Europium(III) complexed by HPSEC size-fractions of a vertisol humic acid: Small differences evidenced by time-resolved luminescence spectroscopy

Article abstract of DOI:10.1016/j.saa.2010.12.075

The size fractionation of a humic acid (HA) by high performance size exclusion chromatography (HPSEC) was used as a proxy for the filtration effect during HA transport through a porous medium with minimum specific chemical interactions. The modification of the Eu(III)-HA complexes' formation with the different size-fractions, as compared to the bulk HA, was studied in time-resolved luminescence spectroscopy (TRLS). Clear modifications in Eu(III)-HA complexes' structures were shown and related to the molecular characteristics of the separated size-fractions. The properties of most of size-fractions did not induce a major alteration of the affinity towards Eu(III). Only the most hydrophilic fractions eluted in the tail of the chromatographic peak, representing about 11% of total fractions-weight, gave some significantly different parameters. Using a simplistic complexation model, it was found that the available complexation sites decreased with the size reduction of humic fractions.

Full text of DOI:10.1016/j.saa.2010.12.075

Spectrochimica Acta Part A 78 (2011) 1173–1179
Contents lists available at ScienceDirect
Spectrochimica Acta Part A: Molecular and
Biomolecular Spectroscopy
journal homepage: www.elsevier.com/locate/saa
Europium(III) complexed by HPSEC size-fractions of a vertisol humic acid: Small
differences evidenced by time-resolved luminescence spectroscopy
Pascal E. Reillera,∗, Julien Brevetb, Antonio Nebbiosoc, Alessandro Piccoloc
a CE Saclay, CEA/DEN/DANS/DPC/SECR, Laboratoire de Spéciation des Radionucléides et des Molécules, Bâtiment 391, P.C. 33, F-91191 Gif-sur-Yvette Cedex, France
b Université d’Evry Val d’Essonne, Laboratoire Analyse et Modélisation pour la Biologie et l’Environnement, CNRS UMR 8587, Bâtiment Maupertuis, Boulevard Franc¸ ois Mitterrand,
F-91025 Evry Cedex, France
c Dipartimento di Scienze del Suolo, della Pianta, dell’Ambiente e delle Produzioni Animali, Università di Napoli Federico II, Via Università 100, I-80055 Portici, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
The size fractionation of a humic acid (HA) by high performance size exclusion chromatography (HPSEC)
was used as a proxy for the filtration effect during HA transport through a porous medium with minimum
specific chemical interactions. The modification of the Eu(III)–HA complexes’ formation with the different
size-fractions, as compared to the bulk HA, was studied in time-resolved luminescence spectroscopy
(TRLS). Clear modifications in Eu(III)–HA complexes’ structures were shown and related to the molecular
characteristics of the separated size-fractions. The properties of most of size-fractions did not induce a
major alteration of the affinity towards Eu(III). Only the most hydrophilic fractions eluted in the tail of the
chromatographic peak, representing about 11% of total fractions-weight, gave some significantly different
parameters. Using a simplistic complexation model, it was found that the available complexation sites
decreased with the size reduction of humic fractions.
Received 26 July 2010
Received in revised form
13 December 2010
Accepted 25 December 2010
Keywords:
Humic acid
Europium
Fractionation
Complexation
Luminescence
Fluorescence
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
size-exclusion chromatography (HPSEC) [17,18], electrospray ion-
ization mass spectrometry [19,20], electrophoresis [21], or nuclear
Humic substances (HS), mainly composed of humic acids (HA)
and fulvic acids (FA), are the main surrogate of natural organic mat-
ter. They play an important role in the transport of lanthanides (Ln)
and actinides (An) in the environment [1–4]. The exact nature of
the Ln/An–HS interactions during transport and especially during
sorption mechanisms is still unclear. Particularly, the fractiona-
tion of natural organic matter on mineral surfaces introduces some
difficulties in modelling sorption of metals in ternary system, i.e.,
metal/HA/surface. In the case of FA, the modelling approach was
partly successful, although the FA’s sorptive behaviour may be
explained with electrostatic considerations [5,6], and assumed to
be more or less similar to that of simple organic molecules [7].
The FA sorption is stable or tends to decrease with ionic strength
[8–10]. Conversely, HA have a different behaviour since a sorption
increase on oxides is occurring [8,10,11]. Hence, it is still difficult
to propose sound interpretations of HA effects on the modelling of
ternary systems in a wide parametric space [12,13].
magnetic resonance [22,23], on the fact that HS are formed
of nanometric sub-entities that associate in larger bodies in
the case of HA. The cohesion of these associations is assured
through short range interactions, which can be largely modi-
fied in contact with strong sorbing mineral surfaces like Fe- or
Al-oxides [24–27].
During transport in a particular environment, HA associations
can be submitted to different interactions. First, a ‘simple’ physical
filtration effect, which can alter the HA structure as by performing
a size-exclusion ‘sorting out’ different constituents making up the
associations. Second, a sorptive fractionation induced by chemical
interactions with minerals in the environment, which induces a
separation of constituents based on their affinity to the mineral
phase [24–27]. In fact, it is difficult to obtain these parameters
independently. Christl et al. showed that the functionality [28]
and metal binding [29] of different size-fractions of a HA from
ultrafiltration were comparable within a narrow range. Conversely,
Hur and Schlautman [25], Reiller et al. [27] and Janot et al. [30]
showed that clear modification occurred on mass-distribution and
acid–base properties of HA after sorptive fractionation. In addition,
Claret et al. [26] noted that the influence of sorptive fractiona-
tion of HS on ␣-Al2O3 was important from the evolution of the
chemical environment of europium(III) by time-resolved lumines-
cence spectroscopy (TRLS). And recently, Janot et al. [31] showed
The intimate structure of HS is still a matter of debate. Nev-
ertheless, a consensus is emerging from small-angle scattering
methods [14], atomic force microscopy [15,16], high-performance
∗
Corresponding author. Tel.: +33 1 6908 4312; fax: +33 1 6908 5411.
E-mail address: pascal.reiller@cea.fr (P.E. Reiller).
1386-1425/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.saa.2010.12.075

1174
P.E. Reiller et al. / Spectrochimica Acta Part A 78 (2011) 1173–1179
that the chemical environment of Eu(III) was greatly affected dur-
ing sorption in a ternary Eu(III)/HA/␣-Al2O3 system. The relative
importance of the physical filtration effect compared to the sorp-
tive fractionation that occurs simultaneously during transport is
still to be evaluated.
The aim of this work was to study the differences in chemical
environment of Eu(III), if any, when complexed by HPSEC-
separated humic size-fractions, as a proxy for the physical filtration
effect on HA during transport through a porous medium with min-
imum specific chemical interactions.
2. Materials and methods
2.1. Humic acid
The humic acid from Ginchi (GHA) was extracted from a recently
cultivated forested vertisol of Ethiopian highlands (09◦01ꢀN,
38◦20ꢀE, 2300 m asl, see Ref. [32], for further details). Soil proper-
ties, isolation and purification procedures, and GHA characteristics
are reported elsewhere [32,33]. The previously freeze-dried GHA
was suspended in distilled water and titrated to pH 7 with a CO2-
free solution of 0.5 mol L−1 KOH by an automatic titrator (VIT 90
Videotitrator, Radiometer, Copenhagen, Denmark) under N2 atmo-
sphere and stirring. After having reached the constant pH 7, the
solution containing potassium-humates was left under titration
for 2 more hours, filtered through a Millipore 0.45 ␮m, and freeze-
dried. Potassium humates were pretitrated at pH 7 to limit HPSEC
interferences.
Fig. 1. HPSEC profile of the Ginchi humic acid. The vertical bars mark the time inter-
vals during which the ten fractions were collected (a), and mass of fractions (bar
graph, left hand side ordinate) and cumulated percentage (square, right hand side
ordinate) of the ten fractions of Ginchi HA separated by preparative HPSEC (b).
2.2. Preparative HPSEC separation
The mobile phase for HPSEC consisted of a 242 mg L−1 Na2HPO4,
248 mg L−1 NaH2PO4·2H2O, 820 mg L−1 CH3CO2Na, milli-Q water
solution at pH 7 and added with 300 mg L−1 NaN3 as bacterio-
static agent. The same solution was used to dissolve the potassium
humates to a concentration of 600 mg L−1. The humic solution was
filtered through glass microfibre filters (Whatman GF/C) and loaded
into a rheodyne rotatory injector, equipped with a 5 mL sample
loop. The HPSEC system consisted of a Gilson autosampler model
231, a Gilson 305 pump, a preparative Biosep SEC-S-2000 (600 mm,
21.2 mm id) column, preceded by a Biosep SEC-S-2000 guard
column (78.0 mm, 21.2 mm id) both from Phenomenex (USA), a
Gilson 116 UV detector set at 280 nm, and a Gilson FC205 fraction
collector, to automatically collect humic fractions in continuous.
The elution flow-rate was set at 1.5 mL min−1 and all chromato-
graphic runs were automatically recorded by a Unipoint Gilson
Software. Ten fractions were separated with HPSEC elution: within
a 37.5–135.0 mL interval, nine fractions were collected changing
vial every 7.5 mL, whereas 30.0 mL was collected for the tenth frac-
tion. The ten isolated size-fractions were first freeze-dried to reduce
their volume, resuspended in 5 mL of deionised water, dialyzed
(Spectra/Por 6 dialysis tube, 1 kDa MW cut-off) against deionised
water, and freeze-dried again. Out of 110 injections of the HA
solution (330.0 mg), the total weight recovered in the ten isolated
fractions was about 90% (297.5 mg) of the initial injected weight.
The preparative HPSEC profile of the Ginchi HA showed two main
peaks (Fig. 1a). The first one (fractions G1–G4) encompassed the
majority of the mass distribution (55%, see Fig. 1b). Fraction G5
shows an increase in absorbance but it does not result in an increase
in mass percentage relative to fraction G4.
cause of the signal broadening in the chromatogram. Moreover,
one must not forget that our separation of humic fractions was
done in the preparative mode (5 mL of loading volume), that has an
even lower resolution than analytical HPSEC, as reported elsewhere
[22,23,36]. The preparative HPSEC has the goal to obtain quantita-
tive amounts of humic fractions which can be subjected to further
analysis more than determining the humic molecular size distribu-
tion. In the preparative mode also the notations of void and total
volumes loose the importance that is commonly attributed to ana-
lytical HPSEC. The important issue is to obtain highly reproducible
chromatograms in order to isolate meaningful and quantitative
fractions during the repetitive chromatographic runs. In prepar-
ative HPSEC mode, adsorption is concomitant with size-exclusion
separation to a larger extent (mL vs. ␮L of loading volume) than in
analytical HPSEC, thereby showing a longer tailing chromatogram.
2.3. Solution
Europium(III) stock solution was obtained from the dissolu-
tion of Eu2O3 (Johnson Matthey, 99.99%) in HClO4. It is assumed
that the humic complexation is complete with C(Eu) = 10 ␮mol L−1
and C(HS) = 200 mg L−1 at pH 5 [37], even if the ratios between
Eu(III) and the number of available sites in the different frac-
tions can be slightly different [28,29,38]. The ionic strength was
fixed with 0.1 mol L−1 (NaClO4), and pH 5 was adjusted using
freshly prepared NaOH and HClO4. The dissolution of GHA, bulk
and fractions, was done at pH 10 overnight and then adjusted
to pH 5. The pH measurements were done using a combined-
glass electrode (Radiometer Analytical XC111) calibrated for its
linear response with a 0.01 mol L−1 HClO4 solution, an equimo-
lar 0.02 mol L−1 NaH2PO4/Na2HPO4 solution and an equimolar
0.02 mol L−1 Na2CO3/NaHCO3 solution, all containing NaClO4 to
keep [Na+] constant at 0.1 mol L−1 (pH = 2.0, 6.8, and 9.9, respec-
Typically, the HPSEC chromatograms of humics reported in lit-
erature do not show the resolution that is proper of homogeneous
materials, since the large number of heterogeneous molecules,
which is accounted to be more than 10,000 in humic acids
[27,34,35], and their absorption spectra overlapping is the very

P.E. Reiller et al. / Spectrochimica Acta Part A 78 (2011) 1173–1179
1175
5
7
by Suwannee River FA (SRFA), and HA (SRHA) spectra previously
D
F
1
0
5
7
D
F
0
7
obtained [48]. It can be noted that the 5D0 → F0 transition, which
0
is not apparent in the Eu3+ spectrum for magnetic and electric
reasons, becomes important in Eu(III)–HS complexes due to loss
of symmetry [50,51]. The transition is slightly red-shifted in the
Eu(III)–GHA complex as compared to Eu(III)–SRFA and –SRHA com-
plexes. The maximum and full-width at mid-height (w) of this
transition are obtained from a fitting with a Lorentzian–Gaussian
SRFA
SRHA
GHA
Eu
peak [52], Ii = Imax,i
:
ꢀ
ꢁ
2
exp −0.5((ꢀi − ꢀ0,i)/ꢁLG,i
)
Ii = Imax,i
(1)
((ꢀi − ꢀ0,i)/ꢁLG,i
)
2 + 1
576
580
584
588
592
596
600
where Imax,i is the maximum intensity of the ith peak, ꢀ0,i is the
wavelength of this maximum, and ꢁLG,i is the ‘standard deviation’,
Wavelength (nm)
Fig. 2. Eu(II) 5D0
C(HS) = 200 mg L−1
→
7F0 and 5D0
→
7F0 transitions at C(Eu) = 10−5 mol L−1
,
which can be related to wi = 1.46 ꢁLG,i
.
,
pH 5, I = 0.1 mol L−1 (NaClO4), ꢀexc = 394.6 nm, D = 10 ␮s,
As shown by Reiller and Brevet [48], the mono-component fit of
the transition is satisfactory in consideration of the existing signal-
to-noise ratio (Fig. S1 of SI). The peak maximum gave the value of
ꢀEu-GHA = 579.26 nm with ꢁLG = 0.75 nm. The red-shift of the peak
maximum was significant in comparison to ꢀEu–SRHA = 579.03 nm
and ꢀEu–SRHA = 578.98 nm [48]. Application of the relation between
the number of coordinated ligands and wave number of the
W = 300 ␮s, 1800 lines mm−1 grating for different humic substances. Eu3+ (dot-dash),
Eu(III)–SRFA (dotted), and –SRHA (dashed) from Reiller and Brevet [48], and Eu-GHA
(full line, this study).
tively). The electrode filling solution was modified with NaClO4
0.1 mol L−1, NaCl 10−2 mol L−1 to prevent KClO4 precipitation in
the frit of the electrode. The pH 5 value was chosen to limit the
formation of hydroxo and carbonato complexes of Eu(III).
5
7F0 → D0 transition ꢂ0 (17263.5 cm−1), i.e., CN = 0.237 ꢃꢂ + 0.638
[53], yielded a CN = 3.6 0.7. This was slightly larger than for both
Eu–SRFA and –SRHA complexes, i.e., CN = 2.0 and 1.6 ( 0.7), respec-
tively [48]. This suggests that the chemical environment for Eu(III)
in complexes with GHA may be somewhat different from those
with Suwannee River extracts.
2.4. Time-resolved luminescence spectroscopy
Europium(III) has been used to probe its laser-induced lumines-
cence properties in contact with humic substances. This technique
has been used to study the Ln(III)/HS interactions, and the rationale
was to either determine interaction constants [37,39,40], or gather
information on the Ln/An(III) chemical environment [26,41–48].
The ꢁLG value, resulting in wGHA = 1.1 nm, compared well with
those obtained with Suwannee River extracts, i.e., ꢁLG,SRFA = 0.71 nm
(wSRFA = 1.04 nm) and ꢁLG,SRHA = 0.72 nm (wSRFA = 1.05 nm) [48] or
other humic extracts [42]. This further confirmed the large distri-
bution of this transition in humic solution, as compared to simple
molecules [54,55], or virus binding sites [56], and also revealed the
general large distribution of chemical environments in HS.
The observed luminescence corresponds to the 5D0 → F0 transi-
7
tion (electric and magnetic dipole forbidden, maximum around
7
580 nm), the 5D0 → F1 transition (magnetic dipole, maximum
7
around 593 nm), and the 5D0 → F2 ‘hypersensitive’ transition
7
7
The peak ratios 5D0 → F0/5D0 → F1 are also noticeably differ-
ent between Eu(III)–SRFA and –SRHA complexes and Eu(III)–GHA
complex. The value of the maximum intensity ratio of the peaks,
I(7F0/7F1) = 0.77, is lower than the one obtained for Eu(III)–SRFA
and –SRHA, i.e., 0.87 and 0.99, respectively [48]. This suggests that
the loss of symmetry in the case of Eu(III)–GHA is lesser than
for Eu(III)–SRFA and –SRHA. However, the area ratio (trapezoid
method) of Eu(III)–GHA, i.e., A(7F0/7F1) = 0.134, is comparable to
those of Eu(III)–SRFA and –SRHA that showed values of 0.139 and
(electric dipole, maximum around 615 nm). These emission lines
come from transitions of the 5D0 excited state to the ground 7Fj
manifold [49].
The excitation laser beam was generated by a 355 nm tripled
output of a Continuum Nd-YAG laser, coupled to an optical
parametric oscillator system (Panther II, Continuum, USA). The
wavelength was tuned to 394 nm, which corresponds to the
7F0 → L6 transition of Eu(III), providing about 1 mJ of energy in
5
7
0.150, respectively [48]. This illustrates that 5D0 → F0 FWMH is
a 5 ns pulse with a repetition rate of 10 Hz. After inner conversion,
the 5D1 level is transferring energy to the 5D0 level and, after 10 ␮s
delay, mainly the transitions to the 7Fj manifold can be observed.
Additionally, HS are able to absorb the laser emission at 394 nm
and part of the absorbed energy is transferred from the 3␲␲* triplet
level to the central europium ion. The time-resolved luminescence
signal is collected at 90◦ and focused into a Acton spectrometer (slit
1 mm) using either a 600 lines mm−1 or a 1800 lines mm−1 grating.
The signal is collected during a gate width W = 300 ␮s, after a gate
delay D = 10 ␮s after excitation by the laser flash. Emission spectra
were recorded using a CCD camera cooled at −15 ◦C.
7
slightly greater than Eu(III)–SRHA and –SRHA, and that 5D0 → F1
transition of Eu(III)–GHA is visually narrower than the Eu(III)–SRHA
transition. The magnetic dipole transition 5D0 → F1 is not sup-
7
posed to be largely influenced by the subtle changes in symmetry
[57] that seem to exist among the different chemical environments
provided by the different humic fractions. Only a close inspection
at 1800 lines mm−1 permits to reveal these modifications directly
in aqueous solution at ambient temperature.
7
3.2. 5D0 → F2 at 1800 lines mm−1
As for previous studies [47,48] the photo-degradation [43] of
both the original humic acid and HPSEC fractions can be neglected
under our conditions.
7
The 5D0 → F2 hypersensitive transition of the Eu(III)–GHA com-
plex is shown in Fig. 3, as compared to Eu(III)–HS complexes
previously obtained with Suwannee River HA and FA [47,48],
Gorleben AH, and Leonardite HA [47]. The modification of the
transition’s structure is even more spectacular, as compared to
3. Results and discussions
7
3.1. 5D0 → F0 and 5D0 → F1 transitions at 1800 lines mm−1
5D0 → F1. The most outstanding difference among the different
7
7
complexes was the shoulder ca. 612 nm that was more prominent
for Eu(III)-SRFA than for any other extract. The HS from purely
aquatic medium, i.e., SRHA (Fig. 3), and Kleiner Kranichsee bog
The 5D0 → F0 and 5D0 → F1 transitions of the Eu(III)–GHA
7
7
complex are presented in Fig. 2, together with Eu(III) complexed

1176
P.E. Reiller et al. / Spectrochimica Acta Part A 78 (2011) 1173–1179
4
Leonardite
SRFA
SRHA
Gohy573
GHA
Ginchi
Ginchi 1
Ginchi 2
Ginchi 3
Ginchi 4
Ginchi 5
Ginchi 6
Ginchi 7
Ginchi 8
Ginchi 9
Ginchi 10
3
2
1
0
-3,5
-3,0
-2,5
-2,0
-1,5
-1,0
-0,5
605
610
615
620
625
630
log (CHA g/L)
Wavelength (nm)
Fig. 4. Comparison of the complexometric titration curves of C(Eu) = 10−5 mol L−1
with increasing concentration of GHA fractions, pH 5, I = 0.1 mol L−1 (NaClO4), based
on area ratios.
Fig. 3. Eu(III) 5D0
→
7F2 transitions at C(Eu) = 10−5 mol L−1, C(HS) = 200 mg L−1, pH
5, I = 0.1 mol L−1 (NaClO4), ꢀexc = 394.6 nm, D = 10 ␮s, W = 300 ␮s, 1800 lines mm−1
grating for different humic acids. Eu–HS complexes with Eu(III)–SRFA (empty
square), –SRHA (empty triangle), –Leonardite HA (empty circle), and –Gorleben HA
(filled circle) from Brevet et al. [47], and Eu(III)–GHA (filled diamond, this study).
omitting the charges:
C(EuHA)
Eu + HA ꢀ EuHA,
ˇ(L mg−1) =
(2)
C(Eu) C(HA)
HA and FA (Germany, see Fig. 4 in Ref. [47]), exhibited mostly
the same transition shape, with a lower 612 nm shoulder. Finally,
Leonardite HA and Gorleben HA, together with commercial Aldrich
HA, showed a narrow-shaped transition with the lowest intensity
for 612 nm shoulders.
The origin of these different 612 nm shoulders is not easily set-
tled. One can be tempted to attribute this shoulder a Stark level. A
Stark level can be identified in solid state or in solution at low tem-
perature (4 K). In aqueous solution under ambient temperature we
do not have the possibility to decompose the signal in overlapping
Stark levels which are broadened with temperature. In addition,
HS are composed of aggregates of molecules which offer distribu-
tions of functionality that lead to overlapping. Hence, the origin of
the 612 nm shoulder is not easily settled and would need further
developments that are out of the scope of this study.
Our results show that the 612 nm shoulder of GHA is compara-
ble to other ‘terrestrial’ HA extracts. Nevertheless, the wavelength
span greater than 615 nm is not comparable to Leonardite HA nor to
Gorleben HA (see Ref. [47], and reference therein). This part of the
spectrum is positioned in between the two families of Eu(III)–HS
complexes. One can interpret these differences in peak’s shapes
with the degree of humification [58]. Since GHA is originated from
a forest soil, its diagenesis is presumably more advanced than in
Suwannee River, although it should be less advanced than in a lig-
nite likeLeonardite HA, orin anoxidizedextract from asedimentary
deposit like Gorleben [59].
where C(HA) = Cc C(HA)tot − C(EuHA).
The obtained parameters, namely the complexation constant
ˇ, and complexation capacity of the extracts Cc (mmol g−1), are
reported in Fig. 5. This model may appear too simplistic, but (i) the
amount of the different fractions was not sufficient to perform reli-
able acid–base titrations [30], and (ii) the application of generic
parameters, as those selected otherwise [28,29,61], may not be
directly applied to our case. As by our approach, the magnitude of
log ˇ (complexation strength) was not the most influencing param-
eter, while it was instead the number of available complexation
sites that explained most of modifications in GHA separated frac-
tions. Acomplexation constantwith amean value oflog ˇ = 6.7 0.7
(2ꢁ) can be proposed for the bulk GHA and the ten different size-
fractions. This value is in agreement with other determination
under comparable total europium concentration and pH [62–65].
A mean value of 0.18 0.05 mmol g−1 was calculated for the Cc
complexing capacity of bulk GHA, and the G1–G8 fractions. The Cc
values of 0.11 0.01 and 0.09 0.02 mmol g−1 were obtained for
the G9 and G10 fractions, respectively.
9
8
7
6
5
4
3
2
1
0
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
3.3. Comparisons of the complexometric titrations
log
The complexometric titration curves for the different extracts
are reported in Fig. 4. These curves were obtained from the evo-
Cc (mmol/g)
lution of the 5D0 → F2/5D0 → F1 peak area ratios [26,39]. Three
groups of fractions can be proposed. First, the fractions for which
complexation properties were mostly identical to the original bulk
GHA, i.e., G2–G4, G7, and G8. Second, G1, G5, and G7 showed only
slightly lower complexation strengths, as compared to bulk GHA.
Third, the most retarded G9 and G10 fractions appeared increas-
ingly different from bulk GHA. It can then be seen that most of
fractions could be fairly well compared with the bulk GHA, except
for the ‘smallest’ or ‘lightest’ fractions, which were significantly
different in terms of complexation strength.
7
7
The interactions can be quantified using a simplistic model [60],
where humic substances are considered as mixture of discrete sites
Fig. 5. Comparison of complexation parameters for the different HPSEC fraction of
GHA obtained from the titration curves in Fig. 4.

P.E. Reiller et al. / Spectrochimica Acta Part A 78 (2011) 1173–1179
1177
4.00
3.80
3.60
3.40
3.20
3.00
Smaller-sized HPLC-fractions are known to be increasingly
enriched in carbohydrate-like molecules, whereas a progressive
decrease in the amount of aromatic structure was found [23,36].
Thus, the observed decrease of A(7F2/7F1) may be due to both an
increasing number of carbohydrate molecules, exhibiting much
lower A(7F2/7F1) value [66], and a decreasing number of aro-
matic molecules [23]. This varying molecular distribution during
GHA fractionation modifies the indirect excitation of the central
3
europium(III) atom via the ␲␲* triplet level. We must also pre-
cise here, that no assumption is made hitherto on the functionality
of these aromatic molecules. The influence of aromatic moieties
on Eu(III) luminescence do not mean that europium(III) is linked
to phenolic moieties at pH 5, which is unlikely (ca. 1% of the total
Eu-HS complex following Ref. [67] for a generic HA), but that aro-
matic moieties contribute to the excitation. The energy transfer
could be due to a resonance phenomenon or due to an exchange
mechanism, i.e., Förster resonance energy transfer (FRET) or Dex-
ter energy transfer, respectively [68]. FRET can be operative over
distances up to 10 nm depending on the particular donor–acceptor
pair. The energy transfer according to Dexter requires an overlap
of the electron orbital of the donor and the acceptor. As the dis-
tance between the ligands and Eu(III) is relatively short, the Dexter
energy transfer could be favoured. So far, in the case of HA, the link
between the binding sites and the chromophores engaged in the
energy transfer has not been demonstrated unequivocally. Thus,
the resonance phenomena cannot be ruled out.
0.13
0.12
0.11
0.10
0.09
7
The evolution of the 5D0 → F2 transition was further checked
using a grating of 1800 lines mm−1. The main difference among
fractions was the 612 nm shoulder, as already discussed for G5 and
G9, although no systematic evolution could be revealed (data not
shown).
GHA G1 G2 G3 G4 G5 G6 G7 G8 G9 G10
Sample
Fig. 6. Evolution of A(7F2/7F1) (up) and A(7F0/7F1) (down) of Eu(III) complexed
by the different Ginchi fractions: C(Eu) = 10−5 mol L−1, C(HS) = 200 mg L−1, pH 5,
I = 0.1 mol L−1 (NaClO4), 600 lines mm−1 grating. Original data in Fig. S4 of SI.
3.5. Luminescence decay-times analyses of GHA
The luminescence decays of the Eu(III)–HS complexes are gen-
erally described by a bi-exponential function, which is described
for our fully integrative system by:
These findings may be related to the modification of HS dur-
ing the size fractionation. The low molecular-weight molecules are
assumed to be more retarded in the gel structure. When the low
molecular weight molecules finally reaches the outlet, their com-
plexation strength seems to be weaker than that of materials that
has move more freely through the gel. This effect is nevertheless
less intense in this case than observed in a previous work after the
sorptive HS fractionation on ␣-alumina [26].
n
n
ꢃ
ꢄ
ꢅ
ꢄ
ꢅꢀ
ꢄ
ꢅꢁ
ꢂ
ꢂ
D+W
t
D
W
ꢄi
F
=
Fio exp
−
dt =
Fioꢄi exp
−
1 − exp
−
ꢄ
ꢄ
i
i
D
(3)
i=1
i=1
ꢆ
ꢄ
ꢅꢀ
ꢄ ꢅꢁ
ꢄ
ꢅꢀ
ꢄ
ꢅꢁ
−D
W
D
W
F = Fo x1ꢄ1 exp
1 − exp
+(1 − x1)ꢄ2 exp
−
1 − exp
−
ꢄ
ꢄ
ꢄ
ꢄ
1
1
2
2
where F is the luminescence signal, Fo, and ꢄi are the initial lumi-
nescence and the decay time of compionent i, respectively, x1is the
proportion of the first decay in the global signal, D is the delay
after the laser excitation, and W is the gate width. The possible
influence of photochemical reactions of humic substances on the
complexation strength can be neglected out under our experimen-
tal conditions as seen previously [48,69].
3.4. Evolutions of spectra for the different fractions
The chemical environments provided by the different size-
fractions were also estimated from changes in luminescence
spectra of diffe−re1nt complexes. Spectra of solutions contain-
The decay time evolution was obtained from the 5D0 → F2 peak
7
area for different Eu(III)–GHA complexes at delays between 10
and 505 ␮s (Fig. S3 and Table S1 of SI). The inevitable correla-
tion between the parameters is shown in Table S1 of the SI. This
showed, as for other extracts [40,46–48], a bi-exponential decay
with ꢄ1 = 44 2 ␮s and ꢄ2 = 184 3 ␮s. The bi-exponential decay
can only happens if two deexcitation processes from two differ-
ent excited states are occurring. In our case the deexcitation states
seem to be very close in structure [48], and ꢄ1, seemed to be char-
acteristic of a fast reaction in the excited state for Ln(III) and An(III)
[48,70,71]. Hence this ‘species’ only occurs during the excitation
and is not involved in the environmental reaction without direct
excitation.
Conversely, ꢄ2 could be interpreted as a function of Eu(III) chem-
ical environment in humic-complexes. The differences in spectra at
low and high delay were evidenced in Reiller and Brevet [48]. The
application of the relation between ꢄ2 and the number of water
molecules in the first hydration sphere proposed by Kimura et al.
ing 10 ␮molEu L
and 200 mgHA L−1 were acquired using the
600 lines mm−1 grating. The different spectra are shown in
Fig. S2 of SI, and the different A(7F2/7F1) and A(7F0/7F1) values are
reported in Fig. 6. The shape of both 5D0 → F1 and 5D0 → F0 tran-
sitions did not evolve. This was expected for the former magnetic
transition, and suggests a small if not absent modification in the
centro-symmetry of complexes from the latter transition. While
the evolution of A(7F0/7F1) was absent, A(7F2/7F1) decreased from
3.6–4.0 to 3.2 with decreasing size of separated fractions (Fig. 6).
This effect cannot be attributed to a difference in the saturation
of humic sites since it appeared to be already attained (Fig. 4).
Although log ˇ and Cc values were almost identical for bulk GHA
and the G1–G8 fractions, the symmetry around Eu(III) was clearly
and progressively evolving with changes in fractions’ sizes. This
could be interpreted as a gradual modification in the composition
of the different fractions.
7
7

1178
P.E. Reiller et al. / Spectrochimica Acta Part A 78 (2011) 1173–1179
250
200
150
100
50
100
80
60
40
20
0
2
1
0
Ginchi Ginchi 1 Ginchi 2 Ginchi 3 Ginchi 4 Ginchi 5 Ginchi 6 Ginchi 7 Ginchi 8 Ginchi 9 Ginchi 10
bulk
Samples
Fig. 7. Evolution of fluorescence decay times of Eu(III) complexed by the different size-fractions as compared to bulk Ginchi HA, C(Eu) = 10−5 mol L−1, C(HS) = 200 mg L−1, pH
5, I = 0.1 mol L−1 (NaClO4), 600 lines mm−1 grating. Original data in Table S1 of SI.
[72], yielded a number of 5 0.5 remaining water molecules out
of 9. This is in agreement with the coordination number of 3.6, as
determined from ꢀmax(7F0). It is to be pointed out that the particular
dependence of Eu(III)–HS complexes on the number of remaining
water molecules in the first coordination sphere, has not been ver-
ified up to now. Moreover, the possibility of internal quenching
leading to a different expression is possible [73–75].
The fitting of the luminescence decays obtained for the Eu(III)
complexes is reported in Table S1 of the SI. The evolution of decay
times with different GHA size-fractions are reported on Fig. 7. In
line with previous studies in aqueous solution [46–48], ꢄ1 did not
show a major trend. We also verified this lack of dependence in
Eu–HA sorbed on ␣-alumina [31]. If this fast decay time was due to
a fast exchange, then its lack of modification with origin or mode
of preparation of humic size-fraction would be explainable. The
very slight decrease observed for ꢄ2 could eventually be reconciled
with a difference of one H2O molecule in the first hydration sphere.
However, the uncertainty of this determination (ꢁ = 0.5) prevents
the reliability of the trend. Hence, there is not a major difference
between all these samples.
fractions were progressively eluted in HPSEC. Here, we noted
that the evolution of A(7F2/7F1) followed closely the expected
change in aromatic content during the HPSEC separation of size-
fractions [22]. In fact, we found an initial decrease for the first
size-fraction relative to the bulk sample, and then a peak followed
by a monotonous decrease. Comparison with HPSEC separation
from a lignite humic acid is more difficult since the smallest-sized
fractions were not collected and analyzed [23].
Conte et al. [22,23] also found that the smaller molecular-
sized fractions were chemically alike and different from previously
eluted fractions [22]. Our complexation results indicate that frac-
tions G9 and G10 behaved differently from the rest of size-fractions.
Our findings also relate the decrease in number of complexation
sites to the decrease of carboxylic carbons as shown by Conte et al.
[22], even though the evolution is not as spectacular. We noted that
the lower-size fractions used by Conte et al. [22] were also poorer
in aromatic molecules than other fractions.
Finally, the G9 and G10 fractions were the least abundant among
the GHA fractions, i.e., 4% and 7% in weight, respectively. Hence,
from a complexation point of view, 89% of fractions behaved simi-
larly to the bulk GHA, despite the evident modifications found for
the Eu(III) complexation environment.
3.6. Link with the structure of HPSEC-fractions
It is to remind here that as humic substances are operationally
defined, they are not classes of molecules, and conclusions on one
extract should be verified on other specific extracts before general-
isation. This work is dealing with physical filtration effects, which
are only part of a larger picture that includes sorptive fractionation
under less pressure and longer contact time. Nevertheless, from
the point of view of physico-chemical properties, the wide existing
data bases of acid–base properties and complexation parameters
[61,67,76,77], as well as Ln/An(III) luminescence spectra [47,48],
one can advance, without being too much optimistic, that there is
not major differences between humic acid samples from wherever
they are. Furthermore, from a structural point of view, the differ-
ences between origins of humic acids can be very thin [78], and
can depend on the dissolution procedure [79]. It is also noteworthy
that we cannot consider the obtained HPSEC fractions as individual
molecules, but rather as aggregates of different (lower) degrees.
Conte et al. [22,23] have shown the structural modification of
size-fractions during HPSEC separation for two different humic
acids. Enrichment in aliphatic carbon and a decrease in aromatic
carbon were assessed by 13C-CPMAS NMR spectra as humic size-
4. Conclusion
Structures of the Eu(III) complexes of Ginchi HA and of its
HPSEC-fractions, probed using TRLS, are slightly but significantly
different. Nevertheless, these differences are not sufficient to
changes the decay time of the Eu(III) luminescence within the com-
plexes. Hence, even if the HPSEC processes induce changes within
the structure of the fraction as compared to the bulk HA, ca. 90% of
the Eu(III) complexes with the HA fractions are very much similar
and could be considered as equivalent to bulk HA in a first approx-
imation. Only ca. 10% of HA fractions can be considered to present
a somewhat stronger affinity for Eu(III). They represent the lighter
fractions which are retarded in the structure of the immobile phase
of HPSEC.
HPSEC can be considered as a forced migration process which
minimizes chemical interactions between the mobile (HS) and the
immobile phase, i.e., mostly physical interactions are in stakes in a
dynamic system. Hence, one can also consider that even if the HS
fractions submitted to this physical ‘sorting out’ are different, they
share mostly the same type of interaction strength with the bulk

P.E. Reiller et al. / Spectrochimica Acta Part A 78 (2011) 1173–1179
1179
HA. These subtle structural modifications are less important than
those evidenced on oxides [25–27]. Particularly, the modification
of the Eu(III) environment is clearly less important after a ‘physical’
sorptive fractionation as compared to a ‘real’ or ‘chemical’ sorptive
fractionation.
[32] R. Spaccini, A. Zena, C.A. Igwe, J.S.C. Mbagwu, A. Piccolo, Biogeochemistry 53
(2001) 1–22.
[33] R. Spaccini, J.S.C. Mbagwu, P. Conte, A. Piccolo, Geoderma 132 (2006) 9–19.
[34] V. Moulin, P. Reiller, B. Amekraz, C. Moulin, Rapid Commun. Mass Spectrom. 15
(2001) 2488–2496.
[35] A. Piccolo, M. Spiteller, Anal. Bioanal. Chem. 377 (2003) 1047–1059.
[36] A. Piccolo, P. Conte, E. Trivellone, B. Van Lagen, Environ. Sci. Technol. 36 (2002)
76–84.
Acknowledgment
[37] L. Marang, P.E. Reiller, S. Eidner, M.U. Kumke, M.F. Benedetti, Environ. Sci. Tech-
nol. 42 (2008) 5094–5098.
[38] I. Christl, H. Knicker, I. Kogel-Knabner, R. Kretzschmar, Eur. J. Soil Sci. 51 (2000)
617–625.
[39] J.C. Dobbs, W. Susetyo, F.E. Knight, M.A. Castles, L.A. Carreira, L.V. Azarraga,
Anal. Chem. 61 (1989) 483–488.
This work was financed through the MRISQ (DEN/DISN/AvC) and
RBPCH (DEN/DISN/RSTB) projects of CEA.
[40] K.H. Chung, W. Lee, Y. Cho, G. Choi, C. Lee, Talanta 65 (2005) 389–395.
[41] J.W. Thomason, W. Susetyo, L.A. Carreira, Appl. Spectrosc. 50 (1996)
401–408.
Appendix A. Supplementary data
[42] H.S. Shin, G.R. Choppin, Radiochim. Acta 86 (1999) 167–174.
[43] J.M. Monsallier, F.J. Scherbaum, G. Buckau, J.I. Kim, M.U. Kumke, C.H. Specht,
F.H. Frimmel, J. Photochem. Photobiol. A 138 (2001) 55–63.
[44] C.W. Li, G.V. Korshin, Chemosphere 49 (2002) 629–636.
[45] M.U. Kumke, S. Eidner, T. Kruger, Environ. Sci. Technol. 39 (2005)
9528–9533.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.saa.2010.12.075.
References
[46] L. Marang, S. Eidner, M.U. Kumke, M.F. Benedetti, P.E. Reiller, Chem. Geol. 264
(2009) 154–161.
[1] J.F. McCarthy, W.E. Sanford, P.L. Stafford, Environ. Sci. Technol. 32 (1998)
3901–3906.
[2] Y. Fujikawa, J. Zheng, I. Cayer, M. Sugahara, H. Takigami, A. Kudo, J. Radioanal.
Nucl. Chem. 240 (1999) 69–74.
[3] P.H. Santschi, K.A. Roberts, L.D. Guo, Environ. Sci. Technol. 36 (2002)
3711–3719.
[4] O. Pourret, G. Gruau, A. Dia, M. Davranche, J. Molenat, Aquat. Geochem. 16
(2010) 31–59.
[5] I. Heidmann, I. Christl, R. Kretzschmar, Geochim. Cosmochim. Acta 69 (2005)
1675–1686.
[6] L.P. Weng, W.H. Van Riemsdijk, T. Hiemstra, Geochim. Cosmochim. Acta 72
(2008) 5857–5870.
[7] C.P. Schulthess, J.F. McCarthy, Soil Sci. Soc. Am. J. 54 (1990) 688–694.
[8] M.A. Schlautman, J.J. Morgan, Geochim. Cosmochim. Acta 58 (1994) 4293–4303.
[9] J.D. Filius, D.G. Lumsdon, J.C.L. Meeussen, T. Hiemstra, W.H. van Riemsdijk,
Geochim. Cosmochim. Acta 64 (2000) 51–60.
[10] L.P. Weng, W.H. van Riemsdijk, L.K. Koopal, T. Hiemstra, Environ. Sci. Technol.
40 (2006) 7494–7500.
[11] E.M. Murphy, J.M. Zachara, S.C. Smith, J.L. Phillips, T.W. Wietsma, Environ. Sci.
Technol. 28 (1994) 1291–1299.
[12] A.W.P. Vermeer, J.K. McCulloch, W.H. van Riemsdijk, L.K. Koopal, Environ. Sci.
Technol. 33 (1999) 3892–3897.
[13] H. Lippold, J. Lippmann-Pipke, J. Contam. Hydrol. 109 (2009) 40–48.
[14] E. Tombácz, J.A. Rice, S.Z. Ren, ACH Models Chem. 134 (1997) 877–888.
[15] R. Österberg, K. Mortensen, A. Ikai, Naturwissenschaften 82 (1995) 137–139.
[16] E. Balnois, K.J. Wilkinson, J.R. Lead, J. Buffle, Environ. Sci. Technol. 33 (1999)
3911.
[17] A. Piccolo, P. Conte, A. Cozzolino, Eur. J. Soil Sci. 50 (1999) 687–692.
[18] A. Piccolo, P. Conte, A. Cozzolino, Soil Sci. 166 (2001) 174–185.
[19] G. Plancque, B. Amekraz, V. Moulin, P. Toulhoat, C. Moulin, Rapid Commun.
Mass Spectrom. 15 (2001) 827–835.
[20] A. These, M. Winkler, C. Thomas, T. Reemtsma, Rapid Commun. Mass Spectrom.
18 (2004) 1777–1786.
[21] J.F.L. Duval, K.J. Wilkinson, H.P. van Leeuwen, J. Buffle Environ. Sci. Technol. 39
(2005) 6435–6445.
[22] P. Conte, R. Spaccini, A. Piccolo, Anal. Bioanal. Chem. 386 (2006) 382–390.
[23] P. Conte, R. Spaccini, D. Smejkalova, A. Nebbioso, A. Piccolo, Chemosphere 69
(2007) 1032–1039.
[24] B. Gu, J. Schmitt, Z. Chem, L. Liang, J.F. McCarthy, Environ. Sci. Technol. 28 (1994)
38–46.
[47] J. Brevet, F. Claret, P.E. Reiller, Spectrochim. Acta A 74 (2009) 446–453.
[48] P.E. Reiller, J. Brevet, Spectrochim. Acta A 75 (2010) 629–636.
[49] J.-C.G. Bünzli, Luminescent probes, in: J.-C.G. Bünzli, G.R. Choppin (Eds.), Lan-
thanides Probe in Life, Chemical and Earth Sciences – Theory and Practice,
Elsevier, 1989.
[50] B.R. Judd, Phys. Rev. 127 (1962) 750–761.
[51] G.S. Ofelt, J. Chem. Phys. 37 (1962) 511–520.
[52] C.W. McNemar, W.D. Horrocks Jr., Appl. Spectrosc. 43 (1989) 816–821.
[53] G.R. Choppin, Z.M. Wang, Inorg. Chem. 36 (1997) 249–252.
[54] W.D. Horrocks Jr., D.R. Sudnick, Acc. Chem. Res. 14 (1981) 384–392.
[55] M.F. Yeh, T. Riedener, K.L. Bray, S.B. Clark, J. Alloys Compd. 303 (2000) 37–41.
[56] S.E. Burroughs, G. Eisenman, W.D. Horrocks Jr., Biophys. Chem. 42 (1992)
249–256.
[57] F.S. Richardson, Chem. Rev. 82 (1982) 541–552.
[58] S.A. Visser, J. Soil Sci. 15 (1964) 202–219.
[59] G. Buckau, R. Artinger, P. Fritz, S. Geyer, J.I. Kim, M. Wolf, Appl. Geochem. 15
(2000) 171–179.
[60] V. Moulin, J. Tits, C. Moulin, P. Decambox, P. Mauchien, O. de Ruty, Radiochim.
Acta 58/59 (1992) 121–128.
[61] C.J. Milne, D.G. Kinniburgh, E. Tipping, Environ. Sci. Technol. 35 (2001)
2049–2059.
[62] M.S. Caceci, Radiochim. Acta 39 (1985) 51–56.
[63] G. Bidoglio, I. Grenthe, P. Qi, P. Robouch, N. Omenetto, Talanta 38 (1991)
999–1008.
[64] Y. Yamamoto, Y. Takahashi, H. Shimizu, Chem. Lett. 34 (2005) 880–881.
[65] R. Kautenburger, H.P. Beck, J. Chromatogr. A 1159 (2007) 75–80.
[66] G. Plancque, Y. Maurice, V. Moulin, P. Toulhoat, C. Moulin, Appl. Spectrosc. 59
(2005) 432–441.
[67] C.J. Milne, D.G. Kinniburgh, W.H. van Riemsdijk, E. Tipping, Environ. Sci. Tech-
nol. 37 (2003) 958–971.
[68] N.J. Turro, Modern Molecular Photochemistry, University Science Books,
Sausalito, CA, USA, 1991.
[69] J.M. Monsallier, W. Schüßler, G. Buckau, T. Rabung, J.I. Kim, D. Jones, R. Keepax,
N. Bryan, Anal. Chem. 75 (2003) 3168–3174.
[70] M. Freyer, C. Walther, T. Stumpf, G. Buckau, T. Fanghänel, Radiochim. Acta 97
(2009) 547–558.
[71] I. Billard, K. Lützenkirchen, Radiochim. Acta 91 (2003) 285–294.
[72] T. Kimura, Y. Kato, H. Takeishi, G.R. Choppin, J. Alloys Compd. 271/273 (1998)
719–722.
[73] S. Kuke, B. Marmodee, S. Eidner, U. Schilde, M.U. Kumke, Spectrochim. Acta A:
Mol. Biomol. Spectrosc. 75 (2010) 1333–1340.
[74] W.D. Horrocks Jr., D.R. Sudnick, J. Am. Chem. Soc. 101 (1979) 334–340.
[75] R.M. Supkowski, W.D. Horrocks Jr., Inorg. Chim. Acta 340 (2002) 44–48.
[76] E. Tipping, Cation Binding by Humic Substances, Cambridge University Press,
Cambridge, U.K., 2002.
[25] J. Hur, M.A. Schlautman, J. Colloid Interface Sci. 264 (2003) 313–321.
[26] F. Claret, T. Schäfer, J. Brevet, P.E. Reiller, Environ. Sci. Technol. 42 (2008)
8809–8815.
[27] P. Reiller, B. Amekraz, C. Moulin, Environ. Sci. Technol. 40 (2006) 2235–2241.
[28] I. Christl, R. Kretzschmar, Environ. Sci. Technol. 35 (2001) 2505–2511.
[29] I. Christl, C.J. Milne, D.G. Kinniburgh, R. Kretzschmar, Environ. Sci. Technol. 35
(2001) 2512–2517.
[30] N. Janot, P.E. Reiller, G.V. Korshin, M.F. Benedetti, Environ. Sci. Technol. 44
(2010) 6782–6788.
[31] N. Janot, M.F. Benedetti, P.E. Reiller, Environ. Sci. Technol., accepted for publi-
cation. http://dx.doi.org/10.1021/es102592a.
[77] P. Reiller, Radiochim. Acta 93 (2005) 43–55.
[78] M.J. Avena, A.W.P. Vermeer, L.K. Koopal, Colloids Surf. A 151 (1999) 213–224.
[79] J.P. Pinheiro, A.M. Mota, J.M.R. d’Oliveira, J.M.G. Martinho, Anal. Chim. Acta 329
(1996) 15–24.