice

environment, which produces a vibrational relaxation process. The lattice environ-

ment, however, is not able to accept the larger energy difference needed to lower

the ion pair to the ground electronic state and, therefore, the ion pair may survive

long enough to undergo spontaneous emission, releasing the remaining excess of

energy as radiation. This electronic transition, denoted as ﬂ uorescence, is repre-

sented in accordance with the Frank – Condon principle by a vertical line. The ﬂ uo-

rescence occurs at a frequency lower than that of the absorption process and the

difference between the frequencies is called the Stokes shift.

The light emission normally occurs after some vibrational energy has been dis-

sipated into the surroundings; consequently the ﬂ uorescence intensity depends

on the ability of the lattice environment (or the surrounding gas - phase molecules)

to accept the electronic and vibrational quanta. Thus, it has been observed that

molecules with widely spaced vibrational levels, such as molecular oxygen, are able

to accept the large quantum of electronic energy and quench the ﬂ uorescence. On

the other hand, ﬂ uorescence emission may be increased by decreasing the tem-

perature, for instance working at liquid nitrogen temperature because, at low

temperatures, lattice vibrations are less favored.

When a spin ﬂ ip occurs for an electron in the S

state, an excited triplet state

) is ﬁ lled. The T

state has a lower potential energy than the S

state since

electron – electron repulsions are lower (Figure 2.18 ).

Phosphorescence can be observed only by populating the T

electronic state,

which occurs when the S

potential energy curve that has been populated by

Figure 2.18 Potential energy curves for ground (S

) and

excited singlet (S

) and triplet (T

) states. Reprinted from

ref. [11] , with permission from Elsevier.

absorption phenomena, intersects the T

curve sharing a common geometry. The

mechanism for changing the spin state of the S

is called Intersystem Crossing

( ISC ). Singlet – triplet transitions may occur by spin – orbit coupling and for this

reason the ISC mechanism is expected to take place more efﬁ ciently in a molecule

with heavy atoms or in an ion pair in a oxidic system since the corresponding

spin – orbit coupling constant is large. Upon ISC, the ion pair moves from a vibra-

tional level of S

to an iso - energetic vibrational level of the T

state, where it con-

tinues to release energy into the surroundings, moving down the vibrational levels

until it reaches the lowest vibrational state. From this vibrational level, the ion pair

may return to the ground state (S

) by emitting radiation; this phenomenon is

called phosphorescence. The T

– S

transition is forbidden by the ∆ S = 0 selection

rule and consequently the intensity of phosphorescence is normally lower than

that of ﬂ uorescence. Generally, phosphorescence is enhanced when the material,

such as a catalyst, has highly dispersed metal ions or highly localized emitting

sites. In these cases, the energy transfer becomes less efﬁ cient and there is time

for ISC mechanism to take place, as the singlet excited state steps slowly past the

intersection point (Figure 2.18 ).

Finally, we note that radiative decay transitions have the same nature as absorp-

tion ones and consequently they obey the same selection rules (see Sections 2.1

and 2.2 ).

2.4.2

Quantum Efﬁ ciency and Lifetime

The quantum efﬁ ciency or yield ( Φ e) and lifetime ( τ ) are important parameters

for identiﬁ cation of the emitting species.

The quantum efﬁ ciency, or yield, of photoluminescence ( Φ e) is deﬁ ned as the

ratio of the numbers of photons emitted from an excited species to the number

of photons absorbed. It can be expressed as

ΦΦ

eeeIC e

*=+

()

=kkk kτ

(2.11)

where k

is the radiative rate constant and k

is the sum of the unimolecular rate

constants of non - radiative processes from the excited states. Φ * and τ are the

concentration of species and the experimental lifetime of the excited state respec-

tively. To achieve a high quantum efﬁ ciency it is necessary to minimize k

, for

instance by cooling the sample to low temperature (77 K or 4 K) in order to reduce

the contribution of non - radiative processes.

For a single exponential decay, the radiative lifetime ( τ

), obtained from the

decay curves of the intensity measured as a function of time, is deﬁ ned by the

mean time the species spends in the excited state prior to return to the ground

state. Generally, lifetimes are longer for the upper states reached through a weakly

absorbing transition and also as the energy separation between the ground and

excited states is larger.

2.4 UV-Vis-NIR Photoluminescence Spectroscopy 79

80 2 The Application of UV-Visible-NIR Spectroscopy to Oxides

The lifetime is expressed as

1=+

()

eIC

(2.12)

In the presence of a quenching molecule, the lifetime ( τ ) is given by

τ= + +

[]

()

1 kk k

eICq

(2.13)

where k

is the rate constant of the quenching process and [Q] the concentration

of the quencher.

Under conditions of steady and constant illumination, the concentration of the

excited state is given by

()

Ikk

eIC

(2.14)

but if a quencher molecule is present the concentration becomes

Φ= + +

[]

()

Ik k k

eICq

(2.15)

When the intensity of the excitation light and the concentration of the emitting

species are kept constant, the photoluminescence intensity is proportional to the

concentration of the excited species.

The presence of a quencher molecule decreases the intensity of photolumines-

cence since the ﬂ uorophore is often returned to the ground state during a diffusive

encounter with the quencher. In this process the ﬂ uorophore is not chemically

modiﬁ ed. For collisional quenching the decrease in intensity is described by the

Stern – Volmer equation:

ΦΦ

()()

[]

()

[]

kkk k

qeIC q

QQτ

(2.16)

A wide variety of molecules can act as quenchers, including oxygen, halogens,

amines and electron - deﬁ cient molecules.

2.4.3

General Remarks on Methodologies Applied for PL Measurements

A PL spectrum reports the emission intensity as a function of the wavelength of

the emitted light when the excitation wavelength and the intensity of the exciting

light are ﬁ xed at constant values. In some cases, it is possible to observe the vibra-

tional ﬁ ne structure in the photoluminescence band, especially if the spectra are

collected at 77 K. As well as the emission spectrum, the excitation spectrum can

also be collected. In this case, the intensity of the emission at a ﬁ xed wavelength

is plotted as a function of the wavelength of the excitation light, which varies with

the extinction coefﬁ cient of the absorbing species. Therefore, the excitation spec-

trum is similar to the absorption one; the advantage of collecting excitation spectra

compared to absorption measurements is their greater sensitivity, even for low

concentrations of photoluminescent species.

Details on the instruments available for collecting PL spectra are reported in

refs. [11, 78] . One of the main points to be stressed is that, because of the high

sensitivity of the PL technique, the sample cell has to be made from high - quality

fused silica with no impurities, such as Suprasil.

PL is typically collected at 90 ° to the transmitted light for transparent liquid

samples, whilst in the case of pellets or powered samples it is necessary to collect

the spectra in front - face geometry, meaning that a swing - away mirror is positioned

to allow the collection of the PL of the sample at an angle 45 ° > θ > 22 ° (depending

on the instrument) with respect to the incident radiation. When oxide catalysts

have to be studied it is important to collect the spectra in vacuo to remove any

gases that might act as a quencher molecules, such as oxygen.

2.4.4

Characterization of Oxide Catalysts by PL

2.4.4.1 Insulating Oxides: the Case of AEO

The possibility of using PL to study insulating oxide catalysts is limited by the

position of the absorption bands that can be used for excitation. For this reason,

most PL investigations on such types of material were focused on alkaline earth

oxides that, as reported above, exhibit excitonic absorptions in the near - UV. In this

respect, the background knowledge in this area is based on the series of investiga-

tions carried out in the late 1970s and early 1980s on MgO, CaO, SrO and BaO

[3, 79 – 83] . These studies demonstrated that the emission and excitation spectra of

AEOs are extremely rich in information, owing to the presence of several bands

associated with ions with different local coordination on the surface.

The PL spectra were actually very sensitive to the overall surface structure

and this allowed the study of the behavior of each type of luminescence center

upon thermal treatment or adsorption of probe molecules. These studies have also

shown that, by the way of an energy - transfer process, emission can arise from

surface sites that are not necessarily those that absorbed light in the ﬁ rst step of

the PL phenomenon. For instance, at 300 K, the energy absorbed by 5 - and 4 -

coordinated sites is transferred to the 3 - coordinated ones, whilst at 77 K, the

energy - transfer process is largely suppressed and the original emission proﬁ les of

4 - and 5 - coordinated centers can be observed [41] .

From these early studies, the PL technique, which was initially aimed at conﬁ rm-

ing the observation of surface sites, was gradually extended to explore other related

aspects, such as surface structure, decay of the excited states and surface reactivity.

In addition, the adsorption of quencher molecules, such as O

and H

, on AEOs

has allowed the clariﬁ cation of the nature of the luminescence sites present at the

surface [81] . In particular, in the case of H

adsorption on SrO, a change in the

excitation band shape was observed. H

may react at different rates with species

absorbing in different parts of the excitation band, producing a change in the band

shape. No corresponding change in band shape was observed in the emission

spectrum. By contrast, O

adsorption did not change the shape of either the excita-

tion or the emission bands; only a decrease in intensity was observed. This

2.4 UV-Vis-NIR Photoluminescence Spectroscopy 81

82 2 The Application of UV-Visible-NIR Spectroscopy to Oxides

spectroscopic evidence suggested (i) that O

and H

are adsorbed on different sites,

(ii) that the absorption and emission processes occur at different, but closely

related, sites on the surface and (iii) that energy can be transferred along the

surface. In general, O

adsorption is associated with regions of the surface that are

oxygen - deﬁ cient or have a local excess of cations, whilst hydrogen is expected to

react with areas that are cation - deﬁ cient or have a local excess of oxygen ions. In

the case of SrO, most of the intensity loss of the emission was not reversible and

must be strongly held on the surface, probably as hydroxyls. Using analysis of

such spectroscopic behavior, it was proposed that the sites on the surface respon-

sible for the light emission are oxygen - deﬁ cient or have a local excess of cations,

whereas the sites responsible for the light absorption are cation - deﬁ cient or have

a local excess of oxygen ions [81] .

The luminescence band of the AEO, outgassed at 1200 K, was shifted to lower

energy as the cation size increased, as reported in Table 2.4 and in Figure 2.19 .

The excitation spectra were at lower energy than the band gap of the corresponding

bulk materials and were similar to the absorption spectra [44, 45] . The excitation

processes were, therefore, associated with anions and cations in low coordination

located at the surface and led to the formation of excitons (electron - hole pairs):

MO MO

LC LC LC LC

22+− +−

+→hν

(2.17)

where LC denotes low coordination.

Surface excitons require less energy in their formation than bulk excitons, owing

to the reduced Madelung constant of the coordinatively unsaturated ions at the

surface [81] .

Furthermore, analysis of the excitation spectra of AEOs has revealed that the

excitation bands are characterized by several components, which is evidence that

centers with different coordination are present on the surface. The lower the

energy, the lower the coordination number. It was established that the three main

components identiﬁ ed in the excitation spectra were associated with three surface

states with different coordination of the ion involved: O

2−

on extended (001) faces,

2−

on the edges and O

2−

on the corners. In the bulk lattice of an AEO the ions

are octahedrally coordinated.

Table 2.4 Peak positions of the absorption, excitation and

emission spectra of the alkaline earth oxides.

Oxide Absorption (eV) Excitation (eV) Emission (eV)

MgO 5.70; 4.58

> 5.40; 4.52

3.18

CaO 5.52; 4.40

> 5.40; 4.40

3.06

SrO 4.64; 3.96 4.43; 3.94 2.64

BaO 3.60; 3.22 3.70 2.67