Sequestration, fluorometric detection, and mass spectroscopy analysis of lanthanide ions using surface modified magnetic microspheres for microfluidic manipulation

Article abstract of DOI:10.1021/ja9035253

Several methods for rapid sequestration, fluorometric detection, and the subsequent mass spectroscopic analysis of lanthanide ions using surface modified polystyrene magnetic microspheres are demonstrated. Mixed-ligand antenna complexes of Eu³⁺ in which one of the ligands is attached to the surface of the microspheres have been used as a means for the sequestration, immobilization, and detection of these ions. Using the ion-exchange properties of these microspheres, this scheme has been extended to the detection of nonluminescent ions. The principles of these assays form the basis for operation of a portable microfluidic device for general analytical and nuclear forensics applications and indicate the manner in which the established methods of analytical chemistry, such as liquid-liquid extraction and ion-exchange chromatography, can be adapted for such miniature devices.

Full text of DOI:10.1021/ja9035253

Published on Web 10/08/2009
Sequestration, Fluorometric Detection, And Mass
Spectroscopy Analysis of Lanthanide Ions Using Surface
Modified Magnetic Microspheres for Microfluidic Manipulation
Ilya A. Shkrob,*,† Michael D. Kaminski,*,† Carol J. Mertz,† Paul G. Rickert,†
Mark S. Derzon,‡ and Kamyar Rahimian§
Chemical Sciences and Engineering DiVision, Argonne National Laboratory, 9700 South Cass
AVe, Argonne, Illinois 60439 and MEMS Technologies and Organic Materials Departments,
Sandia National Laboratories, Albuquerque, New Mexico 87185
Received April 30, 2009; E-mail: shkrob@anl.gov
Abstract: Several methods for rapid sequestration, fluorometric detection, and the subsequent mass
spectroscopic analysis of lanthanide ions using surface modified polystyrene magnetic microspheres
are demonstrated. Mixed-ligand antenna complexes of Eu3+ in which one of the ligands is attached to
the surface of the microspheres have been used as a means for the sequestration, immobilization,
and detection of these ions. Using the ion-exchange properties of these microspheres, this scheme
has been extended to the detection of nonluminescent ions. The principles of these assays form the
basis for operation of a portable microfluidic device for general analytical and nuclear forensics
applications and indicate the manner in which the established methods of analytical chemistry, such
as liquid-liquid extraction and ion-exchange chromatography, can be adapted for such miniature
devices.
Introduction
Liquid-liquid extraction and ion-exchange chromatography
are common methods for preconcentration and separation of
metal ions for subsequent trace element analyses, but these
methods are unsuitable for microfluidic miniaturization of
analytical procedures.1 Surface modified magnetic microspheres
can replace these methods provided that new techniques for
sequestration and detection of the metal ions on these magnetic
microspheres are demonstrated.2,3 Ideally, the sequestration
should provide the means for transporting the selected ion
through the microfluidic device followed by sensitive and
nondestructive detection. In the present work, several schemes
for sequestration and detection of lanthanide ions using the
† Chemical Sciences and Engineering Division, Argonne National
Laboratory.
Figure 1. Conceptual scheme of a mixed ligand antenna assay for
‡ MEMS Technologies, Sandia National Laboratories.
sequestration and TRLF detection of Eu3+ ions.
§ Organic Materials Departments, Sandia National Laboratories.
(1) (a) Smistrup, K.; Lund-Olsen, T.; Hansen, M. F.; Tang, P. T. J. Appl.
Phys. 2006, 99, 08P102. (b) Deng, T.; Prentiss, M.; Whitesides, G. M.
Appl. Phys. Lett. 2002, 80, 461. (c) Morozov, V. N.; Groves, S.; Turell,
M. J.; Bailey, C. J. Am. Chem. Soc. 2006, 129, 12628.
magnetic microspheres are reported. These magnetic micro-
spheres have been used as a versatile platform for immobiliza-
tion of trace amount (<1 ppb) of fission products that are
subsequently detected using time-resolved laser fluorescence
(TRLF) spectroscopy, and their isotopic composition is deter-
mined using mass spectrometry. These assays serve as the basis
for operation of a portable microfluidic device for nuclear
forensics that is developed at Argonne and Sandia National
Laboratories. To shorten this work, the details of the assays,
the methods, and the synthetic and conjugation protocols are
given in the Supporting Information. While Eu3+ was chosen
to demonstrate the methods, the approach is not limited to the
luminescent ions.
(2) (a) Nun˜ez, L.; Buchholz, B. A.; Vandegrift, G. F. Sep. Sci. Technol.
1995, 30, 1455. (b) Kaminski, M. D.; Nun˜ez, L.; Visser, A. E. Sep.
Sci. Technol. 1999, 34, 1103. (c) Nun˜ez, L.; Kaminski, M. D.
CHEMTECH 1998, 28, 41. (d) Shaibu, B. S.; Reddy, M. L. P.; Prabhu,
D. R.; Kanekar, A. S.; Manchanda, V. K. Radiochem. 2006, 94, 267.
(3) (a) Gru¨ttner, C.; Bo¨hmer, V.; Casnati, A.; Dozol, J.-F.; Reinhoudt,
D. N.; Reinoso-Garcia, M. M.; Rudershausen, S.; Teller, J.; Ungaro,
R.; Verboom, W.; Wang, P. J. Magn. Magn. Mater. 2005, 293, 559.
(b) Bo¨hmer, V.; Dozol, J.-F.; Gruttner, C.; Liger, K.; Matthews, S. E.;
Rudershausen, S.; Saadioui, M.; Wang, P. Org. Biomol. Chem. 2004,
2, 2372. (c) Gru¨ttner, C.; Rudershausen, S.; Matthews, S. E.; Wang,
P.; Bo¨hmer, V.; Dozol, J.-F. Eur. Cell. Mater. 2002, 3, 48. (d)
Matthews, S. E.; Parzuchowski, P.; Garcia-Carrera, A.; Gru¨ttner, C.
Chem. Comm. 2001, 417.
9
10.1021/ja9035253 CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 15705–15710 15705

A R T I C L E S
Shkrob et al.
Table 1. Properties of “Smooth” Polystyrene Magnetic Microspheres Supplied by Spherotecha
mean
diameter, µm
surface concentration of
groups, per nm2
total number
of groups,b × 108
volume concentration,
a
molar concentration
of groups, µeq/La
type
area,b × 107 nm2
surface group
× 109 cm-3
PMS-20
CMS-30
AMS-40
2.5
4.7
3.3
2.0
3.4
7.0
2.8
1.24
0.42
CO2
NH2
4.2
1.7
1.4
1.8
290
80
a For a 2.5 wt % stock solution. b Per microsphere.
The general approach (Figure 1) is similar to the well-known
DELFIA (dissociation enhanced lanthanide fluorescence)
immunoassay,4,5 in which Eu3+ forms a luminescent complex,
EuIIIL2A3, where L is a neutral ligand, such as trin-octylphos-
phine oxide (1) and A is an antenna ligand, such as the base of
2-thenoyltrifluoroacetone (TTA, pKa ≈ 6.32; Figure 1). The
antenna ligand absorbs 355 nm laser light and the excitation
was used as a detergent. The choice of this surfactant was
dictated by the unusually high (for a nonionic detergent) critical
micelle concentration (cmc) of 0.25 wt %. This permitted full
coverage of the magnetic microspheres by the nonionic detergent
without introducing the micelles in the bulk of the solvent, so
the only hydrophobic environment for the extraction of the
luminescent complex is provided by the detergent-covered
surface of the microspheres. These MEGA10-stabilized poly-
styrene magnetic microspheres were fully suspendible and stable
for months.
7
energy is transferred to Eu3+ that emits at 620 nm via F2 r
5D transition, with the emission lifetime of 0.5-0.75 ms.5 This
0
long-lived luminescence allows background-free detection of
as little as 1 ppt of Eu3+ by means of TRLF, as it discriminates
against the short-lived fluorescence from the organic molecules.4
In our scheme, the neutral ligand is attached to the surface of
the microspheres. Because water molecules in the coordination
sphere of Eu3+ quench the luminescence, the complex must be
isolated in a hydrophobic environment in order to inhibit ligand
exchange with the aqueous solution. One of the challenges of
adapting the extraction protocols to the magnetic microsphere
platform is that the hydrophobic ligands at the surface cause
aggregation of the modified microspheres, which is undesirable
for microfluidic manipulation. In order to keep the microspheres
fully suspended, these must be covered by nonionic detergents.
Since such surfactant molecules tend to form micelles in the
solution that extract hydrophobic Eu3+ complexes, the system
has to be carefully designed to bias the extraction of the complex
to the hydrophobic surface of the magnetic microspheres.
Another challenge is quenching of the luminescence by the
magnetic microspheres. Since it interferes with the TRLF
detection, it was necessary to search for the magnetic micro-
spheres that do not quench the luminescence on the submilli-
second time scale. All but one brand of commercially available
magnetic microspheres were shown to be efficient luminescence
quenchers on this time scale due to the presence of dispersed
ferric ions near the surface that serve as the sinks for the
excitation energy. These ferric ions are residues from the
incorporation of magnetic nanoparticles into the magnetic
microspheres during their synthesis. Spherotech Inc. supplies
polystyrene magnetic microspheres that have a 1-µm protective
overcoat of polystyrene that isolates these dispersed ferric ions,
and only such “smooth” magnetic microspheres were shown to
support long-lived luminescence from the f-elements. These
carboxylated and aminated magnetic microspheres also have
sulfonate groups at the surface that can be used for ion exchange.
The relevant properties of these magnetic microspheres are given
in Table 1. To prevent the aggregation of these surface modified
microspheres, 0.1% decanoyl-N-methylglucamide (MEGA10)
The experimental details, synthetic procedure, and assay
protocols are given in section 1S of the Supporting Information.
Results
1. DELFIA-Like Volume Assays. In the DELFIA assay,4 the
luminescent complex of Eu3+ is formed inside the core of a
nonionic micelle; the hydrophobicity of the core prevents rapid
quenching of the luminescence by coordinated aqua ligand. The
same approach can be used for sequestration and detection of
Eu3+ using the magnetic microspheres. In the simplest approach,
10-13-10-6 M Eu3+ ions (10-12 M ) 1.5 ppt) were extracted
from 1 mL of 10-4 M HNO3 solution by covalently attached
aminocarboxylic acids on ∼107 microspheres. The signal was
linear with Eu3+ concentration in the entire concentration range.
The best results were obtained using iminodiacetic acid (IDA)
modified BcMag-IDA silica magnetic microspheres from Bio-
Clone Inc. (1 µm) and ethylenediaminetetraacetic acid (EDTA)
modified MagaCell-EDTA cellulose magnetic microspheres
(Cortex Biochem, 10 µm). For the former microspheres,
partitioning coefficients up to 2.3 × 104 for the extraction of
Eu3+ were obtained. The Eu3+-loaded silica microspheres
exhibited a sorption capacity of ∼0.65 µeq per g of the
microspheres (0.14 ions per nm2) with a binding constant ꢀ ≈
4 × 107 M-1
.
These microspheres were magnetically separated and the
luminescence was developed by addition of 0.1% Triton X-100
solution containing 0.5 mM TTA and a neutral coligand. The
luminescent complex was extracted into the micellar phase, and
the detection of 10-12 M of Eu3+ was demonstrated in this
fashion. The coligand that was most efficient in luminescence
enhancement was octyl(phenyl)-N,N′-di(iso-butyl) carbamoyl
methyl phosphine oxide (2). Good results were also obtained
using 1 and triphenylphosphine oxide. Other polyethylene glycol
terminated nonionic detergents (RTX-100, Igepal CO520, Brij
35 and 36T, Triton X series detergents from 207 to 705) and
cationic detergents (such as cetyl triethylamine bromide) yielded
comparable results,6,7 with the luminescence enhancement
reaching maximum right above the cmc. The nonionic detergents
based on carbohydrate derivatives, such as dodecyl-ꢀ-D-mal-
toside, exhibit such strong affinity for the phosphine oxides that
(4) (a) Hemila¨, I. J. Alloys Comp. 1995, 225, 480. (b) Diamandis, E. P.;
Christopoulos, T. K. Anal. Chem. 1990, 62, 1149A. (c) Oliva, M. A.;
Olsina, R. A.; Masi, A. N. Analyst 2005, 130, 1312.
(5) (a) Nishioka, T.; Fukui, K.; Matsumoto, K. In, Handbook on the
Physics and Chemistry of Rare Earths; Gschneidener, Jr., K. A.,
Bunzli, J.-C., Pecharsky, V. K., Eds.; Elsevier: Amsterdam, the
Netherlands, 2007; Vol. 27; pp 171. (b) Sabbatini, N.; Guadigli, M.;
Lehn, J.-P. Coord. Chem. ReV. 1993, 123, 201. Leonard, J. P.;
Gunnlaugsson, T. J. Fluor. 2005, 15, 585. Brunet, E.; Juanes, O.;
Rodriguez-Ubis, J. C. Curr. Chem. Biol. 2007, 1, 11.
(6) Keelan, J. A.; France, J. T.; Barling, P. M. Clin. Chem. 1987, 33,
2292.
(7) Degan, P.; Abbondandolo, A.; Montagnoli, G. J. Biolum. Chemilum.
2005, 5, 207.
9
15706 J. AM. CHEM. SOC. VOL. 131, NO. 43, 2009

Sequestration of f-Ions Using Magnetic Microspheres
A R T I C L E S
Scheme 2. Spacers for Neutral Ligands Shown in Scheme 1
Scheme 1. Covalently Attached Neutral Ligands for Eu3+
Extraction at the Microsphere Surface
even water-soluble trimethyl and triethyl phosphine oxides form
the luminescent complexes in their interior. The typical lumi-
nescence times were 650-720 µs. The sensitivity of the method
can be improved further by increasing the volume of the sample
and the contact time.
2. Surface Mixed-Ligand Complex Assays. The detriment of
the DELFIA-like volume assay is that Eu3+ needs be stripped
off the microsphere for TRLF and mass spectrometric detection,
whereas it is preferable that the ion remains bound to the
magnetic microspheres for the ease of microfluidic manipulation.
This immobilization can be achieved either by physisorption
and/or covalent binding of the ligands. In these assays, (1-3)
× 107 cm-3 of the “smooth” polystyrene magnetic microspheres
were used to sequester the Eu3+ ions in 10-4 M HNO3. All of
the systems reported below exhibited extraction efficiencies of
95+% for pH 4-7 and [Eu3+] < 10-5 M, but the extraction
efficiency decreased dramatically at pH < 3. This pH dependence
allows stripping of Eu3+ from the magnetic microsphere by
addition of 0.01-1 M HNO3. Conversely, Eu3+ can be removed
into the aqueous phase by addition of a strong water-soluble
ion-binding complexant, such as 10-4-10-3 M disodium salt
of EDTA.
“Smooth” magnetic microspheres (Spherotech, model PMS-
20, 2.3 µm) were impregnated by 2, 1, or their 1:2 mixture with
tributyl phosphate in 20% NaCl solution containing 0.01%
MEGA10 (∼2.7 ligands per nm2), which is equivalent to a
monolayer coverage. The extracted luminescent complex was
in a highly hydrophobic environment, as the luminescence
lifetime was 350-430 µs. The magnetic microspheres were fully
suspendible in these saline solutions. These microspheres
extracted Eu3+ very efficiently and produced luminescence
within 15 s after TTA was introduced into the analyte. The
luminescence persisted after several cycles of magnetic separa-
tion followed by suspension of the magnetic microspheres in
0.1% MEGA10.
of the Supporting Information, respectively. Complexation to
TTA was used to develop the luminescence of the complex in
less than 1 min. Most of these luminescent complexes have
relatively short luminescence lifetime of 180 to 400 µs (Table
2), as these complexes include 1-2 water ligands (see below).
However, these relatively short lifetimes are still sufficiently
long to allow background-free detection of Eu3+ using TRLF.
The sensitivity limit was ≈5 × 10-11 M, and the linearity of
the luminescence signal persisted over 5-6 decades in the Eu3+
concentration (Figure 2). For ligand 5a directly attached to the
surface, the sorption capacity for Eu3+ was ∼50 µeq per g of
the microspheres and the stability constant ꢀ of the complex
was ∼1.4 × 107 M-1
.
In a typical assay, 300 µL of Eu3+ solution in 10-4 M nitric
acid was mixed with 300 µL of 1-3 mM TTA and then 30 µL
of 2.5 wt % of the modified magnetic microspheres were added,
and the mixture was gently stirred for 30 s. The luminescence
was measured at intervals of 30 s, with periodic agitation of
the solution. The luminescence was fully developed in 1-2 min
and then gradually subsided to 50-60% of the initial value due
to the slow aggregation of the magnetic microspheres. The
magnetic microsphere solution was magnetized and the lumi-
(8) Arnaud-Neu, F.; Bo¨hmer, V.; Dozol, J.-F.; Gru¨ttner, C.; Jakobi, R. A.;
Kraft, D.; Mauprivez, O.; Rouquette, H.; Schwing-Weill, M.-J.; Simon,
N.; Vogt, W. J. Chem. Soc., Perkin Trans. 1996, 2, 1175.
(9) (a) Delmau, L. H.; Simon, N.; Dozol, J.-F.; Eymard, S.; Tournois, B.;
Schwing-Weill, M.-J.; Arnaud-Neu, F.; Bo¨hmer, V.; Gru¨ttner, C.;
Musigmann, C.; Tunayar, A. Chem. Comm. 1998, 1627. (b) Hersch-
bach, H.; Brisach, F.; Haddaoui, J.; Saadioui, M.; Leize, E.; Van
Dorsselaer, A.; Arnaud-Neu, F.; Bo¨hmer, V. Talanta 2007, 74, 39.
(c) Babain, V. A.; Alyapyshev, M. Yu.; Karavan, M. D.; Bo¨hmer, V.;
Wang, L.; Sokolova, E. A.; Motornaya, A. E.; Vatsouro, I. M.;
Kovalev, V. V. Radiochem. Acta 2005, 93, 749. (d) Boehme, C.; Wipff,
G. Inorg. Chem. 2002, 41, 727. (e) Arduini, A.; Bo¨hmer, V.; Delmau,
L. H.; Desreux, J. F.; Dozol, J.-F.; Garcia Carrera, M. A.; Lambert,
B.; Musigmann, C.; Pocini, A.; Shivanyauk, A.; Ugozzoli, F.
Chem.sEur. J. 2000, 6, 2135. (f) Schmidt, C.; Saadioui, M.; Bo¨hmer,
V.; Host, V.; Spirlet, M.-R.; Desreux, J. F.; Brisach, F.; Arnaud-Neu,
F.; Dozol, J.-F. Org. Biomol. Chem. 2003, 1, 4089.
Alternatively, the extracting agents (Scheme 1) were co-
valently attached to the magnetic microsphere using the amide
conjugation protocol. The general structure was microsphere
surface-spacer-(L)n, (Figure 1) where we used polyamines,
polyethylene glycols, polyamidoamide dendrimers, and poly-
mers as spacers (Scheme 2) to which one or several ligands
were attached. The ligands were 2-like moieties (3),3,8,9 digly-
colamide moieties (4),10 and bidentate imide moieties (5a and
5b) shown in Scheme 1. The simplest of these surface
modifications involved reacting the aminated microspheres with
(PhO)2P(O)Cl in dimethyl formamide; the detailed synthetic and
conjugation protocols are given in section 1S, Parts D and E,
(10) Jan´czewski, D.; Reinhoudt, D. N.; Verboom, W.; Hill, C.; Allignol,
C.; Duchesne, M. T. New J. Chem. 2007, 32, 490. Matloka, K.;
Regalbuto, M.; Guelis, A.; Vandergrift, G. F.; Scott, M. J. J. Chem.
Soc., Dalton Trans. 2005, 3719.
9
J. AM. CHEM. SOC. VOL. 131, NO. 43, 2009 15707

A R T I C L E S
Shkrob et al.
Table 2. Relative Luminescence Yield in a Series of Trialsa for
lamides) that are known to be efficient extracting agents in ion-
exchange resins in this pH range.11,12 Clustering of such
chelators failed to improve the extraction efficiency significantly,
as was also observed by others.3 Due to the high sensitivity of
TRLF, reliable detection is still possible even at this high acidity,
following the pH adjustment by tris(hydroxymethyl) ami-
nomethane. The presence of 1 M NO3- in the solution reduces
the luminescence yield ∼3×, due to luminescence quenching
by the nitrate anion. Using another method, the Eu3+ ions were
extracted from 1 M HNO3, the magnetic microspheres were
separated, washed with 0.1 M imidazole buffer (pH 6) and then
the luminescence was developed by addition of TTA in 0.1%
MEGA10. This protocol resulted in poorer detectability limits,
but it was still possible to detect 10-8 M of Eu3+. The details
of these low-pH adapted protocols are given in section 1S, Part
F, of the Supporting Information.
Surface Modified Magnetic Microspheres
magnetic microsphere
spacer
ligand
TRLF signalb
lifetime, µs
AMS-40
AMS-40
CMS-30
AMS-30
CMS-30
CMS-30
5a
3
5b
4
5a
3
4
4
5b
4
3
4
45
63
91
35
25
80
27
26
73
95
43
36
100
71
200
220
280
220
213
275
250
240
240
260
280
270
310
270
8
8
8
CMS-30
AMS-40
AMS-40
CMS-30
CMS-30
PMS-20/1
PMS-20/2
9
10c
12
13
13
a The conditions of the trials: 150 nM Eu3+ and 1.5 mM TTA in 600
µL of 5 × 10-5 M HNO3 with 0.12 wt % of the surface modified
microspheres. b Arbitrary units. c n ) 6.
Following the sequestration and the TRLF detection, the
isotope composition for the metal ion can be detected using
mass spectrometry, which is essential, inter alia, for obtaining
forensics information from the samples. Three mass spectrom-
etry methods have been demonstrated. Using the first method,
5-50 ppb of sequestered Eu3+ ions ((0.3-3) × 10-7 M) were
stripped off the magnetic microspheres using 1 M HNO3 and
the supernatant was analyzed using inductively coupled plasma
mass spectrometry. For the DELFIA-like solution assay, quad-
rupole electrospray ionization mass spectrometry was used to
+
detect the Eu(TTA)2(2)2 complex in the gas phase, with the
detection limit of 100 ppb. The 151Eu/153Eu isotope ratio was
determined with the accuracy of 0.1% using both of these
methods in the indicated concentration ranges. Alternatively,
we used laser desorption/ionization mass spectrometry. Focused
337 nm laser light was used to photoexcite the surface of the
microspheres, which obviated the need of acid stripping. The
laser excitation of the magnetically separated microspheres
resulted in the direct release of Eu3+ ions into the gas phase.
The detection limit of ∼0.1 ppb has been achieved, with 0.1%
accuracy of isotope ratios at 50 ppb. The details of these mass
spectrometry measurements are given in section 1S, Part G,
Supporting Information.
Figure 2. The concentration plot for TRLF signal for a magnetic
microsphere assay using AMS-40-5a microspheres. Linearity of TRLF
detection for magnetic extraction of Eu3+ by AMS-40 microspheres modified
by 5a in the presence of TTA.
nescence of the supernatant was measured to estimate the
background contribution from the bulk of the solvent (which
was typically negligible, < 1%). At the lowest concentration,
30 µL of 10 mM disodium salt of EDTA was added to the
solution following the measurement in order to quench the
luminescence and acquire the background signal. This residual
signal was subtracted from the TRLF signal. The magnetically
separated microspheres were suspended in 0.6 mL of 0.5-1.5
mM TTA in 0.1% MEGA10 or a 1:1 mixture of 0.1% MEGA10
and 10% NaCl. For some of the designs, the magnetization/
resuspension cycle can be repeated 5-10× before the lumi-
nescence yield begins to decrease. For most of the designs, 2-3
of such cycles did not decrease the TRLF signal by more than
50%. While TTA enhances the surface binding of the Eu3+ ions,
the presence of the coligand is not necessary during the
sequestration. To demonstrate that, the Eu3+ ions were first
extracted by 30 µL of 2.5 wt % microspheres from 300 µL of
10-4 M nitric acid solution; the microspheres were magnetically
separated and then suspended in 600 µL of 0.5 mM TTA in
0.1% MEGA10 (or 10% NaCl, for “physisorption” assays). This
pre-extraction method yielded luminescence that was similar
or even higher than described above.
3. Ion-Exchange Surface Assay. While the TRLF approach
allows detection of luminescent f-ions, such as Eu3+ and Tb3+
,
ion exchange properties of the microspheres can be used to
detect nonluminescent ions at the concentrations that are
comparable to the loading capacity of the microspheres. The
microspheres considerably differ in their capacity as ion
exchangers, and this capacity can be further tuned by changing
the ionic strength. For the standard 0.25 wt % solution of the
magnetic microspheres in 10-4 M HNO3 containing 150 nM
Eu3+, the extraction efficiency of unmodified PMS-20, AMS-
40, and CMS-30 microspheres is 90%, 70%, and 98%. In the
identical solution containing 10% NaCl, this extraction efficiency
is reduced to 43%, 0%, and 17%, respectively. The maximum
capacity of PMS-20 microspheres for Eu3+ in these solutions
was estimated at 0.24 mM. Other brands of the microspheres
also demonstrated high capacity at low ionicity, but only
functionalized microspheres, such as MagaCell-EDTA (Cortex)
and ligand-conjugated AMS-40 and CMS-30 microspheres,
exhibited high extraction efficiency at high ionicity. For good
reproducibility, we found it necessary to have nonionic detergent
(11) Horwitz, E. P.; McAlister, D. R.; Bond, A. H.; Barrans, R. E. SolVent
Extr. Ion Exch. 2005, 23, 9.
At pH < 3, the extraction efficiency of the surface-modified
microspheres rapidly decreases with acidity, and at pH ) 0 this
efficiency is only 1-3%, even for the chelators (e.g., diglyco-
(12) Horwitz, E. P.; Chiarizia, R.; Dietz, M. L.; Diamond, H.; Nelson, D.
Anal. Chim. Acta 1993, 281, 361.
9
15708 J. AM. CHEM. SOC. VOL. 131, NO. 43, 2009

Sequestration of f-Ions Using Magnetic Microspheres
A R T I C L E S
(e.g., 11 and 12), and dendrimers (e.g., 13) did not show
significant advantages over the simpler designs, either in terms
of the extraction efficiency or luminescence enhancement. The
best covalently attached ligand magnetic microsphere designs
were at least 50-75% as efficient as the “physisorption” assays,
but without the tendency to aggregate at low ionic strength.
The greatest efficiency was shown by the magnetic microspheres
in which the ligands were directly attached to the carboxyl and
amino groups of the magnetic microspheres; the addition of
spacers, whether hydrophilic or hydrophobic, had either weak
effect or reduced the luminescence yield. The magnetic micro-
sphere designs in which the polymers/dendrimers were first
conjugated to the magnetic microspheres and then exhaustively
labeled by the ligands yielded the largest luminescence enhance-
ment with the longest luminescence lifetimes among the
conjugates, suggesting efficient shielding of the luminescent
complex from the aqueous solvent. All four classes of the
ligands (3-5) performed equally well, although the conjugates
of 4 and 5 demonstrated greater long-term chemical stability
than the conjugates of 3.
In all cases, the lifetime of the luminescence did not exceed
350-400 µs, often being as short as 200-250 µs, suggesting
the presence of a luminescence quenching ligand. To estimate
how many water ligands were involved, we replaced H2O water
by D2O in extraction of Eu3+ by AMS-40-5b microspheres. The
isotope replacement resulted in the increase of the lifetime from
216 to 295 µs, from which it was estimated, using the method
of Nwe et al.,13 that on average 1.5 water molecules are attached
to the Eu3+ ion. Thus, the short luminescence lifetimes are the
consequence of the relative permeability of the “hydrophobic”
surface layers to water. Increasing the hydrophobicity of the
coatings is problematic, as it leads to aggregation of the magnetic
microspheres. Furthermore, the modified microspheres still have
sulfonate groups at their surface increasing the polarity (and,
subsequently, the water content) of this surface.
Figure 3. The release of Eu3+ from preloaded magnetic microspheres upon
the addition of Gd3+. The Eu3+ ions were detected using TTA/2 in Triton
X-100 micelles. The release of Eu3+ from preloaded CMS-30 microspheres
upon the addition of Gd3+. The Eu3+ ions were detected using TRLF in
Triton X-100 micelles.
in the solution, as the aggregation properties of the magnetic
microspheres changed as the latter were loaded with metal ions.
In the implementation of ion-exchange assay shown in Figure
3, 250 µL of the 0.07% suspension of CMS-30 carboxylated
magnetic microspheres in 10-4 M HNO3 also containing 0.1%
Triton X-100 were equilibrated with 4.3 µM Eu3+ and then
mixed with 250 µL of Gd3+ in 10-4 M HNO3. After 1 min
stirring, the solution was magnetized and 50 µL of the
supernatant was mixed with 1.5 mL of 0.5 mM TTA and 0.5
mM 2 in 0.1% Triton X-100. The luminescence of Eu3+ was
determined in this micellar solution and compared to the
luminescence of the standard Eu3+ solution. Due to the great
excess of the ligands in the micellar solution, the competition
between Eu3+ and Gd3+ for these ligands was negligible. As
seen from Figure 3, the luminescence signal is linear with [Gd3+]
to 30 µM, though there is a weak luminescence background
from free Eu3+ ions in the solution without Gd3+ ions, due to
the extraction equilibrium (discussed below).
This assay can be carried out using TRLF detection of Eu3+
on the secondary microspheres, by diluting the supernatant
containing the released Eu3+ ions and then performing one of
the surface assays using the modified magnetic microspheres.
Therefore, using the combination of the ion exchange micro-
spheres and the mixed-ligand surface assays described above,
it is possible to use these magnetic microspheres for detection
of other ions than Eu3+. Although the identity of these ions
cannot be ascertained, their concentration can be estimated.
The performance of the ion-exchange assays (Figure 3) can
be rationalized in terms of the single-type f-ion binding site on
the surface of the microspheres. Let [L] be the volume
concentration of such binding sites and [M1,2] and [m1,2] be the
free ion and total concentrations of the luminescent (index 1)
and nonluminescent (index 2) f-ions, such as Eu3+ and Gd3+
,
respectively, and [LM1,2] be the volume concentration of the
bound ions. Then the concentrations can be determined from
the following:
Discussion
The “physisorption” assays for Eu3+ yielded the greatest
luminescence enhancement among the surface assays examined.
The detriment of these assays was that high salinity must be
maintained to prevent the aggregation of the magnetic micro-
spheres. In solutions of low ionic strength, there was an ongoing
aggregation of the magnetic microspheres that gradually de-
creased the luminescence yields and shortened the acquisition
of the TRLF signal to a few minutes after mixing the reagents.
Another problem was stripping the metal ions for the subsequent
mass spectrometry analysis. Once the luminescent complex was
formed at the surface at pH 4-7, re-extraction of Eu3+ into the
aqueous phase using 0.01-1 M HNO3 was slow and incomplete.
Upon the increase in the acidity, these magnetic microspheres
rapidly aggregated and the luminescent complex was trapped
inside these aggregates. Since the addition of dispersing agents
(e.g., detergents) removed the 2 into the aqueous phase, the
covalently attached ligands provided a more robust platform.
For these surface-modified magnetic microspheres, clustering
of neutral ligands using branched spacers (e.g., 8), polymers
[M1,2] ) m1,2/(ꢀ1,2[L] + 1)
[L] + [M1] + [M2] ) [L]0 + [m1] + [m2]
(1)
where ꢀ1,2 are the corresponding binding constants. Solving these
equations first for Eu3+, using the experimental data on
extraction of Eu3+, allowed us to determine the maximum
capacity [L]0 for these ions and the binding constant ꢀ1 and
then use these parameters to estimate the optimum loading [m1]/
[L]0 for the maximally linear response, i.e., the maximally linear
initial range of the plot of [M1] vs [m2]. For ꢀ1 ≈ ꢀ2, numerical
simulations indicate this optimum is at [m1]/[L]0 ≈ 0.65. For
the assay shown in Figure 3, this analysis gives the estimates
of [L]0 ≈ 4.6 µM, ꢀ1[L]0 ≈ 60, and ꢀ2[L]0 ≈ 1. The first two
estimates imply that the range of the reliable detection is 2-30
µM (0.5-6.0 [L]0), in agreement with our results. As the
(13) Nwe, K.; Richard, J. P.; Morrow, J. R. J. Chem. Soc., Dalton. Trans.
2007, 5171.
9
J. AM. CHEM. SOC. VOL. 131, NO. 43, 2009 15709

A R T I C L E S
Shkrob et al.
equilibria do not depend on the concentration of the micro-
spheres (eq 1), this assay can be adapted to lower concentrations
of Gd3+ by proportionally decreasing the concentration of the
CMS-30 microspheres and Eu3+. For each such concentration,
the linearity window is the same, in the units of [L]0.
chip devices. These approaches, therefore, provide a way for
adapting the standard methods of liquid-liquid extraction and
ion-exchange chromatography to microfluidic designs, by means
of heterophase chemistry on suspendible microspheres. While
the main thrust of this article was on the manipulation and
detection of luminescent lanthanide ions, nonluminescent ions
can be detected using the same schemes by taking advantage
of ion-exchange properties of these sulfonated microspheres.
We anticipate that these advances will speed up the adoption
of microfluidics for general analytical and nuclear forensics
applications.
Conclusions
We developed and demonstrated three approaches for se-
questration, time-resolved luminescence detection, and subse-
quent mass spectrometry analysis of lanthanide ions using
surface modified magnetic microspheres. One approach was
using a DELFIA-type volume assays in which the ions are
sequestered by aminocarboxylic ligands at the surface of the
microspheres and the luminescence is developed in the volume
bulk, inside the hydrophobic core of the micelles. The second
approach is covering the surface of the microsphere with a
monolayer of a hydrophobic, amphiphilic ligand, keeping the
microsphere in the aqueous solution suspendable through the
control of the salinity. The third approach is covalent linking
of the ligands either directly to the microsphere surface or the
“spacers” attached to this surface, and creating the hydrophobic
environment conducive to the formation of the luminescent
complex by covering this modified surface with nonionic
detergent.
Acknowledgment. The work at Argonne was performed under
the auspices of the Office of Science, Division of Chemical Science,
US-DOE under Contract No. DE-AC-02-06CH11357 and DTRA
Contract No. 08-43151. K.R. thanks S. P. Meserole, J. E. Reich,
and A. Allen for their help with the mass spectrometry measurements.
Supporting Information Available: Section 1S: Details of (A)
TRLF measurements; (B) materials; (C) magnetic microsphere
impregnation protocols; (D) synthetic methods; (E) conjugation
protocols; (F) high-acidity adaptation of the assay; and (G) mass
spectrometry measurements (PDF format). The material is
available free of charge via the Internet at http://pubs.acs.org.
As these microspheres can be manipulated magnetically, these
assays can be readily miniaturized and adapted for lab-on-the-
JA9035253
9
15710 J. AM. CHEM. SOC. VOL. 131, NO. 43, 2009