A colorimetric chemodosimeter for Pd(II): a method for detecting residual palladium in cross-coupling reactions

Article abstract of DOI:10.1016/j.tet.2008.04.105

A colorimetric chemodosimeter (SQ1) for the detection of trace palladium salts in cross-coupling reactions mediated by palladium is described. Decolorization of SQ1 is affected by nucleophilic attack of ethanethiol in basic DMSO solutions. Thiol addition is determined to have an equilibrium constant (K_eq) of 2.9×10⁶ M^-1, with a large entropic and modest enthalpic driving force. This unusual result is attributed to solvent effects arising from a strong coordinative interaction between DMSO and the parent squaraine. Palladium detection is achieved through thiol scavenging from the SQ1-ethanethiol complex leading to a color 'turn-on' of the parent squaraine. It was found that untreated samples obtained directly from Suzuki couplings showed no response to the assay. However, treatment of the samples with aqueous nitric acid generates a uniform Pd(NO₃)₂ species, which gives an appropriate response. 'Naked-eye' detection of Pd(NO₃)₂ was estimated to be as low as 0.5 ppm in solution and instrument-based detection was tested as low as 100 ppb. The average error over the working range of the assay was determined to be 7%.

Full text of DOI:10.1016/j.tet.2008.04.105

Tetrahedron 64 (2008) 8271–8278
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
A colorimetric chemodosimeter for Pd(II): a method for detecting residual
palladium in cross-coupling reactions
,
Ronald J.T. Houk c, Karl J. Wallace b, Himali S. Hewage a, Eric V. Anslyn a
*
a Department of Chemistry and Biochemistry, The University of Texas, 1 University Station, A5300, Austin, TX 78712, United States
b Department of Chemistry and Biochemistry, The University of Southern Mississippi, 118, College Drive, Hattiesburg, MS 39406, United States
c Sandia National Laboratories, Mail Stop 9291, Livermore, CA 94551-0969, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
A colorimetric chemodosimeter (SQ1) for the detection of trace palladium salts in cross-coupling re-
actions mediated by palladium is described. Decolorization of SQ1 is affected by nucleophilic attack of
Received 22 January 2008
Received in revised form 18 March 2008
Accepted 25 April 2008
ethanethiol in basic DMSO solutions. Thiol addition is determined to have an equilibrium constant (Keq
)
of 2.9ꢀ106
M
ꢁ1, with a large entropic and modest enthalpic driving force. This unusual result is attrib-
Available online 30 April 2008
uted to solvent effects arising from a strong coordinative interaction between DMSO and the parent
squaraine. Palladium detection is achieved through thiol scavenging from the SQ1–ethanethiol complex
leading to a color ‘turn-on’ of the parent squaraine. It was found that untreated samples obtained directly
from Suzuki couplings showed no response to the assay. However, treatment of the samples with
aqueous nitric acid generates a uniform Pd(NO3)2 species, which gives an appropriate response. ‘Naked-
eye’ detection of Pd(NO3)2 was estimated to be as low as 0.5 ppm in solution and instrument-based
detection was tested as low as 100 ppb. The average error over the working range of the assay was
determined to be 7%.
Published by Elsevier Ltd.
1. Introduction
highly acidic samples in palladium detection tends to corrode the
cones faster than normal. Hence, new, milder methods for the
Palladium complexes represent some of the more useful facili-
tators of organic transformations known. Palladium(II) salts, such
as PdCl2,1 Pd(OAc)2,2 and PdCl2(PPh3)2,3 are predominantly used as
oxidizing reagents, as well as precatalysts for cross-coupling
reactions. The wide array of commonly used reactions catalyzed by
these complexes,4 such as Suzuki, Heck, and aromatic amination
reactions, is processes that would otherwise be infeasible or
impractical. Many of these methodologies are widely utilized in
pharmaceutical research and development for the discovery and
production of drugs.5 However, governmental restrictions on the
levels of residual heavy metals in end products are very strict.
Typical contamination levels of palladium remaining in the organic
phase after experimental work up range from 5 to 100 ppm.4 Due to
its utility as well as its inherent stickiness, palladium poses a diffi-
cult challenge both for its detection and removal. The most popular
current method for detecting the presence of such metals utilizes
ICP-MS to vaporize the metal ions and obtain a quantitative mass
spectrum. Though very precise even in the ppt and ppq range, the
instruments are somewhat expensive to run, and the need for
sensitive detection of trace palladium are desirable.
Exploitation of reversible covalent bond formation for use as
a tool in the detection of analytical targets has become an area of
much interest in the last few years.6 Of particular interest to the
research presented in this paper are a class of organic dyes known
as squaraines, which have been shown to undergo a decolorizing
electrophilic addition from nucleophiles such as cyanide and vari-
ous thiols.7,8 Squaraines, such as SQ1 shown in Scheme 1, represent
a unique class of organic dyes with a peculiar electronic structure,
which gives rise to strong intramolecular charge transfer character
in both the ground and excited states.9 These charge transfer states
lead to strong, sharp absorption in the long visible and near IR
spectrum giving squaraines an intense blue color. The electronic
structure of the central four-membered ring can be described as
a cyclobutadienyl dication. The two electrons in the aromatic
dication reside as a singlet in a single low energy molecular orbital.
Since Lewis structures are insufficient to convey this situation,
resonance forms similar to that shown in Scheme 1 are often used.
Upon addition of a nucleophile to this electron deficient ring, the
charge transfer is disrupted, and the complex is effectively
decolorized.
Based on the work using thiols, we postulated that a sufficiently
thiophilic metal, such as palladium, would be capable of scavenging
* Corresponding author. Tel.: þ1 512 471 0068; fax: þ1 512 471 7791.
E-mail address: anslyn@ccwf.cc.utexas.edu (E.V. Anslyn).
0040-4020/$ – see front matter Published by Elsevier Ltd.
doi:10.1016/j.tet.2008.04.105

8272
R.J.T. Houk et al. / Tetrahedron 64 (2008) 8271–8278
O
O
O
O
R
R
R
R
R
R
2
S
HS
N
N
N
N
R
N
Base
R
H-Base
O
O
R
R
R
H
R
R
S
N
N
Pd(II)
Pd-SEt
R
R
O
N
O
SQ1:SEt
R
SQ1
Scheme 1. Decolorization of SQ1 with ethanethiol followed by palladium(II) scavenging.
the thiol. With the thiol removed, the conjugation and charge
transfer (CT) character of the parent squaraine would be restored
(Scheme 1). Concurrent with our work in this area, Soto and
co-workers reported a similar system for the selective detection of
mercury in aqueous media.10 Whereas Soto’s work shows selec-
tivity for mercury in aqueous solutions, our own work in organic
media suggests that little selectivity among heavy metal(II) salts is
achieved. Preliminary results show that the current method is
sensitive not only to Pd(II) but also several other commonly used
transition metals such as tin and common pollutants such as lead,
cadmium, and of course mercury. For the application we envision,
however, the lack of specificity will not be detrimental, as generally
only one or two different metals will be present in a sample taken
from a pharmaceutical process plant. Herein, we report the de-
velopment of a simple colorimetric chemodosimeter assay using
UV–vis spectroscopy and the ‘naked eye’ to detect Pd(II) levels in
contaminated products of cross-coupling reactions.11
Figure 1. 1H NMR of SQ1 with ethanethiol in the presence of DBU in CHCl3-d.
[SQ1]0¼37 mM; [DBU]0¼37 mM; [EtSH]0¼185 mM.
2. Results and discussion
SQ1:SEt was first generated by reaction of 1 equiv each of SQ1
and ethanethiol in DMSO facilitated by 0.75 equiv of the base
2,8,9-trimethyl-2,5,8,9-tetraaza-1-phosphabicyclo[3.3.3]undecane,
VB (a Verkade base).
2.1. Understanding the SQ1:SEt complexation
Though several reports have now been published on the
interactions of squaraines with various nucleophiles, very little is
known about the affinity of the nucleophile–squaraine in-
teraction.7,8,12 Since work with cyanide anions and thiols, the
interaction was perceived to be complete and not at equilibrium.
However, this study has found that in organic media, this ‘dosi-
metric’ system is dynamic and has equilibrium-like properties. Thus,
an understanding of this equilibrium is of considerable importance.
The use of this base and its stoichiometry, rather than DBU, will
be discussed in more detail below. The absorbance of unbound SQ1
was measured at regularly increasing temperature from 288.5 to
317.8 K to monitor the change in the complex formation (Fig. 2).
Higher temperatures were not used due to the low boiling point of
ethanethiol.
2.1.1. 1H NMR studies
We conducted a series of 1H NMR spectroscopic studies to fully
characterize the nucleophilic addition of the thiol to the central
four-membered ring (see Supplementary data). These studies, as
well as those previously reported literatures, have shown that
squaraine molecules desymmeterize upon the addition of nucleo-
philes (Fig. 1).12 The * and þ signs show the desymmetrization after
the addition of 1 equiv of ethanethiol. For the conditions to be
optimal for the nucleophilic attack in chloroform a base was added
to facilitate deprotonation of the thiol for greater nucleophilicity. In
these preliminary studies, the base 1,8-diazabicyclo[5.4.0]undec-7-
ene (DBU) was employed.
The temperature dependent equilibrium constants were
derived as follows. The concentration of free SQ1 was calculated
from the absorbance at 656 nm based on its extinction coefficient.
The bound complex SQ1:SEt and free thiol concentrations were
extrapolated from the concentration of free SQ1 assuming a 1:1
association of thiol with the squaraine (confirmed from the 1H NMR
spectroscopic studies, see Fig. 1 and Supplementary data). With
these concentrations the equilibrium constants at each tempera-
ture could be calculated as Keq¼[SQ1:SEt]/([SQ1][RSH]). The plot of
the natural log of these values versus 1/T gives the plot shown in
2.1.2. van’t Hoff analysis
Further studies on the thermodynamics of the SQ1:SEt complex
were carried out through the use of van’t Hoff techniques. A van’t
Figure 2B. The
was found to be ꢁ2.6 kcal/mol whereas
tercept is 20.7 cal/mol/K. Thus at 298 K, this association is primarily
entropy driven, with T
S0 contributing 6.2 kcal/mol versus only
D
H0 determined from the slope of the van’t Hoff plot
S0 taken from the in-
D
Hoff analysis garners a plot of ln Keq versus 1/T, which gives
D
H0
from the slope of the graph and
D
S0 from the intercept. In this case,
D

R.J.T. Houk et al. / Tetrahedron 64 (2008) 8271–8278
8273
10ꢁ5 M. Initial studies were conducted in a DMSO-based solution
containing a small amount of chloroform for squaraine solubility
purposes. As before, the SQ1:SEt complex was prepared by allowing
a thiol (ethanethiol) to react with a solution of SQ1 in an one-to-one
ratio in the presence of a suitable base (DBU) to prepare a complex
in situ that had a concentration of 47 mM. Upon completion of the
reaction, the absorbance at 656 nm is greatly reduced or ‘switched
off’. Under these conditions, formation of SQ1:SEt is kinetically slow
and takes approximately 24 h to come to equilibrium. The SQ1:SEt
complex must be formed prior to Pd(II) analysis and can be stored
for use. When stored in the dark, the solution is stable for roughly
48 h and is able to retain a reproducible response to Pd(II).
2.2.1. UV–vis titrations
Two palladium salts were used in the initial study. Pd(OAc)2 and
PdCl2(PPh3)2 were chosen due to their extensive use in cross-cou-
pling reactions. Titrations of both palladium species into solutions
of SQ1:SEt formed with DBU are shown in Figures 3 and 4.
Upon the addition of the palladium(II) salts, the band at 317 nm
decreases and the band at 656 nm increases, switching ‘on’ the
color. An isosbestic point is observed near 344 nm indicating the
interconversion of two distinct species. The isotherms obtained for
the titration experiments are sigmoidal in shape at low Pd(II)
concentrations because at equilibrium thiol complexation to the
squaraine is not complete. Thus, small amounts of uncomplexed
thiol exist in solution and the initial palladium that is added binds
first to the ‘free’ thiol. Once the free thiol has been bound the
remaining palladium then scavenges thiol away from the SQ1:SEt
complex and turns on the colorimetric response. The saturation at
1 equiv of palladium, evident in Figure 3A, suggests that the species
responsible for the absorbance at 373 nm is a monothiolated
palladium species.
Figure 2. van’t Hoff analysis of SQ1:SEt at 1.2ꢀ10ꢁ5 M. (A) Spectral data of free SQ1 at
temperatures from 288.5 to 317.8 K. (B) van’t Hoff plot of ln Keq versus 1/T.
ꢁ2.6 kcal/mol from the enthalpy term. The overall
D
G0 at 298 K
is ꢁ8.8 kcal/mol with an association constant of 2.9ꢀ106 Mꢁ1. Sol-
vation effects are commonly applied to explain entropy
complexations.13
Mixed solvent studies on squaraines with non-complexing and
complexing solvents have shown that at high concentrations of the
coordinating solvent, a preferential solvation phenomenon oc-
curred in which a 2:1 or 3:1 solvent–solute complex is forming.14
Hence, if the introduction of a nucleophile to the electron deficient
core of SQ1 serves to break up a solvent–solute complex, the sol-
vent release upon binding would be more entropically favorable
than the cost of bringing the squaraine and the nucleophile
together. It is also reasonable to imagine that the formation of this
bond would not be greatly enthalpically favorable. The bond
dissociation energies of C–S single bonds are generally quite low,
usually less than 75 kcal/mol, compared even to aliphatic S–H
bonds, which are usually around 83 kcal/mol.15 Couple these effects
with the decrease in conjugation and charge transfer interactions
that accompany the formation of SQ1:SEt, it becomes clear that any
enthalpic gain in this reaction would be low. Of course, the equi-
libria constants involved in the van’t Hoff also involve proton
transfer from the thiol to the base, and this reaction will influence
the entropy of the reaction. Perhaps the most striking observation
is how rare it is to find entropically driven associations in
supposedly non-competitive media such as DMSO.13
2.2. Pd(II) determination
Based on Beer’s law and van’t Hoff analyses, in order to keep the
absorbance intensity within the limits of Beer’s law, the total con-
centration of SQ1 both bound and unbound should be less than
Figure 3. Titration of Pd(OAc)2 into a solution of SQ1:SEt at 2.35ꢀ10ꢁ5 M in DMSO.
SQ1:SEt complexation was facilitated with 1 equiv of ethanethiol and 2 equiv of DBU.

8274
R.J.T. Houk et al. / Tetrahedron 64 (2008) 8271–8278
from the quenched or worked up Suzuki reaction showed no turn-
on at all.
Scheme 2.
During the course of a palladium catalyzed reaction, much of the
palladium is converted to Pd(0), and upon quenching, all remaining
active catalyst is either precipitated as palladium black or tied up in
bulky, intractable ligands created during the reaction. These pal-
ladium species would be insoluble and therefore undetectable by
this chemodosimeter. Our remedy to this problem was to generate
a uniform Pd(II) species via oxidation or ligand exchange with nitric
acid to form the highly soluble nitrate salt. To characterize an assay
for use with nitric acid digested samples, several titrations were
performed with Pd(NO3)2.
2.4. Pd(NO3)2 characterization
Initially, the SQ1:SEt complex was formed as described above
using a 1:9 v/v chloroform/DMSO solvent mixture in the presence of
2 equiv of DBU. The titration of Pd(NO3)2 into a 24 mM solution of
SQ1:SEt is shown in Figure 5. Interestingly, saturation occurred at
Figure 4. Titration of Pd(PPh3)2Cl2 into a solution of SQ1:SEt at 2.35ꢀ10ꢁ5 M in DMSO.
SQ1:SEt complexation was facilitated with 1 equiv of ethanethiol and 2 equiv of DBU.
Interestingly, the isotherm for Pd(PPh3)2Cl2 does not plateau but
instead a slight decrease in absorbance is seen at 656 nm as more
Pd(II) salt was added. The formation of a covalent Pd–S bond lib-
erates a chloride anion. However, from previous studies on nucleo-
philic addition to squaraines, the probability that this chloride is
responsible for the subsequent ‘re-quenching’ of SQ1 is small.8
A more likely scenario is that the formation of the Pd–S bond
significantly reduces the electrophilicity of the Pd center causing
one or both of the triphenylphosphine ligands to dissociate. The
PPh3 group thus indirectly displaced by the thiol is itself nucleo-
philic and attacks the electron deficient four-membered ring sys-
tem of SQ1. This result was confirmed by preparing a solution of
SQ1 and adding 1 equiv of PPh3 with the observation of a color
‘turn-off’. The ultimate sensitivity for these two Pd(II) salts using
this method was roughly 10ꢁ6 M.
2.3. Testing unknowns
It was apparent that this system could be used to quantitatively
detect Pd(II) salts. In order to test the viability of the system, cali-
bration curves were first created using several solutions containing
known quantities of Pd(OAc)2. To perform real sample tests, several
Suzuki coupling reactions were run as shown in Scheme 2. The
Suzuki coupling was chosen because it uses the palladium(II)
acetate precatalyst and a non-nucleophilic carbonate base. When
aliquots were taken directly from the reaction and administered to
the SQ1:SEt solutions, the result was a moderate turn-on of the
656 nm band corresponding to free SQ1 (see Supplementary data).
However, the calculated concentration did not match well with the
calibration curve. Furthermore, administration of samples taken
Figure 5. Titration data for Pd(NO3)2 into SQ1:SEt (1.18ꢀ10ꢁ5 M in DMSO) formed
with ethanethiol and 2 equiv DBU. (A) Spectra. (B) Isotherm showing apparent 2:1
thiol/palladium interaction. Dotted linedextrapolation of initial slope; dotted and
dashed linedabsorbance at full recovery.

R.J.T. Houk et al. / Tetrahedron 64 (2008) 8271–8278
8275
0.5 equivofpalladium(II) nitratesuggesting a 2:1 thiolate/palladium
interaction. Initially, we attributed this result to the increased elec-
trophilicity of the nitrated palladium center. However, upon further
review of the spectral data, we observed that full turn-on of SQ1 was
not achieved even after reaching saturation. A plot of the calculated
concentration of SQ1 from absorbance at 635 nm (3z99,000) versus
the concentration of palladium(II) added should have a slope of 2
based onthe saturationpointof the titration and a2:1 stoichiometry.
However, the slope was found to be 1 suggesting 1:1 binding. In fact,
extrapolation of the initial slope on the isotherm in Figure 5 (dotted
line) to the point at which full turn-on is predicted by the extinction
coefficient (dotted–dashed line) gives 1:1 binding.
To solve this quandary, we re-evaluated the choice of base, DBU,
and discovered that we had made a poor selection. In aqueous
solution, the pKa of H-DBUþ is roughly 12 and that of a typical al-
iphatic thiol is close to 10.16 Hence, in water DBU is a suitable base
to produce the thiolate anion. However, the polar aprotic envi-
ronment of DMSO drastically changes this relationship. In the cat-
ionophilic DMSO medium, the pKa of Hþ-DBU does not differ very
much from that in water; however, the pKa of a thiol is dramatically
increased to near 18.17
Thus, with DBU in DMSO, the complex formation occurs either
by general base catalysis or involves an exceedingly low concen-
tration of preformed thiolate anion. This revelation explains why
the complex formation was kinetically slow under these conditions.
In addition to being inadequate as a base, it was also discerned that
DBU was too nucleophilic, and when added to SQ1 in excess (4þ
equivalents), fairly rapid decolorization ensued. Our first instinct
then led us to assume that the discrepancy between the saturation
isotherm and Beer’s law equivalency arose from these non-optimal
properties of DBU.
Figure 6. Titration of Pd(NO3)2 into 8.8ꢀ10ꢁ6 M SQ1:SEt in DMSO formulated with
ethanethiol and Verkade base VB. Titration was conducted after 4 h reaction time
between squaraine SQ1 and ethanethiol.
2.5. The Verkade base
To resolve both the basicity and nucleophilicity issues we turned
to the so-called ‘super bases’. In particular, we chose the tricyclic
phosphatrane base, VB also known as a Verkade base.18 Reported to
have a protonated pKa of 26.8 in DMSO, VB is basic enough to
deprotonate an aliphatic thiol in DMSO.19 Several tests were con-
ducted to determine the best ratio of base to thiol to achieve both
facile complex formation and full recovery of SQ1 upon in-
troduction of analytes. Using 0.75 equiv of VB gave the best results
upon addition of Pd(NO3)2. Scheme 1 above shows the equilibrium
between two possible forms of the SQ1:SEt complex. VB is able to
facilitate this equilibrium because its pKa lies between that of the
thiol and the carbon acid shown. Hence, using a sub-stoichiometric
amount of VB helps to drive the complex formation forward. This
stoichiometry was also advantageous in that full recovery of the
parent squaraine was achievable if the SQ1:SEt complex was given
4 h to equilibrate. A typical titration using this method is shown in
Figure 6. For this new system, titrations were performed at a lower
concentration of SQ1:SEt (8.8ꢀ10ꢁ6 M) to maintain Beer’s law
linearity over the complete titration range. In addition to the
expected full recovery of SQ1 absorbance after a 4-h SQ1:SEt
equilibration time, a return to 1:1 Pd(II) to thiol binding was
observed. Yet, as now described, the most accurate Pd(II) sensing
system is achieved after allowing an even longer equilibration time
for SQ1, thiol, and base.
2.6. Calibration curves for DMSO–Verkade system
Samples of known Pd(NO3)2 concentration were prepared and
tested against the titration data to determine its usefulness as
a calibration curve for unknown samples. Unfortunately, in all cases
using a 5 min wait period between sample injection and mea-
surement, the samples tested fell short of their theoretical turn-on
quantity. The discrepancies between the titration data and the
known Pd(II) samples arise from a kinetic effect. Throughout the
course of a titration such as the one shown in Figure 7, each aliquot
of Pd(NO3)2 injected is given a 5 min interval, during which equil-
ibration was found to be achieved because with each data point in
a titration, only a small shift in equilibrium occurs. While the
titration data is a good measure of the overall turn-on one should
expect, when applied to a sample with an intermediate concen-
tration, a longer time is needed because a larger shift in the
equilibrium is occurring.
To better understand the time required to allow equilibration
with Pd(NO3)2 at intermediate concentrations, we turned to time
course plots. For this determination, several Pd(NO3)2 samples of
increasing concentration were assigned fresh solutions of com-
plexed SQ1:SEt, and upon injection of a 20 mL aliquot of the Pd(II)
If the SQ1, VB, and ethanethiol mixture was allowed to come to
full equilibrium over a 12–15 h period, the Pd(NO3)2 titration data
were the most stable from trial to trial. However, longer equili-
bration time led to a slightly lower color recovery and a slightly less
than 1:1 thiol/palladium stoichiometry (approximately 0.9:1).
Figure 7 gives typical Pd(II) titration data with the Verkade system
in DMSO when the titration is performed after allowing 12–15 h for
the formation of the SQ1:SEt complex.
sample, the color turn-on was monitored over time at 656 nm. The
normalized kinetic traces are shown in Figure 8. As expected, the
time to full equilibrium was found to increase as the concentration
of Pd(II) increased until saturation is achieved. The inset in Figure 8
shows the region at which the traces reach the half time to full
turn-on. The line drawn at 0.5 normalized absorbance units in the
inset shows that as the concentration of the Pd(II) sample increases,
the half time to completion becomes steadily longer until the

8276
R.J.T. Houk et al. / Tetrahedron 64 (2008) 8271–8278
to reaction completion, the implication is that for unknown sam-
ples containing less than a full equivalency of palladium, the half
time to full turn-on could be used as a secondary validation in
concentration determination.
This understanding of the kinetic exchange of the thiol between
the squaraine core and the palladium analyte allows one to perform
a more accurate titration for calibration purposes by allowing for
the full equilibration to occur between each point as well as during
real sample testing. This means, that at a minimum, approximately
15 min should be given before reading the Pd(II) concentration. To
test this equilibration method, a calibration curve was created just
prior to determinations of several arbitrary samples of known
palladium concentration and samples taken from the cross-cou-
pling reaction of Scheme 2. The injections, both during the gener-
ation of the calibration curve and the analysis of the unknowns,
were allowed to equilibrate for 15 min. The isotherm derived from
titration is shown in Figure 9. The sigmoidal Richards type 1 curve
fit analysis is shown and has a high degree of fidelity.
Two determinations were conducted at each unknown sample
concentration, and the absorbance values were then applied to the
function derived from the curve fit of the isotherm. Table 1 shows
that in this fashion, the average error over the concentration range
tested was 7%. At low concentrations, the error tended to be low,
whereas higher concentrations led to over estimation. However,
the extent of the error is fairly uniform with an average relative
deviation of roughly 4%. Furthermore, based on these de-
terminations, colorimetric detection of palladium as low as
105ꢂ7 ppb in DMSO solution has been reliably performed.
Figure 7. Titration of Pd(NO3)2 into 8.8ꢀ10ꢁ6 M SQ1:SEt in DMSO formulated with
ethanethiol and Verkade base VB. Titrations were conducted after 12–15 h reaction
time between squaraine SQ1 and ethanethiol.
Figure 9. Isotherm and curve fit of kinetically monitored titration. Isotherm was fit
using a sigmoidal Richards type 1 function with the Origin graphing software.
Table 1
Data for arbitrary known samples tested against a fully equilibrated titration iso-
therm of SQ1:SEt [SQ1:SEt]¼1.17ꢀ10ꢁ5 M
Figure 8. (A) Normalized kinetic traces of calibration curve data using a series of
Sample Known [Pd] [Derived #1] [Derived #2]
Average % error
samples of increasing palladium concentration. Below 1 equiv are shown in black and
1
2
3
4
1.62Eꢁ06
2.70Eꢁ06
4.33Eꢁ06
6.49Eꢁ06
1.62Eꢁ06
2.87Eꢁ06
4.74Eꢁ06
7.36Eꢁ06
1.48Eꢁ06
2.58Eꢁ06
4.51Eꢁ06
7.07Eꢁ06
4.6
5.3
6.8
above shown in red. The inset shows the half time. [SQ1:SEt]¼1.2ꢀ10ꢁ5
; [Pd]
range¼9ꢀ10ꢁ7 to 1.5ꢀ10ꢁ5
.
11.1
isotherm of maximum absorbance reaches saturation. Past this
saturation point, the half time stabilizes briefly then begins to
decrease because increasing palladium concentration increases the
rate at which the system achieves equilibrium. Over the concen-
tration range tested (0.1–1.5ꢀ10ꢁ5 M Pd(NO3)2), full equilibration
times ranged from 2.5 to 15.5 min, while the half times ranged from
5 to 84 s. The traces shown in red are those of samples containing
greater than 1 equiv of Pd(II) and show a quickening trend. Since
a large concentration range gives a regular increase in the half time
Average error
Average relative deviation
6.9
3.6
3. Conclusions
In conclusion, we have developed a sensitive chemodosimeter
for palladium(II) using the highly chromogenic squaraine class of
organic dyes. Thermodynamic analysis of the interaction of the

R.J.T. Houk et al. / Tetrahedron 64 (2008) 8271–8278
8277
thiol nucleophile with the squaraine dye shows a unique entropi-
cally driven association to form the decolorized complex in DMSO.
We attribute this anomalous behavior to the release of a highly
ordered solute–solvent complex upon addition of the thiol. Though
similar thiol–squaraine systems have shown selectivity for mercury
in aqueous media, the extension of this methodology into organic
media has allowed for response to other thiophilic metals. Through
careful consideration of many different aspects, calibration of this
system has been achieved for palladium, with detection tested as
low as 100 ppb. Though our focus has been on palladium due to its
wide range of uses in organic synthesis, preliminary observations
have suggested the system could be useful for determination of
other metals such as cadmium, tin, and lead, as well as others.
4.1.3. van’t Hoff analysis
Five hundred microliters of the above prepared solution of
SQ1:SEt with the Verkade base and 500 L DMSO were placed in
a standard 1 cm path length cuvet. UV–vis spectra were collected
upon equilibration of the SQ1:SEt at iteratively increasing tem-
peratures. The temperature was set using a built-in Peltier appa-
ratus and independently monitored in cuvet using a FisherBrand
digital K-thermocouple. The absorbance readings at 656 and
635 nm were used for separate determinations of the thermody-
namic parameters.
m
4.1.4. Kinetics and calibration curve determination (Fig. 2)
SQ1:SEt solutions were prepared as described in the titrations.
Four milligrams of Pd(NO3)2 was dissolved in 10 mL DMSO to give
4. Experimental
4.1. General
a 1.5ꢀ10ꢁ3 M solution. Fifteen vials labeled 1–15 were given 10
mL-
increasing amounts of the Pd(II) stock solution such that vial 1
contained 10
charged with DMSO to bring the total volume in each vial to 150
Directly before each kinetics determination, each vial was charged
mL and vial 15 contained 150
mL. The vials were then
mL.
1H and 13C NMR spectra were recorded on a Varian Unity Plus
300 MHz spectrometer in CHCl3-d. All spectra are recorded at
ambient temperatures. UV–vis experiments were performed on
Beckman DU-70 and DU-800 UV–vis spectrophotometers. Low- and
high-resolution mass spectra were measured with a Finnigan TSQ70
and VG Analytical ZAB2-E instruments, respectively. Compound
SQ1 was synthesized according to the reference method (see Sup-
plementary data).20 All chemicals and reagents were brought from
Aldrich or Fluka and used without further purification. DMSO was
degassed via displacement with N2 and dried over molecular sieves
for at least 6 h prior to use. Dilutions and aliquots were performed
using FisherBrand Finnpipette autopipets calibrated by mass.
with 150 mL of the SQ1:SEt solution to give a total volume of 300 mL
and SQ1:SEt concentration 1.2ꢀ10ꢁ5 M. Directly before each
kinetics determination, each vial was charged with 250
mL of the
prepared SQ1:SEt solution to give a total volume of 500
m
L and
concentration of SQ1:SEt of 1.2ꢀ10ꢁ5 M.
The cuvet was charged with 500 mL of the SQ1:SEt solution and
500
m
L of DMSO to give 1.2ꢀ10ꢁ5 M solutions of SQ1:SEt. The UV–
vis sample holder was kept at a constant temperature of 25 ꢃC by
a built-in Peltier apparatus. Once the sample had equilibrated to
temperature (5–10 min depending on ambient temperature), an
initial wavelength scan was collected. The UV–vis was then
switched into kinetics mode and set to acquire. A timer set to count
4.1.1. UV–vis titrations in DMSO with DBU
down 5 s was on hand. A 20 mL aliquot of the vial being sampled
A stock solution of SQ1 (6.0ꢀ10ꢁ4 M) was prepared by dissolv-
was injected into the cuvet while simultaneously starting the 5 s
timer. The cuvet was shaken vigorously to mix and replaced in the
cell holder within the 5 s countdown. The kinetics collection at
656 nm was started at the completion of the 5 s countdown and the
absorption was monitored until it leveled off. Upon completion,
another wavelength scan was collected for verification of the final
absorbance value. The kinetic traces were then normalized such
that the maximum absorbance was set equal to 1. Five seconds was
added to the start of the trace to account for the time from injection
to the start of data collection.
ing SQ1 (3 mg, 6 mmol) in 10 mL 1:9 CHCl3/DMSO. This stock so-
lution was then used to prepare a 4.7ꢀ10ꢁ5 M solution of SQ1 using
pure DMSO. A separate stock solution of ethanethiol (2.7ꢀ10ꢁ3 M)
and 2 equiv of DBU was also prepared in pure DMSO. This second
solution was then used to prepare a 4.7ꢀ10ꢁ5 M solution of etha-
nethiol. Equal volumes (2 mL) were added together and left for 24 h
to form a theoretical 2.35ꢀ10ꢁ5 M solution of SQ1:SEt. A 1 mL
aliquot of the SQ1:SEt complex was transferred to the UV–vis cuvet.
A separate solution of the Pd(II) salts was prepared at 10 times
palladium concentration and 10 mL aliquots were added and the
spectrum was recorded 5 min after each aliquot injection.
Acknowledgements
4.1.2. UV–vis titrations in DMSO with Verkade base
We thank Chris Welch of Merck Pharmaceuticals for many
helpful discussions. Funding for this research was provided by
Merck Pharmaceuticals, the Welch Foundation (F-1151), and the
NIH (GM077437).
A 4.7ꢀ10ꢁ5 M solution of SQ1 was prepared in DMSO analo-
gously to the previous method. A separate solution of 4.7ꢀ10ꢁ5 M
ethanethiol and 0.75 equiv (3.53ꢀ10ꢁ5 M) 3.31 was also prepared.
Complex SQ1:SEt formation was achieved by combining one part
each of the above solutions with two parts DMSO. The resulting
concentrations of SQ1 and ethanethiol were each 1.18ꢀ10ꢁ5 M.
Decolorization proceeded quickly, though the solution was allowed
to come to equilibrium for 12–15 h prior to use. The complex thus
prepared and stored over molecular sieves was stable for titration
use up to 48 h.
Supplementary data
It includes synthesis and purification of SQ1, Beer’s law analysis
of SQ1, and 1H NMR titration analysis of SQ1–ethanethiol in-
teraction. This material is available free of charge via the internet at
http://pubs.acs.org. Supplementary data associated with this article
can be found in the online version, at doi:10.1016/j.tet.2008.04.105.
The palladium(II) nitrate titrant solution was prepared by di-
lution of 2.5 mg of Pd(NO3)2$2H2O in 10 mL DMSO. One hundred
microliters of this solution is combined with 750
mL of the SQ1:SEt
References and notes
complex solution and 150 L DMSO to give a 1 mL solution con-
m
taining roughly 10 equiv of Pd(II) to squaraine. A standard, quartz,
3 mL volume, 1 cm path length UV–vis cuvet was charged with
1. Liesen, P. J.; Varma, R. S.; Naicker, K. P. Tetrahedron Lett. 1999, 40, 2075–2078.
2. Liu, W.-J.; Xie, Y.-X.; Liang, Y.; Li, J.-H. Synthesis 2006, 860–864.
3. Baxter, J. M.; Steinhuebel, D.; Palucki, M.; Davies, I. W. Org. Lett. 2005, 7, 215–218.
4. Buchwald, S. L.; Mauger, C.; Mignani, G.; Scholzc, U. Adv. Synth. Catal. 2006, 348,
23–39.
750
m
L of the 1.18ꢀ10ꢁ5 M solution of SQ1:SEt and 250
mL DMSO
giving a final concentration of 8.82ꢀ10ꢁ6 M. The titration was
performed by administering successive 10
solution.
mL aliquots of the Pd(II)
5. Arvela, Riina K.; Leadbeater, N. E.; Collins, Michael J. Tetrahedron 2005, 61,
9349–9355.

8278
R.J.T. Houk et al. / Tetrahedron 64 (2008) 8271–8278
6. Mohr, G. G. Sensors Actuators B 2005, 107, 2–13.
13. Schmidtchen, F. P. Coord. Chem. Rev. 2006, 250, 2918–2928.
14. Law, K. Y. J. Phys. Chem. 1987, 91, 5184–5193.
15. CRC Handbook of Chemistry and Physics, 76th ed.; CRC: Boca Raton, FL, 1995, pp
9-(65-66).
7. Ros-Lis, J. V.; Garcia, B.; Jimenez, D.; Martinez-Manez, R.; Sancenon, F.;
Soto, J.; Gonzalvo, F.; Valldecabres, M. C. J. Am. Chem. Soc. 2004, 126, 4064–
4065.
8. Ros-Lis, J. V.; Martinez-Manez, R.; Soto, J. Chem. Commun. 2002, 2248–2249.
9. Bigelow, R. W.; Freund, H. J. Chem. Phys. 1986, 107, 159–174.
10. Ros-Lis, J. V.; Marcos, M. D.; Martinez-Manez, R.; Rurack, K.; Soto, J. Angew.
Chem., Int. Ed. 2005, 44, 4405–4407.
16. Gordon, A. J.; Ford, R. A. The Chemist’s Companion; John Wiley and Sons: New
York, NY, 1972; pp 57–59.
17. Chrisment, J.; Nicole, D.; Delpuech, J. J. C. R. Seances Acad. Sci., C: Sci. Chim. 1978,
286, 541–544.
11. Optical sensors for Pd(II) are rare, but for two examples, see: Lee, D. Y. K.; Lam,
M. H. W.; Wong, W.-Y. New J. Chem. 2002, 26, 330–335; Song, F.; Garner, A. L.;
Koide, K. J. Am. Chem. Soc. 2007, 129, 12354–12355.
12. Ros-Lis, J. V.; Martinez-Manez, R.; Rurack, K.; Sancenon, F.; Soto, J.; Spieles, M.
Inorg. Chem. 2004, 43, 5183–5185.
18. Lensink, C.; Xi, S. K.; Daniels, L. M.; Verkade, J. G. J. Am. Chem. Soc. 1989, 111,
3478–3479.
19. Laramay, M. A. H.; Verkade, J. G. J. Am. Chem. Soc. 1990, 112, 9421–9422.
20. Tian, M.; Furuki, M.; Iwasa, I.; Sato, Y.; Pu, L. S.; Tatsuura, S. J. Phys. Chem. B
2002, 106, 4370–4376.