Room temperature phosphorescence in the liquid state as a tool in

Analytica Chimica Acta 488 (2003) 135–171

Review

Room temperature phosphorescence in the liquidstate as a tool in analytical chemistry

Jacobus Kuijt, Freek Ariese, Udo A.Th. Brinkman, Cees Gooijer∗Department of Analytical Chemistry and Applied Spectroscopy, Vrije Universiteit Amsterdam,

de Boelelaan 1083, 1081 HV Amsterdam, The Netherlands

Received 28 April 2003; accepted 27 May 2003

Abstract

A wide-ranging overview of room temperature phosphorescence in the liquid state (RTPL1) is presented, with a focuson recent developments. RTPL techniques like micelle-stabilized (MS)-RTP, cyclodextrin-induced (CD)-RTP, and heavyatom-induced (HAI)-RTP are discussed. These techniques are mainly applied in the stand-alone format, but coupling withsome separation techniques appears to be feasible. Applications of direct, sensitized and quenched phosphorescence are alsodiscussed. As regards sensitized and quenched RTP, emphasis is on the coupling with liquid chromatography (LC) and capillaryelectrophoresis (CE), but stand-alone applications are also reported. Further, the application of RTPL in immunoassaysand in RTP optosensing—the optical sensing of analytes based on RTP—is reviewed. Next to the application of RTPLin quantitative analysis, its use for the structural probing of protein conformations and for time-resolved microscopy oflabelled biomolecules is discussed. Finally, an overview is presented of the various analytical techniques which are based on

Abbreviations:1-BrN, 1-bromonaphthalene; 2-BrN, 2-bromonaphthalene; 4-MSA, 4-maleimidylsalicylic acid; 6Br2N, 6-bromo-2-nap-hthol; �-CDI, 6-iodo-6-deoxy-�-cyclodextrin; �-CDI, 6-iodo-6-deoxy-�-cyclodextrin; A, adenine; acac, acetylacetone; AOT, di-2-ethyl-hexylsulfosuccinate sodium salt; AP, alkaline phosphatase; Br-�-CD, heptakis (6-bromo-6-deoxy-�-cyclodextrin); BSA, bovine serumalbumin; C, cytosine; CCD, charge-coupled device; CD, cyclodextrin; CD-RTP, cyclodextrin-induced room temperature phosphorescence;CE, capillary electrophoresis; ChO, cholesterol oxidase; CPB, cetylpyridinium bromide; CPK, coproporphyrin I ketone; CPKTEE,coproporphyrin I ketone tetraethyl ester; CTA+, cetyltrimethylammonium ion; CTAB, cetyltrimethylammonium bromide; CTAC,cetyltrimethylammonium chloride; cyclic AMP, cyclic 3′,5′-adenosine monophosphate; CZE, capillary zone electrophoresis; dansyl chloride,5-dimethylaminonaphthalene sulfonyl chloride; DMF, dimethylformamide; DNA, deoxyribose nucleic acid; DOM,n-dodecyl-�-d-maltoside;EA, electron affinity; EDTA, ethylenediaminetetraacetic acid; ELISA, enzyme-linked immunosorbent assay; ET, energy transfer; ferron,8-hydroxy-7-iodo-5-quinolinesulfonic acid; FIA, flow injection analysis; G, guanine; HAI-RTP, heavy atom-induced room temperaturephosphorescence; HAP, heavy atom perturber; HRP, horseradish peroxidase; IC, inner conversion; IP, ionization potential; ISC,intersystem crossing; LC, liquid chromatography; LED, light-emitting diode; LOD, limit of detection; MEKC, micellar electrokineticchromatography; MIP, molecularly imprinted polymer; MS-RTP, micelle-stabilized room temperature phosphorescence; NS, naphthalenesulfonate; OEPK, octaethylporphine ketone; PAH, polycyclic aromatic hydrocarbon; Ph4TBP, meso-tetraphenyltetrabenzoporphyrin; PMP,pinacolyl methylphosphonate; RNA, ribose nucleic acid; R.S.D., relative standard deviation; RTPL, room temperature phosphorescence in theliquid state; SDBS, sodium dodecylbenzene sulfonate; SDS, sodium dodecyl sulfate; S/N, signal-to-noise ratio; SS-RTP, solid surface roomtemperature phosphorescence; T, thymine; TBP, tetrabenzoporphyrin; UV, ultraviolet; VASS, variable-angle synchronous scanning

∗ Corresponding author. Tel.:+31-204447540; fax:+31-204447543.E-mail address:[email protected] (C. Gooijer).

1 Although the abbreviation RTPL is frequently used in the literature, the specific RTPL techniques are generally designated by RTP, forinstance, CD-RTP. Therefore, in the introductory sections the term RTPL will be used, while in the sections on the specific techniques the termRTP will be applied.

0003-2670/03/$ – see front matter © 2003 Elsevier B.V. All rights reserved.doi:10.1016/S0003-2670(03)00675-5

136 J. Kuijt et al. / Analytica Chimica Acta 488 (2003) 135–171

the closely related phenomenon of long-lived lanthanide luminescence. The paper closes with a short evaluation of thestate-of-the-art in RTP and a discussion on future perspectives.© 2003 Elsevier B.V. All rights reserved.

Keywords:Room temperature phosphorescence; Luminescence; Lanthanides; Liquid state; Analytical applications

1. Basic aspects of RTPL

1.1. Introduction

In the past decades, room temperature phospho-rescence in the liquid state (RTPL) has evolved intoa sensitive and versatile tool in analytical chem-istry, based on many different detection principlesand solvent systems suitable for the generation ofRTPL [1]. In this review, developments from the pastdecade will be highlighted. Topics of interest includemicelle-stabilized room temperature phosphorescence(MS-RTP), cyclodextrin-induced RTP (CD-RTP) andthe recently developed heavy atom-induced RTP(HAI-RTP) technique. In addition, attention will bedevoted to the interesting subject of RTP optosensing,and to the versatile coupling of indirect RTP modes(sensitized and quenched phosphorescence) to variousseparation techniques. Tryptophan RTP will be shownto be a valuable and sensitive tool for the structuralanalysis of proteins. Finally, the subject of long-livedlanthanide luminescence will be discussed. This tech-nique is applied in optosensing and immunoassays,for the detection of biomolecules, and as a detectionmethod in liquid chromatography (LC) and capillaryelectrophoresis (CE). Low-temperature phosphores-cence (Shpol’skii phosphorescence spectroscopy) andsolid surface (SS)-RTP—neither of which are RTPLtechniques—will not be discussed. For a review onShpol’skii phosphorescence spectroscopy, the readeris referred to[1], while SS-RTP is dealt with in[2].

In this review, emphasis will be on fundamental as-pects of RTPL that are important for an assessment ofits present and future potential, and on novel applica-tions and detection principles. The sensitivity of theelectronically excited triplet state to external factorsgenerally requires efficient protection from dissolvedmolecular oxygen as well as blocking of non-radiativedecay. Hence, much research is still devoted to findingways to induce or enhance RTP. As will be seen in thisreview, the range of conditions compatible with RTPL

keeps expanding, thereby broadening the applicationrange of RTPL.

It should be noted that the sensitivity of the (ex-cited) triplet state to external factors also creates pos-sibilities for detection, especially in the field of RTPoptosensing, quenched phosphorescence detection andthe structural probing of proteins. Consequently, theapplication range of RTPL is larger than the group ofcompounds that displays native phosphorescence: anycompound or solvent parameter that is able to influ-ence the phosphorescence signal, either positively ornegatively, can be detected.

1.2. Theory

Phosphorescence of organic molecules can be de-fined as the radiative transition originating from thelowest excited triplet state, T1, to the (singlet) groundstate, S0. In contrast to fluorescence (singlet-to-singlettransition), phosphorescence is a spin-forbidden pro-cess. Nevertheless, it can be observed under specificconditions, due to internal or external spin–orbit cou-pling which mixes pure singlet and triplet states toproduce states with a mixed character in spin multi-plicity [3,4].

The photophysical processes relevant to directlyexcited RTP are presented inFig. 1. In order to ob-tain direct phosphorescence, the phosphorophore isexcited by light of the appropriate wavelength. Afterexcitation, rapid vibrational deactivation and internalconversion (IC)2 occur. Fluorescence will thereforegenerally be emitted from the lowest vibrational levelof the lowest electronically excited singlet state, S1.Many organic molecules also show IC/vibrationalrelaxation to the S0 ground state. In addition, a signif-icant fraction of the excited molecules will undergointersystem crossing (ISC) to T1. The intersystemcrossing quantum efficiency,ϑISC, can be enhancedby internal or external spin–orbit coupling—the

2 For simplicity, this process is not depicted inFig. 1; it is mostimportant for excited singlet states above the S1 state.

J. Kuijt et al. / Analytica Chimica Acta 488 (2003) 135–171 137

Fig. 1. Schematic representation of photophysical processes rele-vant to directly excited RTP. S0, singlet ground state; S1, lowestexcited singlet state; T1, lowest excited triplet state; exc, excita-tion; flu, fluorescence; v, vibrational relaxation; ISC, intersystemcrossing; q, bimolecular quenching; phos, phosphorescence.

so-called heavy atom effect. However, population ofthe triplet state is not the only requirement for obtain-ing RTP: its subsequent deactivation by phosphores-cence should also be efficient, i.e. it should be able tocompete with the other deactivation pathways open tothe triplet state—deactivation by non-radiative decayand bimolecular quenching from dissolved molec-ular oxygen or other quenchers. In this context, itis important to note that strong spin−orbit couplingusually also enhances the rate constant of phosphores-cence,kp, resulting in reduced competition from theother deactivation pathways. Alternatively, the rela-tive importance of RTP can be increased by reducingthe rates of the non-radiative deactivation pathways.This can be achieved by enhancing the rigidity of theenvironment—for example, by inclusion of the phos-phorophore into micelles or CDs—and by efficientdeoxygenation of the sample. The phosphorescencequantum yield,φp, or ϑISC × ϑp, is given by

φp = ϑISC × ϑp = kISC

kISC + kf + knf + ∑kq,f [Q]

× kp

kp + knp + ∑kq,p[Q]

(1)

wherekISC is the intersystem crossing rate constant,kf and kp are the rate constants of fluorescence andphosphorescence, respectively,knf and knp the rateconstants of non-radiative decay, and

∑kq,f [Q] and∑

kq,p[Q] the sums of all effective (unimolecular)quenching rate constants of fluorescence and phos-phorescence, respectively. Due to the slow kineticsof phosphorescence compared to fluorescence, bi-

molecular quenching of the triplet state by molecularoxygen (kq ca. 109 M−1 s−1) is highly efficient, incontrast to collisional quenching of the excited singletstate. Therefore, dissolved molecular oxygen shouldbe removed efficiently in order to achieve a largevalue for ϑp and, thus, forφp. Model calculationsbased onEq. (1) readily show that direct RTPL willbe shown only by a limited number of compounds.For a molecule displaying phosphorescence,kp istypically 1 s−1 or less. Therefore, for such a moleculea phosphorescence quantum yield of ca. 0.5 can beachieved only if the oxygen concentration is below10−9 M—even if it is assumed thatϑISC equals unityand that the other triplet state deactivation pathwayscan be neglected. However, such a low oxygen con-centration is not easily maintained in practice. Forexceptional phosphorophores,kp may be as large as102 s−1. Using the same assumptions as above, thismeans thatφp will be about 0.5 even when the oxygenconcentration is 10−7 M—a concentration level thatis easily achieved when using standard equipment.

As noted above, there are two important modes ofindirect phosphorescence that can be used for detec-tion purposes, sensitized RTP and quenched RTP. Anexample of sensitized RTP is given inFig. 2. This

Fig. 2. Spectra of 5× 10−6 M 2,6-naphthalenedisulfonic acid indemineralized and distilled water: (a) fluorescence (1) and sen-sitized phosphorescence (2) excitation spectrum, with emissionwavelengths of 346 and 513 nm, respectively; (b) fluorescenceemission spectrum; (c) sensitized phosphorescence emission spec-trum (10−4 M biacetyl). The emission spectra were obtained withexcitation at 230 nm. Figure taken from[5].


figure clearly illustrates the large wavelength differ-ence, generally observed in sensitized RTP, betweenthe excitation spectrum (a) and the emission spec-trum (c). Obviously, the fluorescence emission band(b) is at a much shorter wavelength, which resultsin a much smaller wavelength difference for fluores-cence. Sensitized RTP can be obtained if a sensitizer(analyte) molecule is excited that is able to providetriplet–triplet energy transfer (ET) to the phospho-rophore. Obviously, for efficient ET to take place therate constant,kET, should be large. This implies thatthe triplet level of the donor molecule should be higherthan the triplet state of the acceptor[3,4,6]. In addi-tion, there should be significant spectral overlap be-tween the phosphorescence emission spectrum of thedonor and the S0–T1 absorption spectrum of the accep-tor. Even if these conditions are fulfilled, efficient EToccurs only in case of sufficiently long donor tripletlifetimes and high acceptor concentrations—as can beconcluded fromEq. (2)

ϑET = kET[A]

(τD0 )−1 + kET[A]

(2)

whereτD0 is the triplet lifetime of the donor and [A]

the acceptor concentration. Assuming thatkET =1010 M−1 s−1, [A] = 10−2 M, and tD0 = 0.1�s—such values are frequently encountered in sensitizedRTP experiments—ϑET is already close to its maxi-mum value of unity[5]. The intensity of sensitizedphosphorescence is given by

Ip(sens) = 2.303I0(λex)εD[D]�ϑD

ISCϑETϑAp (3)

where I0(λex) is the intensity of the excitation light,while the next three symbols represent the extinctioncoefficient of the donor, its concentration, and the op-tical path length, respectively, and the final three theintersystem crossing efficiency of the donor, the ef-ficiency of energy transfer, and the phosphorescenceefficiency of the acceptor, respectively[5]. In manycases biacetyl has been used as the phosphorophore:strong enhancement compared to direct RTP can thusbe achieved since biacetyl has a very low absorptiv-ity over its entire spectrum[7], whereas in favorablecases the donor (analyte) may have a high extinctioncoefficient.

Bimolecular quenching is not only a problem thathas to be solved in RTP-based methods. It can alsobe used for detection; in this case, the directly excited

or sensitized RTP from exceptionally strong phos-phorophores like biacetyl or bromonaphthalenes isdecreased due to the collisional deactivation providedby the analytes (quenchers). Quenching is based ei-ther on triplet−triplet energy transfer (requirementsare identical to those for sensitized RTP) or on elec-tron transfer. Triplet−triplet energy transfer is gen-erally considered to proceed via the exchange (orcollisional) mechanism proposed by Dexter[3,4], amechanism that essentially requires molecular contact(typical critical interaction radius,R0, ca. 10 Å)[4].In most cases, the long-range (coulombic) mechanismaccording to Förster should be ruled out, because itis spin-forbidden (only interactions in which bothspins are preserved are allowed, e.g. singlet−singletenergy transfer). However, as will be discussed, un-der certain conditions the Förster mechanism canalso be relevant for quenching interactions involv-ing a change in spin, even though this is forbidden[3,4].

Electron transfer reactions are based on the oxi-dizing and reducing properties of the quencher andthe phosphorophore. In a simplified approach3 forground-state molecules in the gas phase, the abilityto release an electron depends on the (first) ioniza-tion potential (IP), while the ability to accept anelectron is given by the electron affinity (EA). Forefficient electron transfer to occur, the energy re-leased should be larger than the energy required forionization of the donor (IP), i.e. the reaction shouldbe exothermic. Donor molecules like amines can beoxidized quite easily, since they have non-bondingnitrogen orbitals at a relatively high energy level.In addition, photo-excited molecules are usuallymuch more susceptible to redox reactions than theirground-state counterparts. This implies that exciteddonor molecules are oxidized more easily (IP∗ < IP),while the propensity for excited acceptor moleculesto accept electrons is also enhanced (EA∗ > EA) [8].

Regardless of the specific mechanism involved,the quenching interaction will be governed by thewell-known Stern−Volmer equation for bimolecular

3 For electron transfer reactions in solution, redox potentials haveto be used, while also solvation energy terms have to be takeninto account. Nonetheless, the variations in the redox potentialsoften parallel those in the values of IP and EA, i.e. a plot of redoxpotential versus IP or EA will generally be linear.


quenching of luminescence4

I0

I= 1 + kq[Q]τ0 (4)

wherekq is the rate constant for bimolecular quench-ing, [Q] the quencher (analyte) concentration, andτ0 the triplet lifetime in the absence of quenchers.According to Eq. (4), even low concentrations ofquenchers will exhibit efficient quenching if the tripletlifetime, τ0, and the associated quenching rate con-stant,kq, are sufficiently high; for collisional quench-ing (triplet−triplet energy transfer, electron transfer),kq is at its maximum5 if the quenching interaction isdiffusion-controlled, i.e. if there is no energy barrierinvolved. In such a case,kq equalskdiff , which isapproximated by the Debye equation[3,4]

kdiff ≈ 8RT

3000η(5)

whereR is the gas constant,T the temperature, andη the viscosity of the solvent. In aqueous solutions atroom temperature,kdiff is about 1010 M−1 s−1, whichenables sensitive detection of many compounds byquenched RTP. This is readily illustrated by a modelcalculation based onEq. (4). If the lifetime of thephosphorophore in the absence of analytes (quenchers)is 10−4 s, an analyte concentration as low as 10−7 Mwill cause a 10% reduction of the phosphorescenceintensity.

2. Micelle-stabilized RTP

2.1. Introduction

As early as 1980, Cline Love et al.[10] exploredthe use of sodium dodecyl sulfate (SDS) micelles forthe inclusion of phosphorophores (the test analytes,

4 Strictly speaking, this equation is valid only under steady-stateconditions, i.e. when continuous illumination is applied. Fortime-resolved detection, an adapted equation has been proposed[9]. However, if the delay time applied before detection is shortcompared to the gating time (which normally is the case), theequation is also correct to a fair approximation for gated detectionsystems.

5 The quenching rate constants measured for fluorescencequenching by the Förster mechanism often exceed this maximum(kq, ca. 1011 M−1 s−1), due to the large critical interaction radiigenerally applying in this case (R0, ca. 50 Å)[3,4].

naphthalene, pyrene and biphenyl), in order to reducethe rate of non-radiative decay and, thus, extend theapplication range of direct RTP to analytes withkp ≤1 s−1. The heavy atom perturbers (HAP), Ag(I) andTl(I), were added to the micellar solutions to increasethe ISC efficiency,ϑISC, and—more importantly—thephosphorescence efficiency,ϑp, thereby enhancing theRTP quantum yield and intensity. In all instances, noRTP was obtained in the absence of HAP ions. Dueto the ability of the metal ions to replace sodium ionsfrom the Stern layer of the micelles, a close proxim-ity of HAP and phosphorophore was obtained, whichis important for a strong enhancement of ISC. Forcompleteness it is noted here that in some unfavorablecasesknp will be enhanced more strongly thankp; ob-viously, φp will not be enhanced in such cases by theaddition of HAP ions.

It should also be stressed that micelles do not pre-vent quenching by molecular oxygen, apart from thereduction in kq due to the increased viscosity (cf.Eq. (5)). Therefore, high RTP quantum yields andintensities require efficient deoxygenation of the sam-ple. The use of nitrogen purging for deoxygenationin micellar systems is complicated by the continuousformation of foam during purging. Therefore, alter-native deoxygenation methods have been developed.In 1986, the use of sodium sulfite as an oxygen scav-enger in micellar solutions (of naphthalene) was pro-posed[11]. The low oxygen concentrations requiredto obtain MS-RTP were achieved by the consumptionof oxygen according to

2SO32− + O2 → 2SO4

2− (6)

Sodium sulfite is not only applied for deoxygenationpurposes in virtually all MS-RTP studies, but also intechniques such as CD-RTP and HAI-RTP (Sections3 and 4, respectively).

Other deoxygenation methods have also been re-ported for RTP in the liquid state: Zn(s)/HCl andNa2CO3/HCl systems were used to remove oxygenfrom cyclodextrin (CD)-containing solutions, basedon the purging action of the evolved gases (H2 andCO2, respectively)[12]. Deoxygenation was reportedto be rapid, but required strongly acidic conditions(pH < 1). Moreover, the RTP signal was stable onlyover a limited period of time (at least 10 min), sincethe purging action ceases when the reaction stops.


Table 1Determination of naproxen and propanolol in pharmaceutical preparations by MS-RTP and LC

Pharmaceutical preparation Declared content (mg) Found (mg)

MS-RTP LC

NaproxenAntalgin (tablets, Syntex Latino) 275 290 277.2Naproval 250 (capsules, Valles Mestre/Farma 86) 250 217 225Proxen (capsules, Berenguer-Infale) 250 263 251.4Proxen (vials, Berenguer-Infale) 817 909 767

PropanololBetadipresan (tablets, Fides) 100 80 93Betadipresan-DIU (tablets, Fides) 100 73 100.3Sumial 10 (tablets, ICI-Farma) 10 10.3 10.4Sumial Retard (capsules, ICI-Farma) 160 155 166

Adapted from[14].

Therefore, these methods will probably not findwidespread application in RTP-based techniques.

2.2. Stand-alone measurements

There are many examples of MS-RTP detectionbased on measurement with a conventional lumines-cence spectrometer. Due to the inherent selectivity ofMS-RTP, detection of target analytes in pharmaceu-tical and agricultural samples is frequently possiblewithout (extensive) sample preparation. SDS has beenused most often (concentration, 10–100 mM). In mostcases, ca. 20 mM thallium nitrate was used as HAP and10–20 mM sodium sulfite for deoxygenation. Someexamples are given below.

Nafronyl was determined in Praxilene tablets with anominal content of 100 mg; analysis yielded 104.2 mgwith an R.S.D. of 2.4%[13]. Naproxen and propanolol[14] were detected in pharmaceutical products withsatisfactory accuracy (Table 1), although the presenceof hydralazine in Betadipresan resulted in quenchingof phosphorescence and, thus, rather low values for thepropanolol concentration. For comparison, the sam-ples were also analyzed using LC with UV absorp-tion detection. In general, the LC data were somewhatcloser to the declared concentrations. Naproxen wasalso determined in human serum and urine (recovery,97–101%), spiked with the analyte at concentrations(20–100 mg/l) typically found within 12 h after appli-cation of the drug[15].

The pesticide napropamide was determined incommercial pesticide formulations and in spiked soil,

tomato and pepper samples[16]. The determinationof 1-naphthaleneacetic acid in spiked canned pineap-ple was also reported[17]. In both cases, satisfactoryrecoveries were achieved.

Some metal ions, in particular Al(III), Ga(III) andIn(III), can be detected by RTP after complexationwith 8-hydroxy-7-iodo-5-quinolinesulfonic acid (fer-ron) due to the ensuing enhancement ofϑISC andϑp

[18].6 Detection of the metal−ferron (1:3) complexeswas performed using several surfactants. For the de-tection of Ga(III), neutral Brij-35 and the cationicsurfactants cetylpyridinium bromide (CPB) andcetyltrimethylammonium bromide (CTAB) providedthe best detectability. The latter surfactant provided alimit of detection (LOD) of 5 ng ml−1, an R.S.D. of 4%(at 50 ng ml−1) and linearity over two orders of mag-nitude. Detection is selective since only a few metalsgive stable complexes. Nevertheless, metal ions likeCu(II), Ni(II), Co(II) and Fe(II) can interfere becausethey are able to quench RTP. Fortunately, these ionscan be masked with 9,10-phenanthroline or EDTA,while the signal from Al−ferron can be suppressed byadding fluoride ions. Alternatively, mixtures of Ga(III)and Al(III) can be analyzed by using Kalman filter-ing [18]. The remaining problem—the presence ofIn(III) in Ga(III)-containing samples—was addressedin a subsequent study by using time-resolved detec-tion to discriminate between the two RTP lifetimecomponents (126 and 158�s for In(III) and Ga(III),

6 Although this method is an example of MS-RTP, it can alsobe classified as an RTP optosensing technique.


respectively)[19]. Thus, binary mixtures could gen-erally be analyzed with 10% accuracy or better.

Detection of Pt(II) after complexation with ferronor ferron-like compounds such as 8-quinolinol and5-sulfo-8-quinolinol, was performed in the presenceof several cationic (CTAC, CTAB) or neutral (Brij-35,Triton X-100, DOM) surfactants for micelle formation[20]. The best result (LOD, 6 ng ml−1) was obtainedwith n-dodecyl-�-d-maltoside (DOM). In anotherstudy, Pd(II) ions were detected using the RTP of itscoproporphyrin III complex after cloud-point precip-itation [21]. After heating of the sample to 95◦C, theTriton X-100 phase containing the analyte separatedfrom the aqueous phase, which resulted in a 10-foldenhancement of the detectability (LOD, 2 nM).

2.3. Stopped-flow mixing

A convenient and easily automatable procedure toobtain MS-RTP was reported by Panadero et al.[22]:by using a stopped-flow mixing technique, they ob-tained RTP from the pesticide carbaryl within 2–3 safter mixing of the solutions (Fig. 3). In the set-upused (Fig. 4), a syringe solution containing thalliumnitrate, sodium sulfite and buffer, and a syringe so-

Fig. 4. (a) Top and (b) side view of stopped-flow device. Figure taken from[23].

Fig. 3. Kinetic curves obtained for stopped-flow mixing experi-ments using carbaryl as the test analyte. Analyte concentrations(mg/l): (1) 0.1; (2) 0.5; (3) 1.0; (4) 2.0. Figure taken from[22].

lution containing SDS, sodium sulfite, buffer andcarbaryl were mixed. Although the RTP signals areobtained much faster than in conventional MS-RTPbatch measurements, it should be noted that both solu-tions contain sodium sulfite. That is, the time neededto obtain RTP is not the deoxygenation time but,rather, the time the thallium ions need to replace thesodium ions from the micellar Stern layer[22]. Both


the slope (before the curve starts levelling off) andthe amplitude (signal after levelling off) were usedfor analysis. With the same set-up, RTP from dipyri-damole[23] and naproxen[24] was obtained within1–10 s. In all cases, LODs were in the 10–20 ng ml−1

range [22–24]. Naproxen was successfully deter-mined (recovery, 91−106%) in serum spiked withinthe therapeutical doses (25–70 mg/l)[24], while theanalysis of Asasantin 75 and Persantin 100 tabletsrequired no sample pretreatment—apart from sim-ple dissolution of the tablets and dilution in 0.1 MSDS[23].

A different syringe filling scheme than the oneconsidered above was reported in[25]: one syringesolution contained SDS and the analyte (nafronyl),and the other one, a buffer with sodium sulfite andthallium nitrate. Since the deoxygenant and HAP areseparated from the analyte and SDS before mixing,the time to obtain phosphorescence (5 s for full sig-nal) includes not only the time for the thallium ionsto replace the sodium ions from the micelles, but alsothe time required for adequate deoxygenation. Thatis, the stopped-flow mixing technique is by far thefastest way to obtain RTP to date.

2.4. Sensitized RTP

As noted before, bimolecular interactions are notseriously hindered by the inclusion of the analytesin micelles. Therefore, sensitized RTP—a techniquerequiring frequent and efficient collision of donor andacceptor molecules—can also be applied in micel-lar systems. The acridine dyes trypaflavine, acridineyellow and acridine orange were used to generatesensitized phosphorescence from several polycyclicaromatic hydrocarbons (PAHs) included in SDS mi-celles [26], with sodium sulfite as deoxygenant andTl(I) as HAP. Selective detection of some targetPAHs in the presence of other PAHs was possibledue to the differences in the T1 energy levels. For in-stance, pyrene (T1, 16,900 cm−1) could be selectivelydetermined—with selectivity factors ranging from350 to 5000—in the presence of fluorene, naphthaleneand phenanthrene (T1, 21,000–22,000 cm−1) uponsensitization by trypaflavine (T1, 17,800 cm−1). Forthese PAHs, the selectivity was improved about 2-, 7-and 35-fold, respectively, compared to fluorescenceand direct MS-RTP.

Sensitized RTP from biacetyl included in di-2-ethylhexylsulfosuccinate sodium salt (AOT) reversed mi-celles was also studied[27]. In these micelles, thehydrophobic tails are directed towards the apolar (orlow-polarity) bulk solution, while the ionic groupssurround the water pool present in the micellar cen-ter. Both the sensitizer,�-naphthylacetic acid, andbiacetyl are present in the micellar water pools. Theenergy transfer efficiency,ϑET (cf. Eq. (2)) was foundto decrease with an increasing relative size of thewater pool,W = [H2O]/[AOT], at higher values ofW, due to the increased average distance betweendonor and acceptor molecules; at low values ofW,the decrease inϑET was absent, most probably be-cause of the concomitant decrease in microviscosity.Compared to (normal) SDS micelles, a 13-fold higherRTP signal was obtained. Another advantage of us-ing AOT reversed micelles is that no foam is formedduring purging with nitrogen[27]. Sensitized RTPfrom biacetyl included in AOT reversed micelles wasalso obtained using either sodium naphthoate as thesensitizer or cetyltrimethylammonium ions (CTA+)as co-surfactant and ‘anchored’ naphthoate counte-rions as the sensitizer[28]. Comparable results wereobtained with both sensitizers.

2.5. Novel micellar systems

In the last decade, studies on micro-emulsion-RTPand block-copolymer-based micelles were published.The use of micellar agents other than SDS should beexplored since a higher sensitivity and analyte sol-ubility may thus be achieved. In addition, foamingproblems associated with nitrogen purging may bereduced.

Water-soluble copolymers of 1-vinylnaphthaleneand methacrylic acid were used for the first time in1991 to generate RTP (from pyrene and benzophe-none)[29]. With this type of micelles, foam formationduring nitrogen purging was reported to be much lessproblematic. Poly(ethylene oxide)−poly(propyleneoxide)−poly(ethylene oxide) block-copolymer mi-celles and mixed-block-copolymer/SDS micelles werealso used[30]. In the block-copolymer micelles, RTPwas slightly enhanced compared to free-solution RTP.With the mixed aggregates, RTP was significantly en-hanced even below the critical micellar concentrationsof both species.


Lin et al. [31] studied the RTP obtained from2-bromonaphthalene (2-BrN) in SDS rod-like mi-celles. At high electrolyte concentrations, a changefrom spherical to rod-shaped micelles was reported totake place. The generation of RTP was considered toresult from a dramatic increase of the microviscosityinside the micelles during the sphere-to-rod transition.It was reported that no RTP was obtained from 2-BrNin spherical micelles, but a convincing explanationfor this observation was not presented.

Corroborating earlier work by Ramos et al.[32],various analytes such as PAHs[33–36], carbaryl[37]and the plant growth regulators 2-naphthoxyaceticacid and 1-naphthaleneacetamide[38] were deter-mined by micro-emulsion RTP. In all instances, thal-lium nitrate was used as HAP. The micro-emulsionsconsist of SDS micelles (size, ca. 10–20 nm) in wa-ter, with higher linear alcohols (typically 0.01–0.02%1-butanol or 1-pentanol) as co-surfactants[32–38]and linear alkanes (0.01–0.02%n-heptane)[32–36]or 0.02% dichloromethane[37,38] incorporated in themicellar core. Because the particles are quite small,no scattering of visible light occurs and the solutionsare clear to the eye. It was stated that an advantage ofusing micro-emulsions instead of normal micelles isthe faster and better solubilization of non-polar com-pounds. Furthermore, an increased protection againstoxygen quenching was reported. However, deoxy-genation still appeared to be necessary in all cases andthe phosphorescence lifetimes of, typically, 1–3 mswere not significantly longer than those reported forconventional micellar systems[32,35].

Carbaryl was successfully determined in spikedsoil samples[37]; the LOD in standard solutions was24 ng ml−1. Naphthalene and phenanthrene were de-tected with LODs of 17 and 12 ng ml−1, respectively[36]. For acenapthene, an LOD of 5 ng ml−1 was re-ported and the analyte was successfully determinedin spiked (75 ng ml−1) seawater[33]. Variable-anglesynchronous scanning (VASS) phosphorometry wasapplied in combination with time-resolved detectionin order to detect PAHs in mixtures containing threeto five PAHs[34,35]; the PAHs were successfully de-termined in spiked coffee[34] and spiked road-dustsamples[35]. Derivative VASS micro-emulsion phos-phorometry was applied to the determination of2-naphthoxyacetic acid and 1-naphthaleneacetamidein spiked soil[38]. In most of the above cases, the

spiking levels were not indicated, but probably theywere quite high (mg/l level). Therefore, far-reachingconclusions about the selectivity achieved in real-lifeanalysis should not be drawn from these data.

2.6. Concluding remarks

Most recent MS-RTP studies either merely confirmthe findings from earlier reports or examplify the ap-plication of MS-RTP in real-life analysis. However,there are also some new developments: stopped-flowmixing was shown to be a convenient and rapid way toobtain RTP, which should find a more widespread use.Sensitized RTP in micelles provides improved selec-tivity compared to direct MS-RTP and fluorescence,but the enhancement is too small to enable directanalysis of most samples encountered in practice. Inaddition, novel MS-RTP techniques based on rod-likemicelles, micro-emulsions, reversed micelles, and mi-celles from co-block polymers were developed. An ad-vantage of the latter two micellar systems is that thereis no foaming during nitrogen purging. However, itshould be noted here that the use of sodium sulfite fordeoxygenation completely eliminates foaming prob-lems. The reported enhancements of detectability forthe novel micellar systems are rather limited, exceptfor the AOT reversed micelles used with sensitizedRTP. Such micelles may also be useful for couplingwith LC, since they are compatible with high con-centrations of organic modifier. Despite the progressthat has been made, the overall impression is thatthe advantages offered by most of the new MS-RTPtechniques discussed above are rather limited.

3. Cyclodextrin-induced RTP

3.1. Introduction

Another approach to obtain direct RTP uses cyclo-dextrins (CD) for inclusion of the phosphorophores.It was found in the 1980s that inclusion of a phospho-rophore in the hydrophobic cavity of a CD inducesRTP under deoxygenated solvent conditions, or at leastenhances RTP compared to homogeneous solutions.By using the differences in phosphorescence lifetimeand intensity between free and complexed phospho-rophores, Turro et al.[39] studied the kinetics of


inclusion, using bromo- and chloronaphthalenes as thephosphorescent probes. Addition of acetonitrile resul-ted in the formation of ternary complexes, and in anenhanced RTP intensity and lifetime for 1-bromonaph-thalene and 1-chloronaphthalene. It was also foundthat quenching by the aqueous-phase quencher nitritewas reduced five-fold upon addition of�-CD. An-other early study[40] was dedicated to the RTP fromn-(4-bromo-1-naphthoyl)alkyl trimethylammoniumbromide surfactants included in�- and�-CDs. Here,the surfactant part of the phosphorophore is partlyincluded in the cavity with the positively chargedend coiling at the mouth of the CD, thus providingadditional protection against molecular oxygen andCo(NH3)63+—a purely aqueous-phase quencher.

The inclusion of guest molecules in the cavity ofhost CDs and the requirements for obtaining directRTP have been studied extensively[41]. It is wellknown that a neat fit of the guest molecule and hostcavity is important for obtaining strong (van derWaals) binding of apolar guest molecules. For polarcompounds, electrostatic interactions such as hydro-gen bonding are important as well[41]. RTP fromincluded phosphorophores is obtained—or enhancedcompared to homogeneous solutions—due to the ef-fective shielding from dissolved molecular oxygenand the more rigid environment (high microviscosity)provided by the cavity of the CD. The phosphores-cence can also be enhanced by co-inclusion of aHAP; due to its close proximity to the analyte af-ter binding, both the ISC and the phosphorescencerate constants will be increased. Co-inclusion of asensitizer molecule instead of HAP will provide sen-sitized RTP; this technique will also be discussedhere.

3.2. Ternary complexes of halogenatedphosphorophores

In the early studies on CD-RTP[39,40], phos-phorescence was obtained from binary complexes ofhalogen-containing phosphorophores and CD mole-cules. As discussed above, the addition of acetonitrileresulted in stronger RTP due to the formation ofternary complexes. Many other compounds able toform ternary complexes with CDs and (halogen-con-taining) phosphorophores have been used since then,including various linear, branched and cyclic alcohols

[42–44] and cationic, anionic or neutral surfactants[44–46]. Strong RTP was obtained even under aeratedconditions.

Among the alcohols studied, the branched andcyclic alcohols (tert-butanol and cyclohexanol, respec-tively) provided the strongest RTP, which suggeststhat steric factors dominate the protection againstquenching by dissolved oxygen[43]. Probably, linearalcohols have the apolar tail in the cavity and the hy-droxy group outside, while branched alcohols favorbinding at the outside of the CD. In the latter case, thephosphorophores are effectively protected from dis-solved oxygen due to capping of the CD mouth[47].In general, the RTP signal from the binary complexeswas strongly enhanced with increasing concentrationsof alcohol, except for the less soluble alcohols, whichdisplace the analyte from the CD cavities at highalcohol concentrations[43].

The enhancement was used to develop an RTP-basedoptosensing scheme for the detection of alcohols inthe low mM range[42]. Here, the RTP of a binaryglucosyl-�-CD:1-bromonaphthalene (1-BrN) com-plex dramatically increased (up to 104-fold) in thepresence of various branched and non-branched al-cohols. The highest sensitivity again was obtainedfor tert-butanol and cyclohexanol; the primary andsecondary alcohols showed poorer detectability.

Among the surfactants used to enhance the RTPof 1-BrN:�-CD complexes, CTAB, sodium dodecyl-benzene sulfonate (SDBS) and polyethylenetert-octylphenyl ether were observed to provide much higherRTP signals than Tween-20, SDS and CPB[46]. Thiscan be partly explained by the bulky head groups ofthe former surfactants: their long hydrophobic chainis partly included in the cavity, while the part with thepolar head group is coiling at the mouth of the cavity[46]. A bulky group in this position provides betterprotection against molecular oxygen.

Escandar and Muñoz de la Peña[44] optimized theconcentrations of�-CD and surfactants (SDS, TritonX-100) or alcohols (cyclopentanol, cyclohexanol and1-pentanol) in order to obtain maximum RTP fromternary 1-BrN complexes in aerated solutions. Theoptimization was straightforward in case of the alco-hols, but more complicated for the surfactants due tothe existence of multiple species and the formationof micelles above the critical micellar concentration.It was observed in several studies[44–46] that the


RTP initially increases with increasing surfactantconcentration, but then rapidly decreases to zeroagain. This is probably due to the formation of mi-celles that start to compete with the CDs for 1-BrN;the formation of micelles was indicated by a decreaseof the surface tension of the solution, which reached aminimum above a certain concentration of surfactant[45,46].

3.3. Ternary complexes of non-halogenatedphosphorophores

The group of halogen-containing phosphorophoresis rather small and a more versatile technique shouldalso enable the detection of non-halogenated analytes.Such a technique, based on the formation of ternarycomplexes, is discussed here.

With non-halogenated phosphorophores deactiva-tion due to phosphorescence is slow. As noted before,bothϑISC andϑp are enhanced in most cases if a heavyatom-containing molecule is brought into close prox-imity to the phosphorophore, which results in strongerRTP. It was found that ternary complex formation ofCD, phosphorophore, and HAP is a convenient wayto achieve this goal: due to the small volume in whichphosphorophore and HAP are organized, strong RTPwas obtained from the ternary complexes of severalPAHs and related compounds[48–52]. The HAP usedin [48]—1,2-dibromoethane—not only enhancedϑISCand kp, but also caused increased protection againstquenching by molecular oxygen, most probably dueto the formation of suspended microcrystals (also seeSection 3.4). Thus, RTP was obtained even in aer-ated solutions, although higher signals were obtainedafter purging with nitrogen. The residual phosphores-cence obtained under aerated solvent conditions waslower for polar compounds such as nitrogen hetero-cyclics than for less polar compounds such as PAHs,because of the lower complexation constants andhigher exit rates[49]. Interestingly, selectivity wasderived from the relative sizes of the phosphorophoreand CD cavity, in presence of the HAP as a secondguest[48].

The intrinsic selectivity of CD-RTP was also ad-dressed in a study on the ternary complex of fluorenewith �-CD and 3-bromo-1-propanol[50]: due to thehigh complexation constant (K = 2290 M−1), fluo-rene could be detected in the presence of a 0.5–200-

Table 2Toleration ratios of several aromatic compounds in the determina-tion of fluorene (50 ng ml−1) by RTP and fluorescence

Species added Toleration ratio (species tofluorene (m/m))

Phosphorescence Fluorescence

Biphenyl 200 59-Bromofluorene 100 0.1Naphthalene, acenaphthene 40 0.05, 0.1Anthracene, benzo[a]pyrene,

benzo[e]pyrene20 0.5, 0.2, 0.5

Pyrene 4 1Fluoranthene, 1-naphthol 2 0.1, 0.1Benzo[k]fluoranthene 2 11,2:5,6-Benzanthracene,

1,2-benzanthracene2 0.5, 1

Acridine, dibenzofuran,dibenzothiophene

1 0.02

Carbazole, phenazine,triphenylene

1 0.5, 0.2, 0.2

2-Naphthol, 2-bromofluorene 1 0.1, 0.1Phenanthrene, chrysene 0.5 0.1, 0.1

Adapted from[50].

fold concentration of other PAHs (Table 2). Thisselectivity was 2–1000-fold higher than that achievedwith fluorescence measurements of fluorene includedin �-CD.

Another interesting selectivity effect based on themolecular dimensions of the analytes was observed byDeLuccia and Cline Love[53] who explored the sen-sitized phosphorescence emitted by biacetyl included,together with a sensitizer, in�-, �- or �-CD. Due tothe confinement of sensitizer and phosphorophore tothe cavity of the CD, the effective concentrations ofboth were increased and RTP was strongly enhanced.The actual enhancement was found to be dependenton the relative sizes of sensitizer and cavity: biphenyl,naphthalene, and phenanthrene were detected in�-, �-and�-CD with increasing sensitivity, while chryseneand triphenylene were detected only when complexedwith �-CD.

3.4. Analyte:CD:precipitant complexes

Several studies deal with ternary complexes consist-ing of analyte,�-CD and precipitating agent[54–57].Due to the addition of the precipitating agent—generally an apolar compound without a heavy atom—


a microcrystalline suspension with microcrystals ofca. 1�m diameter is formed in the aqueous solution[55]. Remarkably strong and long-lived phosphores-cence can be observed from analytes incorporated insuch microcrystalline suspensions even under aeratedconditions. In this respect, the method may be com-pared with the technique introduced by Weinbergerand Cline Love[58], in which RTP was obtained fromaqueous suspensions of PAHs that were previouslydissolved in methanol, THF or dimethylformamide(DMF). It should be noted that no HAP was involvedin most cases; in combination with the aggregationinduced by the precipitating agent this results in ultra-long lifetimes under deoxygenated as well as aeratedconditions[54,55].

Jin et al. [56] studied the RTP of some PAHsand nitrogen heterocyclics, induced in ternaryanalyte:�-CD:precipitant complexes. The precipitantstested includedn-alkanes, bromobenzene and cy-clohexane. Likewise, Nazarov et al.[54,55] studiedthe RTP from the test phosphorophores naphthalene,deuterated naphthalene and phenanthrene. Decalin,n-hexane, chloroform, cyclohexane and isooctanewere used as precipitating agents. Under deoxy-genated conditions at room temperature, the RTPlifetimes were about 2–15 s, while even in aerated so-lutions (at 274 K) the lifetimes exceeded 100 ms[54].Further measurements indicated that, due to the for-mation of microcrystals, the value of the bimolecularquenching rate constant,kq, for molecular oxygen wasreduced 106-fold, i.e.kq was about 103 M−1 s−1 [55].

A study on ternary complexes of 6-bromo-2-naphthol(6Br2N), �-CD and the precipitants cyclohexane,n-hexane, chloroform and benzene was also performed[57]. To investigate if the formation of microcrystalsis a prerequisite for obtaining RTP, experiments wereperformed with cyclohexanol and 2-bromoethanol asthe third component of the ternary complex. Withthese compounds, no precipitation occurred and noRTP was obtained, indicating that precipitation isindeed required.

Due to the formation of microcrystals, the pro-tection against dissolved molecular oxygen is betterthan with conventional, non-precipitating complexes.Therefore, rather high residual RTP signals wereobtained in aerated solutions, i.e. compared with de-oxygenated solutions about 20% of the RTP was stillpresent for nitrogen heterocyclics[56]. All quoted

studies reported that by far the largest RTP en-hancements were achieved with cyclohexane as theprecipitant. This may be a space-matching effect:due to the neat fit of both guests, maximal displace-ment of water occurs and the rigidity is increased,which will cause a decrease of non-radiative decay[56].

3.5. α-CD complexes

Hamai [59,60] studied the RTP of solutions con-taining 6Br2N and�-CD. In such solutions, binarycomplexes as well as ternary 2:1�-CD:6Br2N com-plexes are present, depending on the concentrationof �-CD. From a plot of RTP versus the calculatedconcentration of the ternary 2:1�-CD:6Br2N com-plex it was concluded that RTP is obtained fromthe ternary complexes only; in other words, efficientprotection of the triplet state is provided only aftercomplexation with two�-CD molecules[60]. RTPfrom the 2:1 �-CD:6Br2N complex was obtainedeven in aerated solutions, but the triplet lifetime sub-stantially increased, from 160 to 290�s, after purgingwith nitrogen [59]. In another study on the samecomplexes[61], a 13-fold enhancement was reportedafter purging. The addition of sodium sulfite (fordeoxygenation; cf.Eq. (6)) resulted in the quench-ing of phosphorescence. It is doubtful if this effectcan be attributed to bimolecular quenching, since thesulfite concentration used was less than 1 mM andeffective protection by the CDs should be expected.In contrast to experiments involving binary�- and�-CD complexes, co-complexation with alcohols suchas 2-bromoethanol, 2-bromopropanol, 2,3-dibromo-propanol, cyclopentanol and benzylic alcohol did notenhance RTP. On the contrary, quenching occurreddue to displacement of the phosphorophore from therather small�-CD cavity.

The (water-soluble) 2:1�-CD:6Br2N complexwas used for RTP optosensing of the temperature byBrewster et al.[62]. With increasing temperature, theternary complex became less stable, which resulted ina decrease of the RTP lifetime since only the ternarycomplex yields phosphorescence; over the tempera-ture range from 1.6 to 60◦C, the lifetime was reducedfrom 1.25 to 7.5�s. Due to the high sensitivity of thesensing system differences as small as 0.1◦C couldbe measured.


3.6. Heavy atom-derivatized CD complexes

As noted above, the addition of a HAP is nec-essary to obtain RTP from non-halogen analyte:CDcomplexes, except when precipitant ternary complexesare involved. Interestingly, halogen-derivatized CDscan be used as an alternative to these halogenatedsecond-guest molecules. This was first demonstratedas early as 1985 by Femia and Cline Love[63]. Theauthors obtained RTP from PAHs such as naphthalene,fluorene, phenanthrene and pyrene, included in hep-takis (6-bromo-6-deoxy-�-cyclodextrin (Br-�-CD))—synthesized by replacing the primary hydroxyls ofthe sugar groups with bromine. In order to solubilizeBr-�-CD, a mixture of water and a low-polarity sol-vent, dimethylformamide, had to be used. MaximumRTP was obtained when using 80 vol.% DMF. In theabsence of water, no RTP was obtained because in thatcase the hydrophobic driving force for inclusion intothe Br-�-CD cavity is too low.

The photophysical properties of analytes includedin iodine-substituted CDs were also studied[64,65].Compared to the corresponding parent CD complex, a20% enhancement of RTP was obtained for the 6-iodo-6-deoxy-�-cyclodextrin (�-CDI):2-chloronaphthalenecomplex[64]. For the inclusion complexes of 6Br2Nand 3-bromoquinoline with 6-iodo-6-deoxy-�-CD(�-CDI), a slight (10–18%) reduction of RTP was obs-erved compared to the corresponding 2:1�-CD:analytecomplexes[65]. Probably, the low degree of substi-tution (one iodine atom per CD) in the above cases[64,65] seriously compromizes the ability of�- and�-CDI to induce RTP. On the other hand, it ef-fects a much better water solubility than with Br-�-CD (degree of substitution, 7). It should also benoted that a better indication of the potential of theiodine-derivatized CDs to induce RTP would proba-bly be obtained if experiments with non-halogenatedphosphorophores were performed. The use of CDsderivatized with several heavy atoms as well as ionicgroups to provide adequate water solubility shouldgive better results. Interestingly, such CDs could si-multaneously be used for separation and detectionpurposes, for instance in electrokinetically drivenseparations. By contrast, due to the slow dissocia-tion kinetics associated with ternary complexes—consisting of CD, analyte and HAP—such complexesare probably not suitable for application in separa-

tion techniques because of severe peak broadening[66]. In conclusion, more research will be requiredfor a full assessment of the analytical potential ofhalogen-derivatized CDs.


CD-RTP has been obtained from several types ofbinary and ternary inclusion complexes. Based onthe RTP enhancement observed after co-inclusion ofseveral alcohols, an optosensing scheme for alcoholswas proposed. Ternary precipitant complexes of PAHshave ultralong phosphorescence lifetimes due to theefficient protection from dissolved molecular oxy-gen. Interestingly, as will be seen inSection 5, suchcomplexes can be used for RTP-based optosensing.Halogen-derivatized CDs were used to enhance theRTP from included phosphorophores. However, thepotential of such derivatized CDs for the detection ofnon-halogenated analytes has not been fully exploredyet and deserves further study, particularly concern-ing the eventual coupling of CD-RTP with separationtechniques. Finally, it should be noted that—in con-trast to MS-RTP—almost none of the many CD-RTPstudies was dealing with the application of CD-RTPin real-life analysis. Such studies, however, will berequired to establish the practicability of CD-RTP; atpresent, no evaluation can be given.

4. Heavy atom-induced RTP

4.1. Introduction

For a long time it was generally accepted thatRTP in homogeneous liquid solutions is restrictedto a few exceptional compounds such as biacetyland brominated naphthalenes. For all other com-pounds, the presence of some kind of protectivemedium such as micelles or CDs was believed tobe required. However, recently RTP was also ob-tained (under deoxygenated conditions) from severalphosphorophores (analytes) in homogeneous solu-tions, by applying HAI-RTP7 [67]. This technique

7 It has been suggested in[67] that the designation ‘non-protected(fluid)-RTP’ more adequately describes the specific nature of thistechnique. We fully agree with this argument, but prefer to use theacronym ‘HAI-RTP’ since it is commonly used in the literature.


requires the presence of a HAP at rather high con-centrations. 5-Dimethylaminonaphthalene sulfonylchloride (dansyl chloride) and 16 dansylated aminoacids were studied as target analytes; for dansylchloride an LOD of 5 ng ml−1 was obtained. In thisstudy, TlNO3 (50 mM) was used as HAP; other saltssuch as KI, KBr, and TlHCOO were reported not toinduce RTP. A neutral pH was suitable for deoxy-genation by sodium sulfite (10 mM) and for obtainingRTP.

Table 3Survey of LODs, triplet lifetimes, and type of HAP used for the detection of various compounds by HAI-RTP

Compound HAP LOD (ng ml−1) Triplet lifetime (�s) Reference

Substituted naphthalenesDansyl chloride TlNO3 5 − [67]Dansylamide TlNO3/KIa 86 222 [68]Dansylated amino acids TlNO3 − 263−640 [67]

1-Naphthylacetic acid TlNO3/KIa 15 236 [68]TlNO3 110 418 [69]

Carbaryl TlNO3/KIa 11 180 [68]TlNO3/KIa 3 180 [70]

�-Naphthalene acetamide TlNO3/KIa 8 251 [68,71]Naproxen TlNO3/KIa 18 1028 [68,72]

Nafronyl TlNO3/KIa 12 156 [68]TlNO3/KIa 3 300 [73]

�-Naphthoxyacetic acid TlNO3 2 340 [74]KI 2 350

�-Naphthoxyacetic acid TlNO3/KIa 31 – [75]TlNO3 2 1230 [74]KI 2 670

Naphazoline KI 44 632 [76,77]1-Naphthoxylactic acid TlNO3/KIa 10 – [78]Naftopidil TlNO3 21 475 [79]1-Hydroxy-2-naphthoic acid TlNO3 28 240 [69]1-Amino-5-naphthalene-sulfonic acid TlNO3 7 464 [69]

PAHsAcenaphthene TlNO3/KIb 65 417 [77]Fluorene TlNO3/KIa 15 130 [68]

Nitrogen heterocyclics2-(4-Thiazolyl)-benzimidazole TlNO3/KIa 15 89 [68]Carbazole TlNO3/KIa 4 255 [68]Tryptamine TlNO3/KIa 20 90 [68]Tryptophan TlNO3/KIa 11 166 [68]Indole-3-butyric acid TlNO3/KIa 25 145 [68]9-Hydroxy-4-methoxyacridine TlNO3/KIa 21 91 [68]

a HAP used to obtain the given LODs and triplet lifetimes, i.e. optimal HAP.b Optimal HAP not specified.

4.2. Application range and analytical merits

Soon after its introduction, HAI-RTP from othersubstituted naphthalenes, as well as from nitrogenheterocyclics and some PAHs, was also reported; anoverview is given inTable 3. For all substituted naph-thalenes, the RTP emission spectra were very similarsince the RTP originates from the naphthalene partof the molecule. The LODs inTable 3range from afew to about 100 ng ml−1. Therefore, trace analysis of


real-life samples will often require preconcentration.PAHs such as naphthalene, anthracene, phenanthreneand pyrene[69] give only weak RTP signals (noLODs were presented), which was attributed to lowkp values (a�–�∗ transition instead of an n–�∗ tran-sition is involved in this case). However, for somePAHs (fluorene and acenaphthene) rather low LODswere achieved (Table 3).

The linear detection range is generally limited toabout one order of magnitude. This may be acceptablein routine analyses of pharmaceutical products, but itis a clear disadvantage in, for instance, environmen-tal analysis where analytes can be present at widelydifferent concentrations.

TlNO3 (25–250 mM) and KI (typical concentra-tion, 1 M) were used in all cases to induce intersystemcrossing (Table 3). KI was most suitable for gener-ating RTP with nitrogen heterocyclics; for the sub-stituted naphthalenes, the optimal HAP is dependenton the individual compound. Sodium sulfite was usedfor deoxygenation in all studies, at concentrationsof 1–15 mM [67–79]. At lower sulfite concentra-tions longer stabilization times were needed[70–78],while at higher concentration levels the RTP intensitywas lower [72,78], possibly due to sulfite-inducedquenching. Stable RTP signals were generally ob-tained within a few minutes and the triplet lifetimeswere in the range of ca. 0.1−1 ms (Table 3), which isquite close to the lifetimes generally encountered inMS-RTP and CD-RTP.

Due to the inherent selectivity of RTPL, in severalcases analysis could be performed without (extensive)sample pretreatment. For instance, the determinationof 1-naphthoxylactic acid in spiked serum and urine(200 ng ml−1) was performed successfully (recovery,97%) after mere dilution of the sample[78]. Likewise,naproxen[72] and nafronyl[73] were determined inpharmaceutical formulations. It should be noted thatthe composition of the latter samples is not very com-plex and analyte concentrations were quite high; sam-ple pretreatment will remain necessary in many othercases.

An advantage of HAI-RTP is the intrinsic simplicityof the method: no ‘pseudophase’ is involved and thesolutions are stable and clear. Moreover, the methodis directly applicable to biological samples contain-ing high sodium ion concentrations, in contrast toMS-RTP which suffers from displacement of the thal-

lium ions from the micelles by the sodium ions presentin the sample[78].

Some authors mentioned that the technique, incontrast to MS-RTP and CD-RTP, is well suited forcombination with flow injection analysis (FIA), LCand CE [67,69]. Indeed, as noted before, CD-RTPcannot be combined with separation methods unlesshighly dynamic CD complexes are involved. How-ever, MS-RTP detection has been applied in LC inthe past[80], despite the foaming problems asso-ciated with nitrogen purging. Moreover, the use ofsodium sulfite for deoxygenation, as introduced byDıaz-Garcıa and Sanz-Medel[11], eliminates thisproblem. Finally, it should be noted that the com-bination of separation methods and HAI-RTP itselfwill be seriously hindered by organic solvent effects:several commonly used solvents such as acetone andmethanol strongly reduce the RTP intensity even atthe 1 vol.% level or—as is the case with the additionof 1 vol.% acetonitrile—extend the required stabi-lization time to more than 30 min[74]. Therefore,although 20% acetone[67] or 10% methanol[73]can be tolerated in specific cases—possibly depend-ing on the analyte—the coupling of HAI-RTP andseparation techniques seems most promising for sepa-ration procedures aimed at polar or ionic compounds,which do not require high concentrations of organicsolvent.


HAI-RTP is a recently introduced method for ob-taining RTP in which no protective medium is needed.Many substituted naphthalenes and nitrogen hetero-cyclics show RTP in the presence of Tl(I) or I− ions(HAP) and sodium sulfite (deoxygenation). HAI-RTPcan therefore be regarded as a useful method for ex-panding the application range of direct RTP. Alsohere, studies involving real-life samples are requiredto establish the practicability of HAI-RTP. The cou-pling with separation methods may be simplified sinceno micelles or CDs are involved. On the other hand,the coupling may be limited to aqueous-phase sepa-rations, since organic solvents were found to stronglyreduce phosphorescence in several HAI-RTP experi-ments. More research on the nature and extent of thesolvent effect will be required, however, before a fullevaluation can be given.


5. Indirect phosphorescence detection

5.1. Introduction

All RTP methods that were discussed above, exceptfor some examples of MS-RTP, are based on directphosphorescence, i.e. phosphorescence obtained afterdirect excitation of the phosphorophore. In general,these methods require not only efficient sample de-oxygenation but also efficient protection of the ex-cited triplet state by CDs, micelles, or another kind oforganizing medium—with HAI-RTP as the exception.

Some special compounds, such as biacetyl andvarious bromonaphthalenes, can yield strong RTP inliquid solutions even in the absence of a protectivemedium. Based on these phosphorophores, indirectRTP modes like sensitized and quenched phosphores-cence can also be used in free solution. This openswide perspectives, since under these conditions theindirect RTP methods can easily be combined witha FIA set-up or with a liquid separation method.8 Ofcourse, low oxygen concentrations in the LC eluentsor CE buffers are required to obtain indirect RTP; inaddition, for LC analysis the samples also have to bedeoxygenated. Hence, adapted set-ups have to be used.

Already in the early 1980s it was shown that manycompounds, including (poly)chlorinated biphenylsand naphthalenes[81], as well as biacetyl[82], canbe detected by LC with sensitized phosphorescencedetection. On the other hand, the coupling of LCwith quenched RTP allowed for the detection ofchloroanilines, chlorophenols, nitroaromatics, nitro-gen heterocyclics, sulfur-containing compounds andinorganic ions[83]—all of which are compounds ex-hibiting high kq values (cf.Eq. (4)). Obviously, bothin sensitized and quenched RTP, the LODs dependon the RTP signal provided by the phosphorophore.It is interesting to note that strong phosphorescencefrom biacetyl or bromonaphthalenes can also be ob-tained in (homogeneous) aqueous solutions, althoughthe strongest RTP is obtained in non-polar solvents.This does not only imply that aqueous eluents canbe used in LC, but also, and more importantly, thatthe coupling of (indirect) RTP detection and CE is in

8 Indirect RTP is also used in many cases for batch measure-ments, for instance, for optosensing of molecular oxygen, as willbe discussed inSection 6(RTP optosensing).

principle possible. The recent development of such atechnique will be highlighted inSection 5.2.

5.2. Sensitized and quenched RTP detection in CE

Due to the short optical pathlengths and small de-tection cell volumes available in CE, optical methodssuch as absorption generally offer limited detectabil-ity; therefore, alternative detection strategies are ofprime interest. An intrinsic difficulty associated withthe coupling of RTP detection and CE is the gen-eration of molecular oxygen at the anode (injectionside) during electrophoresis, due to the electrolysis ofwater. In the first study on sensitized and quenchedphosphorescence detection in CE[5], the separationvoltage polarity was reversed, i.e. a negative volt-age was applied during electrophoresis. Accordingly,injection of samples and deoxygenation of the an-odic buffer was performed outside the CE instrument(Fig. 5). To maintain low oxygen levels, an LC pumpwas used to continuously pump the previously deoxy-genated buffer through the anodic (24�l) interface,which alternatingly functioned as the buffer and thesample vial. In later studies[84,85], a simplifiedset-up was used which reduced the start-up time from2 h to 20 min. However, the first set-up may wellprovide easier on-line coupling with LC for sam-ple pretreatment or analyte enrichment. Interestingly,as can be seen inFig. 6 (traces a and b, respec-tively), sensitized (biacetyl) phosphorescence wasobtained for non-substituted naphthalene sulfonates(NS), whereas (near) diffusion-controlled quenchingof biacetyl phosphorescence was observed for thehydroxy- and/or amino-substituted NS; LODs werein the 4× 10−7 to 5× 10−8 M range.

In quenched RTP, the signal is governed by theStern−Volmer equation (cf.Eq. (4)). Hence, thedegree of quenching induced by a certain analyte,present at a certain concentration, is independent ofthe height of the phosphorescence background. There-fore, LODs can be improved by enhancing the RTPbackground signal, which increases its signal-to-noiseratio (S/N). Unfortunately, the molar absorbance ofbiacetyl is very low over its entire absorption spec-trum [83]. While this is an advantage in sensitizedRTP, where direct excitation has to be avoided, it isa clear disadvantage in quenched RTP (Fig. 6, traceb). For this reason, 1,5-dinaphthalenesulfonate was


Fig. 5. Instrumental set-up for CE with sensitized and quenchedphosphorescence detection, consisting of an ultrasonic bath and anitrogen gas supply for deoxygenation, an LC pump for pumpingthe CE buffer and the sample (1 ml/min), a luminescence detector,and a CE system. Sample introduction in the CE capillary isperformed by hydrodynamic injection at a fixed time interval afterswitching the six-way valve. Inset: interface with the injection endof the capillary and the positive electrode inserted in it. Figuretaken from[5].

used in a subsequent study to sensitize biacetyl priorto measuring its (quenched) phosphorescence[86].Combined with the use of the total emission mirrorof the luminescence spectrometer, this provided a25-fold enhancement of S/N and, thus, of the LODs(Fig. 7). Several buffers commonly used in CE, i.e. bo-rate, phosphate, malonate, acetate and succinate (pH4.7–9.0), were found to be compatible with quenchedRTP detection. However, the method could not be usedat extreme pH values, because enolization of biacetyloccurs in both strongly acidic and basic solutions.Another water-soluble phosphorophore was thereforeused in later studies (vide infra). The method wasused to detect substituted NS, nitrophenols, hydrox-ybenzoic acids, thiocarbamates and sulfur-containingamino acids, with LODs down to 10 nM[86].

In both of the above studies, a pulsed xenon lampwas used to enable time-resolved detection (delay

Fig. 6. Electropherograms of five NS. Buffer: 25 mM borate, pH8.5, containing 0.02 M biacetyl. Peaks: (1) oxygen; (2) 1-naphthale-nesulfonic acid; (3) 4-amino-5-hydroxy-2,7-naphthalenedisulfonicacid; (4) 2,6-naphthalenedisulfonic acid; (5) 4,5-dihydroxy-2,7-na-phthalenedisulfonic acid; (6) 1,3,7-naphthalenetrisulfonic acid. De-tection: (a) sensitized RTPL, concentration of NS: 10−5 M; (b)quenched RTPL, concentration of NS: 2.5×10−6 M. Figure takenfrom [5].

time, 50�s; gating time, 0.5–1.0 ms). This enabled theeasy removal of scatter and fluorescence and causedan increase of S/N of the long-lived RTP backgroundsignal (τ0, ca. 70�s).

CE-quenched phosphorescence was also applied tonon-derivatized, low-molecular-weight amino acids,which are not readily detected with high sensitivity bymost other detection techniques[87]. Since quenchedphosphorescence detection of amino acids is basedon electron transfer originating from the deprotonatedamino groups, the pH of the separation buffer shouldbe higher than 9. Since biacetyl cannot be used atpH > 9, a pH-insensitive and water-soluble com-pound, 1-bromo-4-naphthalenesulfonic acid, was usedas phosphorophore instead. Also in this case, nearlydiffusion-controlled bimolecular quenching was ob-served, with LODs close to 10 nM[87].

To expand the applicability of phosphorescence de-tection in CE to neutral analytes, the compatibility ofquenched RTP detection and cyclodextrin-based elec-trokinetic chromatography (CD-EKC) was studied


Fig. 7. Electropherogram of mixture of 22 NS (10−6 M each). The NS are indicated by the positions of their functional groups on the naph-thalene skeleton, i.e. amino/hydroxy/sulfonic acid. Buffer: 25 mM boric acid, pH 9.0, 0.02 M biacetyl, 1×10−3 M 1,5-naphthalenedisulfonicacid. Injection volume equivalent to 288 mbar. Figure taken from[86].

[84]. Nitro-aromatics were chosen as test analytes,because the nitro group provides (near) diffusion-controlled quenching. A priori, it may be expected thatinclusion of the analytes in the (charged) CDs will re-duce the sensitivity of quenched RTP by reducing thebimolecular quenching rate constants. However, dueto the low capacity factors generally encountered inCD-EKC, a large fraction of the analytes (quenchers)will be present in the aqueous phase. Consequently,

the resulting bimolecular quenching rates were foundto be high and LODs low (ca. 20 nM).

The experiments discussed above were all per-formed using conventional light sources. The moststraightforward method to improve LODs in CE-quenched phosphorescence may seem to be enhance-ment of the continuous phosphorescence backgroundby using a laser for excitation. Indeed, it was observedthat the RTP background was significantly enhanced


when using a small-size, 7.8 kHz-pulsed Nd–YAGlaser with a highly stable output and a home-built de-tector instead of a conventional luminescence detector[85]. However, the LODs were not better than withthe lamp-based system: under low irradiance condi-tions the triplet lifetime was long (τ0, ca. 300�s), butquenching between two subsequent laser pulses wasrestricted to a short period of time only (ca. 130�s),which reduced the observed quenching efficiency. Athigh irradiance, the efficiency of bimolecular quench-ing was strongly reduced due to the competitionoffered by triplet−triplet annihilation. When using acylindrical lens for excitation (1 mm detection win-dow), LODs similar to those obtained by pulsed lampexcitation (10 nM) were achieved[85]. Therefore,laser-induced RTP will be more appropriate thanlamp-based RTP (4 mm detection window) if highlyefficient separations are involved.


Quenched phosphorescence detection was shown tobe an interesting detection mode for CE and CD-EKC.Its LODs are in the 10−7 to 10−8 M range, whichis beyond the possibilities of many other techniques,especially for compounds with poor chromophores;absolute detection limits are about 1 fmol. Interest-ingly, such LODs were obtained by using conven-tional, lamp-based detection systems. The use of apulsed, solid-state laser did not improve the detectabil-ity. However, more research will be required to fullyassess the usefulness of laser-based detection. The rel-ative simplicity and high sensitivity of the method maywell render it suitable for detection in chip-based sep-aration systems; in terms of detectability, favorable re-sults should be expected compared to commonly usedindirect detection methods such as indirect absorptionand indirect fluorescence.

6. RTP optosensing

6.1. Introduction

Although some examples of RTP-based optosens-ing were already given inSections 2.2, 3.2, and 3.5,the topic no doubt deserves a more comprehen-sive evaluation. RTP optosensing methods, based on

various detection schemes, have been developed todetermine analytes such as oxygen (transition) metals,lanthanides, organic compounds, and enzyme sub-strates like glucose. Most RTP optosensing detectionschemes are based on FIA set-ups, but probe-typeexperiments are also used. According to Sanz-Medel[88], an important advantage of the FIA mode is thepossibility of using irreversible chemistry to achieveselective binding prior to detection. Low-cost au-tomation involving integrated sample pretreatment orenrichment can also be achieved quite easily[89].Costa-Fernández and Sanz-Medel[89] suggested thatthe development of sensing phases based on molec-ularly imprinted polymers (MIP) can be useful foroptosensing purposes, since MIP sensing phases arerobust, possess high operational stability, and chem-ical (acid, organic solvents) and thermal resistance.Moreover, MIP sensing phases allow for selectiveinteraction prior to detection and can, in principle, bedesigned for any target analyte while their productionfor low-molecular-weight compounds is almost rou-tine [90,91]. Some examples[92–94] of using MIPphases for sensitized and quenched luminescence de-tection will be discussed below. Interestingly, theseexamples point to the feasibility of using MIP phaseswith sensitized and quenched RTP detection.

6.2. Oxygen sensing

The sensing of molecular oxygen based onluminescence9 quenching is regarded as one of themost typical and widespread optosensing applications.In all luminescence quenching detection schemes,a—preferably strong and low-noise—luminescencebackground signal is generated, which is dynami-cally quenched in the presence of molecular oxygenor other quenchers (cf.Eq. (4)). Although sensing isperformed both in the gaseous and the liquid phase,the present discussion is restricted to detection in theliquid phase.

In most cases, Ru(II)−�-diimine complexes (life-time, several hundreds of nanoseconds to tens of mi-croseconds) or Pt(II)– or Pd(II)−porphyrin complexes(lifetime, hundreds of microseconds) are used as lu-minophores. Advantages are their long luminescence

9 The term ‘luminescence’ is used in this case since not onlyRTP but also (long-lived) fluorescence is involved here.


lifetimes, convenient excitation in the visible range,and large Stokes’ shifts (wavelength difference be-tween excitation and emission maxima). Next to stan-dard intensity measurements, lifetime measurementsbased on time-resolved detection and phase modula-tion are used[95]. The use of RTP intensity is straight-forward but is not as reliable as the other methods,which are independent of changes in illumination andcollection efficiencies (light source intensity[96–98],local differences in transparency of the samples andchanges in optical properties of the tissues studied[99,100]), variations in the luminophore concentration(photobleaching[96,98,99], circulation concentrationof the luminophore in tissues[99]), inhomogeneitiesof the sensor phases[97], and membrane positioningin probe-type experiments[98].

As regards the luminophore and the matrix used toimmobilize it, the new Pt(II)− and Pd(II)−porphyrinketone dyes provide a 10-fold increased photochem-ical stability compared to Pt(II)−octaethylporphine[101]. In addition, favorable overlap with yellow andorange light-emitting diodes (LEDs) was achievedand, thus, convenient excitation in the visible rangefor sensing purposes[101]. The position of the ab-sorption bands can also be important for the detectionof oxygen in tissue, since other chromophores presentin tissue can limit the penetration depth of the excita-tion light to less than 1 mm[102].

As for the matrix, materials such as polymer films,sol–gel phases, zeolites and siloxanes have beentested[95]. A good matrix for inclusion of the sensingmolecules is chemically inert and optically transpar-ent, possesses photochemical and thermal stability,and shows negligible intrinsic fluorescence. Manypolymer and sol–gel phases largely fulfil these re-quirements[89]. The porosity of the sensing phase isalso important since the quencher should be able tointeract with the immobilized luminophore[103,104].

Dıaz-Garcıa et al.[105] used FIA with four flow-through cells packed with an Al–ferron doped sol–gelphase to sense dissolved molecular oxygen in water.An intensified charge-coupled device (CCD) cameraand fiber-optics were used for detection and exci-tation/collection, respectively. Other sensing phaseswere also applied in the FIA format: anion-exchangeresin beads with immobilized Al−ferron, incapsulatedin a highly oxygen-permeable silicone rubber filmprovided high operational stability (no substantial de-

crease in RTP intensity during 20 h of continuous oper-ation) and response times of about 2.5 min for sensingin the liquid phase[103]. Sol–gel glasses doped withAl–ferron were also used; high operational and stor-age stability were reported and an LOD of 6 ng ml−1

in water was obtained[104].Lähdesmäki et al.[106] applied bead injection spec-

troscopy for the determination of oxygen consumptionby adherent cell cultures. For detection, microcarrierbeads with immobilized Pt(II)−porphyrin were tran-siently retained at the bottom of a flow-through micro-scope chamber. The bead format was used to ensurefast consumption of oxygen. Three minutes beforeand during RTP intensity measurements the flow wasstopped so that the oxygen in the buffer was depleted.

Probe-type sensing was also applied in many in-stances: Papkovsky et al.[98] used the well-platereader accessory of a standard luminescence spec-trometer for oxygen detection in aqueous samples.The Pt(II)−octaethylporphine complex used for optos-ensing was contained in a polystyrene film depositedat the bottom of the well-plates. With a compara-ble set-up, the metabolic activity (oxygen consump-tion) of living yeast cells on microtiter plates wasmonitored[107].

Several groups performed in vivo oxygen determi-nations based on RTP quenching: the partial oxygenpressure was determined in prepared male Wistar ratsthat were allowed to breathe ambient air, using invivo RTP lifetime measurements[96]. RTP was ob-tained from a Pd(II)−porphyrine solution, which wasinjected intravenously. Non-invasive oxygen determi-nation in arterioles and venules of unanesthetisizedhamsters was achieved by using an implanted trans-parent skin fold chamber[100]. ThepO2 values foundwere similar to those measured using microelectrodes.Spatial in vivo measurements of oxygen levels in ratliver tissue during photodynamic therapy were alsoreported[99]; the oxygen levels near the excitationfiber were observed to decrease from 20 to 2 mmHg.Photosensitizer and phosphorophore were injectedprior to measurements. The optical fibers used for ex-citation and detection were closely spaced (a few mmapart) and positioned just above the exposed tissue,i.e. the technique is non-invasive (Fig. 8). It shouldbe emphasized that RTP intensity measurements can-not be used in this case because of problems relatedto photodegradation and variations in the circulation


Fig. 8. Set-up for time-resolved RTP detection of oxygen levels during photodynamic therapy. The excitation pulse is focused into the probefiber which also collects phosphorescence emission. Phosphorescence is reflected from a dichroic mirror onto an avalanche photodiode vialong pass filters. The distance between the irradiation and probe fiber can be adjusted. Figure taken from[99].

concentration levels of the RTP probe, and changes inthe optical properties of the tissue. In contrast, RTPlifetime can be used without such problems.

6.3. Heavy atoms

FIA was combined with RTP for the detection ofseveral metal ions[108–110]; the method is based onthe RTP obtained from a complex of the metal witha ferron-like compound, due to the ensuing enhance-ment ofϑISC andϑp (also seeSection 2.2). The phos-phorescent complex was transiently retained on ananion-exchange resin, which is stable for 2–3 months,in the flow-cell of the luminometer. Regeneration ofthe resin was performed by injecting a small volumeof a strongly acidic (HCl) solution. Sodium sulfite wasused for deoxygenation.

Using 8-hydroxy-7-quinolinesulfonic acid as theligand, Pb(II) was determined in seawater spiked at aconcentration of 6–100 ng ml−1 (recovery, 94–109%)[108]. An LOD of 0.1 ng ml−1 and a linear rangeof three to four decades were achieved for stan-dard solutions. Some selectivity against other metalswas reported (Table 4), but the interference of these

metal ions should be studied more thoroughly inorder to fully understand its nature and extent. Like-wise, Al(III) was detected (LOD, 2 ng ml−1) by theRTP obtained from its complex (Al−ferron) with8-hydroxy-7-iodo-5-quinolinesulfonic acid (ferron)[109]. The method showed selectivity against manyions at levels normally present in dialysis fluidsand was successfully applied to the determinationof Al(III) in such fluids; Fe(III) interferes but canbe masked by adding 9,10-phenanthroline. RTP op-tosensing was also applied to the detection of Hf(IV)using ferron or 8-hydroxy-5-quinolinesulfonic acidas the ligand[110]. With the latter ligand moderateselectivity against Ga(III), In(III), and Zr(IV) wasobtained: thus, reliable determination of Hf(IV) waspossible in the presence of a 5–30-fold excess.

Iodide is also a well-known enhancer of ISC inorganic compounds. Of course, simultaneously withthe enhancement ofϑISC, the fluorescence quantumyield will be reduced upon complexation of iodide.Using the luminescence of immobilized 8-hydroxy-5-quinolinesulfonic acid, both the reduction of fluores-cence and the enhancement of RTP were applied tomonitor iodide (Fig. 9). LODs were 5 and 10 mg/l,


Table 4Effect of metal ions on RTP optosensing of lead (100 ng ml−1),with 8-hydroxy-7-quinolinesulfonic acid as complexing agent

Metal ion Concentration (mg/l) Intensity (%)

– – 100

Cd(II) 0.2 1081 107

Mg(II) 1 942 88

150 88

Al(III) 0.25 1001 91

Hg(II) 0.2 1030.4 94

Zn(II) 0.4 971 92

Cu(II) 0.2 900.4 77

Co(II) 0.2 1010.4 86

Fe(III) – –a

Mn(II) 0.1 79

Ca(II) 5 9986 173

Taken from[108].a Addition of 50 mg/l 1,10-phenanthroline in the buffer elimi-

nates interference of 5 mg/l Fe(III).

respectively[111]. The method was over 200-foldmore sensitive for iodide than for other halides. TheRTP signal was strongly reduced by organic solventssuch as ethanol and acetone, but the fluorescencesignal was not significantly influenced even at con-centrations as high as 25 vol.%. Fe(III) and Cu(II)ions interfered in both detection modes and should beremoved prior to analysis.

6.4. Lanthanide–ligand complexes

Some trivalent lanthanide ions show narrow lumi-nescence emission bands with up to millisecond life-times. However, detection based on direct excitationis quite unfavorable since their absorptivities are verylow. Because of this, the use of sensitized lanthanideluminescence from lanthanide−ligand complexes is abetter option, which enables detection of the rare-earthmetals as well as the ligands.

Fig. 9. Response for optosensing of iodide in standard solutionusing RTP and quenched fluorescence detection. Figure taken from[111].

Eu(III) and Tb(III) are important rare-earth metalions present in the earth’s crust, while gadoliniumchelates are being evaluated for their use as contrast-enhancing agents in magnetic resonance imaging,which requires analytical methods for the determi-nation of Gd(III) in biological fluids and tissues.Sensitized lanthanide luminescence was applied ina FIA set-up for the determination of Eu(III)[112],Gd(III) [113] and Tb(III) [114]. For detection, theluminescent lanthanide−ligand complex was tran-siently immobilized on a resin in a 25�l flow-cell.Because the energy transfer process is intramolecu-lar by nature, oxygen does not affect the emission.Immobilization of the complex on the resin resultedin increased luminescence intensities and lifetimes,which was attributed to the increased rigidity of theenvironment and removal of water from the first coor-dination sphere[113]. The best results were obtainedfor Eu(III) and Tb(III) with LODs of 1 and 3 nM,respectively.

Conversely, ligands such as tetracyclines[115] andanthracyclines[116] were detected using the lumines-cence of their complexes with Eu(III). Immobilization


of the tetracyclines resulted in four-fold lower LODs(0.25–0.4 ng ml−1) than with liquid-phase detection;for the anthracyclines, LODs of 6−9 ng ml−1 wereachieved. The method was successfully applied to thedetermination of these compounds in spiked urine (re-covery, 96−101%) and pharmaceutical formulations(accuracy better than 0.3%)[115,116].

6.5. Biosensing

The experimental set-ups for the detection ofmolecular oxygen can also be used for the detectionof enzyme substrates like glucose[98,107,117–119],phenolic compounds[120] and cholesterol[121].In the case of glucose, RTP was obtained after theconsumption of oxygen due to the enzymatic reaction

glucose+ O2 + H2Oglucose oxidase→ gluconic acid+ H2O2 (7)

which implies that the obtained RTP is a measureof the glucose concentration. Also in this case, bothFIA [117–121]and probe-type[98,107]analyses weredeveloped. With FIA, air-equilibrated carrier streamswere used to guarantee equal initial oxygen concentra-tions for all samples; the response times were gener-ally 2–5 min. With the probe experiments for glucosedetection, steady-state signals were obtained in lessthan 2 min[98] or even in 10–15 s[107]. The sampleswere exposed to the ambient air, so that the oxygenlevels at the membrane result both from consumptionat the membrane and diffusion from the bulk phase.Both with FIA and probe experiments, fiber-opticswere used for excitation and collection of the emission.

Papkovsky et al.[117] published a study on RTPoptosensing of glucose, based on a FIA set-up con-sisting of a peristaltic pump, injection loop, glucoseoxidase column and flow-through cell. Water-solublePd(II)− and Pt(II)−porphyrins were used to obtainRTP; with the Pt(II) complex detection was possibledown to ca. 0.5 mM. Other workers used a FIA set-upwith a glucose oxidase mini-column positioned be-fore the flow-cell [118]. The cell was packed withan anion-exchange resin to immobilize the phospho-rophore (Al−ferron). An LOD of 80�M was obtainedfor glucose and the method was successfully applied toserum, fruit juice, and lemonade samples which wereinjected after mere dilution. No interference from fruc-

tose, saccharose, or lactose was observed, while theinterference from galactose (due to galactosidase im-purities in glucose oxidase) was only minor. Ovchin-nikov et al. [119] used two designs: a FIA set-up inwhich the glucose oxidase was immobilized directlyon the oxygen membrane sensor and a set-up usinga glucose oxidase microparticle column with a sens-ing membrane at the column outlet; the latter providedthe best detectability (LOD,<0.1 mM); RTP lifetimeswere measured using phase modulation.

Probe-type experiments for glucose determinationwere performed using the same set-up as for oxygendetection[98,107] (cf. Section 6.2), but glucose oxi-dase was immobilized on the sensing membranes inaddition to the phosphorophore. With RTP intensityand lifetime measurements, LODs of 20 and 50�Mwere obtained, respectively[98]. Phase modulationwas also used to determine RTP lifetimes[107]; inthis case, detection in serum was possible down to0.3 mM glucose.

Polyphenolic compounds were detected in a FIAset-up including an immobilized laccase enzyme col-umn and an oxygen-sensing membrane with Pt(II)−octaethylporphine[120]. The best results were ob-tained for di- and polyphenols (LODs down to0.2 mM). The method was successfully applied to thedetermination of the total polyphenol content in tea(expressed in mM hydroquinone). Likewise, the de-termination of cholesterol was studied using a choles-terol oxidase (ChO) column in front of a flow-cellwith Al−ferron on an anion-exchange resin[121]; theLOD was 50�M. A system using a flow-through cellwith mixed ChO and Al−ferron phases was also usedbut found to be less sensitive. To ensure proper sol-ubilization of cholesterol, a hexane/chloroform (95:5(v/v)) carrier was used. The method was applied to thedetermination of cholesterol in butter and eggs; theresults were consistent with those of the Boehringertest (visible/UV enzymatic method). Unfortunately,the operational stability of the ChO-column was lim-ited to about 60 assays.

6.6. Various sensing schemes

An interesting FIA-RTP optosensing method forthe determination of anionic surfactants—in particularfor SDBS—was presented by Badıa and Dıaz-Garcıa[122]. The active RTP sensing phase consisted of


Fig. 10. Schematic presentation showing Al−ferron/BSA active phase conditioning and RTP optosensing of anionic surfactants: (a) RTPspectra of Al−ferron in deoxygenated solution, both in the absence and presence of SDBS. Figure taken from[122].

Al−ferron and bovine serum albumin (BSA) im-mobilized on a strongly basic anion-exchange resin(Fig. 10). Under deoxygenated conditions, the pres-ence of SDBS was observed to generate RTP due tosurfactant-induced stretching of the BSA molecules.This unfolding was considered to result in RTP dueto the increase in local microviscosity and protectionagainst molecular oxygen. No RTP was obtained inthe presence of cationic, non-ionic or amphoteric de-tergents. Consequently, selective detection of anionicdetergents was achieved, although some (cationic andzwitterionic) detergents were observed to quench RTPby about 40% when present in three-fold excess.

Sensing of DNA was performed using the RTP en-hancement obtained from a Pd(II)−porphine complexin the presence of double-stranded DNA—most proba-bly also due to increased rigidity[123]. Sodium sulfitewas used for deoxygenation. A good selectivity againstRNA and (surface-active) albumin was reported and

the LODs were in the attomolar range. The resultsobtained in the DNA analysis of human colon tissuewere consistent with the results from UV absorptionmeasurements at 265 nm.

Optosensing schemes based on luminescencequenching have already been discussed inSections6.2 and 6.5. In some cases, the ability of a certaincompound to quench RTP depends on solvent condi-tions, or it can be influenced by complexation with ananalyte of interest. Jin et al.[124] reported an interest-ing pH optosensing scheme in which the quenchingof RTP was enhanced when the pH increased, thusenabling monitoring of pH in the range 4–8. The en-hancement of quenching due to ET was attributed tothe increased overlap of the RTP emission spectrumof the ternary phosphorophore:�-CD:cyclohexanecomplex and the combined absorption spectrum ofthe pH-sensitive dyes (phenol red, bromocresol purpleand bromophenol blue); no direct influence of the pH


on the RTP was observed. The technique was appliedin the FIA format, using several polymeric resins anda sol–gel phase for immobilization of the moleculesinvolved in sensing. The best results in terms of re-sponse time, sensing phase stability and sensitivitywere obtained when using the sol–gel phase.

A similar principle was reported for the detectionof Hg(II) [125]: complexation of Hg(II) with dithi-zone was observed to effect a distinct shift in theabsorption spectrum of dithizone, which resulted inincreased spectral overlap with the emission spec-trum of the 1-BrN:�-CD:cyclohexane complex and,thus, in increased quenching of RTP. An LOD of14 ng ml−1 was achieved, with moderate selectivitytowards a number of metal ions. Cu(II) and Ag(I)strongly interfered since they can also form strongcomplexes with dithizone[125].

Interestingly, ET between the precipitant phospho-rescent complex and the acceptor molecules appearedto be efficient enough for detection purposes in bothstudies, although dissolved molecular oxygen was notinhibiting (less than 5% difference between deoxy-genated and aerated solutions). This suggests that along-range, coulombic energy transfer process may beinvolved. Although such a process is spin-forbiddenin this specific case, it can become important whenother routes are blocked[3]. This may well be thecase for the precipitant ternary complexes considered:due to the highly rigid environment of the phospho-rophores in such complexes and adequate shieldingfrom molecular oxygen, non-radiative deactivationwill be largely inhibited so that extremely long tripletlifetimes of, typically, several seconds are observed[54,55].

Temperature, pH andpCO2 were monitored on mi-crotitre plates using sensor films with various Ru(II)complexes[97]. Due to the rather short luminescencelifetime of the Ru(II) complexes (0.33–4�s), thepresence of molecular oxygen in the sample did notinhibit detection. Also in this case, optosensing wasbased on modulation of ET. Variation of pH causedchanges in the protonation state and absorption spec-trum of the acceptor, bromothymolblue, and, thus,differences in spectral overlap with the excited statedonor. Sensing ofpCO2 was performed in a similarway. Temperature-sensing was based on the increasedthermal deactivation of the luminescent states athigher temperatures. The methodologies were not

tested with real-life samples. Nevertheless, such testswill be required to evaluate their selectivity.

6.7. Molecularly imprinted polymers

A few examples of quenched fluorescence detec-tion and sensitized lanthanide luminescence detec-tion are presented here to illustrate the potential ofmolecularly imprinted polymer (MIP) phases for sen-sitized and quenched RTP detection. In one study, thefluorescence of the functional monomer, trans-4-[p-(N,N-dimethylamino)styryl]-N-vinylbenzylpyridin-ium chloride was quenched upon binding of cyclic 3′,5′-adenosine monophosphate (cyclic AMP)[92]. Thatis, the functional monomer was involved in recog-nition as well as detection. Problems were the longresponse time of about 30 min and the observed hetero-geneity of the binding sites with regard to accessibilityand binding affinity. Likewise, the fluorescence of aZn(II)−porphyrin signalling functional monomer wasquenched upon binding of 9-ethyladenine (Fig. 11)[93]; as regards structural analogues, no binding ofadenine was observed, while 2- and 4-aminopyridinewere bound 37- and 13-fold less strong, respecti-vely. As a final example, sensitized luminescence ofEu(III), that was partly complexed by a functionalmonomer, was used to detect pinacolyl methylphos-phonate (PMP), a hydrolysis product of the nerve gassoman[94]. The MIP sensing phase was applied atthe tip of an optical fiber. LODs as low as 1 and 7 ng/lwere obtained for a bench-top and portable set-up,respectively. Although selectivity undoubtedly is animportant characteristic of MIP phases, the observedselectivity against organophosphorus pesticides,which may interfere with the determination of PMP, ismainly based on the detection part (unique emissionband at 610 nm) and only partly on selective binding.


In the last decade, RTP optosensing methods basedon principles such as bimolecular quenching, enhance-ment of ISC, sensitized lanthanide luminescence, andmodulation of ET have been developed. In addition,some analytes of interest were detected by interac-tions enhancing the rigidity of their environment. TheET modulation technique is based on variations inthe spectral overlap of the RTP emission spectrum


Fig. 11. Schematic representation of molecular imprinting of 9-ethyladenine (9EA) using a Zn(II)−porphyrin complex (1) and methacrylicacid (MAA) as a functional monomer. Figure taken from[93].

and the absorption spectrum of the quenchers(triplet-to-singlet energy transfer). It is not yet knownto which extent the inclusion of the phosphorophoresin such complexes reduces the bimolecular quenchingrate constants. However, the quenching was shown tobe adequate for analytical purposes and—of particu-lar interest from a practical point of view—dissolvedoxygen hardly plays any role.

The main attractivity of optosensing schemes de-rives from their selectivity, for instance, in biosensingof enzyme substrates, detection of DNA, and lumines-cence detection after complexation. However, somemethods will not be selective enough as such for alltypes of samples. For instance, although no seriousinterferences should be expected in the detection ofoxygen in air, its detection in aqueous samples maywell be hindered by organic or inorganic quenchers.Apart from selectivity, improved sensitivity and sens-ing phase stability are still required. That is, RTP op-tosensing is still in its infancy, despite the interestingdetection principles that have been demonstrated. Fi-nally, the examples of MIP sensing phases using sen-sitized and quenched luminescence suggest that thereare also analytical possibilities for MIP sensing phasesbased on sensitized and quenched RTP detection.

7. RTP measurement of biomolecules

7.1. Introduction

As was mentioned inSection 6.6, RTP can be usedfor the detection of (double-stranded) DNA in aqueous

solutions. In this section, the application of phospho-rescent labels for the detection of biomolecules suchas proteins, antigens, RNA and DNA—particularly in(immuno)assays and time-resolved microscopy—willbe discussed. The use of tryptophan RTP lifetime as astructural probe for proteins will also be considered.

7.2. RTP labels and assays

Fluorescent labels are by far the best choice for thesensitive detection of many classes of compounds, butunder certain conditions, i.e. when strongly autofluo-rescent material is involved, the use of labels basedon long-lived luminescence can be advantageous. Inthis case, improved sensitivity of assays may result,since time-resolved detection can be applied. In thepast, almost exclusively lanthanide labels were usedfor derivatization[126]; more recently, various met-alloporphyrins were also applied. These compoundsnot only exhibit long phosphorescence lifetimes, butalso large Stokes’ shifts, high absorptivities and highphosphorescence quantum yields.

An overview of the photophysical properties ofvarious metalloporphyrins is given inTable 5. ForPt(II)– and Pd(II)–porphyrin ketones, the extinctioncoefficients were ca. 45,000–170,000 M−1 cm−1 andthe phosphorescence quantum yields ca. 0.01–0.12;the emission maxima were significantly red-shiftedtowards the range of ca. 760–800 nm. The pho-tostability of these labels was reported to be sig-nificantly higher than for other porphyrin deriva-tives [127]. Pd(II)– and Lu(III)–tetrabenzoporphyrin(TBP), and Pd(II)–tetraphenyltetrabenzoporphyrin


Table 5Photophysical data for selected phosphorescent porphyrin ketone–metal complexes

Compound Absorption bands RTP quantum yield

B(0,0) (Soret band) Q(0,0)

λmax (nm) ε (M−1 cm−1) λmax (nm) ε (M−1 cm−1)

Pt(II)–OEPKa 398 86200 592 55100 0.12Pt(II)–CPKTEEa 397 82700 592 45700 0.11Pt(II)–CPKb 394 n.m. 588 n.m. 0.08Pd(II)–OEPKa 410 82600 603 53500 0.01Pd(II)–CPKTEE n.m. n.m. n.m. 0.01Pd(II)–CPKb 406 n.m. 599 n.m. 0.008Zn(II)–OEPKa 422 170000 622 52100 <0.0005Pd(II)–TBPc 407 179000 606 105000 0.079Lu(III)–TBPc 426 95500 627 30900 0.035Pd(II)–Ph4TBPc 442 119000 628 51000 0.08

OEPK, octaethylporphine ketone; CPK, coproporphyrin I ketone; CPKTEE, coproporphyrin I ketone tetraethyl ester; TBP, tetrabenzopor-phyrin; Ph4TBP, meso-tetraphenyltetrabenzoporphyrin; n.m., not measured.

a Solvent: chloroform,[127].b Solvent: methanol,[127].c Solvent: DMF,[102].

(Ph4TBP) were also studied[102]. The highest ab-sorptivities and RTP quantum yields were obtainedfor Pd(II)–TBP (Table 5); emission was in the rangeof ca. 770–800 nm.

Pd(II)−coproporphyrins were used in a competi-tive, universal immunoassay for the detection of in-sulin (down to the sub-nM level) and other antigens[128]. With an extinction coefficient of 200,000M−1 cm−1 and a phosphorescence quantum yield,φp,of 0.17 the Pd(II)−coproporphyrins are among themost powerful RTP labels available (cf.Table 5).

Monofunctional p-isothiocyanatophenyl deriva-tives of Pt(II)− and Pd(II)−coproporphyrin I wereused for the covalent labelling of proteins[129]. Theantibody–label conjugates essentially retained the ac-tivity and long-term storage stability of the unboundlabel; labels and antibody–label conjugates were eval-uated in solid-phase immunoassays and found to behighly sensitive (detection down to 1–10 pM) com-pared to many other assays. Finally, dansyl chloridecan be used as an RTP label for amino acids[130].Using SDS micelles, MS-RTP from the dansylatedamino acids was observed for pH 7–10. For deoxy-genation 0.016–0.02 M sodium sulfite was used, while0.03–0.05 M TlNO3 was used as HAP. The authorssuggested that the RTP detection of dansylated aminoacids may be coupled with some separation technique

involving SDS micelles. In a more recent study[67],the potential of dansyl chloride for RTP labelling inhomogeneous solutions, in the presence of high con-centrations of HAP, was addressed (cf.Section 4.1).

Phosphorescent metalloporphyrins can also be usedfor labelling of biomolecules in time-resolved mi-croscopy; the application of time resolution is advanta-geous since many tissues are strongly autofluorescent[131,132]. N-Hydroxysuccinimide ester porphyrinderivatives were synthesized and bound to the antibod-ies avidin, streptavidin, and neutravidin. Glucose andglucose oxidase were added for deoxygenation[131].The streptavidin−porphyrin derivative conjugateswere used to study 28S rRNA in HeLa cells and CD4membranes of lymphocytes[131]. In a further study[132], label−antibody conjugates were used to detectprostate-specific antigen, glucagon, human andragonreceptor, p53 and gluthathione transferase, again inautofluorescent tissues. In addition, metallocopropor-phyrin-labelled dUTPs were synthesized and appliedto the detection of DNA in metaphase spreads. In allthese cases, the contrast was significantly improvedcompared to fluorescence microscopy.

Another interesting study reported on phospho-rescence anisotropy measurements, which were per-formed to monitor the binding of antibodies to aBSA antigen labelled with eosin-5′-isothiocyanate


[133]. Upon binding of the antibody—monoclonalanti-BSA (mouse IgG 2a)—the degree of phospho-rescence anisotropy from the eosin-5′-isothiocyanatelabel was enhanced. In other words, after excita-tion with linearly polarized light some polarizationwas retained during RTP emission because the ro-tational movements were strongly reduced. Thus, adirect immunoassay format was achieved. A (pulsed)dual-channel phosphorimeter was used to monitorboth polarizations simultaneously. Consequently, themeasuring time was shortened and photodegradationproblems were reduced compared to earlier, sequen-tial studies[134]. An advantage of phosphorescenceanisotropy over fluorescence anisotropy is the reducedinterference from the excitation pulses. In addition,when using phosphorescence anisotropy rotationalmovements in the micro-to-millisecond timescale canbe measured, while with fluorescence anisotropy oneis restricted to the nanosecond timescale. Therefore,the phosphorescence anisotropy technique enablesmonitoring of much larger molecules[133].

7.3. RTP lifetime as a structural probe

The potential of tryptophan RTP lifetime measure-ments for probing conformational changes in pro-teins was first recognized in 1985 by Strambini andGonnelli [135]. The authors suggested that the tryp-tophan phosphorescence lifetime is mainly influencedby differences in the rigidity (microviscosity) of theindole environment. Changes in the accessibility ofproteins to molecular oxygen were supposed to playonly a minor role, if any, since proteins in solution arequite permeable to molecular oxygen, independentof their actual conformation. More importantly, inall probing studies considered the dissolved oxygenwas efficiently removed by purging with high-purityargon or nitrogen. Because of the strong dependenceof the tryptophan RTP lifetime on the rigidity of theenvironment, RTP lifetime is a very sensitive probethat can detect subtle conformational changes[136].

As a first example, tryptophan phosphorescencelifetime (Trp 109) and activity measurements wereused to monitor structural changes inEscherichia colialkaline phosphatase (AP) induced by the addition ofseveral denaturants (EDTA, acid, and guanidine·HCl)[137]; RTP lifetimes were reduced upon unfolding ofthe protein. An initial response was obtained within

Fig. 12. Enzymatic activity and RTP intensity of alkaline phos-phatase during incubation in 5 mM EDTA. Figure taken from[137].

seconds after addition of the denaturant, but the to-tal denaturation process took up to several hours(Fig. 12). In general, several binding interactions,including hydrophobic and electrostatic interactions,van der Waals forces, hydrogen bonds, and covalentbonds between cysteins, determine the structure of aprotein. Experiments involving a variety of denatu-rants may therefore help to indicate which interactionsare most important in specific cases. The experimentswith AP indicated a complex behavior which in-volved several slowly interconverting conformationswith different RTP lifetimes and activities. These twomeasured parameters did not correlate in all cases.The phosphorescence decay could best be describedby a lifetime distribution even in the initial state[137]. However, upon unfolding the distribution wasobserved to broaden, which indicates increased het-erogeneity, i.e. the presence of several conformations.

In another study, changes in pH were used to induceconformational changes in the protein azurin, whichwere studied by monitoring the RTP from the trypto-phan residue[138]. Two different, slowly interconvert-ing, conformations were shown to exist, as indicatedby the two RTP lifetime components (417 and 592 ms)observed in the phosphorescence decay curves. Thepre-exponential factors appeared to change with pH(one going up and one going down); the changes inconformation were ascribed to the protonation of theazurin His 35 residue at low pH. Interestingly, thisis not a denaturation study since the conformationalchange at low pH is required for the functioning ofthe protein.


Finally, tryptophan phosphorescence was used toperform structural assessments on the bovine heart mi-tochondrial F1-ATPaseε-subunit, which in vitro canstill hydrolyse ATP[136]. Measurements of the band-width of the (vibrationless) 0–0 band at low tempera-ture, as well as the heterogeneous RTP decay kineticsin fluid solution, indicated the existence of multiple,rather stable conformations of theε-subunit. The ex-posure of theε-subunit to 6 M urea, which is stronglydissociating, did not result in shorter decay times; thisstrongly suggests that the tryptophanyl side-chain wasnot solvent-exposed upon dissociation. By contrast,the conformational changes accompanying the bind-ing of Mg−ATP were clearly indicated by changesin the phosphorescence lifetime distribution—mostnotably in a decrease of the observed lifetimes.


RTP labels are useful probes, especially for the ap-plication of time-resolved detection in immunoassays.In addition, time-resolved microscopy is a promisingtechnique to monitor labelled biomolecules in stronglyautofluorescent tissues. Phosphorescence anisotropyis an interesting technique, since it provides a di-rect assay format for monitoring antigen−antibodybinding. Finally, tryptophan lifetime measurementscan be considered a valuable tool in the structuralanalysis of proteins, enabling the detection of subtleconformational changes.

8. Lanthanide luminescence

8.1. Introduction

A discussion of lanthanide luminescence is relevantbecause long-lived lanthanide luminescence showsstrong similarities with RTP, although the nature ofthe radiative transitions is fundamentally differentfor both luminescence modes. As regards the simi-larities, long luminescence lifetimes are involved inboth types of luminescence, so that time-resolveddetection can be applied simply by using pulsedxenon sources for excitation. Sensitized lanthanideluminescence has the inherent advantage of showingvery large Stokes’ shifts: the difference between theexcitation wavelength of the organic sensitizers and

the emission wavelength of lanthanide luminescencegenerally exceeds 250 nm, which is rather similar tothe Stokes’ shifts typically observed in RTP. More-over, all modes present in RTP (direct, sensitized andquenched mode) are, in principle, also available inlanthanide luminescence.

In contrast with phosphorescence, which is based onemission from the T1 to the S0 state of an organicmolecule, the luminescence from Eu(III) and Tb-(III)—these lanthanides are used almost invariably—involves emission from the5D0 (and, to some extent,5D1) and 5D4 levels to the7F ground-state level.Moreover, these atomic emission bands are verynarrow (a few nanometers) due to the protectednature of the lanthanide f shells. In addition, in asensitizer−lanthanide complex intramolecular energytransfer is fast (kET, ca. 1010 s−1 [139]). Both factorsstrongly reduce the quenching efficiencies for dis-solved molecular oxygen, so that in practice sampledeoxygenation is not required at all. Instead, quench-ing by water due to energy transfer from the excitedlanthanide to the surrounding water molecules—resulting in vibrational excitation of water—plays aprominent role[140].

Although many analytical applications of lan-thanide luminescence involve the use of a separationtechnique, analytical methods based on immunoas-says or involving batch measurements or FIA, havealso been reported. Sensitized lanthanide lumines-cence obtained from Eu(III) or Tb(III) is used mostoften, with the sensitizer being coordinated by thelanthanide ion; occasionally, quenched lanthanide lu-minescence detection is applied. The present overviewis far from exhaustive; it is intended to give an indi-cation of the most important methods based on lan-thanide luminescence, and of its potential in practicalanalysis.

8.2. Stand-alone measurements

Due to the selective nature of sensitized lanthanideluminescence detection (synthetic) mixtures and sam-ples that are not too complex can be analyzed with-out any previous separation, especially when specialscanning techniques are used. For example, second-derivative synchronous scanning fluorimetry wasused for the simultaneous determination of salicylicacid and diflunisal in spiked urine and serum[141];


the analysis results were quite accurate (recovery,87–111%).

Benzoic acid and saccharin were simultaneouslydetermined in soft drinks using a stopped-flow tech-nique (cf.Fig. 4) [142]. One syringe contained Tb(III),Triton X-100, and trioctylphosphine oxine, while theother contained the sample. The Tb(III) luminescencewas enhanced due to inclusion of the Tb(III)−analytecomplex in the mixed aggregate. The extra dimen-sion added (the luminescence curves have characteris-tic slopes and amplitudes) and the inherent selectivityof lanthanide luminescence enabled the direct analy-sis of soft drinks. Glucose, cyclamic acid, caffeine,citric acid and ascorbic acid could be tolerated in a7–100-fold excess. As mentioned by the authors, easyautomation of sensitized lanthanide luminescence de-tection can be achieved in this way.

The enhancement of lanthanide luminescence, dueto ET after binding of Tb(III) ions to the nucleobasesguanine (G), adenine (A), thymine (T) and cytosine(C), can be used for the detection of single-strandedoligonucleotides or single-stranded regions of DNA;double-stranded DNA effects no enhancement, whichhas been attributed to perturbation of the electronicstructure upon base-pair formation[143]. In a recentreport, an approx. 10-fold increase of the luminescenceintensity was obtained for single-stranded oligonu-cleotides (10-mer)[144]. Interestingly, the presenceof single mismatched base-pairs in 10-mer duplexesalso leads to enhancement of Tb(III) luminescence[144]. As is the case with detection of single-strandedoligonucleotides, the enhancement here is due to ETfrom the nucleobases. The largest enhancement wasprovided by GG mismatches followed by GA and CA;low enhancements were observed for TT and GT mis-matches (Table 6).

Detection of higher-ordered sequences in double-stranded DNA could be performed, by using a modi-fied 14-mer single-stranded oligonucleotide with a Tb-(III)-chelating group attached on a flexible hexamethy-lene chain,[143]. In the presence of double-strandedDNA with specific purine tracts, triple-stranded DNAsections were formed—so-called Hoogsteen-typebase-pairing. As a result, the DNA base-sensitized lan-thanide luminescence was quenched two to four-fold,which again was attributed to perturbation of the elec-tronic levels of the DNA bases. Thus, double-strandedoligonuceotides with sections of 14 subsequent T−A

Table 6Influence of single base-pair mismatch on RTP emission intensitywith 25�M Tb3+ and 50�M base

5′-CGCAXTTGCG-3′,3′-GCGTYAACGC-5′

Mismatch Relativeintensity

X Y

A T None 1.0G G GG 9.6A G GA 7.0A C CA 6.7C C CC 6.3C T TC 4.7A A AA 4.5T T TT 3.6G T GT 3.4

Taken from[144].

base-pairs could be detected, using a single-stranded,all-T, 14-mer probe.

Eu(III)−tetracycline chelates were used to detectsingle- as well as double-stranded DNA[145]. Inthis case, tetracycline is the sensitizer both in the ab-sence and presence of DNA and the luminescence isenhanced by ternary complex formation with DNA,most probably due to increased protection againstwater-induced quenching. Fish sperm DNA was suc-cessfully determined in mixtures containing a five-foldexcess of yeast RNA. The method was reported tohave a sensitivity similar to the ethidium bromidemethod, while the selectivity was significantly en-hanced; therefore, the use of Eu(III) chelates maybe preferred over the strongly carcinogenic ethidiumbromide.

8.3. Reversed micelles

As was described inSection 2.4, surfactants mayform reversed micelles in apolar solutions. After ad-dition of a small amount of water, water pools areformed inside the micellar core. The analytical poten-tial of sensitized lanthanide luminescence detectionin AOT reversed micelles, which contained Tb(III) orEu(III) as counterions, was discussed in several papers[146–148]. A strong enhancement of the lanthanideluminescence was achieved, due to fixation of thelanthanide ions at the water/micelle interface, sincethis results in closer proximity of donor (analyte)and acceptor (lanthanide) and, thus, enhances energy


transfer[146]. Since analytes with ionic groups andpolar compounds can penetrate the micellar core moreeffectively than non-polar compounds, the formercan be detected much more sensitively: LODs were10−7 to 10−9 M for polar and 10−5 M for less polarcompounds[147]. The technique was applied bothin stand-alone measurements and in normal-phaseLC for the determination of theophylline in Pri-matene tablets[148]; using the LC method, an LODof 9 ng ml−1 was achieved for standard solutions.Ephedrine, which is also present in Primatene tablets,did not interfere with the detection of theophylline.Therefore, the stand-alone measurements were selec-tive enough to provide accurate results and yielded94–100% of the nominal content—similar to the LCmethod. By contrast, the presence of theobromineinterfered and LC should be used for the detection oftheophylline in theobromine-containing samples.

8.4. Immunoassays and DNA hybridization assays

The long-lived luminescence of lanthanide chelateswas also used in immunoassays and DNA hybridiza-tion assays. Also here, the application of time-resolveddetection resulted in improved detectability. Meyerand Karst[149,150]proposed an enzyme-linked im-munosorbent assay (ELISA) detection scheme, whichutilizes the lanthanide luminescence from ternarylanthanide chelates formed after a reaction catalyzedby horseradish peroxidase (HRP). After enzymaticamplification, HRP catalyzes the dimerization of thesubstrate, 4-hydroxyphenylpropionic acid, after whichthe dimer forms a luminescent ternary complex witha Tb(III)−EDTA chelate; for goat anti-rabbit IgGan LOD of 3 ng ml−1 was achieved. The proposedmethodology was tested in an ELISA for the deter-mination of human anti-gliadin IgA in serum; theresults agreed well with those from ELISAs usingphotometric detection[150].

For the determination of DNA in hybridizationassays, the immobilized target DNA was hybridizedwith a derivatized probe oligonucleotide. After eithera single[151] or a double[152] enzyme amplificationstep, conversion of a suitable substrate, 5′-fluorosalicylphosphate, into the corresponding fluorosalicylicatewas catalyzed by bound AP. After complexation of5′-fluorosalicylate with Tb(III)–EDTA, long-livedlanthanide luminescence was obtained. The assay

with two amplification steps, showed a 10-fold im-proved S/N over the one using a single amplificationstep[152].

8.5. Liquid chromatography

The coupling of sensitized lanthanide lumines-cence detection and LC was first reported by DiBellaet al. [140]. The lanthanides Tb(III) an Eu(III) wereadded in the post-column mode using a second LCpump and a mixing tee followed by a mixing coil.The quenching by water molecules could be reducedby the post-column addition of acetate ions, whichremove water from the first coordination sphere ofthe lanthanides.

The analytes to be detected should be able bothto form a complex with the lanthanide ion and totransfer their excitation energy to it via an intramolec-ular pathway [153,154]. This seriously limits theapplication range. Nevertheless, there are a numberof useful applications, e.g. the analysis of orotate,theophylline, tetracyclines, steroids, nucleic acidsand thiols[153]. A typical example, which serves tohighlight the advantages of the technique, is the de-termination of orotate in urine[155]. For separation,ion-pairing reversed-phase LC was used; Tb(III) wasadded post-column. The method was sensitive (LOD,10 nM) as well as selective: no sample pretreatmentwas necessary, except for a simple pH adjustment,filtration and dilution of the sample.

A disadvantage of post-column addition is theinevitable dilution of the analyte[153,154], but insome cases the use of post-column addition is nec-essary in order to achieve optimum conditions forboth separation and detection (complexation)[153].As an alternative to the pre- or post-column addi-tion of lanthanide ions, the detection of carboxylicacid-containing analytes was performed using a detec-tion cell containing immobilized (partly complexed)Tb(III) [154]. The reported LODs (ca. 2�M) wereabout 10-fold higher than with post-column addi-tion of Tb(III), but it was noted by the authors thatconsiderable enhancement of detectability may beanticipated with an improved set-up.

Interestingly, a pre-column derivatization reactionwith 4-maleimidylsalicylic acid (4-MSA), normallyused to obtain fluorescent products from thiols,was extended to generate long-lived lanthanide


Fig. 13. LC chromatograms ofl-cysteine (10−6 M) derivatized with 4-maleimidylsalicylic acid in 10-fold diluted urine using fluorescencedetection (right) and time-resolved Tb(III) luminescence detection (left). Figure taken from[156].

luminescence[156]. After reaction of the thiols with4-MSA, the fluorescent product was complexed withTb(III) and long-lived luminescence was obtainedwhile the fluorescence was strongly suppressed. Com-pared with fluorescence detection, enhanced selectiv-ity and a 5-fold improved sensitivity were obtainedfor glutathione andl-cysteine (LOD for both com-pounds, 1.5× 10−7 M). The method was successfullyapplied to the determination ofl-cysteine in urinespiked at a concentration of 10−6 M (Fig. 13); apartfrom 10-fold dilution and filtration, no sample pre-treatment was used.

Dynamic quenching of Eu(III) and Tb(III) lumines-cence was also applied as a highly selective detectionmode in FIA and LC[157], with different responsesfor Eu(III) and Tb(III). When using Eu(III), nitrite,sulfite, chromate, Fe(CN)6

3− and Fe(CN)64− couldbe detected with good sensitivity; no response wasobserved for halides, thiocyanate, sulfate and nitrate.When using Tb(III), thiocyanate and Pt(Cl)4

2− could

additionally be monitored, while sulfite could not bemeasured due to precipitation. For nitrite, an LOD of50 nM was achieved in LC.

Finally, ligand exchange was also applied for de-tection in LC: organic phosphates, which are able tocompete with the original ligand, acetylacetone (acac),for the Tb(III) ions, were detected with satisfactoryLODs of 30–70 nM[158].

8.6. Capillary electrophoresis

The combination of sensitized lanthanide lumi-nescence detection and micellar electrokinetic chro-matography (MEKC) was first studied by Milofskyet al. [159], who used the steroids testosterone andprogesterone as test analytes and a HeCd laser forexcitation. Likewise, orotic acid was determined inurine, without any sample pretreatment (apart fromdilution, filtration, and pH adjustment to 6.0)[160].Compared to LC, the run time was reduced more


Fig. 14. Schematical design of a post-column reactor for CE.Figure taken from[139].

than 10-fold to as little as 1.5 min. At the same time,the LOD increased only five-fold (to 50 nM), whichis quite acceptable in view of the optical pathlengthavailable for detection (75�m). In both MEKC stud-ies, Tb(III) was added to the separation buffer; there-fore, the capillary had to be conditioned for ca. 5 hevery week since the lanthanide ions strongly interactwith the negatively charged capillary wall[160].

Such problems can be avoided by applyingpost-column addition of the lanthanide ions. With thisapproach, capillary zone electrophoresis (CZE) wasused to determine catecholamines and related com-pounds in urine[139]; for standard solutions, LODsof about 10−7 M were achieved. In addition to Tb(III),EDTA was added in order to induce the formation ofa stable ternary complex at high pH. The post-columnreactor was filled with a Tb(III)/EDTA solution andcontained a porous tube surrounding the closelyspaced ends of the separation and detection capillar-ies (Fig. 14). The solution was introduced into thedetection capillary by applying a 30 mbar pressure on

both the reactor and the inlet vial (the latter for com-pensation). A similar reactor, but without the poroustubing and with a space of about 30–50�m betweenthe capillaries, was used for the determination of di-flunisal and salicylic acid[161]; the distance betweenthe two capillaries should be carefully controlled inorder to avoid peak broadening. The post-columnreactor solution contained Tb(III) and CTAB; thelatter compound was added to enhance the lumines-cence by compartmentalization and protection againstwater.

Dynamic quenching of (sensitized) lanthanide lu-minescence was also used as a detection mode in CZE[162]. The luminescence signal of the Tb(III)−acaccomplex was effectively quenched by nitrite, whichresulted in LODs of 200 and 3 nM under normal andstacking injection conditions, respectively. Unfortu-nately, the peaks of other ions such as chromate wereseverely broadened. In addition, the surfactants nonyl-and dodecylsulfate were detected by the decrease inluminescence caused by the electrophoretic displace-ment of acac by the non-luminescent surfactants,while EDTA was detected by the ligand-exchangemode; unfortunately, the peaks were also broadenedin the latter case.


Relevant applications of lanthanide luminescencedetection have been discussed here; they range fromcoupling with separation techniques like LC and CEto application in immunoassays and stand-alone mea-surements; high selectivity and analyte detectabilityare the most notable advantages of lanthanide lu-minescence, as proven in several examples involv-ing real-life samples. Interestingly, this lumines-cence mode can also be used for the detection ofdouble-stranded DNA and higher-ordered DNA sec-tions, as well as for single-stranded DNA and singleDNA base-pair mismatches. Since RTP can only beused for the detection of double-stranded DNA (cf.Section 6.6), Therefore, lanthanide luminescence ap-pears to be a more versatile method for the analysis ofDNA than RTP, although the latter technique providesmore selective detection of double-stranded DNA inthe presence of RNA. Finally, long-lived lanthanideluminescence also has potential for application inMIP sensing phases, as was discussed inSection 6.7.


9. Evaluation and future perspectives

In this review, various RTPL techniques were dis-cussed with emphasis on recent developments. Al-though some useful new methods were reported inMS-RTP and CD-RTP, progress in these fields is ratherlimited. In almost all cases, direct RTPL is still appliedwithout coupling to a separation technique. Because ofthe rather high selectivity of RTPL, such an approachcan be used successfully provided that the sample ma-trix is not too complex. Nevertheless, the practicabilityof stand-alone MS-RTP and CD-RTP measurementsfor real-life samples appears to be rather limited. Thisalso applies for the newly developed HAI-RTP tech-nique, which provides a useful addition to MS-RTPand CD-RTP.

The coupling of direct RTP techniques with a sep-aration technique appears to be feasible. However,certain problems have to be solved: in MS-RTP,foaming problems are generated by purging with ni-trogen unless alternative micellar systems are used;as an alternative, chemical deoxygenation can be ap-plied. In CD-RTP, coupling with a separation methodmay be hindered by slow mass transfer, especiallywhen stable ternary complexes are involved. Heavyatom-derivatized, water-soluble CDs may offer a so-lution here. In the case of HAI-RTP, such couplingseems to be simpler since no ‘pseudophase’ is in-volved. However, in this case coupling may be lim-ited to separation techniques that use aqueous-phasebuffers or eluents, because of the negative effectsinduced by several organic solvents. For a full evalu-ation of this effect further study will be required. Inany case, the coupling of direct RTP with separationtechniques will undoubtedly increase its value as atool in analytical chemistry.

For the indirect RTP modes (sensitized andquenched RTP), successful coupling with LC and CEhas been reported and was shown to provide highsensitivity. In this context, it should be noted thatquenched phosphorescence detection provides lowconcentration detection limits in CE even for ana-lytes with poor chromophoric properties. In addition,extremely low absolute detection limits (ca. 1 fmol)were obtained, due to the small capillary dimensions.

RTP optosensing is an interesting and rapidlyexpanding field of research, which—based on manydifferent detection principles—allows for the determi-

nation of various analyte classes as well as for mea-surement of solvent parameters such as temperatureand pH. Also in this case, selectivity (as well as sen-sitivity and sensing phase stability) remains a criticalissue, since no separation is involved. Nevertheless,in several cases, depending on the optosensing prin-ciple and the sample matrix, adequate selectivity wasachieved for reliable analysis. In this context, the useof MIPs for selective binding may become of interest.

The use of labels with long luminescence lifetimes—either providing RTP or lanthanide luminescence—provides clear advantages in immunoassays andmicroscopic monitoring of biomolecules, especiallywhen a strongly fluorescent background is present. Inaddition, tryptophan RTP lifetime measurements havebecome a valuable tool in the structural analysis ofproteins. A more widespread use of such labels andprobes in the near future can therefore be anticipated.

Compared to RTP (sensitized) lanthanide lumines-cence has been applied more widely in LC and CE.It provides enhanced selectivity compared to absorp-tion detection—as was proven in the analysis of sev-eral real-life samples. Its use in immunoassays willundoubtedly continue even though porphyrin RTP la-bels were recently introduced as an interesting alterna-tive. The application of lanthanides for the detection ofsingle-stranded DNA and DNA single mismatches ap-pears promising and conveniently complements RTP,which is able to detect double-stranded DNA only.

It should be realized that most RTP studies dis-cussed in this review do not involve real-life (complex)samples. Consequently, the potential of RTP for selec-tive detection in complex samples (e.g. soil, biota) can-not be assessed yet. Nevertheless, the most promisingRTP methods to handle real-life samples will undoubt-edly involve coupling with a separation technique. Thecoupling of quenched RTP and CE has already beenshown to provide useful separations as well as LODsin the low ng ml−1 range.

References

[1] R.J. Hurtubise, Phosphorimetry: Theory, Instrumentation,and Applications, VCH Publishers, New York, 1990.

[2] R.J. Hurtubise, Anal. Chim. Acta 351 (1997) 1.[3] J.A. Barltrop, J.D. Coyle, Principles of Photochemistry,

Wiley, New York, 1978.[4] N.J. Turro, Molecular Photochemistry, Benjamin, New York,

1967.


[5] J. Kuijt, U.A.Th. Brinkman, C. Gooijer, Anal. Chem. 71(1999) 1384.

[6] J.J. Donkerbroek, A.C. Veltkamp, A.J.J. Praat, C. Gooijer,R.W. Frei, N.H. Velthorst, Appl. Spectrosc. 37 (1983) 188.

[7] J.J. Donkerbroek, C. Gooijer, N.H. Velthorst, R.W. Frei,Anal. Chem. 54 (1982) 891.

[8] G.J. Kavarnos, Fundamentals of Photoinduced ElectronTransfer, VCH Publishers, New York, 1993.

[9] R.A. Baumann, M. Schreurs, C. Gooijer, N.H. Velthorst,R.W. Frei, Can. J. Chem. 65 (1987) 965.

[10] L.J. Cline Love, M. Skrilec, J.G. Habarta, Anal. Chem. 52(1980) 754.

[11] M.E. Dıaz-Garcıa, A. Sanz-Medel, Anal. Chem. 58 (1986)1436.

[12] H.R. Zhang, J. Zhang, Y.S. Wei, W.J. Jin, C.S. Liu, Anal.Chim. Acta 357 (1997) 119.

[13] J.A. Murillo-Pulgarın, A. Alañón-Molina, P. Fernández-López, Anal. Chim. Acta 382 (1999) 77.

[14] I.Rapado-Martınez, R.M. Villanueva-Camañas, M.C. Garcıa-Alvarez-Coque, Analyst 119 (1994) 1093.

[15] T. Pérez-Ruiz, C. Martınez-Lozano, V. Tomás, J. Carpena,J. Pharm. Biomed. 17 (1998) 719.

[16] J.A. Murillo-Pulgarın, L.F. Garcıa-Bermejo, J. Agric. FoodChem. 50 (2002) 1002.

[17] A. Segura-Carretero, C. Cruces-Blanco, F. Alés-Barrero, A.Fernández-Gutiérrez, J. Agric. Food Chem. 46 (1998) 561.

[18] Y.M. Liu, M.R. Fernández de la Campa, M.E. Dıaz-Garcıa,A. Sanz-Medel, Mikrochim. Acta 1 (1991) 53.

[19] Y.M. Liu, J.I. Garcıa-Alonso, M.R. Fernández de la Campa,M.E. Dıaz-Garcıa, A. Sanz-Medel, Mikrochim. Acta 1(1991) 199.

[20] H. Hoshino, M. Suzuki, M. Kan’no, T. Ohmachi, T.Yotsuyanagi, Anal. Chim. Acta 407 (2000) 71.

[21] S. Igarashi, K. Endo, Anal. Chim. Acta 320 (1996) 133.[22] S. Panadero, A. Gómez-Hens, D. Pérez-Bendito, Anal.

Chem. 66 (1994) 919.[23] A. Muñoz de la Peña, A. Espinosa-Mansilla, J.A. Murillo-

Pulgarın, A. Alañón-Molina, P. Fernández-López, Talanta48 (1999) 1061.

[24] S. Panadero, A. Gómez-Hens, D. Pérez-Bendito, Anal. Lett.28 (1995) 1405.

[25] A. Muñoz de la Peña, A. Espinosa-Mansilla, J.A. Murillo-Pulgarın, A. Alañón-Molina, P. Fernández-López, Analyst123 (1998) 2285.

[26] I.Y. Goryacheva, G.V. Melnikov, S.N. Shtykov, J. Anal.Chem. USSR 55 (2000) 874.

[27] J.W. Xie, J.G. Xu, G.Z. Chen, Spectrochim. Acta A 51(1995) 1909.

[28] L.A. Blyshak, I.M. Warner, Anal. Chem. 62 (1990) 1953.[29] I. Soutar, L. Swanson, Analyst 116 (1991) 671.[30] K. Nakashima, M. Tango, K. Araki, K. Takeuchi,

Spectrochim. Acta A 58 (2001) 123.[31] W.Y. Lin, J.H. Huang, J.F. Hsu, Photochem. Photobiol. A

109 (1997) 25.[32] G.R. Ramos, I.M. Khasawneh, M.L. Garcıa-Álvarez-Coque,

J.D. Winefordner, Talanta 35 (1988) 41.

[33] C. Cruces-Blanco, A. Segura-Carretero, A. Fernández-Gutiérrez, Anal. Chim. Acta 318 (1996) 357.

[34] A. Segura-Carretero, C. Cruces-Blanco, A. Fernández-Gutiérrez, Anal. Chim. Acta 329 (1996) 165.

[35] A. Segura-Carretero, C. Cruces-Blanco, A. Fernández-Gutiérrez, Anal. Chim. Acta 353 (1997) 337.

[36] W.J. Jin, Y.S. Wei, W.S. Duan, C.S. Liu, Anal. Chim. Acta287 (1994) 95.

[37] A. Segura-Carretero, C. Cruces-Blanco, A. Fernández-Gutiérrez, Anal. Sci. 12 (1996) 653.

[38] A. Segura-Carretero, C. Cruces-Blanco, A. Fernández-Gutiérrez, Analyst 122 (1997) 925.

[39] N.J. Turro, J.D. Bolt, Y. Kuroda, I. Tabushi, Photochem.Photobiol. 35 (1982) 69.

[40] N.J. Turro, T. Okubo, C.J. Chung, J. Am. Chem. Soc. 104(1982) 1789.

[41] A. Muñoz de la Peña, M.C. Mahedero, A. Bautista-Sánchez,Analysis 28 (2000) 670.

[42] A. Ponce, P.A. Wong, J.J. Way, D.G. Nocera, J. Phys. Chem.97 (1993) 11137.

[43] M. Santos, G.M. Escandar, Appl. Spectrosc. 55 (2001) 1483.[44] G.M. Escandar, A. Muñoz de la Peña, Appl. Spectrosc. 55

(2001) 496.[45] X.Z. Du, Y. Zhang, Y.B. Jiang, X.Z. Huang, G.Z. Chen,

Spectrochim. Acta A 53 (1997) 671.[46] X.-Z. Du, Y. Zhang, Y.-B. Jiang, L.-R. Lin, X.-Z. Huang,

G.-Z. Chen, Photochem. Photobiol. A 112 (1998) 53.[47] C.H. Evans, M. Partyka, J. Van Stam, J. Incl. Phenom. 38

(2000) 381.[48] S. Scypinski, L.J. Cline Love, Anal. Chem. 56 (1984) 322.[49] S. Scypinski, L.J. Cline Love, Anal. Chem. 56 (1984) 331.[50] M.A. González, M.H. López, Analyst 123 (1998) 2217.[51] M. Hernández-López, M.A. González, M.I. López-Molina,

Talanta 49 (1999) 679.[52] A. Segura-Carretero, C. Cruces-Blanco, A. Fernández-

Gutiérrez, A. Muñoz de la Peña, Appl. Spectrosc. 52 (1998)420.

[53] F.J. DeLuccia, L.J. Cline Love, Talanta 32 (1985) 665.[54] V.B. Nazarov, V.I. Gerko, M.V. Alfimov, Russ. Chem. Bull.

45 (1996) 2109.[55] V.B. Nazarov, V.I. Gerko, M.V. Alfimov, Russ. Chem. Bull.

46 (1997) 1386.[56] W. Jin, Y. Wei, A. Xu, C. Liu, Spectrochim. Acta A 50

(1994) 1769.[57] G.M. Escandar, M.A. Boldrini, Talanta 53 (2001) 851.[58] R. Weinberger, L.J. Cline Love, Spectrochim. Acta A 40

(1984) 49.[59] S. Hamai, J. Phys. Chem. 99 (1995) 12109.[60] S. Hamai, J. Chem. Soc., Chem. Commun. (1994) 2243.[61] A. Muñoz de la Peña, M.P. Rodrıquez, G.M. Escandar,

Talanta 51 (2000) 949.[62] R.E. Brewster, M.J. Kidd, M.D. Schuh, Chem. Commun.

(2001) 1134.[63] R.A. Femia, L.J. Cline Love, J. Phys. Chem. 89 (1985) 1897.[64] S. Hamai, N. Mononobe, Photochem. Photobiol. 91 (1995)

217.[65] S. Hamai, T. Kudou, Photochem. Photobiol. A 113 (1998)

135.


[66] M.J. Sepaniak, C.L. Copper, K.W. Whitaker, V.C. Anigbogu,Anal. Chem. 67 (1995) 2037.

[67] L. Li, Y. Chen, Y. Zhao, A. Tong, Anal. Chim. Acta 341(1997) 241.

[68] A. Segura-Carretero, C. Cruces-Blanco, B. Cañabate-Dıaz,J.F. Fernández-Sánchez, A. Fernández-Gutiérrez, Anal.Chim. Acta 417 (2000) 19.

[69] L.D. Li, J. Zhao, Y.G. Wu, A.J. Tong, Talanta 46 (1998)1147.

[70] A. Segura-Carretero, C. Cruces-Blanco, J.F. Fernández-Sánchez, B. Cañabate-Dıaz, A. Fernández-Gutiérrez, J.Agric. Food Chem. 48 (2000) 4453.

[71] C. Cruces-Blanco, A. Segura-Carretero, M.I. Ramırez-Garcıa, A. Fernández-Gutiérrez, Int. J. Environ. Anal. Chem.75 (1999) 377.

[72] A. Segura-Carretero, C. Cruces-Blanco, M.I. Ramırez-Garcıa, B. Cañabate-Dıaz, A. Fernández-Gutiérrez, Talanta50 (1999) 401.

[73] C. Cruces-Blanco, A. Segura-Carretero, J.F. Fernández-Sánchez, A. Fernández-Gutiérrez, J. Pharm. Biomed. 23(2000) 845.

[74] L.-D. Li, X.-K. Chen, L. Mou, W.-Q. Long, A.-J. Tong,Anal. Chim. Acta 424 (2000) 177.

[75] A. Segura-Carretero, C. Cruces-Blanco, A. Fernández-Gutiérrez, J. Agric. Food Chem. 46 (1998) 3683.

[76] A. Fernández-Gutiérrez, A. Segura-Carretero, B. Cañabate-Dıaz, C. Cruces-Blanco, Appl. Spectrosc. 53 (1999) 741.

[77] A. Segura-Carretero, C. Cruces-Blanco, B. Cañabate-Dıaz,A. Fernández-Gutiérrez, Anal. Chim. Acta 361 (1998) 217.

[78] J.A. Murillo-Pulgarın, A. Alañón-Molina, Fresenius J. Anal.Chem. 368 (2000) 505.

[79] J.A. Murillo-Pulgarın, A. Alañón-Molina, M.T. Alañón-Pardo, Analyst 126 (2001) 234.

[80] R. Weinberger, P. Yarmchuk, L.J. Cline Love, Anal. Chem.54 (1982) 1552.

[81] J.J. Donkerbroek, C. Gooijer, N.H. Velthorst, R.W. Frei,Chromatographia 15 (1982) 218.

[82] E. Blanco-González, R.A. Baumann, C. Gooijer, N.H.Velthorst, R.W. Frei, Chemosphere 16 (1987) 1123.

[83] J.J. Donkerbroek, Liquid Phase Room Temperature Phosphorescence for Continuous-Flow Detection, Ph.D. Thesis, VrijeUniversiteit, Amsterdam, 1983 (Chapter 7).

[84] J. Kuijt, D. Arraez-Roman, F. Ariese, U.A.Th. Brinkman,C. Gooijer, Anal. Chem. 74 (2002) 5139.

[85] J. Kuijt, F. Ariese, U.A.Th. Brinkman, C. Gooijer,Electrophoresis 24 (2003) 1193–1199.

[86] J. Kuijt, U.A.Th. Brinkman, C. Gooijer, Electrophoresis 21(2000) 1305.

[87] J. Kuijt, R. van Teylingen, T. Nijbacker, F. Ariese, U.A.Th.Brinkman, C. Gooijer, Anal. Chem. 73 (2001) 5026.

[88] A. Sanz-Medel, Anal. Chim. Acta 283 (1993) 367.[89] J.M. Costa-Fernández, A. Sanz-Medel, Quım. Anal. 19

(2000) 189.[90] D. Kriz, O. Ramstrom, K. Mosbach, Anal. Chem. 69 (1997)

345A.[91] K. Haupt, Analyst 126 (2001) 747.[92] P. Turkewitsch, B. Wandelt, G.D. Darling, W.S. Powell,

Anal. Chem. 70 (1998) 2025.

[93] J. Matsui, M. Higashi, T. Takeuchi, J. Am. Chem. Soc. 122(2000) 5218.

[94] A.L. Jenkins, O.M. Uy, G.M. Murray, Anal. Chem. 71 (1999)373.

[95] J.M. Demas, B.A. DeGraf, P.B. Coleman, Anal. Chem. 71(1999) 793A.

[96] L.N. Torres, I.P. Torres Filho, Braz. J. Med. Biol. Res. 34(2001) 129.

[97] G. Liebsch, I. Klimant, B. Frank, G. Holst, O.S. Wolfbeis,Appl. Spectrosc. 54 (2000) 548.

[98] D.B. Papkovsky, J. Olah, I.N. Kurochkin, Sens. Actuators,B: Chem. 11 (1993) 525.

[99] B.W. McIlroy, A. Curnow, G. Buonaccorsi, M.A. Scott,S.G. Bown, A.J. MacRobert, J. Photochem. Photobiol. B 43(1998) 47.

[100] I.P. Torres Filho, M. Intaglietta, Am. J. Physiol. 265 (1993)H1434.

[101] D.B. Papkovsky, G.V. Ponomarev, W. Trettnak, P. O’Leary,Anal. Chem. 67 (1995) 4112.

[102] S.A. Vinogradov, D.F. Wilson, J. Chem. Soc., Perkin Trans.2 (1995) 103.

[103] Y.-M. Liu, R. Pereiro-Garcıa, M.J. Valencia-González, M.E.Dıaz-Garcıa, A. Sanz-Medel, Anal. Chem. 66 (1994) 836.

[104] J.M. Costa-Fernández, M.E. Dıaz-Garcıa, A. Sanz-Medel,Anal. Chim. Acta 360 (1998) 17.

[105] J. Dıaz-Garcıa, J.M. Costa-Fernández, N. Bordel-Garcıa, A.Sanz-Medel, Anal. Chim. Acta 429 (2001) 55.

[106] I. Lähdesmäki, L.D. Scampavia, C. Beeson, J. Ruzicka,Anal. Chem. 71 (1999) 5248.

[107] D.B. Papkovsky, T. O’Riordan, A. Soini, Biochem. Soc.Trans. 28 (2000) 74.

[108] B. San Vicente de la Riva, J.M. Costa-Fernández, R. Pereiro,A. Sanz-Medel, Anal. Chim. Acta 359 (1999) 1.

[109] R. Pereiro-Garcıa, Y. Ming Liu, M.E. Dıaz-Garcıa, A.Sanz-Medel, Anal. Chem. 63 (1991) 1759.

[110] J. Li, Z. Zhang, Anal. Lett. 27 (1994) 2769.[111] Y.M. Liu, R. Pereiro-Garcıa, M.E. Dıaz-Garcıa, A.

Sanz-Medel, Anal. Chim. Acta 255 (1991) 245.[112] J. Lu, Z. Zhang, Analyst 120 (1995) 2585.[113] Z. Gong, Z. Zhang, Mikrochim. Acta 126 (1997) 117.[114] Z. Gong, Z. Zhang, A. Zhang, Anal. Lett. 29 (1996) 515.[115] F. Alava-Moreno, M.E. Dıaz-Garcıa, A. Sanz-Medel, Anal.

Chim. Acta 281 (1993) 637.[116] F. Alava-Moreno, M.J. Valencia-González, M.E. Dıaz-

Garcıa, Analyst 123 (1998) 151.[117] D.B. Papkovsky, A.P. Savitsky, A.I. Yaropolov, J. Anal.

Chem. USSR 45 (1990) 1045.[118] M.J. Valencia-González, Y.M. Liu, M.E. Dıaz-Garcıa, A.

Sanz-Medel, Anal. Chim. Acta 283 (1993) 439.[119] A.N. Ovchinnikov, V.I. Ogurtsov, W. Trettnak, D.B.

Papkovsky, Anal. Lett. 32 (1999) 701.[120] D.B. Papkovsky, A.L. Ghindilis, I.N. Kurochkin, Anal. Lett.

26 (1993) 1505.[121] M.J. Valencia-González, M.E. Dıaz-Garcıa, Anal. Chem. 66

(1994) 2726.[122] R. Badıa, M.E. Dıaz-Garcıa, Anal. Chim. Acta 371 (1998)

73.


[123] M. Rosa-Fernández, M.J. Valencia-González, M.E.Dıaz-Garcıa, Anal. Chem. 69 (1997) 2406.

[124] W.J. Jin, J.M. Costa-Fernández, A. Sanz-Medel, Anal. Chim.Acta 431 (2001) 1.

[125] B. San Vicente-de la Riva, J.M. Costa-Fernández, W.J. Jin,R. Pereiro, A. Sanz-Medel, Anal. Chim. Acta 455 (2002)179.

[126] E. Soini, T. Lovgren, Crit. Rev. Anal. Chem. 18 (1987) 105.[127] D.B. Papkovsky, G.V. Ponomarev, O.S. Wolfbeis, Spectro-

chim. Acta A 52 (1996) 1629.[128] E.Y. Mantrova, M.V. Demcheva, A.P. Savitsky, Anal.

Biochem. 219 (1994) 109.[129] T.C. O’Riordan, A.E. Soini, D.B. Papkovsky, Anal.

Biochem. 290 (2001) 366.[130] L.-D. Li, W.-G. Huang, Anal. Chim. Acta 312 (1995) 345.[131] R.R. de Haas, R.P.M. van Gijlswijk, E.B. van der Tol,

H.J.M.A.A. Zijlmans, T. Bakker-Schut, J. Bonnet, N.P.Verwoerdt, H.J. Tanke, J. Histochem. Cytochem. 45 (1997)1279.

[132] R.R. de Haas, R.P.M. van Gijlswijk, E.B. van der Tol, J.Veuskens, H.E. van Gijssel, R.B. Tijdens, J. Bonnet, N.P.Verwoerdt, H.J. Tanke, J. Histochem. Cytochem. 47 (1999)183.

[133] L. Yang, D. McStay, A.J. Rogers, P.J. Quinn, Meas. Sci.Technol. 5 (1994) 1096.

[134] R.D. Ludescher, D.D. Thomas, Biochemistry 27 (1988)3343.

[135] G.B. Strambini, M. Gonnelli, Chem. Phys. Lett. 115 (1985)196.

[136] G. Solaini, A. Baracca, G.P. Castelli, G.B. Strambini, Eur.J. Biochem. 214 (1993) 729.

[137] J.V. Mersol, D.G. Steel, A. Gafni, Biophys. Chem. 48 (1993)281.

[138] J.E. Hansen, D.G. Steel, A. Gafni, Biophys. J. 71 (1996)2138.

[139] R. Zhu, W.Th. Kok, Anal. Chem. 69 (1997) 4010.[140] E.E. DiBella, J.B. Weisman, M.J. Joseph, J.R. Schultz, Th.J.

Wenzel, J. Chromatogr. 328 (1985) 101.[141] E.S. Lianidou, P.C. Ioannou, C.K. Polydorou, C.E.

Efstathiou, Anal. Chim. Acta 320 (1996) 107.

[142] M.P. Aguilar-Caballos, A. Gómez-Hens, D. Pérez-Bendito,Analyst 124 (1999) 1079.

[143] S. Sueda, T. Ihara, B. Juskowiak, M. Takagi, Anal. Chim.Acta 365 (1998) 27.

[144] P.K.-L. Fu, C. Turro, J. Am. Chem. Soc. 121 (1999) 1.[145] Y.-X. Ci, Y.-Z. Li, X.-J. Liu, Anal. Chem. 67 (1995)

1785.[146] A.G. Mwalupindi, L.A. Blyshak, T.T. Ndou, I.M. Warner,

Anal. Chem. 63 (1991) 1328.[147] A.G. Mwalupindi, T.T. Ndou, I.M. Warner, Anal. Chem. 64

(1992) 1840.[148] A.G. Mwalupindi, I.M. Warner, Anal. Chim. Acta 306 (1995)

49.[149] J. Meyer, U. Karst, Analyst 125 (2000) 1537.[150] J. Meyer, U. Karst, Analyst 126 (2001) 175.[151] P. Radovich, S. Bortolin, Th.K. Christopoulos, Anal. Chem.

67 (1995) 2644.[152] P.C. Ioannou, Th.K. Christopoulos, Anal. Chem. 70 (1998)

698.[153] A. Rieutord, P. Prognon, F. Brion, G. Mahuzier, Analyst

122 (1997) 59.[154] Th.J. Wenzel, R. Evertsen, B.E. Perrins, Th.B. Light Jr.,

A.C. Bean, Anal. Chem. 70 (1998) 2085.[155] M. Schreurs, J.P.C. Vissers, C. Gooijer, N.H. Velthorst, Anal.

Chim. Acta 262 (1992) 201.[156] M. Schreurs, C. Gooijer, N.H. Velthorst, Anal. Chem. 62

(1990) 2051.[157] R.A. Baumann, D.A. Kamminga, H. Derlagen, C. Gooijer,

N.H. Velthorst, R.W. Frei, J. Chromatogr. 439 (1988)165.

[158] M. Schreurs, Detection in Column Liquid ChromatographyBased on Lanthanide Ion Luminescence, Ph.D. Thesis, VrijeUniversiteit, Amsterdam, 1992 (Chapter 7).

[159] R.E. Milofsky, M.G. Malmberg, J.M. Smith, J. High Res.Chromatogr. 17 (1994) 731.

[160] R. Milofsky, S. Spaeth, Chromatographia 42 (1996) 12.[161] R. Milofsky, E. Bauer, J. High Res. Chromatogr. 20 (1997)

638.[162] M.W.F. Nielen, J. Chromatogr. 608 (1992) 85.

Documents

Room temperature phosphorescence in the liquid state as a tool in