Structure–photoluminescence relationship in Eu(iii) β-diketonate-based organic–inorganic hybrids. Influence of the synthesis method: carboxylic acid solvolysis versus conventional

Structure–photoluminescence relationship in Eu(III) b-diketonate-basedorganic–inorganic hybrids. Influence of the synthesis method: carboxylicacid solvolysis versus conventional hydrolysis

Lianshe Fu,a R. A. Sa Ferreira,a N. J. O. Silva,a A. J. Fernandes,b Paulo Ribeiro-Claro,b I. S. Goncalves,b

V. de Zea Bermudezc and L. D. Carlos*a

Received 15th March 2005, Accepted 26th May 2005

First published as an Advance Article on the web 21st June 2005

DOI: 10.1039/b503844h

Organic–inorganic hybrids incorporating Eu(nta)3?bpy (where nta and bpy stand for 1-(2-

naphthyl)-4,4,4-trifluoro-1,3-butanedionate and 2,29-bipyridine, respectively) were prepared

either by acetic acid solvolysis or a conventional hydrolysis sol–gel route. The host framework of

these materials, classed as di-ureasil, consists of a siliceous network grafted, through urea cross-

linkages, to both ends of poly(ethylene oxide) chains. The resulting Eu(III)-based di-ureasils were

investigated by small angle X-ray scattering, X-ray diffraction, Fourier transform mid-infrared

spectroscopy, 29Si and 13C nuclear magnetic resonance, and photoluminescence spectroscopy,

with particular attention paid to the effect of the adopted synthesis strategy on the relationship

between structure and emission properties. The dimensions and the degree of condensation of the

siloxane nanodomains depend noticeably on the synthesis route and the overall emission quantum

yield decreases from 15 (conventional hydrolysis) to 6% (solvolysis route). The broad white-light

emission typical of the di-ureasil host was not detected here suggesting, therefore, the activation of

energy transfer channels between the hybrid host’s emitting centres and the Eu(III) ions. As the

first coordination shell of Eu(III) is essentially independent of the synthesis method employed, the

significant decrease in the emission quantum yield for the di-ureasil prepared by acetic acid

solvolysis might be explained by the interaction between the hybrid emitting centres and the nta

ligand levels, favouring a larger non-radiative transition probability.

Introduction

The sol–gel process is a promising technique for the develop-

ment of organic–inorganic hybrids due to its mild reaction

conditions, versatility of processing and potential for mixing

the inorganic and organic precursor components at the

nanometer scale.1,2 When functional active molecules, such

as optical, electronic, magnetic and biological species, are

incorporated into the hybrid structure, functional organic–

inorganic hybrid nanocomposites may be thus synthesized.

In this context, it must be emphasized that the incorporation

of luminescent molecules into organic–inorganic hybrid

matrices has made great progress in both fundamental

luminescence spectroscopic studies and the development of

advanced optical materials.2–6

Although lanthanide complexes exhibit a much more

efficient emission under ultraviolet excitation,7 up to the

present day they have been excluded from practical applica-

tions as tuneable solid-state lasers or phosphor devices due to

their poor thermal stability and mechanical properties.8 In

order to circumvent these shortcomings, the lanthanide com-

plexes can be incorporated into polymers and/or organic–

inorganic matrices using low-temperature soft-chemistry

processes, such as the sol–gel route. Indeed, much work has

been focused on this field to date, and many lanthanide com-

plexes have been incorporated into sol–gel derived matrices

or other solid hosts such as zeolite, layered or mesoporous

matrices.9

The di-urea or di-urethane cross-linked poly(ethylene oxide)

(PEO)–siloxane structures (named di-ureasils or urethanesils,

respectively) are promising hybrids for the fabrication of large

area neutron detectors,10 as nanocomposite gel electrolytes

for dye-sensitized photoeletrochemical cells11 and as efficient

white-light room temperature emitters (quantum yield of

10–20%).12–19 These materials can be prepared through

hydrolysis and condensation of the corresponding organic–

inorganic hybrid precursors obtained from the reaction of the

terminal amine groups of PEO-containing diamines (or the

hydroxyl groups of poly(ethylene glycol) for di-urethanesils)

with the isocyanate group of 3-isocyanatopropyltriethoxy-

silane (ICPTES).12 Alternatively, di-ureasils and di-urethane-

sils can be produced via acetic acid (AA) or valeric acid

solvolysis,13,14 displaying an emission quantum yield 27–35%

higher than that calculated for the analogues synthesised via

conventional sol–gel technique.14 Furthermore, transparent

and optically uniform di-ureasil films doped with a Eu3+

complex with thenoyltrifluoroacetone and 2,29-bipyridine

aDepartamento de Fısica, CICECO, Universidade de Aveiro, 3810-193,Aveiro, Portugal. E-mail: [email protected]; Fax: +351 234424695;Tel: +351 234 424370356bDepartamento de Quımica, CICECO, Universidade de Aveiro,3810-193, Aveiro, PortugalcDepartamento de Quımica and CQ-VR, Universidade de Tras-os-Montes e Alto Douro, 5000-911, Vila Real Codex, Portugal

PAPER www.rsc.org/materials | Journal of Materials Chemistry

This journal is � The Royal Society of Chemistry 2005 J. Mater. Chem., 2005, 15, 3117–3125 | 3117

(bpy) ligands were prepared by AA solvolysis.18 The Eu3+

emission, whose maximum intensity value is approximately

60% of that of Rhodamine-B, results from excitation on the

ligand levels and subsequent intramolecular energy transfer to

the 4f states. Although the organic–inorganic matrix also

seems to contribute to these energy transfer processes, the

nature of this contribution was unclear.18

The white-light photoluminescence (PL) of di-ureasils

results from a convolution of donor–acceptor pairs recombi-

nations that occur in the NH groups of the urea linkages

and in ?O–O–SiM(CO2) oxygen-related defects of the siliceous

nanodomains.3,15–17 Energy transfer between these hybrids’

emitting centres and the Eu3+ ions has been quantitatively

discussed elsewhere.19–21 The activation of these energy

transfer mechanisms noticeably depends on the Eu3+ local

coordination to the carbonyl group of the urea cross-linkages.

Moreover, that activation induces a decrease on the emission

quantum yield (relatively to that of the undoped nanohybrids)

and permits a fine-tuning of the emission chromaticity across

the CIE (Comission Internacionalle d’Eclairage) diagram.19–21

The present work aims at gaining a deeper understanding

of the hybrid host–Eu3+ energy transfer mechanisms by

comparing the luminescence features of Eu(nta)3?bpy-doped

di-ureasils synthesized through conventional hydrolysis and

AA solvolysis sol–gel routes. The emission component

associated with the hybrid host’s emitting centres could not

be detected, clearly suggesting the presence of active energy

transfer channels between them and the Eu3+ ions. The

efficiency of these energy transfer processes should be larger

compared with that of the di-ureasils with Eu(CF3SO3)3, where

the host-related emission is only absent for large amounts of

incorporated salt.19–21 On the other hand, the Eu3+ coordina-

tion in the Eu(nta)3?bpy-doped di-ureasils involves the

carbonyl oxygen of the urea bridges, independently of the

synthesis method adopted. However, the dimension and

the condensation degree of the siloxane nanodomains depend

on the synthesis route. Thus, promoting differences in the

dimensions and structure of the siloxane domains we change

the hybrid host-to-ligands-to-Eu3+ ion energy transfer

channels (essentially those connected with the nta ligand

levels) with the subsequent changes in the overall emission

quantum yield of the hybrids.

Experimental

Materials and synthesis

The diamine a, v-diaminepoly(oxyethylene-co-oxypropylene)

with a molecular weight of about 600 g mol21—corresponding

to approximately 8.5 (OCH2CH2) repeat units and commer-

cially designated as Jeffamine ED-6001, Fluka—was dried

over molecular sieves (4 A, 1.6 mm pellets, Aldrich) before use.

ICPTES (Fluka, 95%) and AA (Aldrich, 99.7%) were used

without further purification. Tetrahydrofuran (THF), chloro-

form (CHCl3) and absolute ethanol (CH3CH2OH) were dried

over molecular sieves at room temperature before use.

The europium complex Eu(nta)3?bpy22 (Scheme 1) was

synthesized by adding a solution of bpy (0.055 g, 0.36 mmol)

in 5 mL of CHCl3 to a solution of Eu(nta)3?2H2O (0.34 g,

0.35 mmol) in 15 mL of CHCl3 at room temperature. The

reaction mixture was stirred for 3 h at room temperature. Then

the solvent was removed, and the resulting solid was washed

with n-hexane. After drying in vacuum, 0.27 g of an orange

powder was obtained (yield = 71%). Elemental analysis,

Fourier transforms infrared (FTIR), Raman and nuclear

magnetic resonance (NMR) spectra confirm that the resulting

compound is the target product.

Anal. calc. for EuC52H32F9N2O6: C, 56.58; H, 2.92; N, 2.54;

Found C, 56.50; H, 2.59; N, 2.90. Selected umax/cm21 (KBr):

2977 (w), 2911 (w), 1640 (s), 1621 (s), 1614 (vs), 1568 (s), 1530

(s), 1509 (s), 1472 (m), 1463 (m), 1384 (m), 1299 (vs), 1198 (s),

1189 (s), 1135 (s), 1123(s), 793 (s), 763 (m), 684 (m), 568 (m).

Selected Raman (cm21) 3059 (m), 2990 (m), 1622 (s), 1595 (s),

1570 (w), 1466 (s), 1430 (w), 1386 (vs), 1356 (w), 1312 (m),

1292 (m), 1218 (m), 1201 (m), 1015 (m), 771 (m), 517 (m), 346

(w), 228 (w). uH (300 MHz, CDCl3, SiMe4) 12.95 (br, 2H, bpy),

10.43 (d, J = 7 Hz, 2H, bpy), 9.74 (br, 2H, Bpy), 8.52 (d,

J = 8 Hz, 2H, Bpy), 7.91 (d, J = 8 Hz, 3H, Naphth), 7.79, 7.76

(s + d, 9H, Naphth), 7.55 (d, J = 8 Hz, 6H, Naphth), 7.42 (d,

J = 8 Hz, 3H, Naphth), 3.45 (s, 3H, CH).

Thompson et al. earlier reported the synthesis of two

different crystalline structures of Eu(nta)3?bpy, one neat, and

the other with an isopropanol crystallization molecule, that

correspond to two different conformational isomers.23

Eu(nta)3?bpy was incorporated into d-U(600) di-ureasils

either by conventional hydrolysis or AA solvolysis. The

resultant luminescent hybrids were designated as d-U(600)–

Eu(nta)3bpy and d-U(600)–Eu(nta)3bpy–AA, respectively. The

Eu(nta)3?bpy content versus the total hybrid’s mass is ca. 4.5%,

corresponding to a ether-type oxygen atoms of PEO chains

per Eu3+ atom ratio equal to ca. 180. The structure of the

di-ureasils is outlined in Scheme 1. For the synthesis of

the d-U(600)–Eu(nta)3bpy or d-U(600)–Eu(nta)3bpy–AA, two

steps were involved. A typical synthetic procedure was as

follows:

Step 1. Synthesis of the di-ureasil precursor d-UPTES(600).

The procedure used was similar to that previously reported in

the literature.14 A volume of 0.40 mL (0.70 mmol) of Jeffamine

ED-6001 was added to a flask, followed by addition of 3.8 mL

Scheme 1 Chemical structure of (a) d-U(600) di-ureasil hybrid and

(b) Eu(nta)3?bpy complex.

3118 | J. Mater. Chem., 2005, 15, 3117–3125 This journal is � The Royal Society of Chemistry 2005

of dried THF in a fume cupboard. A volume of 0.36 mL

(1.40 mmol) of ICPTES was then added to this precursor

solution under stirring. The molar ratio of Jeffamine ED-6001

to ICPTES was 1 : 2. The flask was sealed and the solution was

stirred at room temperature in N2 atmosphere for 24 h. The

grafting process was monitored by infrared through the

observation of the progressive reduction and ultimate dis-

appearance of the band located at 2274 cm21, attributed to the

vibration of the MSi(CH2)3NCO group, and the growth of the

band envelope characteristic of the urea (urethane) group

(1800–1500 cm21 interval).

Step 2. Synthesis of d-U(600)–Eu(nta)3bpy and d-U(600)–

Eu(nta)3bpy–AA by conventional sol–gel method or carboxylic

acid solvolysis process. The THF in the precursor solution

was evaporated under vacuum, and a transparent precursor

d-UPTES(600) oil was thus obtained. For synthesis of

d-U(600)–Eu(nta)3bpy by conventional sol–gel method,

36.2 mg (0.0328 mmol) of Eu(nta)3?bpy was dissolved in

10 mL of CHCl3 and then 0.32 mL (5.48 mmol) of

CH3CH2OH and 0.075 mL (4.17 mmol) of pH = 2.0 HCl

was added to this solution. The molar ratio of ICPTES :

CH3CH2OH : H2O is 1 : 4 : 3. Finally, this mixed solution

was added to the precursor under stirring in air at

room temperature. The solution was further stirred for

24 h. For synthesis of d-U(600)–Eu(nta)3bpy–AA, 36.2 mg

(0.0328 mmol) of Eu(nta)3?bpy was dissolved in 10 mL of

CHCl3, and then 0.32 mL (5.48 mmol) of CH3CH2OH was

added to this solution. The mixture was added to the

precursor. Finally 0.24 mL of AA was added under stirring

in N2 at room temperature, and the mixture was stirred for

24 h. The following procedures are similar to the preparation

of the corresponding pure di-ureasils.

Experimental techniques

Mid-infrared spectra were recorded at room temperature using

a MATTSON 7000 FTIR Spectrometer. The spectra were

collected over the range 4000–400 cm21 by averaging 64 scans

at a maximum resolution of 4 cm21. The compounds were

finely ground (about 2 mg), mixed with approximately 175 mg

of dried potassium bromide (Merck, spectroscopic grade) and

pressed into pellets. Prior to recording the spectra the discs

were stored in an oven under vacuum at 80 uC for several days

in order to reduce the levels of by-product, solvent and

adsorbed water. Consecutive spectra were recorded until

reproducible results were obtained.

X-ray diffraction patterns were recorded using a Philips

X’Pert MPD Powder X-ray diffractometer system. The

powders were exposed to the Cu Ka radiation (l = 1.54 A)

at room temperature in a 2h range (scattering angle) between

1 and 80u. The xerogel samples, analyzed as films, were not

submitted to any thermal pre-treatment.

X-ray scattering measurements were performed using the

synchrotron SAXS beamline of the National Synchrotron

Light Laboratory (LNLS, Campinas, Brazil) with mono-

chromatic (l = 1.608 A) and horizontally focused beam. The

scattering intensity was recorded as a function of the modulus

of the scattering vector q, q = (4p/l)sin(h). The parasitic

scattering intensity from the air, slits and windows was

subtracted from the total intensity.29Si magic-angle spinning (MAS) and 13C cross-polarization

(CP) MAS NMR spectra were recorded on a Bruker Avance

400 (9.4 T) spectrometer at 79.49 and 100.62 MHz, respec-

tively. 29Si MAS NMR spectra were recorded with 2 ms

(equivalent to 30u) rf pulses, a recycle delay of 60 s and at a

5.0 kHz spinning rate. 13C CP/MAS NMR spectra were

recorded with 4 ms 1H 90u pulse, 2 ms contact time, a recycle

delay of 4 s and at a spinning rate of 8 kHz. Chemical shifts are

quoted in ppm from tetramethyl silane (TMS).

The emission, PL, and excitation, PLE, spectra and

lifetime measurements were detected between 10 K and

room temperature on a modular double grating excitation

spectrofluorimeter with a TRIAX 320 emission mono-

chromator (Fluorolog-3, Jobin Yvon-Spex) coupled to a

R928 Hamamatsu photomultiplier, in the front face acquisi-

tion mode. All the photoluminescence spectra were corrected

for optics and detection spectral response. The absolute

emission quantum yields (w) were measured at room tempera-

ture using the technique for powered samples described by

Brill et al.,24 through the following expression:

w~1{rst

1{rx

� �Ax

Ast

� �wst (1)

where rst and rx are the diffuse reflectance (with respect to a

fixed wavelength) of the hybrids and of the standard phosphor,

respectively, and wst is the quantum yield of the standard

phosphor. The terms Ax and Ast represent the area under the

di-ureasils and the standard phosphor emission spectra,

respectively. Diffuse reflectance and emission spectra were

acquired with the experimental setup aforementioned to detect

photoluminescence. In order to have absolute intensity values

BaSO4 was used as reflecting standard (r = 91%). The same

experimental conditions, i.e., the position of the hybrids/

standard holder, excitation and detection monochromator

slits (0.3 mm) and optical alignment, were fixed. To prevent

insufficient absorption of the exciting radiation, a powder

layer around 3 mm was used and utmost care was taken in

order to ensure that only the sample was illuminated, in order

to diminish the quantity of light scattered by the front sample

holder. The standard phosphor used was sodium salicylate

(Merck P.A.), whose emission spectra are formed by a large

broad band peaking around 425 nm, with a constant w value

(60%) for excitation wavelengths between 220 and 380 nm.

Three measurements were carried out for each sample, so that

the presented w value corresponds to the arithmetic mean

value. The errors in the quantum yield values associated with

this technique were estimated within 10%.24

Results and discussion

Small angle X-ray scattering and powder X-ray diffraction

The X-ray diffraction patterns of d-U(600)–AA, d-U(600)–

Eu(nta)3bpy and d-U(600)–Eu(nta)3bpy–AA are shown in

Fig. 1. These spectra show a main broad peak centred at ca.

q = 1.5 A21 associated with order within the siloxane domains,

corresponding to structural unit distances of ca. 4.1 A and

revealing that the compounds are highly amorphous. The


coherent length L associated with this peak was estimated

using a modified Scherrer equation (see, for instance ref. 14)

to be ca. of 9 ¡2 A. This value is identical to the one reported

for similar solvolysis-derived undoped hybrids14 and to the

diameter of the siliceous domains in organically modified

silicates obtained by SAXS.25 The second-order of the above

peak appears as a broad weak hump around 3 A21. The peak

appearing at ca. q = 0.23 A21 in XRD and SAXS patterns has

been assigned to an interference effect between siliceous

domains25 located at the ends of the polymer chain and

spatially correlated at a mean distance of 27 ¡ 1 A. The low

q-range of the SAXS patterns of the d-U(600)–Eu(nta)3bpy

and d-U(600)–Eu(nta)3bpy–AA (inset of Fig. 1) are dominated

by a power law regime: I(q) 3 q2D, with D = 2.6 and 3.2,

respectively. This low q-range scattering can be associated

with the siliceous domains, with D corresponding to their

dimensionality. Therefore, the different D values point out

distinct siloxane nanostructures for d-U(600)–Eu(nta)3bpy and

d-U(600)–Eu(nta)3bpy–AA.

Fourier transform infrared spectra (FTIR)

The FTIR spectra of d-U(600)–AA, d-U(600)–Eu(nta)3bpy,

d-U(600)–Eu(nta)3bpy–AA, and Eu(nta)3?bpy are shown in

Fig. 2. The shoulders around 920 cm21 in all the spectra

provide the evidence that in the di-ureasils the PEO chains

attain complete disorder. Comparing with the undoped di-

ureasil, the intensities of the peaks at around 920 cm21 show a

very slight decrease with the incorporation of the complex,

indicting a very weak Eu3+–PEO chains interaction. The bands

at 1324 cm21 in the spectrum of d-U(600)–AA are charac-

teristic of the amorphous state. The 1354 cm21 bands, ascribed

to the CH2 wagging vibrations, can be clearly seen in all the

spectra.

It is known that the nCO mode of PEO is an excellent tool to

probe the changes undergone by the polymer chains of the

hybrids upon incorporation of the guest compounds.26 In this

spectral region (1180–970 cm21), if complexation of the

cations by the oxygen atoms of the polyether chains occurs,

there will be an evident shift of the strong nCO band to lower

wavenumbers. The fact that peaks at around 1110 cm21,

characteristic of noncoordinated oxyethylene moieties, almost

unchanged in the spectra of the Eu3+-based di-ureasils

suggests that the PEO chains of the host materials persist in

an uncomplexed state.

In order to study in detail the vibrations of the urea groups

(the complex incorporation may induce Eu3+ coordination to

the carbonyl oxygen atoms of the urea cross-links), spectral

deconvolutions to the so called ‘‘amide I’’ (1800–1600 cm21)

and ‘‘amide II’’ regions (1600–1500 cm21) were carried out

using Gaussian band shapes as reported elsewhere.14,27 Three

components were isolated for the ‘‘amide I’’ envelope of

d-U(600)–AA at ca. 1720, 1668 and 1640 cm21 previously.

According to the literature,26,27 the former two components

are due to the vibrations of urea–poly(ether) hydrogen-bonded

structure, whereas the last one is ascribed to the strong

self-associated hydrogen-bonded urea–urea associations.

When Eu(nta)3?bpy was incorporated into d-U(600), either

by conventional sol–gel or AA solvolysis, besides the three

mentioned peaks, a new one at around 1620 cm21 appears,

indicating interactions between the Eu3+ ions and the carbonyl

oxygen atoms of the urea cross-links. Compared to the

undoped di-ureasil, the former three components have a

small red-shift peaking at ca. 1713–1716, 1657–1659 and 1632–

1636 cm21. The absence of an individual band at ca. 1750 cm21

in all the spectra indicates that neither CLO or N–H

groups from urea cross-linkages are left free in the hybrid

materials. The ‘‘amide II’’ mode is a mixed contribution of the

N–H in-plane bending, C–N stretching, and C–C stretching

vibrations. The bands for all the samples in this region appear

at ca. 1565 cm21.

NMR spectra

The 29Si MAS NMR spectra of d-U(600)–AA, d-U(600)–

Eu(nta)3bpy and d-U(600)–Eu(nta)3bpy–AA exhibit broad

Fig. 1 XRD patterns of (a) d-U(600)–AA, (b) d-U(600)–Eu(nta)3bpy

and (c) d-U(600)–Eu(nta)3bpy–AA.

Fig. 2 FTIR spectra of (a) d-U(600)–AA, (b) d-U(600)–Eu(nta)3bpy,

(c) d-U(600)–Eu(nta)3bpy–AA and (d) Eu(nta)3?bpy.


signals characteristic of T1, T2, and T3 units (Fig. 3a).

These sites are labelled using the conventional Tn notation,

(R’Si(OSi)n(OR)32n), where n (n = 1, 2, 3) is the number of

Si-bridging oxygen atoms. The absence of T0 indicates that,

although the polycondensations in most cases do not proceed

to completion, no precursor is left unreacted. In addition, the

absence of signals due to Qn notation (Si(OSi)n(OH)42n)

species between 290 and 2120 ppm, indicates that no cleavage

of the urea groups or Si–C bonds occurred under the

experimental conditions employed and that all the silicon

atoms are covalently connected to carbon atoms. The

dominate environments (at 258 and 266 ppm, respectively)

for all the samples are clearly seen, showing the presence of

two main types of local structures. The very weak shoulders

displayed in these hybrids at ca. 252 ppm are ascribed to T1

sites. Compared to the undoped sample, there is a relative

intensity inversion between T2 and T3 for d-U(600)–

Eu(nta)3bpy and d-U(600)–Eu(nta)3bpy–AA, indicating that

the complex incorporation affected the polycondensation

process. The relative populations of the various silicon sites

quantitatively estimated after deconvolutions are listed in

Table 1. The d-U(600)–Eu(nta)3bpy di-ureasil has T1 and T2

ratios higher than the d-U(600)–Eu(nta)3bpy–AA sample,

indicating that the hydrolysis favours structures with

lower dimension compared with the AA solvolysis, in

agreement with SAXS results. The condensation degree

obtained, c = 1/3(%T1 + 2%T2 + 3%T3), is ca. 68 and 78%,

for d-U(600)–Eu(nta)3bpy and d-U(600)–Eu(nta)3bpy–AA,

respectively (Table 1). Compared to the undoped sample, the

condensation degree decreases, suggesting that there are some

interactions between the complex and the matrix (urea cross-

links) and that the relative larger molecular size for complex

sterically prevents the polycondensation process and results in

the lower condensation degree. This is in agreement with the

results previously obtained when the lanthanide complex was

formed a in silicon–base matrix via a conventional sol–gel

route.28

The 13C CP MAS NMR spectra of d-U(600)–AA,

d-U(600)–Eu(nta)3bpy and d-U(600)–Eu(nta)3bpy–AA are

shown in Fig. 3b. Peak assignments were done previously

elsewhere for undoped di-ureasil.14 The most intense peak at

about 70.6 ppm is attributed to the main-chain carbons of

ethylene oxide, whereas the shoulders at about 75.0 ppm are

originated from the main-chain carbons of propylene oxide.

These two bands are partially overlapped. The peaks at ca.

45.5, 24.5 and 10.8 ppm are characteristics of the (CH2)3

aliphatic chains. The weak peak at approximately 160 ppm is

associated with the CLO groups in the urea-linkages.27,28 The

signal at 18.4 ppm is assigned to different –CH3 groups of the

polymer chains, while the shoulder at 17.4 ppm is ascribed

to the ethoxy groups of the carbons. The spectra of the two

doped hybrids exhibit similar profiles compared with those

of the undoped di-ureasil, with the exception of a decrease in

the intensity of the peaks at ca.160 ppm and those arising

from (CH2)3 aliphatic chains. This supports the interaction

between the complex and the CLO groups of the hybrid

matrix, in agreement with the results obtained from FT–IR

and 29Si NMR.

Photoluminescence measurements

Fig. 4 shows the room temperature PL spectra for several

excitation wavelengths for the d-U(600)–Eu(nta)bpy–AA. The

spectra are mainly composed of a series of straight lines typical

of the Eu3+ energy level structure, assigned to the 5D0 A 7F0–4

transitions. The spectra also present a very low-intensity

emission band in the blue-green spectral region. Such emission

must arise from the ligands triplet levels or from the hybrid

host emitting levels. The undoped di-ureasil presents a large

broad band in the blue spectral region, similar to that

previously observed in other amine-functionalized organic–

inorganic hybrids prepared through solvolysis and hydro-

lysis.3,14–17 This band was previously attributed to the

Fig. 3 29Si MAS NMR spectra (a) and 13C CP/MAS NMR spectra (b) of (i) d-U(600)–AA, (ii) d-U(600)–Eu(nta)3bpy and (iii) d-U(600)–

Eu(nta)3bpy–AA.

Table 1 29Si NMR chemical shifts (ppm), population of different Tn

(n = 1,2,3) species (%), and degree of condensation (%), c, of d-U(600)–AA, d-U(600)–Eu(nta)3bpy and d-U(600)–Eu(nta)3bpy–AA.

Samples T1 T2 T3 c

d-U(600)–AA 255.5 (11) 258.9 (28) 266.9 (61) 8314

d-U(600)–Eu(nta)3bpy 252.5 (31) 258.5 (34) 266.6 (35) 68d-U(600)–Eu(nta)3bpy–AA 252.6 (15) 258.3 (37) 266.4 (48) 78


convolution of donor–acceptor pair recombinations that

occur in the NH groups of the urea linkages and in the

siliceous nanodomains.3,14–17 It was recently proposed that

the mechanism responsible for the NH-related component is

associated with photoinduced proton-transfer between NH2+

and N2 defects, whereas the PL mechanism subjacent to the

component associated with the siliceous nanodomains involves

oxygen-related defects.18 One particular property of such

hybrid-related emission is the strong emission energy depen-

dence on the selected excitation energy, in such a way that an

emission red shift is observed when the excitation wavelength

is increased.3,14–17 Whilst the Eu3+-based di-ureasils’ broad

band appears in the same spectral region as that of the

undoped host emission it is almost independent of the

excitation energy. This suggests that the broad band in

the d-U(600)–Eu(nta)3bpy–AA hybrids may be mainly

ascribed to emission arising from the ligands excited states.

In order to further verify it, the PLE spectra for the Eu3+-

based hybrids were monitored along the large broad band

(see inset of Fig. 4) and compared with that characteristic of

the undoped host.15 The spectra are formed of two low-

intensity peaks at ca. 275 and 320 nm and of a main band

peaking around 395 nm. These PLE characteristics differ

from those of the hybrid host, whose PLE spectra are

formed of a large broad between 300 and 500 nm,3,15

reinforcing that the broad band in the PL spectra of the

d-U(600)–Eu(nta)3bpy–AA must be essentially related with the

ligands excited states.

The non-observation of the emission component associated

with the hybrid host’s emitting centres clearly suggests the

presence of active energy transfer channels between them

and the Eu3+ ions, a claim in agreement with the suggested

interaction between the Eu3+ ions and the urea cross-links

through the carbonyl groups. However, for similar di-ureasils

incorporating Eu(CF3SO3)3 (in which the local coordination

also involves the carbonyl groups) the host-related emission

is only absent for large amounts of incorporated salt

(approximately three times).19–21 Therefore, in the di-ureasils

incorporating Eu(nta)3?bpy the efficiency of the energy

transfer processes should be larger compared with that of

the di-ureasils with Eu(CF3SO3)3. We will return to this

point later.

The PLE spectra of the Eu3+-di-ureasils and of the

Eu(nta)3?bpy complex were also monitored around the more

intense line of the 5D0 A 7F2 transition, as exemplified in

Fig. 5a. The spectra are similar, presenting a large broad band

between 240 and 450 nm with three main components peaking

at ca. 275, 325 and 395 nm, respectively, and one very low-

intensity peak ascribed to an intra-4f6 line. The origin of

these PLE components may be related both to the ligands

and the hybrid host emitting centres. The spectrum of the

Eu(nta)3?bpy complex is red-shifted with respect to that of the

Eu3+-based di-ureasil hybrids, presenting mainly a large broad

band peaking around 395 nm, reinforcing that the incorpora-

tion of the complex in the d-U(600) changed the Eu3+

surroundings. Moreover, the lower relative intensity of the

Eu3+ line (,0.5%) strongly suggests that the metal ions are

essentially excited via an efficient sensitized process rather

than by direct population of the intra-4f6 levels. It is worth

noting that the efficiency of the sensitized process in the

hybrids is almost constant in a very broad UV-Vis spectral

region (250–395 nm).

With the goal of further investigating the changes induced in

the Eu3+ first coordination shell due to complex incorporation

and the influence of the synthesis method, Fig. 5b compares

the Eu3+ PL features of the two hybrids with those of the

complex Eu(nta)3?bpy. All the spectra show the Eu3+ charac-

teristic 5D0 A 7F0–4 transitions. Comparing the emission

Fig. 4 Room temperature PL spectra excited at (1) 272, (2) 325, (3)

370, and (4) 395 nm of d-U(600)–Eu(nta)3bpy–AA. The inset shows

the room temperature PLE monitored at (1) 435, (2) 467, and (3)

500 nm of the same hybrid. The higher intensity in the lower

wavelength region (250–280 nm) is due to correction effects.

Fig. 5 (a) Room temperature PLE spectra monitored around 615 nm

for the (i) d-U(600)–Eu(nta)3bpy, (ii) d-U(600)–Eu(nta)3bpy–AA, and

(iii) Eu(nta)3?bpy; (b) room temperature PL spectra for the previous

samples excited around 382 nm for the hybrids, and around 398 nm for

the complex. The inset shows the 5D0 A 7F0 transition and (circles) the

respective fit to a single Gaussian function.


spectra of the hybrids, no significant changes were observed

in the energy peak position and in the full-width at half-

maximum (fwhm) of each line. The higher number of Stark

components (e.g., 3 and 4 clearly express components

for the 5D0 A 7F1,2 transitions) and the higher intensity of

the 5D0 A 7F2 transition suggests that the Eu3+ local

coordination site in the hybrids has lower symmetry without

an inversion centre. The similarity between the spectra of the

two hybrids points out that the synthesis route did not

significantly affects the Eu3+ first coordination shell. However,

different PL features were observed in the spectra of

Eu(nta)3?bpy, namely in the energy and fwhm of the 5D0 A7F0–4 transitions. For instance, the energy (E00) and the fwhm

(fwhm00) of the non-degenerated 5D0 A 7F0 line transition

were estimated by deconvoluting the PL spectra of the di-

ureasils and the complex, assuming a single Gaussian function.

The analysis to the 5D0 A 7F0 energy is quite important since

E00 is usually correlated with the sum of the nephelauxetic

effects arising from the Eu3+ first neighbours.19–21,29,30 A good

quality fit was obtained for all the spectra, independently to

the excitation wavelength used (see inset in Fig. 5b), revealing

similar results for the di-ureasils, E00 = 17257.6 ¡ 0.1 cm21

and fwhm00 = 26.2 ¡ 0.2 cm21, and lower values for the

Eu(nta)3?bpy precursor complex, E00 = 17232.6 ¡ 0.1 cm21

and fwhm00 = 5.6 ¡ 0.1 cm21. These results unequivocally

suggest one Eu3+ local coordination site in all of the

materials, and the substantial increase of the fwhm00 values

characteristic of the hybrids, relative to that of Eu(nta)3?bpy,

indicates that the incorporation of the complex into the

d-U(600) host induces a higher distribution of Eu3+-closed

symmetry sites. This is additional evidence supporting

the effective interaction between the Eu3+ ions and the di-

ureasil host.

The 5D0 lifetime for all the materials was monitored within

the 5D0 A 7F2 transition and at 395 nm excitation wavelength

for different temperatures between 14 and 300 K. All the decay

curves are well reproduced by a single exponential function,

revealing lifetime values around 0.711 ¡ 0.003 and 0.824 ¡

0.004 ms (14 K) and 0.589 ¡ 0.003 and 0.698 ¡ 0.006 ms

(300 K), for the d-U(600)–Eu(nta)3bpy–AA and d-U(600)–

Eu(nta)3bpy, respectively. For both hybrids, the lifetimes

display the same temperature dependence, being approxi-

mately constant in the temperature range 14–200 K. The

room temperature decay curve of the Eu(nta)3?bpy complex

revealed a typical single exponential behaviour with a

lifetime value around 0.620 ms, which is different from that

obtained for the hybrids, reinforcing the occurrence of

structural changes as the complex is incorporated into the

hybrid host.

Emission quantum yield and 5D0 quantum efficiency

Quantum yields of 15.0 ¡ 1.5 and 6.0 ¡ 0.6% were measured

for d-U(600)–Eu(nta)3bpy and d-U(600)–Eu(nta)3bpy–AA,

respectively. As the Eu3+ first coordination shell is essentially

independent of the synthesis route adopted, the significant

decrease in the emission quantum yield of d-U(600)–

Eu(nta)3bpy–AA might be explained by a different interaction

between the hybrid emitting centres and the nta and bpy

ligands, favouring thus a larger non-radiative transition

probability, compared to that of d-U(600)–Eu(nta)3bpy. In

fact, comparing the three excitation spectra of Fig. 5a with

that of the Eu(nta)3?2H2O complex,31 we can conclude that the

solvolysis process changes essentially the ligands-to-Eu3+

energy transfer channels involving nta ligand levels (wave-

length region lower than 300 nm). In order to further check if

the dependence of the quantum yield on the synthesis route

is basically connected with that modification in the nta

ligand levels, the 5D0 quantum efficiency, q, the radiative, kr,

and non-radiative, knr, transition probabilities were calculated

for the hybrids and compared with those of the Eu(nta)3?bpy

complex.

The lanthanide luminescence quantum efficiency is defined

by the competition between knr and kr processes:

q~kr

krzknr(2)

For the Eu3+, the value of kr can be estimated using20,21

kr~A0{1Bv0{1

S0{1

X6

j~0

S0{j

Bv0{j

(3)

where j represents the final (7F0–6) levels, S is the integrated

intensity of the particular emission lines and hv stands for the

corresponding transition energies. A0A1 is the Einstein’s

coefficient of spontaneous emission between the 5D0 and the7F1 Stark levels. The branching ratio for the 5D0 A 7F5,6

transitions must be neglected as they are not observed

experimentally. Therefore, we can ignore their influence in

the depopulation of the 5D0 excited state. The 5D0 A 7F1

transition does not depend on the local ligand field seen by

Eu3+ ions and, thus, may be used as a reference for the whole

spectrum, in vacuo A(5D0 A 7F1) = 14.65 s21.32 An average

index of refraction of 1.5 was considered for both compounds

leading to A(5D0 A 7F1) ca. 50 s21.20,21,33 Appropriate analysis

of the Eu3+ emission lines yielded to the values collected

in Table 2 for the kr,, knr, and q. The value of knr can be

calculated using the experimental lifetime of the 5D0 state. The

lower q value found for the hybrid prepared via solvolysis is

mainly due to an increase in knr (ca. 60%), meaning that a

more efficient non-radiative channel involving the first excited

triplet state exists in the solvolysis derived hybrid. The higher

temperature dependence of the 5D0 lifetime for this hybrid,

with respect to that found for the di-ureasil prepared through

conventional sol–gel, is an additional argument supporting

the hypothesis that the phonon-assisted back-transfer from

Eu3+ to the triplet states is more important for the solvolysis

derived material.

Table 2 5D0 quantum efficiency (q) for the d-U(600)–Eu(nta)3bpy,d-U(600)–Eu(nta)3bpy–AA and for the Eu(nta)3?bpy complex. Thenumbers in parenthesis indicate the excitation wavelength used.

Eu(nta)3bpy(395 nm)

U(600)–Eu(nta)3bpy(383 nm)

U(600)–Eu(nta)3bpy–AA (370 nm)

kexp/ms21 1.613 1.432 1.699knr/ms21 0.797 0.580 0.925kr/ms21 0.816 0.852 0.774q (%) 50.6 59.5 45.6


It is worth noting, however, that the decrease of the 5D0

quantum efficiency of the di-ureasil prepared through AA

solvolysis (approximately 30% relative to that of the hybrid

synthesised by conventional hydrolysis), is not enough to

describe the larger reduction of its overall emission quantum

yield (approximately 150%).

Conclusions

Di-ureasil organic–inorganic hybrids incorporating

Eu(nta)3?bpy were prepared by AA solvolysis and the

conventional hydrolysis–polycondensation sol–gel route.

SAXS, XRD, FT-IR, 29Si and 13C MAS NMR results

demonstrate an effective interaction between the

Eu(nta)3?bpy complex and the carbonyl groups of the urea

linkages. Moreover the dimension and the degree of condensa-

tion of the siloxane nanodomains noticeably depend on this

synthesis route. The PL spectra of these Eu(III)-based di-

ureasils display essentially the typical 5D0 A 7F0–4 intra-4f6

Eu3+ transitions. Indeed the broad emission band typical of

amine-functionalized hybrid hosts was not detected suggesting,

therefore, the activation of energy transfer channels between

the hybrid host’s emitting centres and the Eu(III) ions. The

efficiency of these energy transfer channels depends on the

synthesis strategy adopted as the overall emission quantum

yield and the 5D0 quantum efficiency strongly decrease for the

di-ureasil prepared through solvolysis, relative to that synthe-

sised by conventional hydrolysis (from 15 to 6% and from 60

to 46%, respectively). Furthermore, as the Eu3+ first coordina-

tion shell is essentially independent of the synthesis method,

the changes detected on the hybrid host-to-ligands-to-Eu3+ ion

energy transfer channels should be primarily induced by the

interaction between the hybrid emitting centres (NH groups of

the urea linkages and oxygen-related defects of the siliceous

nanodomains) and the nta and bpy ligands, favouring,

therefore, larger non-radiative transition probability in the

di-ureasils prepared through AA solvolysis. Thus, the tuning

of the efficiency of the hybrid host-to-ligands-to-Eu3+ ion

energy transfer channels with the subsequent changes in the

overall emission quantum yields might be achieved by

promoting differences in the dimensionally and structure of

the siloxane domains through the embracing of different

synthesis strategies. A suitable choice of ligands that better

sensitize the Eu3+ emission together with a fine control of

the synthesis process attending to the optimization of the

radiative hybrid host-to-Eu3+ energy transfer efficiency,

definitely endorse the design of nanohybrids with better

emission conversion performances and higher absolute quan-

tum yields.

Acknowledgements

The authors acknowledge the assistance of K. Dahmouche,

C. V. Santilli and LNLS staff during SAXS measurements

and the collaboration of J. Rocha for NMR results. This

work was supported by FEDER and Fundacao para a

Ciencia e Tecnologia, POCTI/CTM/46780/02. L.S.F. and

N.J.O.S. thank FCT for post-doctoral (SFRH/BPD/5657/

2001) and PhD grants (SFRH/BD/10383/2002).

References

1 L. Hench and J. K. West, Chem. Rev., 1990, 90, 33.2 C. Sanchez and B. Lebeau, Mater. Res. Soc. Bull., 2001, 26, 377

(Hybrid Organic–Inorganic Materials).3 L. D. Carlos, R. A. Sa Ferreira and V. de Zea Bermudez, in

Handbook of Organic–Inorganic Hybrid Materials andNanocomposites, ed. H. S. Nalwa, American Scientific Publishers,Morth Lewis Way, California, Vol. 1, Ch. 9, 2004, p. 353.

4 H. Maas, A. Currao and G. Calzaferri, Angew. Chem., Int. Ed.,2002, 41, 2495.

5 R. Reisfeld, Struct. Bonding, 2004, 106, 209.6 C. Sanchez, B. Lebeau, F. Chaput and J.-P. Boilot, Adv. Mater.,

2003, 15, 1969.7 G. F. de Sa, O. L. Malta, C. de Mello Donega, A. M. Simas,

R. L. Longo, P. A. Santa-Cruz and E. F. da Silva, Jr., Coord.Chem. Rev., 2000, 196, 165.

8 D. W. Dong, S. C. Jiang, Y. F. Men, X. L. Ji and B. Z. Jiang, Adv.Mater., 2000, 12, 646.

9 (a) L. R. Matthews and E. T. Knobbe, Chem. Mater., 1993, 5,1697; (b) A. C. Franville, D. Zambon, R. Mahiou and Y. Troin,Chem. Mater., 2000, 12, 428; (c) M. Bredol, T. Justel and S. Gutzov,Opt. Mater., 2001, 18, 337; (d) H. H. Li, S. Inoue, K. Machida andG. Adachi, Chem. Mater., 1999, 11, 3171; (e) K. Binnemans,P. Lenaerts, K. Driesen and C. Gorller-Walrand, J. Mater. Chem.,2004, 14, 191; (f) G. D. Qian, M. Q. Wang, M. Wang, X. P. Fanand Z. L. Hong, J. Mater. Sci. Lett., 1997, 16, 322; (g) L. S. Fu,Q. G. Meng, H. J. Zhang, S. B. Wang, K. Y. Yang and J. Z. Ni,J. Phys. Chem. Solids, 2000, 61, 1877; (h) I. L. V. Rosa, O. A. Serraand E. J. Nassar, J. Lumin., 1997, 72–74, 532; (i) Q. H. Xu, L. S. Fu,L. S. Li, H. J. Zhang and R. R. Xu, J. Mater. Chem., 2000, 10,2532; (j) Q. H. Xu, L. S. Li, X. S. Liu and R. R. Xu, Chem. Mater.,2002, 14, 549; (k) M.-S. Zhang, W. Yin, Q. Su and H.-J. Zhang,Mater. Lett., 2002, 57, 940.

10 H. J. Im, C. Willis, A. C. Stephan, M. D. Pawel, S. Saengkerdsuband S. Dai, Appl. Phys. Lett., 2004, 84, 2448.

11 E. Stathatos, P. Lianos, U. Lavrencic-Stangar and B. Orel, Adv.Mater., 2002, 14, 354.

12 (a) V. de Zea Bermudez, L. D. Carlos, M. C. Duarte, M. M. Silva,C. J. R. Silva, M. J. Smith, M. Assuncao and L. Alcacer, J. AlloysCompd., 1998, 275–277, 21; (b) V. de Zea Bermudez, L. D. Carlosand L. Alcacer, Chem. Mater., 1999, 11, 569.

13 T. Brankova, V. Bekiari and P. Lianos, Chem. Mater., 2003, 15,1855.

14 L. S. Fu, R. A. Sa Ferreira, N. J. O. Silva, L. D. Carlos, V. deZea Bermudez and J. Rocha, Chem. Mater., 2004, 16, 1507.

15 L. D. Carlos, V. de Zea Bermudez, R. A. Sa Ferreira, L. Marquesand M. Assuncao, Chem. Mater., 1999, 11, 581.

16 L. D. Carlos, R. A. Sa Ferreira, V. de Zea Bermudez andS. J. L. Ribeiro, Adv. Funct. Mater., 2001, 11, 111.

17 L. D. Carlos, R. A. Sa Ferreira, R. N. Pereira, M. Assuncao andV. de Zea Bermudez, J. Phys. Chem. B, 2004, 108, 14924.

18 R. Moleski, E. Stathatos, V. Bekiari and P. Lianos, Thin SolidFilms, 2002, 416, 279.

19 L. D. Carlos, R. A. Sa Ferreira, V. de Zea Bermudez, C. Molina,L. A. Bueno and S. J. L. Ribeiro, Phys. Rev. B, 1999, 60, 10042.

20 L. D. Carlos, Y. Messaddeq, H. F. Brito, R. A. Sa Ferreira, V. deZea Bermudez and S. J. L. Ribeiro, Adv. Mater., 2000, 12, 594.

21 R. A. Sa Ferreira, L. D. Carlos, R. R. Goncalves, S. J. L. Ribeiroand V. de Zea Bermudez, Chem. Mater., 2001, 13, 2991.

22 R. G. Charles and A. Perrotto, J. Inorg. Nucl. Chem., 1964, 26,373.

23 L. C. Thompson, F. W. Atchison and V. G. Young, J. AlloysCompd., 1998, 275–277, 765.

24 A. Bril and A. W. De Jager-Veenis, J. Electrochem. Soc., 1976, 123,396.

25 (a) K. Dahmouche, C. V. Santilli, S. H. Pulcinelli andA. F. Craievich, J. Phys. Chem. B, 1999, 103, 4937; (b)K. Dahmouche, L. D. Carlos, V. De Zea Bermudez,R. A. Sa Ferreira, C. V. Santilli and A. F. Craievich, J. Mater.Chem., 2001, 11, 3249.

26 V. de Zea Bermudez, R. A. Sa Ferreira, L. D. Carlos, C. Molinaand S. J. L. Ribeiro, J. Phys. Chem. B, 2001, 105, 3378.

27 A. C. Franville, R. Mahiou, D. Zambon and J. C. Cousseins, SolidState Sci., 2001, 3, 211.


28 M. C. Goncalves, V. de Zea Bermudez, R. A. Sa Ferreira,L. D. Carlos, D. Ostrovskii and J. Rocha, Chem. Mater., 2004, 16,2530.

29 S. T. Frey and W. De Horrocks, Jr., Inorg. Chim. Acta, 1995, 229, 383.30 O. L. Malta, H. J. Batista and L. D. Carlos, Chem. Phys., 2002,

282, 21.

31 L. D. Carlos, C. De Mello Donega, R. Q. Albuquerque,S. Alves, Jr., J. F. S. Menezes and O. L. Malta, Mol. Phys.,2003, 101, 1037.

32 M. H. V. Werts, R. T. F. Jukes and J. W. Verhoeven, Phys. Chem.Chem. Phys., 2002, 4, 1542.

33 M. F. Hazenkamp and G. Blasse, Chem. Mater., 1990, 2, 105.


Structure–photoluminescence relationship in Eu(iii) β-diketonate-based organic–inorganic hybrids. Influence of the synthesis method: carboxylic acid solvolysis versus conventional

Documents