Evaluation of binding selectivities of caged crown ligands toward heavy metals by electrospray ionization/quadrupole ion trap mass spectrometry

Article abstract of DOI:10.1021/ac991125t

Electrospray ionization (ESI)/quadrupole ion trap mass spectrometry is used to evaluate the heavy metal binding selectivities of five caged crown ethers and two polyether reference compounds in methanol solution. The binding preferences for Hg²⁺, Pb²⁺, Cd²⁺, and Cu²⁺ were analyzed by comparison of ESI mass spectral intensities with the aim of developing this method for the rapid screening of binding selectivities of new synthetic ligands. The cage compounds preferentially bind Hg²⁺, except for the cage cryptand derivative, which favors Pb²⁺. The preference for Hg²⁺ stems from the favorable positioning of the nitrogen or sulfur atoms for linear coordination of Hg²⁺, whereas the cryptand derivative favors Pb²⁺ because of its larger cavity size. The counterions of the metal salts influence the type of complexes observed in the ESI mass spectra because the strengths of the metal - anion bonds affect retention of the anion in the complexes.

Full text of DOI:10.1021/ac991125t

Anal. Chem. 2000, 72, 2433-2445
Evaluation of Binding Selectivities of Caged Crown
Ligands toward Heavy Metals by Electrospray
Ionization/Quadrupole Ion Trap Mass Spectrometry
Sheryl M. Blair and Jennifer S. Brodbelt*
Department of Chemistry and Biochemistry, The University of Texas at Austin, Austin, Texas 78712
Alan P. Marchand,* Kaipenchery A. Kumar, and Hyun-Soon Chong
Department of Chemistry, University of North Texas, NT Station Box 305070, Denton, Texas 76203-5070
potentiometric, NMR, extraction, or UV-vis methods.4 However,
the conventional methods are not appropriate in all solvent
systems, do not uniformly give structural information, and can
require as much as 100-1000 times more analyte than ESI-MS.
ESI-MS offers an alternative way to estimate binding selectivities
and provides rapid feedback with minimal sample consumption.
The use of electrospray ionization (ESI) mass spectrometry for
the evaluation of selectivity in host-guest complexation has been
validated previously for polyether compounds, such as crown
ethers and cryptands, as well as for the estimation of binding
selectivities involving alkali metals and a few transition metals.2,5
Excellent correlation between the ESI mass spectral data obtained
by monitoring relative intensities of host-guest complexes and
expected binding selectivities was observed for solutions contain-
ing a single host with two metal cations.4
Electrospray ionization (ESI)/ quadrupole ion trap mass
spectrometry is used to evaluate the heavy metal binding
selectivities of five caged crown ethers and two polyether
reference compounds in methanol solution. The binding
preferences for Hg2+, P b2+, Cd2+, and Cu2+ were analyzed
by comparison of ESI mass spectral intensities with the
aim of developing this method for the rapid screening of
binding selectivities of new synthetic ligands. The cage
compounds preferentially bind Hg2 +, except for the cage
cryptand derivative, which favors P b2 +. The preference
for Hg2 + stems from the favorable positioning of the
nitrogen or sulfur atoms for linear coordination of Hg2 +
,
whereas the cryptand derivative favors P b2 + because of
its larger cavity size. The counterions of the metal salts
influence the type of complexes observed in the ESI mass
spectra because the strengths of the metal-anion bonds
affect retention of the anion in the complexes.
In the present study, the binding selectivities of the compounds
in Figure 1 for several heavy metals, including Hg2+, Pb2+, Cd2+
.
and Cu2+, are analyzed in methanolic solutions by comparison of
ESI mass spectral intensities. The primary aim is to report for
the first time the use of ESI-MS for studies involving screening
of novel synthetic ligands for the selective complexation of heavy
metals. The caged crowns are designed to complex heavy metals
for the selective removal of these toxic metals from waste. Heavy
metals accumulate in the environment from many industrial
sources,6 and these heavy metals inhibit metabolism, break
down the tissues of many organs, or result in neurological
poisoning when accumulated in the body.6 Much effort is currently
The development of new methods to explore structure/ binding
selectivity relationships in host-guest complexation has become
increasingly important due to the wide variety of practical
applications of molecular recognition in areas such as drug design
and environmental chemistry.1 The present study focuses on the
measurement of heavy metal binding selectivities of novel syn-
thetic host compounds (Figure 1), caged crowns, by a promising
method2 based on electrospray ionization mass spectrometry (ESI-
MS).3 Binding selectivities have traditionally been measured by
(4) Izatt, R. M.; Bradshaw, J. S.; Pawlak, K.; Bruening, R. L.; Tarbat, B. J. Chem.
Rev. 1 9 9 2 , 92, 1261. Martell, A. E.; Hancock, R. D. Metal Complexes in
Aqueous Solutions; Plenum Press: New York, 1996; Chapter 7, p 217.
Tsukube, H.; Furuta, H.; Odani, A.; Takeda, Y.; Kudo, Y.; Inoue, Y.; Liu, Y.;
Sakamoto, H.; Kimura, K. Determination of Stability Constants. In Compre-
hensive Supramolecular Chemistry; Atwood, J. L., Davies, J. E. D., Macnicol,
D. D., Vogtle, F., Ripmeester, J. A., Eds.; Elsevier Science: New York, 1996;
Vol. 8, Chapter 10, p 425.
(5) Leize, E.; Jaffrezic, A.; Van Dorsselear, A. J. Mass Spectrom. 1 9 9 6 , 31, 537-
544. Wang, K.; Gokel, G. W. J. Org. Chem. 1 9 9 6 , 61, 4693-4697. Young,
D.-S.; Hung, H.-Y.; Liu, L. K. J. Mass Spectrom. 1 9 9 7 , 32, 432-437. Young,
D.-S.; Hung, H.-Y.; Liu, L. K. Rapid Commun. Mass Spectrom. 1 9 9 7 , 11,
769-773. Peiris, D. M.; Yang, Y.; Ramanathan, R.; Williams, K. R.; Watson,
C. H.; Eyler, J. R. Int. J. Mass Spectrom. Ion Processes 1 9 9 6 , 157/ 158, 365-
378.
(1) Vogtle, F., Weber, E., Eds. Host-Guest Complex Chemistry: Macrocycles;
Springer-Verlag: New York, 1985. Atwood, J. L., Ed. Inclusion Phenomena
and Molecular Recognition; Plenum Press: New York, 1988. Kuby, J.
Immunology; W. H. Freeman and Co.: New York, 1982. Koryta, J.; Stulik,
K. Ion Selective Electrodes, 2nd ed.; Cambridge University Press: New York,
1983.
(2) Blair, S. M.; Kempen, E. C.; Brodbelt, J. S. J. Am. Soc. Mass Spectrom. 1 9 9 8 ,
9, 1049-1059. Blair, S. M.; Reddy, G. M.; Marchand, A. P.; Brodbelt, J. S.
J. Mass Spectrom. 1 9 9 8 , 33, 721-728. Brodbelt, J.; Kempen, E.; Reyzer,
M. Struct. Chem. 1999, 10, 213-220. Kempen, E. K.; Brodbelt, J. S.; Bartsch,
R. A.; Jang, Y.; Kim, J. S. Anal. Chem. 1 9 9 9 , 71, 5493-5500.
(3) Fenn, J. B. J. Am. Soc. Mass Spectrom. 1 9 9 3 , 4, 524-535. Cole, R. B., Ed.
Electrospray Ionization Mass Spectrometry: Fundamentals, Instrumentation
& Applications; John Wiley & Sons: New York, 1997.
10.1021/ac991125t CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/25/2000
Analytical Chemistry, Vol. 72, No. 11, June 1, 2000 2433

Figure 1. Cage and reference compound structures.
devoted to the development of new synthetic materials for the
selective extraction of metals from solution,7 and thus evaluation
of the binding properties of novel ligands is an active area of
research.
cury(II) with linear coordination has a diameter of 0.69 Å, but in
its octahedral coordination state, it has a diameter of 1.02 Å.8
Copper, cadmium, and lead have diameters of 0.73, 0.95, and 1.19
Å, respectively, in octahedral binding geometries.9
The analytical strategy presented herein involves mass spec-
trometric examination of numerous solutions containing one or
two host ligands and one or more metals, along with extensive
variation of counterions and ionic strengths, to develop a solid
framework for using the ESI-MS method in the context of heavy
metal binding studies. Experiments with corresponding “refer-
ence” compounds including a cryptand and a diaza-crown ether
were performed to further understand the factors that influence
the metal binding selectivities, such as the preferred coordination
geometry and size of the metal, cavity size of the ligand, and
variation in the metal counterions. The preferred coordination
geometry of mercury is linear, unlike those of the other heavy
metals, which prefer octahedral coordination geometries.8 Mer-
The two reference compounds were chosen because they
possess structures that share features to those of the novel caged
ligands of interest in this study and because they have been
examined previously by conventional methods, thus allowing facile
comparison with the present results. Schwing-Weill et al.10
determined the stability constants for Pb2+, Cu2+, and Cd2+ with
1,7-diaza-15-crown-5 (D15C5), 1,4,10-trioxa-7,13-diazacyclopenta-
decane, and a [2.2.1]-cryptand (221), 4,7,13,16,21-pentaoxa-1,10-
diazabicyclo[8.8.5]tricosane, in methanolic solutions using con-
ventional potentiometric and spectrophotometric methods. It was
found that the host with the smaller binding cavity, D15C5,
preferred to bind with the smaller metal cations in the following
trend: Cu2+ > Cd2+ > Pb2+. Likewise, the larger cavity host, 221,
selectively complexed the larger metal cations in the following
order: Pb2+ > Cd2+ > Cu2+. Binding studies using aqueous
(6) Forstner, U.; Wittmen, G. T. W. Metal Pollution in the Aquatic Environment;
Springer-Verlag: New York, 1979. Salomons, W., Forstner, U., Mader, P.,
Eds. Heavy Metals: Problems and Solutions; Springer-Verlag: New York,
1995. Vernet, J.-P. Heavy Metals in the Environment; Elsevier: New York,
1991.
(7) Faust, S. D.; Aly, O. M. Chemistry of Water Treatment; Butterworths: Boston,
MA, 1983. National Research Council. Innovations in Groundwater and Soil
Cleanup: From Concept to Commercialization; National Academy Press:
Washington, DC, 1997. Freeman, H. M., Ed. Standard Handbook of
Hazardous Waste Treatment and Disposal; McGraw-Hill: New York, 1997.
(8) Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry, 5th ed.; John
Wiley & Sons: New York, 1988.
(9) Lide, D. R., Ed. CRC Handbook of Chemistry and Physics, 74th ed., CRC
Press: Ann Arbor, MI, 1993.
(10) Annaud-Neu, F.; Spiess, B.; Schwing-Weill, M. J. Helv. Chim. Acta 1 9 7 7 ,
60, 2633. Spiess, B.; Annaud-Neu, F.; Schwing-Weill, M. J. Helv. Chim. Acta
1979, 62, 1531. Spiess, B.; Annaud-Neu, F.; Schwing-Weill, M. J. Helv. Chim.
Acta 1 9 8 0 , 63, 2287.
2434 Analytical Chemistry, Vol. 72, No. 11, June 1, 2000

Scheme 1
titrimetric methods were performed by Gramain et al. with the
reference compound 1,7-diaza-15-crown-5 (D15C5).11 Results showed
that D15C5 preferentially complexed Cu2+ and Cd2+ over other
divalent and monovalent metal cations.12 Anderegg et al. deter-
mined the aqueous stability constants for a compound similar to
D15C5, 1,10-diaza-18-crown-6, and found that the mercury complex
had a significantly higher stability constant than the cadmium
complex.12 Marchand et al. studied the alkali metal picrate
extraction behavior of the caged cryptand compound 1 2 (Figure
1) and found that 1 2 selectively extracted Na+ and K+ picrates
in preference to other alkali metals.13
All single host-multiple guest solutions had a concentration ratio
of 1:1:1:1:1. Typical concentrations of hosts or metal salts were
1.5 × 10-4-2 × 10-4 M. Molecular modeling of the host
compounds was performed using SymApps software, which is
based on the MM2 force field. All metal salt compounds were
purchased from Aldrich Chemical Co. (Milwaukee, WI) and used
without further purification. All caged compounds were synthe-
sized in the Marchand group, and the two main synthetic pathways
are outlined in Schemes 1 and 2.
Melting points are uncorrected. Elemental microanalyses were
performed by personnel at M-H-W Laboratories, Phoenix, AZ.
High-resolution mass spectral data for 1 1 were obtained at the
Mass Spectrometry Facility, Department of Chemistry and Bio-
chemistry, The University of Texas at Austin, by using a ZAB-E
double-sector high-resolution mass spectrometer (Micromass,
Manchester, England) that was operated in the chemical ionization
mode.
3 ,5 -Bis(2 -brom oethyl)-4 -oxahexacyclo[5 .4 .1 .0 2 ,6 .0 3 ,1 0
.0 5 ,9 .0 8 ,1 1 ]dodecane (3 ). Method A. A solution of 3,5-divinyl-
4-oxahexacyclo[5.4.1.02,6.03,10.05,9.08,11]dodecane16 (1 , 500 mg, 2.36
mmol) in hexane (10 mL) was cooled to 10 °C via application of
an external ice-water bath. To this cooled solution was added,
with stirring, benzoyl peroxide (57 mg, 0.23 mmol). Hydrogen
bromide gas, generated via reaction of Br2 (3 mL, 58 mmol,
excess) and tetralin (4 mL, 29 mmol, excess), was bubbled
through the cooled reaction mixture during 0.5 h. The resulting
mixture was poured into a separatory funnel and then was washed
sequentially with 10% aqueous NaHCO3 (5 mL) and water (5 mL).
The organic layer was dried (MgSO4) and filtered, and the filtrate
was concentrated in vacuo. The residue was purified via column
chromatography on silica gel by eluting with hexane. Pure 3 (720
mg, 82%) was thereby obtained as a colorless oil: IR (neat) 2980
(s), 1428 (m), 1132 cm-1 (m); 1H NMR (CDCl3) δ 1.51 (AB, J )
EXPERIMENTAL METHODS
All mass spectrometry experiments were performed on a
Finnigan ion trap operating in the mass-selective instability mode
with modified electronics to allow axial modulation.14 The elec-
trospray interface is based on a design developed by Oak Ridge
National Laboratory.15 The Harvard syringe pump system was set
at 2.0-3.0 µL/ min. The needle potential for the methanol solutions
ranged from 3.2 to 3.3 kV for all experiments. The vacuum
chamber was operated at a pressure of 1 mTorr with He. Each
spectrum taken was an average of 25-30 scans. All experimental
conditions were held constant throughout each series of experi-
ments to minimize differences arising from inconsistencies in ion
formation or transmission.
All solutions were prepared in 100% methanol. The chloride
or perchlorate salts were mixed in solution with the host of interest
to form the complexes studied. Host-guest solutions containing
a single host and a single guest had a concentration ratio of 1:1.
(11) Gramain, P.; Fre`re, Y. Nouv. J. Chim. 1 9 7 9 , 3, 53.
(12) Anderegg, G. Helv. Chim. Acta 1 9 7 5 , 58, 1218-1225.
(13) Marchand, A. P.; Alihodzic, S.; McKim, A. S.; Kumar, K. A.; Mlinaric-
Majerski, K.; Sumanovac, T.; Bott, S. G. Tetrahedron Lett. 1 9 9 8 , 39, 1861-
1864.
(14) March, R. E.; Hughes, R. J.; Todd, J. F. J. Quadrupole Storage Mass
Spectrometry; Wiley: New York, 1989.
(15) Van Berkel, G. J.; Glish, G. L.; McLuckey, S. A. Anal. Chem. 1 9 9 0 , 62,
1284-1289.
(16) Marchand, A. P.; Kumar, K. A.; McKim, A. S.; Majerski, K.; Kragol, G.
Tetrahedron 1 9 9 7 , 53, 3467.
Analytical Chemistry, Vol. 72, No. 11, June 1, 2000 2435

Scheme 2
10.2 Hz, 1 H), 1.86 (AB, J ) 10.2 Hz, 1 H), 2.28-2.70 (m, 12 H),
3.39 (t, J ) 7.0 Hz, 4 H); 13C NMR (CDCl3) δ 28.2 (t), 36.2 (t),
43.3 (d), 43.5 (t), 44.0 (d), 47.5 (d), 58.1 (d), 95.0 (s). Anal. Calcd
for C15H18Br2O; C, 48.16: H, 4.85. Found: C, 48.53: H, 5.00.
Method B. A solution of 3,5-bis(2-hydroxyethyl)-4-oxahexacyclo-
[5.4.1.0.2,60.3,100.5,90.8,11]dodecane16 (2 , 300 mg, 1.25 mmol) in
benzene (10 mL) was cooled to 0 °C via application of an external
ice-water bath. To this cooled solution was added, with stirring,
a solution of PBr3 (146 mg, 0.54 mmol) in benzene (3 mL). After
the addition of the brominating agent had been completed, the
external ice-water bath was removed, and the stirred reaction
mixture was allowed to warm gradually to ambient temperature
during 12 h. The reaction was quenched via addition of water (5
mL). The organic layer was separated from the mixture, washed
with water (5 mL), dried (MgSO4), and filtered, after which the
filtrate was concentrated in vacuo. The residue was purified by
column chromatography on silica gel by eluting with hexane. Pure
3 (285 mg, 61%) was thereby obtained as a colorless oil. The IR,1H
NMR, and 13C NMR spectra of this material were essentially
identical to the corresponding spectra obtained for 3 that had been
prepared previously via method A (vide supra).
3 , 5 - B i s ( 2 - m e r c a p t o e t h y l ) - 4 - o x a h e x a c y c l o -
[5 .4 .1 .0 2 ,6 .0 3 ,1 0 .0 5 ,9 .0 8 ,1 1 ]dodecane (4 ).17 To a suspension of
3 (830 mg, 2.31 mmol) in water (10 mL) was added thiourea (458
mg, 6.02 mmol), and the resulting mixture was refluxed for 6 h.
To the hot reaction mixture was added 10% aqueous NaOH (2
mL, 5 mmol), and the resulting mixture was refluxed for an
additional 4 h. The reaction mixture was allowed to cool gradually
to ambient temperature and then was acidified by addition of 10%
aqueous H2SO4 (2.5 mL). The resulting mixture was extracted
with CH2Cl2 (3 × 20 mL). The combined organic extracts were
washed with water (10 mL), dried (MgSO4), and filtered, and the
filtrate was concentrated in vacuo. The residue was purified via
column chromatography on silica gel by eluting with 10% EtOAc-
CH2Cl2. Pure 4 (430 mg, 67%) was thereby obtained as a colorless
oil: bp 138 °C (0.8 mmHg); IR (film) 2981 (s), 2530 (w), 1430
1
cm-1 (m); H NMR (CDCl3) δ 1.40-1.58 (m, 5 H), 1.86 (d, J )
10.2 Hz, 1 H), 2.02-2.58 (m, 14 H); 13C NMR (CDCl3) δ 20.5 (t),
37.6 (t), 41.5 (d), 43.5 (t), 44.2 (d), 47.7 (d), 58.4 (d), 96.4 (s).
Anal. Calcd for C15H20OS2: C, 64.24; H, 7.19. Found: C, 64.30; H,
7.06.
Crown Ether 6 . To a suspension of NaH (37 mg, 1.56 mmol)
in dry THF (10 mL) under argon was added with stirring dithiol
4 (200 mg, 0.71 mmol), and the resulting mixture was stirred at
ambient temperature for 0.5 h. To this mixture was then added
dropwise with stirring during 0.5 h a solution of triethylene glycol
ditosylate (5 , 327 mg, 0.71 mmol) in dry THF (20 mL). The
resulting mixture was refluxed for 2 days and then was allowed
to cool gradually to ambient temperature. The reaction mixture
was extracted with CH2Cl2 (4 × 25 mL). The combined organic
extracts were dried (MgSO4) and filtered, and the filtrate was
concentrated in vacuo. The residue was purified via column
chromatography on neutral alumina by eluting with 10% EtOAc-
CH2Cl2. Pure 6 (260 mg, 43%) was thereby obtained as a colorless
oil: IR (film) 2978 (s), 1428 (m), 1220 cm-1 (m); 1H NMR (CDCl3)
δ 1.51 (AB, JAB ) 10.1 Hz, 1 H), 1.84 (AB, JAB ) 10.1 Hz, 1 H),
1.94-2.10 (m, 6 H), 2.32-2.78 (m, 14 H), 3.60 (s, 4 H), 3.69 (t, J
) 7.6 Hz, 4 H); 13C NMR (CDCl3) δ 27.4 (t), 30.7 (t), 33.1 (t),
41.3 (d), 43.5 (t), 43.8(d), 47.7(d), 58.5 (d), 70.6 (t), 72.0 (t), 95.0
(s). Anal. Calcd for C21H30O3S2: C, 63.92; H, 7.66. Found: C, 63.67;
H, 7.39.
Crown Ether 8 (R ) CH2 P h). To a solution of 3 (740 mg,
1.98 mmol) and 7 18 (542 mg, 1.65 mmol) in CH3CN (44 mL) were
added sequentially Na2CO3 (1.84 g, 16.5 mmol) and NaI (130 mg,
0.875 mmol), and the resulting mixture was refluxed for 24 h.
The reaction mixture was allowed to cool to ambient temperature
and then was concentrated in vacuo. The residue was dissolved
in CH2Cl2 (30 mL), and the resulting solution was washed with
water (3 × 30 mL). The organic layer was dried (MgSO4) and
filtered, and the filtrate was concentrated in vacuo. The residue
(17) Urquhart, G. G.; Gates, J. W., Jr.; Connor, R. Organic Syntheses; Wiley: New
(18) Gatto, V. J.; Arnold, K. A.; Viscariello, A. M.; Miller, S. R.; Morgan, C. R.;
York, 1955; Collect Vol. III, p 363.
Gokel, G. W. J. Org. Chem. 1 9 8 6 , 51, 5373.
2436 Analytical Chemistry, Vol. 72, No. 11, June 1, 2000

was purified via column chromatography on neutral alumina by
eluting with 5:1 EtOAc-hexane. Pure 8 (610 mg, 63%) was
thereby obtained as a colorless oil: IR (film) 2968 (s), 1630 (m),
1450 (m), 1148 (m), 730 (m), 700 cm-1 (m); 1H NMR (CDCl3) δ
1.50 (AB, JAB ) 10.0 Hz, 1 H), 1.78-2.04 (m, 5 H), 2.35 (s, 2 H),
2.50-2.70 (m, 10 H), 3.33 (s, 6 H), 3.45 (t, J ) 6.6 Hz, 4 H), 3.55-
3.62 (m, 8 H), 7.18-7.32 (m, 10 H); 13C NMR (CDCl3) δ 30.3 (t),
41.9 (d), 44.2 (t), 44.4 (d), 48.6 (d), 51.0 (t), 53.6 (t), 59.2 (d), 59.4
(t), 95.4 (s), 127.3 (d), 128.7 (d), 129.4 (d), 140.4 (s). Anal. Calcd
for C35H44N2O3: C, 77.74; H, 8.20. Found: C, 77.53; H, 7.96.
Hydrogenolysis of 8 . To a solution of 8 (900 mg, 1.6 mmol)
in CH3OH (70 mL) was added 10% palladized charcoal catalyst
(200 mg, catalytic amount). The resulting mixture was subjected
to hydrogenolysis by agitation with excess H2(g) at 55 psi in a
Parr hydrogenation apparatus at ambient temperature for 12 h.
The reaction mixture was filtered through Celite, and the filtrate
was concentrated in vacuo. The residue was purified via column
chromatography on neutral alumina by eluting with 15% EtOAc-
CH3OH. Pure 9 (480 mg, 83%) was thereby obtained as a colorless
oil: IR (film) 3347 (br, m), 2982 (s), 1468 (m), 1350 (w), 1135
cm-1 (m); 1H NMR (CDCl3) δ 1.52 (AB, JAB ) 9.9 Hz, 1 H), 1.75-
2.02 (m, 5 H), 2.32 (s, 2 H), 2.41-2.60 (m, 8 H), 2.70-2.83 (m,
8H), 3.50-3.62 (m, 8 H); 13C NMR (CDCl3) δ 31.8 (t), 41.2 (d),
43.4 (t), 43.7 (d), 46.2 (t), 47.6 (d), 49.2 (t), 58.2 (d), 70.0 (t), 71.0
(t), 96.2 (s). Anal. Calcd for C21H32N2O3: C, 69.97; H, 8.95.
Found: C, 69.79; H, 8.97.
mg, 39%) was thereby obtained as a colorless oil: IR (film) 2969
(s), 1461 (m), 1370 (w), 1122 cm-1 (m); 1H NMR (CDCl3) δ 1.03
(t, J ) 6.7 Hz, 4 H), 1.51 (AB, JAB ) 10.5 Hz, 1 H), 1.75-2.05 (m,
5 H), 2.35 (s, 2 H), 2.38-2.62 (m, 10 H), 2.66 (t, J ) 5.4 Hz, 4 H),
2.76 (t, J ) 6.2 Hz, 4 H), 3.55 (m, 10 H); 13C NMR (CDCl3) δ 12.3
(q), 29.0 (t), 41.4 (t), 43.5 (t), 43.8 (d), 47.97 (t), 47.99 (2 C, d),
49.3 (t), 52.5 (t), 58.6 (d), 70.5 (t), 71.1 (t), 94.8 (s). Exact mass
(CI HRMS), amu: calcd for C25H40N2O3, Mr+ 416.303 894; found,
Mr+ 416.303 048.
RESULTS AND DISCUSSION
Five similar caged aza-crown ether derivatives and two aza-
polyether reference compounds (Figure 1) were analyzed for their
selectivities toward heavy metals. The novel, cage hosts each have
a multicyclic cage substituent attached to an aza-polyether ring.
The reference compounds, a crown ether and a cryptand, simulate
the binding cavities of the cage compounds and are primarily used
to undertake detailed studies of the effects of ionic strength,
counterions, and other factors on the measurement of heavy metal
binding selectivities by ESI-MS. Molecular models indicate that
the array of binding sites in the rings of the reference compounds
offer slightly more flexibility than those of the cage compounds.
This greater degree of flexibility is due to the absence of the
cagelike substituent. The cage substituent changes the geometry
of the cavity. It provides additional rigidity, with a concomitant
increase in preorganization compared to the corresponding
noncage-annulated crown ether or cryptand, slightly expands the
cavity size compared to the noncaged analogues, and also adds a
hydrocarbon residue that increases the lipophilicity of the resulting
host system, making the species largely insoluble in water. In
addition, the inclusion of the cage substituent provides sites that
can be used to attach the host ligand to a soluble polymer and
thereby immobilize the host. This approach may be used produc-
tively to facilitate handling of the host-guest complexes and
eventual recovery of the host ligands after use in the extraction
procedures. The guest ions may be stripped from the host-guest
complex, and the polymer-attached host system can then be
recycled.
Crown Ether 1 0 . To a solution of 9 (270 mg, 1.98 mmol)
and 1-bromo-2-methoxyethane (550 mg, 3.96 mmol) in CH3CN
(15 mL) under argon were added sequentially Na2CO3 (740 mg,
3.33 mmol) and NaI (110 mg, 2.18 mmol), and the resulting
mixture was refluxed for 65 h. The reaction mixture was allowed
to cool gradually to ambient temperature. The reaction mixture
was then filtered, and the filtrate was concentrated in vacuo. The
residue was dissolved in CHCl3 (20 mL), and the resulting solution
was washed with water (3 × 20 mL). The organic layer was dried
(MgSO4) and filtered, and the filtrate was concentrated in vacuo.
The residue was purified via column chromatography on neutral
alumina by eluting with 20% EtOAc-hexane followed by continued
elution of the chromatography column with 3% CH3OH-EtOAc.
Compound 1 0 (190 mg, 73%) was thereby obtained as a colorless
This study examines the structure/ binding selectivity relation-
ships of the cage compounds, including such structural aspects
as host cavity size, and both the numbers and types of binding
atoms and ring substituents. Each of the cage and reference
1
oil: IR (film) 2969 (s), 1461 (m), 1351 (w), 1116 cm-1 (m); H
compounds were mixed with heavy metal salts, such as Cd2+
,
NMR (CDCl3) δ 1.46 (AB, JAB ) 9.9 Hz, 1 H), 1.74-2.02 (m, 5
H), 2.28 (s, 2 H), 2.38-3.79 (m, 17 H), 3.26 (s, 3 H), 3.36 (t, J )
5.6 Hz, 2 H), 3.45-3.60 (m, 8 H); 13C NMR (CDCl3) δ 29.6 (t),
31.1 (t), 41.3 (2 C, d), 43.5 (t), 43.78 (d), 43.84 (d), 47.0 (t), 47.5
(d), 47.8 (d), 49.8 (t), 50.2 (t), 53.4 (2 C, t), 58.1 (d), 58.6 (d), 58.8
(d), 70.1 (t), 70.3 (t), 71.2 (t), 71.3 (t), 95.4 (s), 96.6 (s). Anal.
Calcd for C24H38N2O4: C, 68.87; H, 9.15. Found: C, 68.96; H, 9.09.
Crown Ether 1 1 . To a mixture of 9 (210 mg, 0.58 mmol)
and Na2CO3 (250 mg, 2.33 mmol) in CH3CN (10 mL) was added
diethyl sulfate (180 mg, 1.17 mmol), and the resulting mixture
was refluxed for 34 h. The reaction mixture was allowed to cool
gradually to ambient temperature. The reaction mixture was then
filtered, and the filtrate was concentrated in vacuo. Dichlo-
romethane (10 mL) was added to the residue; the resulting
mixture was filtered, and the filtrate was concentrated in vacuo.
The residue was purified via column chromatography on neutral
alumina by eluting with 10% EtOAc-hexane. Compound 1 1 (85
Cu2+, Hg2+, and Pb2+ chlorides or perchlorates, in methanol
solution. The sums of the intensities of the various ligand-metal
complexes in the ESI mass spectra were directly compared to
determine the binding preferences and trends for each compound
with each metal. Because conventional solution methods assume
that the counterion is associated in some way with the host-
metal complex, both the free metal and the metal salt complexes
in each ESI mass spectrum were considered when the binding
selectivities were determined.
Reference Compounds. To evaluate the ability to determine
binding selectivities for heavy metals and to probe the solution
factors that may influence the ESI-MS measurements, a series of
experiments involving two reference compounds, [2.2.1]-cryptand
and 1,7-diaza-15-crown-5, were performed. The binding trends of
these two compounds when studied with metal chloride salts
mimic those seen with the cage compounds. The ESI mass
Analytical Chemistry, Vol. 72, No. 11, June 1, 2000 2437

Table 1. Dominant Complexes Observed in ESI Mass SpectraA
spectral trends observed for the perchlorate solutions do not
always agree with the trends observed for the chloride salt
solutions, as described later.
chlorine atom. The results for the mixed metal solutions are
displayed in Table 2. The reported percentages varied by 10% on
a day-to-day basis, with most of the entries in Table 2 representing
three to six replicates.
Metal Chloride Solutions. The examination of the reference
compounds began with two sets of experiments. The first set
concerned the evaluation of each reference compound mixed in
a 1:1 solution with each metal chloride salt. A series of the spectra
(one for each metal) were then compared to assess the complex-
ation of the host compound under specific solution conditions and
to examine any discrepancies that may have arisen from differ-
ences in the relative ESI response factors of the host-guest
complexes. The spectra of the 1:1 solutions were used to identify
all the ion complexes produced for each combination of host and
metal ion, including charge-reduced complexes and ones that
retain a counterion. The types of complexes formed from the
single host-single metal salt solutions are summarized in Table
1. The second type of experiment involved the analysis of solutions
containing one reference compound mixed with all four metal
chloride salts in the ratio 1:1:1:1:1. The summations of intensities
of both the free metal and the metal chloride complexes, including
isotopic peaks, within each spectrum were directly compared to
determine the selectivity of the reference compound. Examples
of this strategy are shown in Figure 2 for complexation of [2.2.1]-
cryptand. Because a relatively large mass range is compressed,
the individual isotope distributions are not readily discernible in
the spectral figures. However, specific regions of the mass spectra
are expanded to allow scrutiny of the isotopic distributions, as
shown in Figure 2B for the lead-containing complexes of the
[2.2.1]-cryptand. The isotopes are not well resolved for the doubly
charged Pb2+ complexes (separated by 0.5 amu in the mass
spectra) because the quadrupole ion trap only allows nominal unit
A variety of complexes are seen in the spectra containing the
metal chloride salts (Figure 2A). These complexes stem from
complexation of the free metal, such as Cd2+, and of the metal
chloride, such as CdCl+, by the host ligand. In general, the
complexes in which one counterion is retained are the most
abundant for the metal chloride experiments.
The mixed-metal 1:1:1:1:1 solutions reveal the nature of the
competitive equilibria. The data shown in Figure 2A are used to
assess the general trends in binding selectivity (Table 2). For
example, Figure 2A shows that the lead-containing complexes are
favored over the cadmium complexes and the mercury and copper
complexes are obscured in the baseline. The reference compound
1,7-diaza-15-crown-5 (D15C5) is used to imitate the macrocyclic
portion of the smaller cage compounds, 9-11. D15C5 has a cavity
diameter of 1.0 Å.10 With the smaller diameter of its binding cavity,
D15C5 tends to favor complexation with smaller metals, such as
Cu2+ or Cd2+, on the basis of the intensities of the complexes
formed from the 1:1:1:1:1 mixed-metal solutions (Table 2). These
metals can be fully encapsulated within the ring of the D15C5
molecule unlike a larger metal, such as Pb2+. The D15C5 molecule
also prefers to bind Hg2+ over Pb2+. Mercury prefers coordination
by nitrogen atoms over oxygen atoms,8 thus matching well with
D15C5. The complexation of [2.2.1]-cryptand (221) shows a
greater degree of selectivity than that observed for D15C5. 221
has a larger binding cavity (1.1 Å10) and more donor atoms than
D15C5, and therefore it is able to more fully coordinate the larger
metals such as lead.
+
resolution. The isotope distribution for the singly charged PbClO4
complex that is expanded in Figure 2B shows excellent agreement
with that expected for a complex containing one lead and one
Metal P erchlorate Solutions. For the optimal comparison
of binding selectivities estimated from ESI-MS to those obtained
by conventional solution methods, the experiments were repeated
2438 Analytical Chemistry, Vol. 72, No. 11, June 1, 2000

Figure 2. [2.2.1]-Cryptand with all chlorides vs all perchlorates.
Table 2. Comparison of Metal Selectivity Trends in Reference Compounds (1:1:1:1:1 Solutions)
using perchlorate salts because binding constants reported in the
literature10 used these salts. Solutions containing the host and
individual metal and mixed metals (following the procedures
described earlier in this work) were analyzed for each reference
compound with the four metal perchlorate salts mentioned above.
The types of complexes formed are summarized in Table 1. For
the perchlorate salts, the dominant signals are for the free metal
complexes, and the signals for ones that retain a single counterion
are less intense or absent. This contrast to the results obtained
for the solutions containing the chloride salts is addressed in a
later section on counterion effects.
mercury, unlike the trend noted for the analogous chloride
experiments. This discrepancy stems from a counterion effect that
is described in the next section, and it causes the metal selectivi-
ties of D15C5 to differ for the chloride versus the perchlorate and
nitrate metal salts. For the reference compound [2.2.1]-cryptand,
the binding of a larger metal, lead, is still the most favored overall,
as seen in Table 2 and Figure 2B. The binding preferences seen
with the metal chloride solutions are identical to the orders
obtained with the perchlorate solutions for the [2.2.1]-cryptand
host. When the concentration of the host, such as [2.2.1]-cryptand,
is increased relative to the concentration of the metal guests, metal
complexation is suppressed and protonation becomes more
favorable, as seen in Figure 2C. In contrast, when the concentra-
The selectivity trends for the 1:1:1:1:1 perchlorate solutions
are summarized in Table 2. D15C5 favors complexation with
Analytical Chemistry, Vol. 72, No. 11, June 1, 2000 2439

tion of the host, such as [2.2.1]-cryptand, is decreased relative to
the concentration of the metal guests, binding selectivity is
increased and complex formation becomes more dominant, as
seen in Figure 2D. Comparison of Figure 2B to Figure 2D shows
that the preference for complexation of lead is slightly increased,
as expected when a solution environment that enhances binding
selectivity is created (i.e., greater excess of metals and limited
supply of host). The absolute concentrations of the 1:1:1:1:1
solution (i.e., [2.2.1]-cryptand:metals) were also varied to examine
the effect of concentration on the observed selectivities in the ESI
mass spectra. Neither increasing nor decreasing the total con-
centration by a factor of 2 had a large impact on the observed
distribution of complexes. For example, Figure 2E shows the ESI
mass spectrum obtained when the total concentration of the
components in the initial solution is halved. Further increases or
decreases in the initial concentrations lead to loss of the spray
stability due to the high salt content of the solution or reduction
in the ion intensities due to the low concentrations of complexes,
respectively.
Binding constants obtained using conventional potentiometric
methods are available for the two reference compounds, 1,7-diaza-
15-crown-5 and [2.2.1]-cryptand, in methanol with three of the
perchlorate salts, copper, cadmium, and lead perchlorate,10 sum-
marized as log K values in Table 2. In addition to having very
high transition metal binding constants, each of these compounds
also has a large proton affinity, thus explaining the presence of
great intensities of the protonated molecules in the ESI mass
spectra. The orders of the binding constants reported for these
compounds qualitatively agree with the selectivities observed in
the ESI mass spectra (see Table 2) for the metal perchlorate
solutions of 221 and D15C5. The smaller cavity host compound,
D15C5, preferentially complexes Cu2+ over Cd2+ or Pb2+, and
likewise the host with the larger cavity, 221, selectively binds the
Figure 3. Spectral differences due to counterion variation.
charge delocalization caused by ion pairing.20 Thus, a closer
examination of counterion effects on the ESI mass spectra is
warranted. Several differences in the ESI mass spectra are
observed among complexes formed with the perchlorate, chloride,
and nitrate salts. In general, when chloride and perchlorate
solutions of the same concentrations are sprayed on the same
day, the chloride mixtures tend to produce spectra that have
slightly higher intensities than those of the perchlorate solutions.
This general observation suggests that the metal chloride salts
increase the formation of charged complexes, perhaps by a
desolvation effect or change in the conductivity of the solution.
The use of chloride salts promotes formation of complexes that
retain one counterion, such as the (D15C5 + HgCl)+ complex
that is seen in Figure 3A or the (221 + PbCl)+ and (221 + CdCl)+
ions that are seen in Figure 2A. Small amounts of free metal
complexes are also seen occasionally, like the (221 + Pb)2+ and
(221 + Cd)2+ ions in Figure 2A. However, when complexation
with the perchlorate and nitrate salts, the predominant complexes
seen are those formed with the free metal, as illustrated by the
(221 + Pb)2+ complex in Figure 2B and the (D15C5 + Hg)2+
complex in Figure 3B. For the perchlorate and nitrate salt studies,
small amounts of complexes containing one chlorate or nitrate
anion can also be seen, such as the (221 + PbClO4)+ complex in
Figure 2B, but these complexes are much less abundant. More-
over, in comparison to the complexes formed by the D15C5 crown
ether reference, the 221 cryptand shows a greater preference for
retention of one counterion with the heavy metal, especially with
the nitrate salts. Despite its geometric constraints, the cryptand
possesses a slightly larger cavity than D15C5, thus making it more
favorable for retention of a loosely bound counterion. Given that
the primary ring of the 221 cryptand is an 18-membered one, not
larger metal, Pb2+
.
Metal Nitrate Solutions. To further explore the effects caused
by counterion variation on the binding selectivities observed in
the ESI mass spectra, complexation of each reference compound
with heavy metal nitrates was also studied. Solutions containing
the host and individual-metal nitrates and mixed-metal nitrate salts
(following the procedures described earlier) were examined for
each reference compound and the four metal nitrate salts. The
types of complexes formed are summarized in Table 1. Interest-
ingly, D15C5 favors formation of free metal complexes, whereas
all of the 221 complexes incorporate one counterion. For the 1:1:
1:1:1 solutions, D15C5 favors complexation with mercury, similar
to the trend observed in the perchlorate studies (see Table 2).
For the reference compound [2.2.1]-cryptand, binding of a larger
metal, lead, is again the most favorable overall, as seen in Table
2. However, complexation with mercury is also seen in the mixed-
metal nitrate ESI mass spectra.
Counterion Effects. Although the effects of counterions on
host-guest complexation have not been thoroughly scrutinized,
several previous reports have noted that the identity of the
counterion can affect the conformations of the host-guest
complexes due to the degree of ion pairing between the counterion
and the metal, even when the metal is complexed to the host.19
In addition, certain counterions can cause reduction of stability
constants for the host-guest complexes due to the competitive
(19) Kyba, E. P.; Timko, J. M.; Kaplan, L. J.; de Jong, F.; Gokel, G. W.; Cram, D.
J. J. Am. Chem. Soc. 1 9 7 8 , 100, 4555-4568. Doxee, K. M. J. Org. Chem.
1 9 8 9 , 54, 4712-4715.
(20) Smith, R. M.; Martell, A. E. Critical Stability Constants; Plenum Press: New
York, 1976; Vol. 4.
2440 Analytical Chemistry, Vol. 72, No. 11, June 1, 2000

a 15-membered one as in the crown ether reference, the counte-
rion may be able to perch above the cavity more favorably than
in the crown ether. A previous report by Dearden et al.21 showed
that when the 221 cryptand binds larger metals, the 18-membered
face of the cryptand becomes almost planar. This change allows
the metal to have optimal contact with six donor atoms on one
side of the cryptand and prevents close contact with the seventh
donor atom on the opposite side. The counterion is thus able to
perch on the side of the 18-membered ring, promoting retention
of the counterion with reduced steric hindrance even though the
metal is complexed to a bicyclic host.
The differences in the types of complexes formed when various
salts are used arise from the strength of the metal-anion bonds
and the degree of charge separation. The larger, symmetrical
anions, such as the perchlorates and nitrates, allow for the negative
charge to be more delocalized, thus creating a weaker ion pair
between the counterion and the metal.22 This weaker bond found
in the perchlorate and nitrate salts, along with the bulkier nature
of the perchlorate or nitrate anions, enhances the displacement
of the counterions by the host because the host is able to more
effectively compete with the anion for complexation with the
cation. This results in formation of a greater portion of the
complexes involving the free metal ion in the ESI process. Nitrates
and perchlorates are known to be weaker coordinating anions than
the chlorides. For example, the aqueous stability constants20
reported for the chlorides with heavy metals relative to those of
the nitrates or perchlorates with heavy metals are typically 2-4
orders of magnitude greater (i.e., log K values averaging 0.1 for
the nitrates and perchlorates and 1.2-6.7 for the chlorides),
confirming the higher binding energies of the chlorides. With the
chloride salts, the dominant complexes observed are those that
incorporate one chloride ion because a single anion remains more
strongly bound to the metal ion during the desolvation process
of ESI.
Differences in the binding selectivity trends for the two
reference hosts obtained with metal chlorides versus the nitrates
or perchlorates can also be rationalized by the strength of the
ion pairs and the sizes of the metals and cavities. The chloride
salts form tighter ion pairs, and thus the retention of the chloride
ions shifts the observed selectivities of the hosts, making them
preferentially bind to smaller metal ions than observed when the
counterions are not strongly associated (as for the perchlorates
and nitrates). For example, D15C5 prefers to bind mercury over
the slightly smaller metals, copper and cadmium, in the perchlo-
rate and nitrate experiments but prefers copper and cadmium over
mercury in the chloride experiments. Because the chloride binds
to the metal ion in a strong ion pair, thus changing the nominal
size and coordination geometry relative to a free metal or weak
ion pair that dissociates in the ESI process, the coordination of
the smaller metals by D15C5 is more favorable for the metal
chloride experiments.
Figure 4. Effect of increased ionic strength with addition of MgCl2.
way to probe the influence of counterions on metal binding
selectivities. Because counterion effects with respect to heavy
metal complexation of the crown ligands have not been evaluated
previously by conventional solution methods, such as potentiom-
etry or spectrophotometry, correlations with the present results
are not possible.
Ionic Strength Studies. The effect of increasing the ionic
strength of the solution on the binding selectivities observed in
the ESI mass spectra was also investigated. Previous studies
performed by Cole et al. indicated that increasing the ionic
strength of an electrosprayed solution caused the ion intensities
to decrease without causing a notable change in the charge
distribution of the ions.22 Upon the addition of a second salt to a
solution containing a single reference compound and a single
heavy metal salt, the ionic strength of the solution was increased
in the present experiments. The second salt was chosen to contain
a divalent metal that was not expected to bind strongly to the
hosts. For example, an equimolar amount of MgCl2 was added to
a host-salt solution, forming a 1:1:1 solution of host-heavy metal
salt 1-MgCl2. Thus, the total ionic strength of the mixture was
increased by a factor of 2. The divalent metal cation, Mg2+, should
not competitively bind with the host compound in question. For
this study, the reference compound 1,7-diaza-15-crown-5 was
mixed with MgCl2 and also CuCl2, CdCl2, or HgCl2. The 1:1
D15C5-HgCl2 solution produced the spectrum in Figure 4A. The
higher ionic strength solution, such as D15C5-MgCl2-HgCl2 (1:
1:1), resulted in the spectrum in Figure 4B under the same
instrumental conditions. Comparison of the two spectra shows
that increasing the ionic strength of the solution has little effect
on the ion intensities and has no effect on the types of metal
complexes formed. Examination of numerous other solutions
containing the host with pairs or groups of metal salts at a range
of ionic strengths and up to 4 × 10-4 M total additional salt
revealed no changes in the trends in binding selectivities.
Thus, no evidence for a significant ionic strength effect has been
found.
Overall, these results show that the complexation is sensitive
to counterion effects in some cases, stemming from the solution
chemistry or the ESI process, and this feature may provide a new
(21) Chen, Q.; Cannell, K.; Nicoll, J.; Deaden, D. V. J. Am. Chem. Soc. 1 9 9 6 ,
118, 6336-6344.
(22) Wang, G.; Cole, R. B. Anal. Chem. 1 9 9 4 , 66, 3702-3708. Wang, G.; Cole,
R. B. In Electrospray Ionization Mass Spectrometry; Cole, R. B., Ed.; John
Wiley and Sons: New York, 1997; Chapter 4.
Analytical Chemistry, Vol. 72, No. 11, June 1, 2000 2441

Figure 5. 9 with heavy metal chlorides.
Aza-Crown Ether Derivatives. The five cage compounds
were synthesized and purified, and their respective selectivities
toward heavy metal cations were investigated using various
mixtures of the host compounds with metal chloride and perchlo-
rate salts. Solutions that contained only a single host and a single
metal ion guest (1:1) were investigated, as were solutions that
contained a single host and all four metal ion guests (1:1:1:1:1)
(Figure 5). The summations of intensities of both the free metal
and the related metal salt complexes within the latter spectra were
directly compared to determine the selectivities of the cage
compounds.
Single Metal Chloride Solutions. The first set of experiments
involved ESI-MS analysis of solutions containing a 1:1 mixture of
a single metal chloride salt and a single cage compound. The types
of the metal complexes detected in the ESI mass spectra obtained
for complexes that involved the various metal salts were then
compared with one another. Examples of these spectra are shown
in Figure 5 for 9 with CdCl2, PbCl2, HgCl2, and CuCl2, respectively.
Protonated and sodium-cationized complexes dominate the spec-
tra, highlighting the high basicities and sodium affinities of 9 ,
even in the presence of 10-4 M concentrations of metal salts. Since
the solutions each contained only a single metal, selective
complexation could not be monitored directly, but the relative
intensities of the complexes provided a preliminary assessment
of ESI efficiency for well-defined solution conditions. As shown
in Figure 5C, host 9 generates the most intense complex when
binding Hg2+. In contrast, Pb2+ complexes are never observed
(Figure 5B), presumably due to the low Pb2+ affinity relative to
the preference for proton or Na+ complexation. The large size of
the Pb2+ ion may prohibit optimal coordination by 9. The dominant
types of complexes are analogous to those observed for the
reference complexes (i.e., retention of one chloride ion), as
previously summarized in Table 1.
Mixed Metal Chloride Solutions. The second set of experi-
ments involved ESI analysis of solutions having a 1:1:1:1:1 mixture
of the host cage compound and equal amounts of each of each of
the four heavy metal chlorides. In previous work aimed at
measuring metal binding selectivities, the ESI response factors
for different ligand-metal complexes were measured by analyzing
solutions containing well-defined concentrations and comparing
these results to distributions of complexes calculated for solution
equilibria.2 In fact, few discrepancies between the ESI mass
spectral data and the calculated solution equilibrium distributions
of complexes were found for systems containing a single ligand
with a series of similar metals. The good correlations were
attributed to the similar desolvation and spray efficiencies of
complexes containing the same host ligand but different metals.
Since binding constants are not known for the cage compounds
in the present study, calculations of equilibria distributions and
detailed measurements of ESI response factors are not possible.
Thus, the intensities of all related metal complexes (i.e., all doubly
charged complexes plus all those containing a single chloride) in
the ESI mass spectra of the 1:1:1:1:1 solutions were compared
and presumed to qualitatively reflect the distribution of complexes
in the original solutions. An example of one of these spectra can
be seen in Figure 5E for 9 . A direct comparison of the intensities
is made to estimate the binding selectivity of 9 under well-defined
solution conditions. Figure 5E shows that 9 prefers to form
complexes with copper in the mixed-metal solutions. Interestingly,
the spectra in Figures 5A-D indicated that the 9 / Hg2+ complex
2442 Analytical Chemistry, Vol. 72, No. 11, June 1, 2000

Table 3. Comparison of Metal Selectivity Trends in Cage Compounds (1:1:1:1:1 Solutions)
produced the greatest intensity in the series of ESI mass spectra
of the solutions containing a single metal. This complex is not
observed at all in Figure 5E for the mixed-metal solution,
suggesting that the dominant factor that influences the ESI mass
spectra is the competitive equilibria in solution, not the spray
efficiencies of the various metal complexes. The results for the
other mixed-metal solutions are summarized in Table 3.
Three of the cage compounds differ only by the substituents
attached to the nitrogens on the aza-polyether ring, 9 , 1 0 and
1 1 , but each of these has a relatively small cavity. 9 , which has
no pendant groups and is closely modeled by the compound 1,7-
diaza-15-crown-5, tends to be more selective toward the smaller
metals such as copper and cadmium, which it can fully encapsu-
late, rather than lead, which is a larger divalent cation. See Table
3. The smaller metals are preferred because they are better able
to fit inside the cavity of the ring to maximize their binding
coordination.
Substituents added to the polyether rings greatly affect the
binding selectivities of these cage compounds. If an ether
substituent is added, as is the case with 1 0 , the additional oxygen
helps stabilize the complexed metal ion by providing another
binding site, but the compound still prefers to bind with the
smaller copper and cadmium metal cations, which fit better into
the cavity. 1 1 , which has two ethyl substituents added to the
binding ring, does not exhibit any metal complexation upon ESI.
This result is probably due to steric hindrance caused by the
flexible substituents blocking the host’s binding cavity. Thus, the
pendant arms hinder any metals from being encapsulated by the
ring. However, the protonated species of this compound is present
in the mass spectrum, indicating that a small charge-dense ion
may attach to 1 1 .
atoms and also prefers linear coordination,23 it is no surprise to
see that 6 is very selective toward mercury. This preference is
similar to that of its nitrogen analogue, 9 . The placement of the
sulfur atoms on opposite sides of the ring in 6 may enhance the
linear coordination favored by mercury.
1 2 has a larger binding cavity that can be modeled by the
[2.2.1]-cryptand molecule, with a cavity size of 1.1 Å.10 With a
larger cavity and more binding sites, this molecule is able to better
coordinate larger metal ions. Therefore, among the metals used
in the present study, 1 2 selectively prefers to bind with lead,
which has a diameter of 1.19 Å.9
Mixed-Metal P erchlorate Solutions. The next set of experi-
ments involved ESI-MS analysis of solutions having a 1:1:1:1:1
mixture of a host cage compound and equimolar amounts of each
of the four heavy metal perchlorate salts. The intensities of all
the free metal and related metal salt complexes in the observed
ESI mass spectrum were compared with one another. The direct
comparison of the intensities was then used to evaluate the binding
preferences for the metal perchlorate salts. A summary of the
binding selectivity trends for the mixed heavy metal perchlorate
solutions is found in Table 3.
Some of the binding selectivities of the cage compounds
observed for the mixed heavy metal perchlorate solutions are
similar to the selectivities noticed for the mixed-metal chloride
studies, as seen in Table 3. For example, 6 -A favors complexation
with mercury in both the chloride and perchlorate salt solutions,
and 1 2 selectively binds with lead in both cases as well. However,
for the compounds 9 and 1 0 , complexation with mercury is
preferred over that with the larger metal cations, unlike the
selectivities obtained using metal chloride salts. Since the pre-
dominant complexes formed when perchlorate salts are used are
6 is the sulfur analogue of 9 . This host also has a relatively
small binding cavity and tends to selectively bind small metal
cations (Table 3). Given that mercury forms strong bonds to sulfur
(23) Martell, A. E.; Hancock, R. D. Metal Complexes in Aqueous Solutions; Plenum
Press: New York, 1996; p 56. Shannon, R. D. Acta Crystallogr., Sect. A 19 76,
A32, 751.
Analytical Chemistry, Vol. 72, No. 11, June 1, 2000 2443

complexes with the free metal cation, selective complexation with
mercury for these two compounds is not surprising because
mercury has the smallest ionic diameter in its linear coordination
geometry, as stated previously.
Two of the remaining cage compounds, 6 and 1 2 , have
identical metal binding selectivities for the mixed perchlorate
solutions as they did for their mixed chloride solutions. These
results can be seen in a side-by-side comparison in Table 3. For
reasons stated previously, the small sulfur-containing ring of 6
still prefers the small mercury and copper cations and the larger
binding cavity of 1 2 more favorably coordinates the largest metal
used in the study, lead. A comparison of the ESI-MS metal
selectivity trends to reported values from the literature is not
possible because these novel hosts have not been studied
previously. 11 again forms only a protonated complex, rather than
complexing any of the bulkier metal species, due to steric
hindrance.
Quantitation of Binding Selectivities. In previous work,8 it
was found that ESI-MS is a useful tool for quantitative evaluation
of binding selectivities of host-guest systems. Moreover, ESI
mass spectral results correlated well with expected solution
equilibrium distributions of complexes, especially when a single
host binding a series of similar guest cations was analyzed.9 In
the present study, a similar strategy was undertaken to quantitate
the binding selectivities of a reference host compound, D15C5,
and two heavy metal guests, Cu(ClO4)2 and Cd(ClO4)2. The known
solution binding constants for D15C5 with Cu2+ and Cd2+ are log
K ) 9.45 and log K ) 8.72, as reported from potentiometric
measurements.11 In solutions containing 2:1:1 ratios of the host
to two metals, there should be no binding selectivity observed
due to the large magnitude of the binding constants. Using the
method described in our previous work,10 solutions containing
D15C5 and a single heavy metal guest ion (1:1), either Cu(ClO4)2
or Cd(ClO4)2, were sprayed individually to analyze ESI response
factors (Figure 6A,B). The ESI response factors reflect differences
in the ionization efficiencies between the types of complexes
formed for the two metals and are typically related to desolvation
effects. Then a multicomponent solution made by mixing the two
individual solutions, containing D15C5 with Cu(ClO4)2 and Cd-
(ClO4)2 in the ratio 2:1:1, was sprayed (Figure 6C). The ratio of
the intensities of the products, (cadmium complexes)/ (copper
complexes), taken from the latter ESI mass spectrum, was
corrected on the basis of the ESI response factors from the
analysis of the single-metal solutions. The ESI-MS selectivities
were obtained by summing the intensities of all related complexs,
singly and doubly charged, because a more optimal way to
differentially weight the intensities of the singly versus doubly
charged complexes has not been found. The ESI-MS selectivity
ratio for the cadmium versus copper complexation could then be
compared to the theoretical selectivity ratio calculated from known
binding constants. These results are found in Table 4. The
corrected ESI ratios (0.97 when considering all of the types of
metal complexes formed and 1.1 when considering only the free
metal complexes formed) correlate well with the theoretical
selectivity ratio of 1.0.
Figure 6. Complexation of diaza-15-crown-5 with Cu(ClO4)2 and
Cd(ClO4)2.
Table 4. Binding Selectivity of Diaza-15-Crown-5 for
Cu(ClO4)2 and Cd(ClO4)2: ESI vs Theoretical Solution
Results
In similar experiments, the concentrations of the heavy metals
(Cu(ClO4)2 and Cd(ClO4)2) were increased relative to the con-
centration of the host (D15C5), resulting in host:metal ratios of
1:5 and host:mixed-metal solution ratios of 1:2.5:2.5, to evaluate a
more competitive metal binding environment. The results are
displayed in Table 4. When one considers all of the metal
2444 Analytical Chemistry, Vol. 72, No. 11, June 1, 2000

complexes formed, the corrected ESI mass spectral ratio cor-
relates well with the theoretical selectivity ratio based on the
calculated equilibrium distribution. This excellent agreement
shows that the ESI-MS method can be used to estimate heavy
metal selectivities in addition to its well-validated capabilities for
alkali metal complexation.
Estimation of Relative Binding Constants. Using a method
similar to the one described above, the relative heavy metal
binding strengths of the two reference compounds, D15C5 and
221, were qualitatively compared to the binding strengths of the
two most similar cage compounds, 9 and 1 2 , respectively. For
these experiments, a multicomponent mixture containing a refer-
ence compound, the corresponding cage compound, and a heavy
metal salt in a 1:1:2 ratio was sprayed, and the resulting ESI mass
spectrum was analyzed. The intensities of the reference-metal
complexes were then compared to the intensities of the cage-
metal complexes to provide a qualitative estimation of the binding
constant of each cage compound for the given heavy metal relative
to the binding constant of each reference compound. For instance,
for a mixture of D15C5 and 9 with Cu(ClO4)2, the spectrum in
Figure 7A shows only peaks corresponding to 9 -copper com-
plexes. Thus, the greater intensities of the 9 complexes indicate
that 9 has a substantially greater binding constant for copper than
does D15C5. Likewise, in Figure 7B, when the complexation of
D15C5 is compared to that of 9 in a mixture with Cd(ClO4)2, the
intensities of the 9 -cadmium complexes are larger than the
intensities of the D15C5-cadmium complexes by a factor of 1.4.
Once again, this indicates that 9 has a higher binding constant
than D15C5 for cadmium. Finally, the lead binding constants of
the reference compound 221 and the cage compound 1 2 were
compared. In this case, 221 formed more intense lead complexes
than 1 2 by a factor of 1.8, which implies that 1 2 has a lower
binding constant than 221 for lead. Although the ESI-MS method
does not provide absolute binding constants for these latter series
of experiments, further refinements may allow such measurements
in the future.
Figure 7. Complexation of diaza-15-crown-5 and 9 with Cu(ClO4)2
or Cd(ClO4)2.
In similar experiments, the concentrations of the hosts (D15C5
and 9 ) were increased relative to those of the guests (Cd(ClO4)2
or Cu(ClO4)2), resulting in 5:1 host:guest ratios for the single-
host solutions and 2.5:2.5:1 ratios for the mixed-host solutions, to
create a more competitive environment between the hosts for the
metal guest. However, protonation of the hosts proved to be more
favorable than complexation with copper or cadmium, as shown
in Figure 7C,D.
of the metal salt. Stronger metal-anion bonds make it more
difficult to displace the anion from the complex in the ESI pro-
cess, resulting in one counterion being incorporated into the
complex. Further evaluation by both ESI-MS and conventional
solution methods should permit the development of a broader
understanding of the influence of counterion effects on binding
selectivities.
CONCLUSIONS
ACKNOWLEDGMENT
ESI-MS proves to be a rapid method for determining the heavy
Funding for this work was provided in part by grants from
metal binding selectivities of a series of caged compounds and
thus could be used as an efficient option for screening new
synthetic compounds developed for remediation efforts and for
evaluating structure/ binding selectivity relationships. Sample
consumption is minimal, and this method may be applied to many
solvent systems that are amenable to ESI but not to any single
conventional method for the examination of binding selectivities.
Structural modifications of the caged compounds, such as the
addition of a diether bridging group across the nitrogen donors,
cause significant changes in the binding selectivities. The
types of metal complexes formed and the apparent binding
preferences of the crown ligands are affected by the counterion
the National Science Foundation (CHE-9357422), the Welch
Foundation (F-1155) and the Texas Advanced Technology Pro-
gram (003659-0206). A.P.M. thanks the Robert A. Welch Founda-
tion (Grant B-0963) and the Department of Energy (Grant DE-
ID-07-98ER14936) for financial support of this study. In addition,
we thank Dr. I. N. N. Namboothiri for his kind assistance with
the purification and characterization of 4 .
Received for review September 29, 1999. Accepted
February 29, 2000.
AC991125T
Analytical Chemistry, Vol. 72, No. 11, June 1, 2000 2445