New nucleophilic catalysts for bright and fast peroxyoxalate chemiluminescence

Article abstract of DOI:10.1021/ac991339a

Miniaturized detection applications based on chemiluminescence require fast reaction kinetics for optimum performance. In this work, high-intensity light from the analytically useful peroxyoxalate chemiluminescence reaction has been generated at high rates by employing both single-component and dual- component nucleophilic catalysis. 4-(Dimethylamino)pyridine and its derivatives were superior to all other bases in terms of reaction speed and intensity of the generated light and outshone imidazole, which hitherto has been considered as the best catalyst. The light intensity was related to the difference in pK(a) between the 4-aminopyridine catalyst and the leaving group of the reagent, and the optimum ΔpK(a) was found to be close to 0. Similarly, high light intensities were obtained when mixtures of the imidazole analogue 1,2,4-triazole and the strong, nonnucleophilic base 1,2,2,6,6-pentamethylpiperidine acted as catalysts. The mechanism behind this was concluded to be a 'base-induced nucleophilic catalysis', where the ancillary strong base assisted the production of the highly nucleophilic 1,2,4-triazolate anion, which as the actual catalyst then participated in the formation of a more reactive transient reagent. All the investigated catalysts reduced the light yield of the reaction due to base-catalyzed breakdown reactions of the reagents and/or intermediates. The intensity peak maximums of these bright and fast reactions typically appeared after less than 10 ms, whereafter the light decayed to darkness within a few seconds. These reaction characteristics are especially advantageous for sensitive detection applications where the observation volumes and times are limited, e.g., peaks emerging from a capillary-based separation process.

Full text of DOI:10.1021/ac991339a

Anal. Chem. 2000, 72, 1373-1380
New Nucleophilic Catalysts for Bright and Fast
Peroxyoxalate Chemiluminescence
Tobias Jonsson and Knut Irgum*
Department of Chemistry, Umeå University, S-901 87 Umeå, Sweden
band-broadening is to be kept within acceptable limits, the time
constant of the detection system must not exceed one-tenth of
the width of the eluted peak.2 This means that if chemiluminescent
(CL) reactions are to be applied as detection schemes with these
efficient separation techniques, the light can only be collected
during a few milliseconds if the separation efficiency is to be
preserved. By only physically reducing the visible volume of the
detection cell while employing slow CL reactions, the risk of light-
piping from a hidden part of the flow path greatly increases.
Equally unacceptable in miniaturized systems are the extra flow
elements that are demanded by the CL reactions from which the
light intensity slowly climb to its maximum intensity. All these
criteria call for an increase in the rate of light production and an
increased light intensity. A rapid decay of the light can also be
advantageous since the effective cell volume then becomes
independent of the physical volume and only relies on the light
decay kinetics. In liquid chromatography this phenomenon has
been called “chemical band-narrowing”3 although this description
is quite misleading since the effect rather is a reduction of the
detection cell related band-broadening.
The reaction between hydrogen peroxide and an activated
derivative of oxalic acid can produce chemiluminescent light in
the presence of a fluorophore.4-6 This reaction, commonly referred
to as peroxyoxalate chemiluminescence (POCL), has attracted
much attention and has successfully been used as a highly
sensitive detection technique in several procedures developed for
trace level determinations of a large variety of analytes,4 e.g.,
fluorophore-labeled amino acids7 and steroids,8 polycyclic aromatic
hydrocarbons,9 and hydrogen peroxide.10 However, while numer-
ous reports concerning the reaction mechanism have been
Miniaturized detection applications based on chemilumi-
nescence require fast reaction kinetics for optimum
performance. In this work, high-intensity light from the
analytically useful peroxyoxalate chemiluminescence re-
action has been generated at high rates by employing both
single-component and dual-component nucleophilic ca-
talysis. 4 -(Dimethylamino)pyridine and its derivatives
were superior to all other bases in terms of reaction speed
and intensity of the generated light and outshone imida-
zole, which hitherto has been considered as the best
catalyst. The light intensity was related to the difference
in pKa between the 4 -aminopyridine catalyst and the
leaving group of the reagent, and the optimum ∆pKa was
found to be close to 0. Similarly, high light intensities were
obtained when mixtures of the imidazole analogue 1 ,2 ,4 -
triazole and the strong, nonnucleophilic base 1 ,2 ,2 ,6 ,6 -
pentamethylpiperidine acted as catalysts. The mechanism
behind this was concluded to be a “base-induced nucleo-
philic catalysis”, where the ancillary strong base assisted
the production of the highly nucleophilic 1 ,2 ,4 -triazolate
anion, which as the actual catalyst then participated in
the formation of a more reactive transient reagent. All the
investigated catalysts reduced the light yield of the reac-
tion due to base-catalyzed breakdown reactions of the
reagents and/ or intermediates. The intensity peak maxi-
mums of these bright and fast reactions typically appeared
after less than 1 0 ms, whereafter the light decayed to
darkness within a few seconds. These reaction charac-
teristics are especially advantageous for sensitive detec-
tion applications where the observation volumes and
times are limited, e.g., peaks emerging from a capillary-
based separation process.
(2) Scott, R. P. W. Liquid Chromatography Detectors, 2nd ed.; Journal of
Chromatography Library, Vol. 33; Elsevier: Amsterdam, 1986; p 27.
(3) de Jong, G. J.; Lammers, N.; Spruit, F. J.; Brinkman, U. A. Th.; Frei, R. W.
Chromatographia 1 9 8 4 , 18, 129-133.
(4) Stigbrand, M.; Jonsson, T.; Ponte´n, E.; Irgum, K.; Bos, R. Mechanism and
Applications of Peroxyoxalate Chemiluminescenc. In Chemiluminescence
in Analytical Chemistry; La Campan˜ a, A.-M., Baeyens, W. J., Eds.; Marcel
Dekker: New York, in press.
(5) Hadd, A. G.; Birks, J. W. In Selective Detectors: Chemical Analysis; Sievers,
R. E., Ed.; John Wiley & Sons: New York, 1995; Vol. 131, pp 209-240.
(6) Givens, R. S.; Schowen, R. L. In Chemiluminescence and Photochemical
Reaction Detection in Chromatography; Birks, J. W., Ed.; VCH: New York,
1989; Chapter 5, pp 125-147.
Rapid development in the field of liquid separation sciences
toward higher efficiency and miniaturized chromatographic, elec-
trochromatographic, and electrophoretic systems has resulted in
the reduction of the bandwidth of the eluted peaks to typically a
few nanoliters,1 which corresponds to a residence time in the
detector cell of 1 s or less. This requires an adaptation of the
detection systems to stand up to the new and harder demands
on sensitivity and bandwidth conservation. If the extracolumn
(7) Kobayashi, S.; Imai, K. Anal. Chem. 1 9 8 0 , 52, 424-427.
(8) Appelblad, P.; Jonsson, T.; B¨ackstro¨ m, T.; Irgum, K. Anal. Chem. 1 9 9 8 ,
70, 5002-5009.
* Corresponding author: (e-mail) kim@chem.umu.se.
(9) Sigvardson, K. W.; Kennish, J. M.; Birks, J. W. Anal. Chem. 1 9 8 4 , 56,
(1) Cruz, L.; Shippy, S. A.; Sweedler, J. V. In High Performance Capillary
Electrophoresis. Chemical Analysis; Khaledi, M. G., Ed.; John Wiley &
Sons: New York, 1998; Vol. 146, pp 303-354.
1096-1102.
(10) Emteborg, M.; Irgum, K.; Gooijer, C.; Brinkman, U. A. Th. Anal. Chim.
Acta 1 9 9 7 , 357, 111-118.
10.1021/ac991339a CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/26/2000
Analytical Chemistry, Vol. 72, No. 7, April 1, 2000 1373

published,11-16 debate still exists as to the actual mechanism of
the reaction. It is however fairly well established that the preferred
properties of the fluorophore are a low excitation energy and a
low oxidation potential9,13,17 as explained by the chemically initiated
electron-exchange luminescence (CIEEL)18 theory. Because the
intensity of the light represents the rate of the light-producing
reaction, the most favorable oxalic acid derivatives for sensitive
detection are highly activated, and thus reactive, oxalates19 or
oxamides.20 The function of the catalyst usually present is to
promote and speed up the reaction to produce a higher light
intensity. Besides its basic properties, not much is known about
the preferred nature of the catalyst, and the popularity of imidazole
has been based on experimental results21,22 only. Our study23 of
substituted imidazoles and other good leaving groups as catalyst
for this reaction unfortunately did not provide any compound
superior to imidazole, but clearly illustrated the importance of the
nucleophilicity of the catalyst.
The earlier unexplained catalytic efficiency of imidazole was
clarified by the introduction20 of 1,1′-oxalyldimidazole (ODI) as a
stand-alone reagent in POCL. Recent studies of the reaction of
imidazole with activated oxalates24,25 and imidazole-catalyzed
POCL15,16 have confirmed that ODI forms as a transient intermedi-
ate during the reaction. The formation of an intermediate reagent,
more activated and reactive than the primary reagent, is by
definition26 the result of nucleophilic catalysis. The mechanism
is in many aspects similar to the nucleophilic catalyzed hydrolysis
of monocarboylic esters (e.g., 4-nitrophenyl acetate).26,27 The
formation of an intermediate seems to be a common feature of
efficient catalysts for POCL,23 corroborating the conclusion that
the POCL reaction is predominantly subject to nucleophilic
catalysis, rather than general base catalysis. When imidazole is
used as catalyst for the POCL reaction of bis(2,4,6-trichlorophenyl)
oxalate (TCPO), the formation of ODI is the reaction step limiting
the rate of light production15,16 and, thus, the light intensity. The
studies on the reaction between imidazole and TCPO24,25 have also
exposed the cooperative action of two imidazole molecules, where
one acts as a general base catalyst for the other, which acts as a
nucleophile. From the inferior results of catalysis by 1-methylimi-
dazole,23 it can be concluded that this cooperation between two
imidazole molecules is essential for the catalytic efficiency of
imidazole and that the neutral imidazole is too weak a nucleophile
to attack a moderately reactive reagent such as TCPO.
An efficient nucleophilic catalyst must be an unusually effective
nucleophile and the formed intermediate must be unusually
susceptible toward nucleophilic attack.26 The most obvious way
to increase the catalytic efficiency is to use a stronger nucleophile,
but the concept of nucleophilicity toward esters follows no simple
equation28 and is dependent on several diverse properties such
as basicity, polarizability, hydrogen-bonding ability, solvation,
resonance stabilization, and steric effects. Another, maybe less
palpable approach, is to use a mixture of catalysts, where each
compound ideally is dedicated to a specific catalytic task. In the
POCL reaction, where the complexity of the catalytic role is high
and involves features of nucleophilicity, leaving group ability, and
basicity,23 such a design may very well prove to be successful. In
fact, we have recently reported29 on the combined use of
1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) and 1,2,4-triazole as
catalysts for POCL and demonstrated the potential of this system
to outshine the imidazole-catalyzed reactions.
4-(Dimethylamino)pyridine (DMAP) and other 4-(dialkylamino)-
pyridines30-32 are strong nucleophiles and have for long been used
to catalyze acylation reactions. These compounds act specifically
as nucleophilic catalysts, and the mechanism is believed to involve
the formation of acylpyridinium ions, which are stabilized by the
ability to distribute a positive charge over several resonance
structures. These highly reactive ions form loose ion pairs with
the leaving groups from the acyl carbon and can thus be easily
attacked by another nucleophile. DMAP has once been tested33
as a catalyst in the POCL reaction with somewhat promising
results, but the effect was only briefly studied and no comparison
was made with imidazole. Another widely used acylation catalyst
is 1-methylimidazole,34 but its basicity and nucleophilicity is
apparently23 too low for the use as a catalyst in POCL.
In this study, several new compounds have been tested as
catalysts for the POCL reaction with the goal of achieving faster
reactions and higher light intensities to allow for more sensitive
detection applications. Catalysis by strong bases and strong
nucleophiles such as DMAP and mixtures containing two catalysts
are the different approaches that have been used to realize this.
This examination also aims to outline the optimal properties of a
catalyst, or combination of catalysts, to allow for a more rational
design and selection of reagents and catalysts in the future.
(11) Rauhut, Bollyky, L. J.; Roberts, B. G.; Loy, M.; Whitman, R. H.; Ianotta, A.
V.; Semsel, A. M.; Clarke, R. A. J. Am. Chem. Soc. 1 9 6 7 , 89, 6515-6522.
(12) Catherall, C. L. R.; Palmer, T. F.; Cundall, R. B. J. Chem. Soc., Faraday
Trans. 2 1 9 8 4 , 80, 823-834.
(13) Catherall, C. L. R.; Palmer, T. F.; Cundall, R. B. J. Chem. Soc., Faraday
Trans. 2 1 9 8 4 , 80, 837-849.
(14) Orlovic, M.; Givens, R. S.; Alvarez, F.; Matuszewski, B.; Parekh, N. J. Org.
Chem. 1 9 8 9 , 54, 3606-3610.
(15) Stevani, C. V.; Lima, D. F.; Toscano, V. G.; Baader, W. J. J. Chem. Soc.,
Perkin Trans. 2 1 9 9 6 , 989-995.
(16) Hadd, A. G.; Robinson, A. L.; Rowlen, K. L.; Birks, J. W. J. Org. Chem. 1 9 9 8,
63, 3023-3031.
EXPERIMENTAL SECTION
Reagents and Solutions. TCPO was synthesized as previ-
ously described35 or purchased from Aldrich (Steinheim, Ger-
many) or Sigma (St. Louis, MO). ODI was prepared as formerly
outlined,36 under a nitrogen atmosphere and from freshly distilled
(17) Honda, K.; Miyaguchi, K.; Imai, K. Anal. Chim. Acta 1 9 8 5 , 177, 111-120.
(18) Schuster, G. W. Acc. Chem. Res. 1 9 7 9 , 12, 366-373.
(19) Honda, K.; Miyaguchi, K.; Imai, K. Anal. Chim. Acta 1 9 8 5 , 177, 103-110.
(20) Stigbrand, M.; Ponte´n, E.; Irgum, K. Anal. Chem. 1 9 9 4 , 66, 1766-1770.
(21) Hanaoka, N.; Givens, R. S.; Schowen, R. L.; Kuwana, T. Anal. Chem. 1 9 8 8 ,
60, 2193-2197.
(22) Imai, K.; Nishitani, A.; Tsukamoto, Y.; Wang, W.-H.; Kanada, S.; Hayakawa,
K.; Miyazaki, M. Biomed. Chromatogr. 1 9 9 0 , 4, 100-104.
(23) Jonsson, T.; Emteborg, M.; Irgum, K. Anal. Chim. Acta 1 9 9 8 , 361, 205-
215.
(24) Neuvonen, H. J. Chem. Soc., Perkin Trans. 2 1 9 9 5 , 945-949.
(25) Hadd, A. G.; Birks, J. W. J. Org. Chem. 1 9 9 6 , 61, 2657-2663.
(26) Bender, M. L. Mechanisms of Homogeneous Catalysis from Protons to Proteins;
John Wiley & Sons: New York, 1971; pp 147-179.
(28) Jencks, W. P.; Carriuolo, J. J. Am. Chem. Soc. 1 9 6 0 , 82, 1778-1786.
(29) Jonsson, T.; Irgum, K. Anal. Chim. Acta 1 9 9 9 , 400, 257-264.
(30) Ho¨ fle, G.; Steglich, W.; Vorbru¨ ggen, H. Angew. Chem., Int. Ed. Engl. 1 9 7 8 ,
17, 569-583.
(31) Scriven, E. F. V. Chem. Soc., London 1 9 8 3 , 12, 129-161.
(32) DMAP: Update, Reilly Report; Reilly Industries, Indianapolis, IN, 1982.
(33) Orosz, G.; Torkos, K.; Borossa, J. Acta Chim. Hung. 1 9 9 1 , 128, 911-917.
(34) Connors, K. A.; Pandit, N. K. Anal. Chem. 1 9 7 8 , 50, 1542-1545.
(35) Mohan, A. G.; Turro, N. J. J. Chem. Educ. 1 9 7 4 , 51, 528-529.
(36) Murata, S. Chem. Lett. 1 9 8 3 , 1819-1820.
(27) Kirsch, J. F.; Jencks, W. P. J. Am. Chem. Soc. 1 9 6 4 , 86, 833-837.
1374 Analytical Chemistry, Vol. 72, No. 7, April 1, 2000

oxalyl chloride. ODI was stored at 5 °C under a moisture-free
atmosphere in tight-closing vials protected from light. DMAP was
either from Reilly Industries (Indianapolis, IN) or Aldrich. 4-Ami-
nopyridine (98%), 4-pyrrolidinopyridine (PYP; 98%), 1,2,2,6,6-
pentamethylpiperidine (PMP; 97%), DBU (98%), 3-aminofluoran-
thene (3-AFA; 97%), 2,4,6-trichlorophenol (98%), (trimethylsilyl)-
imidazole (“derivatization grade”), and oxalyl chloride (98%) were
obtained from Aldrich. Fluka (Buchs, Switzerland) supplied
triethylamine (p.a.) and morpholine (purum). Hydrogen peroxide
(Perhydrol Suprapur, 30% in water) was from Merck (Darmstadt,
Germany), HPLC-grade acetonitrile (specified to contain <0.01%
water) was from J. T. Baker (Deventer, The Netherlands), and
piperazine (anhydrous) was from Sigma. Water was purified and
deionized to a conductivity of less than 60 nS‚cm-1 using Milli-Q
equipment (Millipore, Bedford, MA).
Reagent solutions of TCPO (1 mM) in acetonitrile were
prepared every second day, while solutions of ODI (1 mM) were
prepared every fourth hour in dried acetonitrile (dried using a
3-Å molecular sieve from KeboLab, Stockholm, Sweden). These
solutions were stored in poly(tetrafluoroethene) (PTFE) bottles
equipped with silica gel-filled drying tubes to enhance the reagent
stability37 and to protect the solutions from ambient humidity
during withdrawal. Stock solutions of 4-aminopyridine, DMAP, and
PYP were prepared every second week. The mixed hydrogen
peroxide/ fluorophore/ catalyst reagent solutions were made daily
from stock solutions of each component and contained 1 mM
hydrogen peroxide, 51 µM 3-AFA, and the catalysts (10 mM unless
otherwise noted) in acetonitrile. Note that the reaction cell
concentrations are half of the concentrations in the reagent
solutions.
Estimation of Basicity. The pKa measurements were per-
formed with a HP8452A diode array spectrophotometer (Hewlett-
Packard, Palo Alto, CA), which had been blanked against a quartz
cuvette with pure acetonitrile. Each base was mixed with 2,4,6-
trichlorophenol at approximate concentrations of 2 and 0.5 mM,
respectively, and the absorbance of the formed 2,4,6-trichlorophe-
nolate ion was measured at the absorbance maximum with a 10-
nm slit window. The absorbtivity of the 2,4,6-trichlorophenolate
ion was determined in an acetonitrile solution saturated with
potassium hydroxide. With the knowledge of the total concentra-
tions of the two species in the cuvette and the measured
concentration of the 2,4,6-trichlorophenolate ion, the relative pKa
values for each base could be calculated. These values were
thereafter calibrated against the known pKa’s38 of morpholine and
triethylamine. The reported data are the averages of two or three
different measurements and had a relative standard deviation of
∼1%.
monitored. The high-sensitivity emission photomuliplier tube of
the instrument was operated at 500 V and mounted directly on
the backside of the cell assembly without any light or wavelength
restrictions. For each reagent composition, three voltage vs time
traces were recorded, overlaid, and averaged at frequencies of
4000, 40, and 4 Hz (and if required also at 0.4 Hz). These three
(or four) traces of different frequencies were linked together and
leveled with the initial 4-kHz trace to allow for evaluation of all
parameters of the light emission profile in a spreadsheet computer
program. The half-life of the light intensity was always measured
from the intensity peak time.
Spectrophotometric Reaction Studies. The studies of re-
agent decomposition were performed at 22 ( 1 °C, either with
the stopped-flow instrument described above operated in the
absorbance mode or with the diode array spectrophotometer
mentioned earlier in which a pneumatically driven HP89054/ 55A
(Hewlett-Packard) stirring module was mounted. When the
stopped-flow instrument was used, equal volumes of a TCPO or
ODI reagent solution and a diluted catalyst solution were injected
into the detection cell chamber. With the spectrophotometer,
appropriate amounts of TCPO or ODI solutions were pipetted into
a quartz cuvette and diluted with acetonitrile, after which base
was added and the measurement immediately started. The
concentrations in the detection chamber of oxalic reagent and
catalyst in acetonitrile were in both cases 0.5 and 5 mM,
respectively. The decay of TCPO was monitored by the accumula-
tion of the 2,4,6-trichlorophenolate ion (330 nm), whereas the
decline of ODI was directly measured at an appropriate wavelength
(330 nm) where the absorbtivities of the catalysts were negligible.
Absorbance changes were monitored with wavelength slit window
of 2-10 nm. Each first-order rate constant was calculated as the
slope of the forced-zero-intercept linear least-squares fit (average
R2 ) 0.99, except for piperazine where R2 ) 0.86) of the first part
of a plot of the natural logarithm of the inverse of the relative
reagent concentration vs time.
WARNING. The degradation product from TCPO, 2,4,6-
trichlorophenol, and the fluorophore 3-AFA are suspected car-
cinogens. All 4-aminopyridines are potentially toxic and should
be handled with due care.
RESULTS AND DISCUSSION
Reaction Conditions. Acetonitrile was chosen as the solvent
due to its frequent use in analytical applications of POCL and its
nearly aprotic properties. The two activated oxalic acid derivatives
ODI and TCPO were selected as model reagents to represent very
reactive and moderately reactive reagents, respectively. Because
of the modest reactivity of TCPO, a base catalyst is required for
its reaction with hydrogen peroxide, whereas ODI is sufficiently
reactive to be used as a stand-alone reagent.20 Equal concentra-
tions of the oxalic reagent and hydrogen peroxide were used
throughout, and the concentration of catalyst was selected to 10
times higher than the other reagents to obtain a fairly constant
catalytic environment. A series of different bases of various
strengths and nucleophilicities were tried as catalysts. The strong
base and efficient acylation catalyst DMAP and its structural
analogues 4-aminopyridine and PYP were tested together with four
other strong bases, morpholine, piperazine, triethylamine, and
PMP, which all have aqueous pKa’s comparable to DMAP; see
Table 1. Triethylamine is known as a base with low nucleophlicity
Stopped-Flow Chemiluminescence. All time-resolved chemi-
luminescence measurements were performed at 25.0 ( 0.1 °C
using a computer-controlled stopped-flow spectrofluorometer
(model DX-17MV, Applied Photophysics, Leatherhead, U.K.) with
the light source off. The two reagent solutions were simulta-
neously introduced into the detection cell at a 1:1 mixing ratio
from two gastight glass syringes. A total volume of 200 µL was
flushed through the cell at a combined flow rate of ∼15 mL/ s
before the flow was abruptly stopped and the light production
(37) Emteborg, M.; Ponte´n, E.; Irgum, K. Anal. Chem. 1 9 9 7 , 69, 2109-2114.
(38) Coetzee, J. F.; Padmanabhan, G. R. J. Am. Chem. Soc. 1 9 6 5 , 87, 5005-
5010.
Analytical Chemistry, Vol. 72, No. 7, April 1, 2000 1375

The data extracted from the chemiluminescence vs time traces
obtained with this system and the different catalysts and reagents
previously mentioned are summarized in Table 2, and representa-
tive emission profiles from some illustrative catalyst/ reagent
combinations are shown in Figure 1. A quick glance at Figure 1
reveals the profound differences between the two reagents TCPO
and ODI. With TCPO, the light intensities were generally lower
and the kinetics were slower. The light yields were, on the other
hand, much higher with TCPO. For these reasons, TCPO is most
appropriate for large-volume detection applications, whereas ODI
is better suited for miniaturized systems which require fast
kinetics. Shifting part of the attention to Table 2 and the
performance of the different catalysts discloses two striking effects
common to all the strong bases. First, the peak time of the light
production was short, typically a few tens of milliseconds,
compared to ∼0.5 s or more with imidazole as catalyst. Second,
the total light yields were poor, at least 1 order of magnitude lower
than with imidazole. The worst light yields were obtained with
morpholine, piperazine, and triethylamine. These reactions emitted
less than 1% of the amount of light produced by the imidazole-
mediated reactions. When comparing the intensities of the emitted
light, the 4-aminopyridines (4-aminopyridine, DMAP, PYP) and
the other bases produced very different results. All 4-aminopy-
ridines gave high light intensities, ranging from one-third of up
to threefold higher than the imidazole-catalyzed reactions, depend-
ing on the reagent. The other bases gave very low light intensities,
especially with TCPO. Although the light intensity half-lifes were
diverse with the TCPO reagent, and all quite short with ODI, the
general tendency was much shorter half-lifes with all strong bases
compared to with imidazole.
Table 1. Dissociation Constants for Different
Compounds in Water and in Acetonitrilea
compoundb
pKaH O
pKaCH CN
2
3
no.
1
2
3
4
5
6
7
8
9
2,4,6-trichlorophenolate
imidazole
4-aminopyridine
DMAP
6.0 40
6.99 41
9.29 42
9.71 42
9.90 30
8.49 41
9.78 41
10.72 41
11.25 43
20.4 44
14.5 45
16.8
17.1
17.6
PYP
morpholine
piperazine
triethylamine
PMP
16.7c
18.4
18.4c
18.4
a
The autoprotolysis constant for acetonitrile is 3 × 10-29 38
.
b The
pKa values for these bases refer to the release of a proton from the
corresponding acid; i.e.; BH f B- + H+. c The literature values for
these compounds were used to calibrate the relative pKa scale in this
study. Literature values for morpholine and triethylamine are 16.61
and 18.46,38 respectively.
and PMP is a very sterically hindered base with essentially no
nucleophilicity.39 For comparison, the rather weak base imidazole,
which has a thoroughly documented21-23 efficiency as POCL
catalyst, was also included in the study. Because the basicities of
these compounds in acetonitrile most certainly differ from that
in water and such data could not be found in the literature, the
relative pKa’s of the bases were probed in dilute mixtures of the
compounds and 2,4,6-trichlorophenol in pure acetonitrile, using
the absorbance of the negatively charged phenoxide ion as
indicator. These data are also presented in Table 1 and reveal
that the 2,4,6-trichlorophenol released from TCPO during the
POCL reaction will be partially deprotonated with strong bases
as catalysts. If, for example, triethylamine is used as a catalyst
for TCPO at a concentration ratio of 10:1 as in this study, ∼20%
of the phenol and 4% of the base is expected to be ionized after
the reaction is completed. This proton transfer can cause rather
peculiar effects on the light emission profile at low base concen-
trations.46
Single-Component Catalysis. During the very first experi-
ments with strong bases as catalysts for the POCL reaction, it
became clear that the light production in many cases was so fast
that the stopped-flow instrument23 failed to detect the peak of the
emission profile. An inert zero-dead volume integrated cyclone
mixer and fountain detection cell was therefore manufactured and
initially utilized29 for studying the rapid kinetics induced by these
catalytic systems. However, with the fastest catalysts it became
apparent that this construction suffered from problems in terms
of insufficient mixing and development of secondary flow paths
within the cell chamber. Consequently, this strategy was aban-
doned and the equipment described above was employed.
The large contrast in light intensities and yields between the
4-aminopyridines and the other strong bases implies that there is
a difference in their catalytic mechanisms, since the difference in
base strength between the compounds is very small. The evident
nucleophilic properties of the 4-aminopyridines, the documented
sensitivity of the POCL reaction to nucleophilic catalysis, and the
high light intensities produced with the 4-aminopyridines com-
pared to with other strong bases together provide significant
circumstantial evidence for nucleophilic catalysis of the POCL
reaction by these compounds. Thus, new and more reactive
derivatives of oxalic acid are expected to be formed during the
course of the reaction. In the UV absorption experiments
discussed below, we also hoped to be able to observe the
formation of such intermediates since this, as previously men-
tioned, would be evidence for nucleophilic catalysis. Unfortunately,
no intermediate formation could be detected with any of the
catalysts, but this does not rule out its existence since the bases
themselves are strong absorbers in most of this UV region and
the concentration of such intermediates most likely is low.
Although this investigation does not aim at disclosing the identity
of the suspected intermediate, a qualified speculation is that the
structure resembles the acylpyridinium ions that have been
observed by others.30
(39) Sommer, H. Z.; Lipp, H. I.; Jackson, L. L. J. Org. Chem. 1 9 7 1 , 36, 824-
828.
(40) Weast, R. C., Ed. Handbook of Chemistry and Physics, 57th ed.; CRC Press:
Cleveland, OH, 1974; p D-151.
(41) Dean, J. A., Ed. Lange’s Handbook of Chemistry, 13th ed.; McGraw-Hill: New
York, 1985.
(42) Essery, J. M.; Schofield, K. J. Chem. Soc. 1 9 6 1 , 3939-3953.
(43) Hall Jr., H. K. J. Am. Chem. Soc. 1 9 5 7 , 79, 5444-5447.
(44) Magonski, J.; Pawlak, Z.; Jasinski, T. J. Chem. Soc., Faraday Trans. 1 9 9 3 ,
89, 119-122.
A closer examination of the 4-aminopyridines and their relative
performance reveals that the kinetics as well as the light intensity
and yield were dependent on the base strength. The light yield,
peak time, and half-life all decreased with increasing base strength
regardless of the reagent. Interestingly, the light intensity in-
(45) Wubbels, G. G.; Sevetson, B. R.; Sanders, H. J. Am. Chem. Soc. 1 9 8 9 , 111,
1018-1022.
(46) Jonsson, T.; Irgum, K., manuscript in preparation.
1376 Analytical Chemistry, Vol. 72, No. 7, April 1, 2000

Table 2. Effect of Base Catalyst and Oxalic Reagent on the Peroxyoxalate Chemiluminescence Reaction Light
Emission Profile in Acetonitrilea
TCPO as reagent
ODI as reagent
max intens,
mV
light yield,
peak time,
half-life,
ms
max intens,
mV
light yield,
peak time,
half-life,
ms
no
catalyst
none
imidazole
4-aminopyridine
DMAP
mV‚s
ms
mV‚s
ms
1
2
3
4
5
6
7
8
9
n/ ab
980
300
620
1400
4
8
38
99
n/ a
11000
1400
1000
900
47
4
90
500
n/ a
4900
3
2
1
98
2
15
35
n/ a
9600
670
83
43
2100
46
75
620
2200
1900
1800
37
39
110
210
4900
4200
390
180
100
11
650
250
8
4
2
9
2
2
1
43000
4000
81
27
11
120
7
14
PYP
morpholine
piperazine
triethylamine
PMP
1
2
9
610
2300
49
a
The uncertainty of the extracted data in this table was estimated to 2-10% relative standard deviation (n ) 12), based on four repeated runs
(each consisting of averaged repeated injections) on three different preparations of DMAP/ hydrogen peroxide/ fluorophore mixed reagent solutions
reacted with one single ODI reagent solution. The yield and half-life showed the largest deviations. Of course, the relative standard deviations of
repeated injections of the same solutions were much smaller, typically below 1%. b n/ a, no detectable chemiluminescence.
Figure 1. Time-resolved peroxyoxalate chemiluminescence with different base catalysts and oxalic reagents. The figure has been split into
three parts with different time resolutions to visualize the substantial impact that the catalysts had on the reaction. The concentrations of the
compounds in the reaction vessel were 0.5 mM TCPO or ODI, 0.5 mM hydrogen peroxide, 5 mM catalyst, and 25 µM 3-AFA as fluorophore. T
and O mean TCPO and ODI, respectively, and the numbers refer to the position of the catalysts in Table 2. Key: (O-1) ODI with no catalyst,
(O-2) ODI with imidazole, (O-3) ODI with 4-aminopyridine, (T-2) TCPO with imidazole, and (T-5) TCPO with PYP.
creased linearly (R2 ) 0.99) with pKa when TCPO was the reagent
but decreased (R2 ) 0.83) when ODI was used. It is notable that
the 4-aminopyridine giving the highest intensity was for both
reagents the one with the smallest difference in pKa relative to
the leaving group. By plotting the observed intensities vs the
difference in pKa (see Figure 2), we can conclude that there exists
an optimum close to zero; i.e., the 4-aminopyridine catalyst and
the leaving group should be of roughly the same basicity to
generate the highest possible light intensity. Thus, each reagent
has its own dedicated catalyst within the 4-aminopyridine series,
which is why it should be beneficial to explore other derivatives
than the ones examined here.
Competing Reactions. The low light yields obtained with all
strong bases implies that, as we previously suggested29 for DMAP,
there exist competing non-light-producing reactions that are
accelerated by these compounds. The negative effect of bases on
the quantum yield of the POCL reaction has been observed in
several previous studies.11,12,14,16,24,37,47 Various different explana-
tions have been presented to account for this observation, e.g.,
formation of a complex between the base and the oxalic reagent,12
catalysis of reagent37 or intermediate16 breakdown and, in semi-
aqueous environments, base-catalyzed hydrolysis of the re-
agent.24,47 Anyhow, such dark reactions would severely limit the
amount of reagents and/ or intermediates available for light
production and may also rationalize the occurrence of a maximum
in light intensity at minimum basicity difference discussed above.
To examine this theory and to investigate the stability of the
activated oxalic acid derivatives in strongly basic environments,
the decomposition rates of the oxalic reagents in the presence of
different strong base catalysts in acetonitrile were studied by UV
spectroscopy. As evident from Table 3, both TCPO and ODI were
unstable in the presence of the bases and decomposed quite
rapidly, with one notable exception: TCPO with PMP. Aside from
this exception, the decays of the reagents were generally slower
with the 4-aminopyridines than with the other strong bases,
despite their ability to produce fast and intense light in the
corresponding POCL reactions. There was no direct correlation
between the basicities of the catalysts (see Table 1) and the
reagent decomposition rates. However, among the 4-aminopy-
(47) DeVasto, J. K.; Grayesky, M. L. Analyst 1 9 9 1 , 116, 443-447.
Analytical Chemistry, Vol. 72, No. 7, April 1, 2000 1377

Figure 3. Schematic representation of the catalytic as well as
destructive effects of nucleophilic compounds (Nu) and bases (Base)
on the peroxyoxalate chemiluminescence reaction. An increase in
the rate of the light-producing reaction (vertical path) will increase
the light intensity unless the relative rate increase of the quantum
yield diminishing trails of reagent deactivation and decomposition
(horizontal paths) is greater.
Figure 2. Chemiluminescence light intensity as a function of
difference in pKa in acetonitrile between the 4-aminopyridine catalyst
and the leaving group from the reagent. The two insets show the
structures of the protonated leaving groups in each set of experiments.
not be true for the 4-aminopyridines because they, as evident from
the POCL emission profiles in Table 2, activate the reagent. The
ODI decomposition appeared to be related to both the nucleo-
philicity and the basicity of the bases since the non-nucleophilic
bases also induced fast reagent decays. Consequently, the 4-ami-
nopyridines must also induce some basicity-related ODI break-
down apart from the reagent activation. Because the base strength
increases in the series 4-aminopyridine, DMAP, PYP, an increase
in the decomposition rates within this series can be expected. The
low light yields obtained with TCPO and the 4-aminopyridines in
the POCL reaction and the decrease in yield with basicity among
these compounds indicate the presence of a corresponding base-
related decomposition reaction of the activated intermediates. With
this evidence, we suggest that the more activated the derivatives
of oxalic acid become, the more sensitive they are toward basic
environments, and the faster they decompose. Thus, fast nucleo-
philically catalyzed POCL reactions inevitably lead to reduced light
yields. The rate of the catalytic reaction of the 4-aminopyridines
with both TCPO and ODI generally increased with base strength
according to Tables 1 and 3. However, we suspect that the base-
catalyzed breakdown reaction dominates over the catalysis reac-
tion and becomes reaction-limiting with ODI as reagent, whereas
the reaction rate of the catalysts themselves are rate-limiting with
TCPO as reagent. Hence, the maximum in intensity at minimum
basicity difference between the catalyst and the leaving group is
caused by a change in the rate-limiting reaction in the multistep
POCL process. Figure 3 represents an attempt to summarize the
discussion above and provide a generalized diagram to schemati-
cally show how the nucleophilic and basic properties of the catalyst
affect the reagent and the intermediates involved in the POCL
reaction. Furthermore, because the relative increase in light
intensity and yield appears to be larger with 4-aminopyridine than
with the other acylation catalysts, a stabilization of the intermediate
in the catalysis reaction can be suspected. Since 4-aminopyridine
has the ability to form hydrogen bonds with other bases via the
amino group or expel a proton from this nitrogen, we believe that
complete or partial neutralization of the intermediate is advanta-
geous for its stability.
Table 3. The Influence of Different Bases on the
Decomposition Rates of Chemiluminescence Reagents
in Acetonitrile
base
K
TCPO, min-1
KODI, min-1
none
4-aminopyridine
DMAP
<0.001
0.18
0.17
0.012
0.25
0.34
0.95
22
830
25
6.4
PYP
0.48
morpholine
piperazine
triethylamine
PMP
9.7a
450a
0.33
0.007
a
With these bases, the two phenolic groups of TCPO appeared to
be released consecutively at two different rates and these decay
constants refer to the initial release rate, which should correspond to
the disappearance of TCPO. The second decay constants with mor-
pholine and piperazine were 0.89 and 6.5 min-1, respectively.
ridines the decay rates generally increased with base strength.
For most catalysts, the consumption rate of the reagent increased
by a factor of 2 when changing from TCPO to ODI, but with
4-aminopyridine the rate increase was only ∼40% and with
triethylamine and PMP the rate increase was as large as 1-2
orders of magnitude. The slow ODI decay induced by 4-aminopy-
ridine promises even higher light intensities in the POCL reaction
at higher concentrations of 4-aminopyridine, which is interesting
considering that the light was particularly intense with this
combination of catalyst and reagent.
The extremely slow TCPO decay that was induced by the
strong sterically hindered non-nucleophilic39 base PMP and the
slow decomposition of TCPO observed with triethylamine indicate
that this moderately reactive reagent is only marginally affected
by basic environments and that the main reason for disappearance
of TCPO is nucleophilic attack by the bases. A nucleophilic
reaction between the reagent and the base must produce either
an activated or a deactivated compound, where the first route
should result in high light intensities and the latter low light
intensities in the POCL reactions. It thus seems that the attainable
light intensities with morpholine and piperazine are limited by
dark nucleophilic reagent breakdown reactions, whereas this could
However, complementary POCL experiments with ODI and a
low concentrations of the catalyst (both 0.5 mM, data not shown)
1378 Analytical Chemistry, Vol. 72, No. 7, April 1, 2000

showed a dependence of light intensity on basicity opposite to
the trend in Table 2, i.e., a slight increase in intensity with base
strength of the 4-aminopyridine derivative. At first this result might
seem to contradict the conclusion above, but when the nucleo-
philic catalyst is not in excess over the reagent, most catalyst
molecules will be engaged in formation of the transient reagent
and only a few are available to degrade the reagent. Thus, this
observation actually supports the theory that a change in the
process limiting the amount of the transient reagent causes the
intensity optimum at ∆pKa ) 0 (cf. Figure 2).
In addition to these experiments, the stability of TCPO was
monitored in the presence of DBU but the decomposition was
almost instantaneous. This explains the very rapid light decay and
inferior light yields that we previously observed29 when DBU was
used as a catalyst for the POCL reaction. This clearly pinpoints
the nucleophilicity of DBU, a compound that has been claimed
to be a useful non-nucleophilic base.48 Others have also come to
the same conclusion.49
Table 4. Dual-Component Catalysis of TCPO
Peroxyoxalate Chemiluminescence with Different
Concentrations of 1,2,4-Triazole and
1,2,2,6,6-Pentamethylpiperidine
triazole, PMP, max intens, light yield, peak time, half-life,
mM
mM
mV
mV‚s
ms
ms
5
5
5
20
20
80
80
0
0.5
5
0.5
2
0.5
4
37a
1700
1500
2000
2100
1600
1900
n/ aa
4100
330
2500
550
1100a
170
17
91
29
6400a
1000
110
760
140
500
41
1500
150
45
7
a
This reaction experienced a pulsed emission profile and these
figures refer to the first one of these pulses. The second pulse
experienced its 16-mV maximum at 280 s and declined so slowly that
the whole emission profile could not be integrated; however, the light
yield up to 1000 s was 15 V‚s and at that time the intensity had
decreased by ∼20% from the maximum value of the second pulse.
Dual-Component Catalysis. As outlined in the introduction,
the catalysis by imidazole involves two different modes: general
base and nucleophilic catalysis.24,25 This was verified25 by the
observation that triethylamine catalyzed the reaction between
imidazole and TCPO. As we were curious if this could be used to
increase the light intensity of imidazole-catalyzed POCL reactions,
4-aminopyridine was tested as an ancillary strong base together
with imidazole using TCPO as reagent. However, the light
intensity remained essentially unchanged and the only effects were
a decreased peak time, a faster light decay rate, and reduced light
yield. Alternative nucleophilic catalysts were therefore tested in
mixtures with different strong bases. An astonishing effect on
TCPO catalysis was observed when the imidazole analogue 1,2,4-
triazole was used together with strong bases such as 4-aminopy-
ridine, DMAP, triethylamine, or DBU. The light intensity was quite
similar with the different cocatalysts, but the peak time, the half-
life, and the yield of the light all decreased with increasing basicity
of the ancillary base. The addition of small amounts of a strong
base caused a more than 10-fold increase in light intensity from
a reaction catalyzed by 1,2,4-triazole. As shown in a preliminary
report describing this effect,29 the light intensity from a reaction
catalyzed by 1,2,4-triazole or imidazole was the same when small
amounts of DBU were added to 1,2,4-triazole. However, the
maximum light intensity was reached ∼100 times faster and
displayed a peak time of a couple of tens of milliseconds with the
1,2,4-triazole/ DBU mixture, compared to several seconds with
imidazole.
Supported by the pKa dependency reported above, the inter-
pretation of the dramatic effect of strong bases on the catalytic
efficiency of 1,2,4-triazole must be a simple acid-base equilibrium
between the two catalysts. The cocatalyst subtracts a proton from
the weak monoacidic base 1,2,4-triazole (aqueous pKa 2.39 41) to
produce the 1,2,4-triazolate anion (aqueous pKa 9.97 41), which is
a strong enough nucleophile to accomplish an attack on TCPO
and effectuate a replacement of the phenolic group(s). However,
unlike the mechanism for imidazole catalysis,24,25 it is not probable
that the proton removal from 1,2,4-triazole and the attack on the
carbonyl carbon occurs in a concerted mechanism. The reaction
between 1,2,4-triazole and TCPO leads to the formation of an
intermediate,23 and it is most likely that this happens also in more
basic environments, only much faster. This intermediate can, due
to the strong analogy to imidazole-catalyzed POCL, be presumed
to be 1,1′-oxalylditriazole (ODT). The intermediate reagent ODT
then reacts further with hydrogen peroxide to actualize the
formation of the key intermediate responsible for the energy
transfer to the fluorophore.
If, as we anticipate, only the basicity of the cocatalyst is
important, it would be favorable to reduce its nucleophilicity to
minimize the effect of reagent breakdown as previously discussed.
Hence, the POCL reaction of TCPO was catalyzed by 1,2,4-triazole
in combination with the non-nucleophilic base PMP, which showed
the slowest rate of reagent breakdown; see Table 3. The results
with this combination of catalysts were very encouraging indeed,
as displayed in Table 4. The light intensities were about twice
those of the imidazole-catalyzed TCPO reaction, and the peak
times were in the range of milliseconds. The total light yield
decreased quite dramatically with the addition of PMP compared
to the reaction only catalyzed by 1,2,4-triazole, demonstrating the
sensitivity of the formed intermediates toward basic environments.
The ratio between the two catalysts and their concentration
relative to the TCPO reagent influenced the light emission profiles
dramatically. Light yield, peak time, and half-life decreased with
increasing concentrations of 1,2,4-triazole and PMP. The light
intensity was dependent on the concentrations of the two catalysts
and first increased but subsequently decreased with catalyst
concentrations. The optimum catalyst-to-reagent ratio was in the
range of 40:1-160:1 and the ratio between the catalysts (1,2,4-
triazole/ PMP) giving the highest light intensity was between 10:1
and 1:1. These optimums are probably due to the existence of
base-catalyzed breakdown (as outlined above) competing with the
light-producing reaction.
Sensitivity to Water. The majority of applications utilizing the
POCL reaction as a detection technique involve water as one of
the solvent components. Water, or rather the hydroxide ion, is
also a nucleophilic competitor to hydrogen peroxide for the
available reagent molecules. In addition, the different environment
in mixed aqueous/ organic solutions will most certainly influence
the characteristics of the species involved in the reaction. It is
(48) March, J. Advanced Organic Chemistry: Reactions, Mechanisms and Struc-
ture, 4th ed.; John Wiley & Sons: New York, 1992; p 1023.
(49) Chakrabarty, M.; Batabyal, A.; Patra, A. J. Chem. Res. 1 9 9 6 , 190-191.
Analytical Chemistry, Vol. 72, No. 7, April 1, 2000 1379

Table 5. Effect of Water on the Catalysis of the TCPO Peroxyoxalate Chemiluminescence Reaction
catalyst system
water, %
max intens, mV
light yield, mV‚s
peak time, ms
half-life, ms
PYP
PYP
PYP
0
5
20
0
5
20
1400
1400
450
1700
1700
960
900
540
93
4100
4100
330
1
1
3
170
70
21
43
87
35
1000
800
100
1,2,4-triazole/ PMP (10:1)
1,2,4-triazole/ PMP (10:1)
1,2,4-triazole/ PMP (10:1)
therefore highly relevant to investigate to what extent the
performance of catalysts subject to testing in this paper are
influenced by the presence of water. The PYP catalysis of TCPO
and the 1,2,4-triazole/ PMP (10:1) catalysis of TCPO were selected
as representatives for the single-component and dual-component
catalysis, respectively. It is evident from Table 5 that the overall
influence of water was greater on PYP catalysis than on 1,2,4-
triazole catalysis, but neither of the two catalyst systems were
appreciably affected by water concentrations up to 5%. However,
in contrast to the 1,2,4-triazole-catalyzed reaction, the light yield
with PYP as catalyst decreased in a solution with 5% water. At a
water content of 20%, both reactions were markedly negatively
influenced and the light intensities were reduced to about 30 and
60% of the original intensities for the PYP and 1,2,4-triazole-
catalyzed reactions, respectively. A strange, and at this point
unexplainable, divergence was observed for the peak time. It
increased with increasing water content with PYP as catalyst but
decreased with 1,2,4-triazole and PMP as catalysts. Another
notable feature of the 1,2,4-triazole/ PMP-catalyzed reactions was
the increased half-life but decreased light yield at 5% water
compared to with no added water. This apparent discrepancy is
simply an illustration of the different light decay kinetics for this
reaction. Initially, the decay rate was slower at 5% than at 0% water
but this changed to a faster decay rate within 50 ms, which caused
the intensity to decline below that of the water free reaction after
∼200 ms. The practical consequence is that small amounts of
water can be beneficial for the light yield of the 1,2,4-triazole/
PMP catalyst system on the short time scale within a small
detection cell.
can form unstable and reactive intermediates. By using more
potent nucleophilic catalysts such as the 4-aminopyridines, the
light intensity can be substantially increased but at the cost of
lower total quantum yield. The obvious drawbacks with the use
of effective and strongly basic nucleophiles are reagent and
intermediate breakdown, which restrict the attainable light intensi-
ties. This can be partially circumvented by choosing a catalyst
that matches the basicity of the reagent’s leaving group or by using
a mixture of two catalysts, in which each catalyst is aimed at a
specific catalytic task. One targeted for nucleophilic catalysis and
one for general base catalysis. Such “base-induced nucleophilic
catalysis” demands the use of a strong non-nucleophilic base and
a quite poor nucleophile which can be activated by the subtraction
of an acidic proton.
However, both these strategies emerge as promising ap-
proaches for miniature flow-through POCL detectors, as they
combine short peak times with high light intensities. Naturally,
this requires the reagents to be merged and efficiently mixed
immediately before entering the detector. Instrumental develop-
ment might thus be necessary before the full potential of these
reactions can be utilized.
ACKNOWLEDGMENT
This work was supported by grants from J. C. Kempes
Minnesfond and Magnus Bergwalls Stiftelse. DMAP was a most
generous gift from Reilly Industries, Indianapolis, IN. We are
indebted to Bjo¨ rn Sjo¨ blom for help on setting up the stopped-
flow equipment and to Malin Stigbrand and Einar Ponte´n for
valuable comments on the manuscript.
Received for review November 22, 1999. Accepted
January 28, 2000.
CONCLUSIONS
The POCL reaction is very sensitive to nucleophlic catalysis,
not only by imidazole but also by other good nucleophiles that
AC991339A
1380 Analytical Chemistry, Vol. 72, No. 7, April 1, 2000