1-s2.0-S003991401400914X-main

7172019 1-s20-S003991401400914X-main

httpslidepdfcomreaderfull1-s20-s003991401400914x-main 19

Green analytical determination of emerging pollutantsin environmental waters using excitationndashemissionphotoinduced 1047298uorescence data and multivariate calibration

Mariacutea del Carmen Hurtado-Saacutenchez a Valeria A Lozano b Mariacutea Isabel Rodriacuteguez-Caacuteceres a Isabel Duraacuten-Meraacutes a Graciela M Escandar bn

a Department of Analytical Chemistry University of Extremadura 06006 Badajoz Spainb Departamento de Quiacutemica Analiacutetica Facultad de Ciencias Bioquiacutemicas y Farmaceacuteuticas Universidad Nacional de Rosario Instituto de Quiacutemica de Rosario

(IQUIR-CONICET) Suipacha 531 Rosario S2002LRK Argentina

a r t i c l e i n f o

Article history

Received 20 October 2014

Accepted 11 November 2014Available online 20 November 2014

Keywords

Emerging pollutants

Photoinduced 1047298uorescenceUnfolded partial least-squaresresidual

bilinearization

Water samples

a b s t r a c t

An eco-friendly strategy for the simultaneous quanti1047297cation of three emerging pharmaceuticalcontaminants is presented The proposed analytical method which involves photochemically induced1047298uorescence matrix data combined with second-order chemometric analysis was used for thedetermination of carbamazepine o1047298oxacin and piroxicam in water samples of different complexity

without the need of chromatographic separation Excitationndashemission photoinduced 1047298uorescencematrices were obtained after UV irradiation and processed with second-order algorithms Only one of the tested algorithms was able to overcome the strong spectral overlapping among the studied

pollutants and allowed their successful quantitation in very interferent media The method sensitivity

in super1047297cial and underground water samples was enhanced by a simple solid-phase extraction withC18 membranes which was successful for the extractionpreconcentration of the pollutants at tracelevels Detection limits in preconcentrated (1125) real water samples ranged from 004 to 03 ng mL 1

Relative prediction errors around 10 were achieved The proposed strategy is signi1047297cantly simpler and

greener than liquid chromatography-mass spectrometry methods without compromising the analyticalquality of the resultsamp 2014 Elsevier BV All rights reserved

1 Introduction

Emerging pollutants are compounds not currently covered byexisting water-quality regulations representing potential threats toecosystems and human health because of their toxic effects [12]They do not need to persist to negatively affect the exposed organ-

isms since their introduction into the environment is continuousespecially those belonging to the pharmaceutical group [1ndash7]

Pharmaceutically active compounds used in both human and vete-

rinary medicine are excreted via feces and urine partly transformedinto glucuronides and sulphates or even unchanged and are suspectedto enter aquatic bodies through the ef 1047298uents of sewage treatmentplants [7ndash11] Therefore continuous efforts are devoted to developappropriate methods for their monitoring and quanti1047297cation in natural

samplesAlthough liquid chromatographyndashmass spectrometry (LCndashMS) is

one of the most commonly applied methods for the determination of pharmaceutical compounds and their degradation products in the

aquatic environment [21213] greener methodologies ie withoutseparations and clean up steps and minimizing the use of organicsolvents are very welcome [14]

Contents lists available at ScienceDirect

journal homepage wwwelseviercomlocatetalanta

Talanta

httpdxdoiorg101016jtalanta201411022

0039-9140amp 2014 Elsevier BV All rights reserved

Abbreviations A antibiotics CBZ carbamazepine CLC capillary liquid chroma-

tography DAD diode array detection DICLO diclofenac DVB divinylbenzene DWdrinking water EC electrophoresis capillary EEPIFM excitationndashemission photo-

induced 1047298uorescence matrix EJCR elliptical joint con1047297dence region EP emerging

pollutants EW environmental water FLU 1047298ufenamic acid GC gas chromatogra-phy HLB hydrophilic-lipophilic balance Horm hormones IBU ibuprofen LC

liquid chromatrography LIF laser induced 1047298uorescence detection LOD limit of

detection LOQ limit of quanti1047297cation MCR-ALS multivariate curve resolution-

alternating least-squares MPs surface-funcionalized magnetic particles MS mass

spectrometry MSMS tandem mass spectrometry MWCN multi-walled carbon

nanotubes MW mineral water NSAI non-steroidal anti-in1047298ammatory OFL

o1047298oxacin OP organic pollutants Pharm pharmaceuticals PARAFAC parallel factor

analysis PIF photoinduced 1047298uorescence PDS polydimethylsiloxane PX pirox-

icam QqLIT quadrupole linear ion trap tandem mass spectrometry QTOF hybrid

quadrupole time-of-1047298ight REC recovery REP relative error of prediction RMSEProot-mean-square error of prediction RSD relative standard deviation RSW

reservoir water RW river water SAL salicylic acid SPE solid-phase extraction

SPME solid-phase microextraction SW sea water TF thin 1047297lm TOF electrospray

time-of-1047298ight TW tap water UPLC ultra-high-performance liquid chromatogra-

phy U-PLSRBL unfolded-partial least-squares with residual bilinearization UV

ultraviolet detection UW underground water WW Wastewater WWE waste-

water ef 1047298uent WWI wastewater in1047298uentn Corresponding author Telfax thorn54 341 4372704

E-mail address escandariquir-conicetgovar (GM Escandar)

Talanta 134 (2015) 215ndash223

7172019 1-s20-S003991401400914X-main


In the present work three emerging pollutants representative of

different groups of therapeutic drugs were investigated the antic-onvulsant carbamazepine (CBZ) the antibacterial 1047298uoroquinoloneo1047298oxacin (OFL) and the non-steroidal anti-in1047298ammatory piroxicam(PX) (Scheme 1) because they are frequently found in environmen-

tal waters They display photo-induced 1047298uorescence (PIF) upon UV irradiation which could allow their quanti1047297cation Relatively fewmolecules are 1047298uorescent and 1047298uorescent photoproducts are even

fewer or parent compounds are photodegradated after UV irradia-tion This may lead to the erroneous conclusion that PIF-based

methods are free from interferences However as presently demon-strated in multicomponent systems the probability of the occur-

rence of interferences signi1047297cantly increases and in principle clean-up and separation procedures are almost unavoidable

Recently our research group quanti1047297ed CBZ as a single analyte inenvironmental waters using the PIF signals after UV irradiation of

acidic solutions in a simple laboratory-constructed reactor [15] Thelack of selectivity was overcome by the second-order advantage of multi-way calibration [16] and pollutant was quantitated in thepresence of unknown sample constituents Second-order data were

obtained as excitationndashemission photoinduced 1047298uorescence matrices(EEPIFMs) and processed by different algorithms although successfulresults were obtained with multivariate curve resolution-alternatingleast-squares (MCR-ALS) [17]

The critical difference of the present report with the earlier workis that the simultaneous resolution of three usual emerging con-taminants which strongly overlap their PIF spectra is presentlyintended with the concomitant change in both data analysis and

results interpretation Further the determinations are performed insolutions containing the analytes and additional pharmaceuticalssuch as ibuprofen (IBU) diclofenac (DICLO) salicylic acid (SAL) and

1047298ufenamic acid (FLU) (Scheme 1) The latter are profusely employed

in our geographical region and may thus be present in real watersand showed 1047298uorescence signals (either in native form or from theirphotoproducts) which signi1047297cantly overlap those of the analytes

Three chemometric algorithms achieving the second-order advan-tage ie parallel factor analysis (PARAFAC) [18] MCR-ALS and

unfolded partial least-squaresresidual bilinearization (U-PLSRBL)

[1920] were applied to process the EEPIFMs Noticeable differences

in the prediction capabilities of the employed algorithms were found

and discussedTo the best of our knowledge it is the 1047297rst time that the selectivity

offered by the chemometric analysis is evaluated for the simulta-neous determination of several analytes using EEPIFMs in very

interfering media The feasibility of determining the three emergentpollutants in real water samples using sustainable resources isdemonstrated

2 Experimental

21 Reagents and solutions

CBZ OFL and PX were purchased from Sigma (St Louis MOUSA) Methanol (MeOH) formic acid and hydrochloric acid (HCl)were obtained from Merck (Darmstadt Germany) IBU DICLO SAL and FLU were of analytical grade and were used as received Stock

standard solutions of individual analytes (4040 mg mL 1 CBZ4200 mg mL 1 PX and 5100 mg mL 1 OFL) were prepared bydissolving an appropriate amount of each compound in methanol

and stored at 4 1C Working analyte solutions of 20 mg mL 1 weredaily prepared by dilution of stock standard solutions in ultrapurewater Ultra pure Milli-Q water was used throughout the work

22 Instrumentation

Fluorescence measurements were performed on an Aminco

Bowman (Rochester NY USA) Series 2 luminescence spectro-photometer equipped with a 150 W xenon lamp EEPIFMs weremeasured in the ranges 246ndash333 nm (each 3 nm excitation) and380ndash480 nm (each 1 nm emission) leading to 29100 matrices

Excitation and emission slit widths were of 8 nm using 100 cmquartz cells The photomultiplier tube sensitivity was 600 V andthe cell temperature was regulated at 20 1C using a thermostaticbath (Cole-Parmer IL USA) EEPIFMs were saved and transferred

to a PC for subsequent chemometric analysisFor the reference chromatographic analysis see Supplementary

material

Scheme 1 Structures of carbamazepine (CBZ) o1047298oxacin (OFL) piroxicam (PX) ibuprofen (IBU) diclofenac (DICLO) salicylic acid (SAL) and 1047298ufenamic acid (FLU)

MdC Hurtado-Saacutenchez et al Talanta 134 (2015) 215ndash 223216

7172019 1-s20-S003991401400914X-main


23 Chemometric algorithms and software

For a brief theoretical description of the applied algorithms seeSupplementary material All routines are written in MATLAB 710

[21] and implemented using the graphical interface MVC2 [22]available on the Internet [23] Design Expert 60 (Stat-Ease Inc)was used for the experimental design

24 Calibration validation and test samples

A calibration set was built with a central composite design in the

concentration ranges between 00 and 60 ng mL 1 for all analytes(Table 1) The corresponding volumes of the aqueous standardsolutions of each analyte were transferred into 200 mL volumetric

1047298asks and 2 mol L 1 HCl was added to the mark These solutions

were transferred to a 10 cm quartz cell and irradiated during20 min in a laboratory-constructed reactor described in a previous

work [15] Finally solutions were cooled to 20 1C and their EEPIFMswere recorded in the conditions described in Section 22

A set of 15 validation samples was prepared and processed in asimilar way having analyte concentrations different from thecalibration ones and selected at random from the correspondingcalibration ranges

With the purpose of evaluating the method in the presence of the interfering pollutants IBU DICLO SAL and FLU which have

1047298uorescence signals (either native or photoinduced) overlappedwith those for the analytes 15 samples were prepared containing

random analyte concentrations in the range 0ndash60 ng mL 1 and highinterferent concentrations 1000ndash3000 100ndash300 3000ndash8000 and1000ndash5000 ng mL 1 (IBU DICLO SAL and FLU respectively) Sincethe highest analyte concentration was about 60 ng L 1 interferents

were between 2 and 130 times more concentrated

25 Water samples

CBZ OFL and PX were analyzed in real waters including river(Paranaacute River Argentina) underground (Funes City and Santa RosaCity Argentina) and tap water (Venado Tuerto City Santa FeArgentina) They were prepared by spiking them with the analytes

at two different concentrations between 008 and 14 ng mL 1 Allsamples were sequentially 1047297ltered through paper and a 20 mmnylon membrane to remove suspended solids To improve thesensitivity most samples were subjected to solid-phase extraction

(SPE) with C18 disks Each disk was previously conditioned with05 mL of MeOH and 1 mL of ultrapure water Aliquots of either

100 mL (analyte concentrations 402 ng mL 1

) or 250 mL (analyte

concentrationso02 ng mL 1) were passed through the disks undervacuum with a 1047298ow rate of 10 mL min1 No pre-concentration wasapplied for concentrations larger than 6 ng mL 1

After elution of the retained organic compounds with 500 mL of

MeOH the extract was collected in a 200 mL volumetric 1047298ask thesolvent was evaporated with nitrogen and the residue wasreconstituted with 2 mol L 1 HCl until the mark This impliespre-concentration degrees of 150 or 1125 depending on the

sample volume Finally the samples were subjected to the proce-dure described above and the analyte concentrations were esti-mated using second-order multivariate calibration

Aliquots of the investigated samples were analyzed by LCndashMS

A similar SPE procedure was applied but after elution of theretained organic compounds with 500 mL of MeOH the extract wascollected in a 200 mL volumetric 1047298ask reconstituted with ultrapure Milli-Q water until the mark and injected in the chromato-

graphic system

3 Results and discussion

31 Preliminary studies

As previously reported [15] CBZ is not 1047298uorescent but emits

1047298uorescence upon UV irradiation in acid media with excitationand emission maxima at 308 and 410 nm respectively ( Fig1A) To

obtain the largest signals optimal working conditions were foundto be 2 mol L 1 HCl and 20 min irradiation time with two 4 Wgermicide tubes separated by 6 cm from each other

On the other hand OFL is natively 1047298uorescent (excitation 290 nmemission 500 nm Fig 1B) When OFL is irradiated in the above CBZconditions 1047298uorescence is observed (excitation 252 nm emission435 nm) ascribed to 1047298uorescent photoproducts

Finally in the case of PX intense 1047298uorescent signals are detectedonly at pHo2 [24] Under UV irradiation in 2 mol L 1 HCl the PX

photoproducts display excitation and emission maxima at 294 nmand 372 nm respectively (Fig 1C) This can be mainly ascribed to

2-aminopyridine which exhibits maxima near the above values [25]Linear relationships between the original analyte concentra-

tions and the obtained 1047298uorescence intensities were corroboratedAmong the three studied analytes CBZ shows the lowest signals

(Fig 1) and consequently the experimental conditions for thequantitative analyses were adjusted in order to optimize the CBZsignals These conditions were indicated above and were main-tained in the subsequent experiments

32 Quantitative analysis

Fig 2A shows the normalized 1047298uorescence spectra for the CBZOFL and PX photoproducts obtained upon UV irradiation under theemployed working conditions It is clear that overlapping occursamong both the excitation and emission spectra which hinders

their direct determination though zeroth-order calibration Theselectivity situation becomes more serious if other 1047298uorescentpollutants are also present (Fig 2B) Therefore with the purpose of overcoming this problem avoiding separation steps second-order

calibration of EEPIFMs and applying algorithms achieving thesecond-order advantage [16] was intended As already indicatedthis advantage implies analyte quantitation in the presence of unsuspected constituents in samples avoiding the requirement of

either interference removal as in zeroth-order calibration or theconstruction of a large and diverse calibration set as in 1047297rst-order

calibration

Table 1

Calibration samples provided by a central composite design

Sample CBZa OFL a PXa

1 300 00 3002 300 600 300

3 300 300 00

4 300 300 600

5 00 300 300

6 600 300 3007 122 122 122

8 122 478 122

9 122 122 478

10 122 478 478

11 478 122 12212 478 478 122

13 478 122 478

14 478 478 478

15 300 300 300

a All concentrations are given in ng mL 1

MdC Hurtado-Saacutenchez et al Talanta 134 (2015) 215ndash 223 217

7172019 1-s20-S003991401400914X-main


321 Calibration and validation samplesAfter UV irradiation EEPIFMs were recorded for the calibration

and validation samples (Fig 3A) where only CBZ PX and OFL arepresent and subjected to chemometric analysis A set of EEPIFMs

can be arranged as a three-way array which in general complies

with the trilinearity conditions [26] and therefore the algorithmof choice for data processing should be PARAFAC [27]

The selection of the optimum number of PARAFAC components

was performed according to Supplementary material This numberwas 4 in validation samples which can be justi1047297ed by the presenceof three analytes and a background signal

Fig 4A illustrates the predicted analyte concentrations when

PARAFAC was applied to the validation set While the predictions forboth OFL and PX are in good agreement with the nominal values theresults for CBZ are poor This conclusion is reinforced by the elliptical

join con1047297dence region (EJCR) test [28] which computes the joint

con1047297dence interval for the intercept and the slope of the found vsnominal concentration plot and checks if the ideal values of 0and 1 are within the ellipse Both OFL and PX comply with the testunlike CBZ The poor PARAFAC recoveries for CBZ may be ascribed to

lack of selectivity ie signi1047297cant spectral overlapping between weakCBZ signals and those for PX (Fig 2A)

MCR-ALS which proved to be the best algorithm for determiningCBZ in natural waters by a similar EEPIFM approach [15] was also

applied Since a signi1047297cant overlapping among the studied analytesoccurs in excitation and emission spectra both augmentation modeswere checked eg column-wise (emission spectral) and row-wise(excitation spectral) The optimum number of MCR-ALS componentswas estimated according to Supplementary material For both aug-

mentation modes the number was four ascribed to three analytes

and a background Non-negativity in both modes was applied and

Fig 2 (A) Normalized excitation and emission photoinduced 1047298uorescence spectra

for CBZ (blue line) OFL (red line) and PX (green line) (B) Comparison with the

normalized spectra for DICLO (dashed-gray line) FLU (dashed-cyan line) IBU

(dashed-pink line) and SAL (dashed-black line) after UV irradiation (For inter-

pretation of the references to color in this 1047297

gure legend the reader is referred tothe web version of this article)

Fig 1 Excitation and emission 1047298uorescence spectra for aqueous CBZ (A) OFL

(B) and PX (C) (dashndashdot-dotted line) and in 2 mol L 1 HCl before (dashed line)

and after UV irradiation (solid line) C CBZfrac14C OFL frac14C PXfrac14600 ng mL 1


7172019 1-s20-S003991401400914X-main


the convergence criterion was 01 (relative 1047297t change for successiveiterations)

MCR-ALS in both augmentation modes showed a similar

behavior to that of PARAFAC rendering good results for OFL andPX (Fig 4B) but unsuitable predictions for CBZ (not shown) Thisfact was also ascribed to the signi1047297cant spectral overlapping

between CBZ and strong PX signals Due to these results PARAFACand MCR-ALS were not applied to more complex samples

U-PLSRBL has been successfully employed in systems requiringprocessing 1047298exibility [27] and was thus applied to the present

system In a 1047297rst phase validation samples were studied withU-PLS and then considerably more complex samples were ana-lyzed using RBL to model the regression residues as a sum of bilinear contributions from the unexpected components

The optimum number of latent variables estimated by cross-validation (see Supplementary material) was four for CBZ and OFL and three for PX Apparently due to the relatively high intensitiesof PX signals U-PLS does not require an additional component to

model the backgroundThe U-PLS analyte predictions are very good in the validation

samples (Fig 4C) even for the con1047298icting analyte CBZ This factanalyte can be justi1047297ed by the use of latent variables which are

1047298exible enough to overcome the problem of the high degree of spectral similarity among certain analytes

From the EJCR test (Fig 4) we conclude that all ellipses includethe theoretically expected point (10) For OFL and PX they are

signi1047297cantly smaller than that corresponding to PARAFAC suggestingbetter precision Table 2 supports this conclusion with a relative errorof prediction (REP) equal to or less than 10 for all analytes

It is important to remark that the limits of detection (LODs) were

calculated according to a novel IUPAC-consistent estimator [29]which adopts the form of a detection interval as shown in Table 2

322 Test samples

The potential interferents IBU DICLO SAL and FLU display signals

which strongly overlap those for the target pollutants (Fig 2B)

Therefore with the purpose of simulating a genuine situation testsamples containing the analytes and the above compounds whichcould be concomitantly present in real samples were analysed

(see Fig 3B)For the test samples U-PLS required in addition to the calibration

latent variables the RBL procedure with two components corre-

sponding to the unexpected constituents Adding more componentsdid not improve the RBL 1047297t Apparently U-PLSRBL considers the

pro1047297les of all interferences as two mathematical components anddistinguishes these combined signals from those of the analytes and

the backgroundFig 5A illustrates the excellent U-PLSRBL predictions for the test

samples and Fig 5B displays the corresponding EJCR tests whichdenote accuracy The analytical performance of is further appreciated

in Table 2 The results are encouraging taking into account that thesimultaneous determination of three analytes is easily and rapidlyperformed in complex matrices Considering that photoreactor geo-metry allows the simultaneous irradiation of four solutions in about

5 min per sample and that EEPIFM measurements are performed in5 min a throughput of about 6 samples per hour is achieved

323 Real water samplesCBZ is one of the most frequently detected drugs in environ-

mental waters all over the world [30] Due to the fact that CBZ isrecalcitrant to various wastewater treatment processes it is consid-

ered as a hydrologic tracer of wastewaters [8] The incomplete CBZremoval when employing on-site wastewater treatments results in ahigh probability of groundwater surface water and 1047297nally drinkingwater contamination [10] It was recently pointed out that CBZ is one

of the six pharmaceuticals most often found in 1047297nished drinkingwaters with levels as high as 06 ng mL 1 [11] CBZ concentrationvalues around 015 ng mL 1 were detected in a municipal waste-water in1047298uent in Waco (Texas USA) [10] A multi-residue analysis of

both human and veterinary pharmaceuticals in surface and treatedwaters from different sites in Catalonia (Spain) revealed values

ranging from 2103

to 016 ng mL 1

for CBZ and up to 031 and

Fig 3 Three-dimensional and contour plots of EEPIFMs for (A) a validation sample 280 ng mL 1 CBZ 320 ng mL 1 OFL and 315 ng mL 1 PX (B) a test sample 78 ng mL 1

CBZ 82 ng mL 1 OFL 168 ng mL 1 PX 3000 ng mL 1 IBU1500 ng mL 1 DICLO 1500 ng mL 1 SAL and 100 ng mL 1 FLU and (C) a spiked underground water after solid-phase extraction (original concentrations are C CBZfrac14034 ng mL 1 C OFL frac14036 ng mL 1 and C PXfrac14034 ng mL 1)


7172019 1-s20-S003991401400914X-main


Fig 4 Plots for the CBZ (blue) OFL (red) and PX (green) predicted concentrations in validation samples as a function of the nominal values (the solid lines are the perfect

1047297ts) and elliptical joint regions (at 95 con1047297dence level) for the slope and intercept of the regression of the corresponding data Black points mark the theoretical

(interceptfrac140 slopefrac141) point (A) PARAFAC (B) MCR-ALS and (C) U-PLS (For interpretation of the references to color in this 1047297gure legend the reader is referred to the web

version of this article)


7172019 1-s20-S003991401400914X-main


025 ng mL 1 for OFL and PX respectively [31] CBZ concentrationsup to 030 ng mL 1 have been detected in Serbian waters similar toother European regions [30] Likewise values ranging from 022 to080 ng mL 1 were reported for OFL and other related antibiotics

CBZ and OFL were found at concentrations up to 12 and 058 ng mL 1respectively in sewage treatment plant ef 1047298uents of several Europeancountries although higher levels of the former (until 63 ng mL 1)were detected in Germany and Switzerland [32] OFL and other two

1047298uoroquinolones have been found in hospital wastewaters (at con-centrations of 006ndash120 ng mL 1) in wastewater treatment plantef 1047298uents (2103

ndash058 ng mL 1) and in surface waters (5103ndash

130 ng mL 1) throughout the world including the United States Italy

Switzerland Finland Sweden Germany China and Australia [4] OFL was one of the three most reported 1047298uoroquinolone antibiotics inChinese surface waters with concentrations up to 51 ng mL 1 [6]

Since the analytes are generally detected as part- and sub-part-

per-billions pre-concentration with C18 membrane-SPE wasapplied The selectivity is provided by the chemometric tool andthe physical separation of target analytes from the matrix is notrequired as in traditional extraction techniques

Because the polarity of the molecules plays a crucial role toachieve ef 1047297cient extraction in the C18 membrane the working pHwas selected on the basis of the pK as of the target compounds Thelatter pK as (23 and 1399 for CBZ [33] 605 and 811 for OFL [34] and

181 and 512 for PX [35]) suggest that the uncharged species for CBZOFL and PX prevail in the pH 3ndash13 7 and 3 respectively In principlethe selection of an optimal pH for the simultaneous retention of thethree analytes is not possible However extracting 250 mL of

synthetic aqueous samples at pH 35 50 and 70 containing030 ng mL 1 of each analyte demonstrated recoveries nearly 100 for the three compounds in all cases This result can be justi1047297edconsidering the small concentration of analytes in the samples

combined with the strong extraction power of the C18 membranesTherefore the neutral pH of real samples was not adjusted before thecorresponding treatment

A recovery study was carried out by spiking four different types of

waters with the analytes in duplicate at two different concentrationlevels following the treatment indicated above Typical EEPIFM plotsof a spiked underground water after preconcentration are shown in

Fig 3C The strong matrix interference is evident However thephysical removal of these interferences is not necessary when using

an appropriate second-order calibration methodology highlightingthe value of the chemometric approach

RBL was also required for real samples with two unexpectedcomponents in most cases Adding more components did notimprove the 1047297t The recoveries (Table 3) are statistically

Table 2

Statistical results for CBZ OFL and PX in validation test and real water samples

using the proposed methodology and U-PLSRBL

CBZ OFL PX

Validation samples

LOD range (minndashmax) 4ndash7 3ndash6 2ndash4

RMSEP 2 3 2REP 7 10 7

Test samplesa


RMSEP 2 3 2

REP 7 9 7

Tap water (Venado Tuerto city)

LOD range (minndashmax)b 016ndash022 007ndash009 005ndash006

LOD range (minndashmax)c 031ndash044 014ndash018 010ndash013

RMSEP 003 007 003

REP 6 11 10Underground water (Santa Rosa city)



RMSEP 004 004 003

REP 9 10 6

Underground water (Funes city)


LOD range (minndashmax)c 043ndash057 017ndash0 21 013ndash017

RMSEP 003 004 005

REP 8 9 12

River water (Paranaacute River)

LOD range (minndashmax)d 5ndash7 5ndash7 3ndash5

LOD range (minndashmax)c 01ndash02 01ndash02 005ndash008RMSEP 1 2 2

REP 8 9 6

LOD limit of detection calculated according to ref 29RMSEP root-mean-square error of prediction

REP relative error of prediction

Values for LOD and RMSEP are given in ng mL

1 Values of REP are giving in The real samples results refer to the original water samples before SPE

a Fifteen samples containing IBU DICLO SAL and FLU as interferentsb Pre-concentration factorfrac14125c Pre-concentration factorfrac1450d Sample without pre-concentration

Fig 5 (A) Plot for CBZ (blue) OFL (red) and PX (green) predicted concentrations by U-PLSRBL in samples containing potential interferences as a function of the nominal

values (the solid line is the perfect 1047297t) (B) Elliptical joint regions (at 95 con1047297dence level) for the slope and intercept of the regression of the corresponding data The black

point marks the theoretical (interceptfrac140 slopefrac141) point (For interpretation of the references to color in this 1047297gure legend the reader is referred to the web version of this

article)


7172019 1-s20-S003991401400914X-main


comparable to those provided by LCndashMS at a 95 con1047297dence level[36] the experimental t -coef 1047297cients for U-PLSRBL in the cases of CBZ (t frac14074) OFL (t frac14035) and PX (t frac14102) favorably compare

with the tabulated value for n1 degrees of freedom [t crit

(0057)frac14236] suggesting that foreign compounds which may be

present in the studied samples do not produce a signi1047297cant

interference in our analysis Finally the good analytical perfor-mance for U-PLSRBL can be appreciated from the statisticalresults shown in Table 2 These results indicate that the REP is

not signi1047297cantly affected by the fact that real samples are beingstudied Besides LODs re1047298ect the bene1047297ts of the pre-concentra-

tion and the possibility of determining the studied analytes at

Table 3

Recovery study of CBZ OFL and PX for spiked water samples using U-PLSRBL algorithm and LCndashMS methoda

Sample CBZ OFL PX

Taken U-PLSRBL LCndashMS Taken U-PLSRBL LCndashMS Taken U-PLSRBL LCndashMS

Tap waterb 023 024 (001) 020 (001) 029 029 (001) 028 (001) 017 016 (004) 017 (001)

[104] [87] [100] [97] [94] [100]

049 054 (002) 060 (001) 072 081 (004) 079 (004) 042 038 (001) 056 (002)

[110] [122] [113] [110] [90] [133]Underground waterc 020 017 (001) 023 (001) 016 018 (002) 019 (001) 013 015 (003) 012 (001)

[85] [115] [113] [119] [115] [92]

034 039 (001) 035 (001) 036 039 (006) 038 (002) 034 036 (002) 027 (001)

[115] [103] [108] [106] [106] [79]

Underground waterd 008 008 (001) 010 (001) 021 021 (004) 021 (001) 046 042 (001) 043 (002)

[100] [125] [100] [100] [91] [93]029 024 (008) 028 (001) 092 09 (01) 100 (001) 063 053 (003) 059 (002)

[83] [97] [98] [109] [84] [94]

River watere 137 14 (2) 136 (01) 614 61 (03) 60 (03) 105 118 (23) 101 (04)[102] [99] [99] [98] [112] [96]

068 075 (009) 078 (001) 102 10 (02) 11 (01) 084 10 (02) 083 (003)

[110] [115] [98] [108] [119] [99]

a Concentrations are given in ng mL 1 standard deviations (mean of two determinations) are given between parentheses and recoveries (between square brackets) are

given in b Venado Tuerto cityc

Santa Rosa cityd Funes citye Paranaacute River

Table 4

Analytical performance of selected methods recently reported for emerging contaminants in natural waters

Pre-treatment Method Compounds Other Medium LODa RSD REP RECb Sample Ref

SPE (Oasis HLB cartridges) LCndashMSMS CBZ OFL

PX

Horm

Pharm

Organic 1103ndash25103 RECfrac14769ndash934 WW [37]

S PE (HLB cart ridg es) LCndashMSMS OFL A Organic 28103ndash

66103

RECfrac1475ndash112 RW WWI WWE [38]

SPE (Oasis HLB cartridges) UPLCndashMSMS CBZ Pharm Organic 285103 RSDfrac1422 RW [39]

RECfrac14993

SPME (silica 1047297ber-polymeric phase)

LCndashDAD CBZ PX NSAI Organic 26ndash3 RSDfrac1446ndash8 RW [40]RECfrac14716ndash1190

SPME (PDSDVBpolyacrylate)

LCndash

DADndash

MCR-ALS

CBZ PX NSAI Organic LOQ frac1410ndash

20 RECfrac1472ndash

1193 RW WW [41]

SPE (Oasis HLB cartridges) UPLCndashQTOFndashMS

OFL A Organic 04 WWE WWI WW RW [42]

SPE (Oasis HLB cartridges) UPLCndashMSMS OFLO A Organic 38103 RDSfrac143ndash25 RECfrac1433ndash

142

RW [43]

On line SPE (Strata-X

column)

LCndashMSMS CBZ OP Organic 07ndash1103 TW RW [44]

SPE (Strata-X column) ECndashLIF OFLO A Organic 09103 RDSfrac1426ndash98

RECfrac141006ndash1066

UW TW [45]

TF-SPME LCndashMSMS CBZ Pharm Organic 2103 RSDfrac1414 RECfrac14837 WW [46]

MPs (PDSMWCN) CLC OFLO A Organic 048 RSDfrac1459 RECfrac14112 MW [47]SPE (Oasis HLB cartridges) UPLCndashQqLITndash

MSMS

CBZ OFL

PX

P harm Or ganic 01103ndash46103 RSDfrac1415ndash159 DW RSW RW SW

WWE WWI

[31]

RECfrac1440ndash124

SPME (modi1047297ed silica gel) LCndashUV CBZ Pharm Organic 236 RECfrac14656 EW [48]

SPE (Oasis HLB cartridges) GCndashMS CBZ Pharm Organic 8103 RECfrac14920 RSDfrac1463 RW [49]

SPE (C18 membranes) PIFndashMCR ndashALS CBZ Aqueous 01ndash2 RECfrac1478ndash117 REPfrac142ndash7 TW RW UW [15]

SPE (Oasis HLB cartridges) UPLCndashQqLITndash

MSMS

CBZ OFL

PX

P harm Or ganic 02103ndash

24103

RSDfrac1425ndash159 RECfrac1456ndash

124

DW RW WE [30]

SPE (Oasis HLB cartridges) LCndashTOFndashMS CBZ OFL EP Organic LOQ frac1425103ndash

31103

RSDfrac141ndash15 REC frac1430ndash

128

RW WW [13]

SPE (C18 membranes) PIFndashU-PLSRBL CBZ OFLPX

Aqueous 004ndash03 REPfrac146ndash12 TW RW UW ThisworkRECfrac1483ndash119

a For comparison concentration units were uni1047297ed to ng mL 1b Relative standard deviation (RSD) relative error of prediction (REP) and recovery (REC) all in


7172019 1-s20-S003991401400914X-main


sub-part-per-billion levels It should be noted that an SPE employ-ing a larger sample volume would allow to decrease the LODeven more

In Table 4 a comparison with selected methods for the determi-

nation of the studied compounds in water samples is performedincluding already cited reports and additional ones [37ndash49] Thegreat advantage of the proposed approach is that it allows thedetermination of the analytes using very simple equipment and

without involving signi1047297cant amounts of organic solvents As aconsequence the experimental time and the errors associated withmultiple experimental steps are substantially diminished working atthe same time under the green chemistry principles

4 Conclusions

A sustainable photoinduced 1047298uorescence method suitable for the

simultaneous determination of CBZ OFL and PX at trace levelswithout the need of chromatographic separation has been devel-oped The method is assisted by second-order chemometric analysisand representing a new example of the power of coupling the partial

least-squares algorithm with residual bilinearization for the resolu-tion of very complex systems The beauty of this procedure is that it

achieves an outstanding selectivity avoiding the use of toxic organicsolvents a fact which is essential for the environmental safety Inaddition the method is fast allowing a sample throughput of about

6 samples per hour On the basis of the obtained results one canassert that the proposed method favorably compares with moresophisticated approaches

Acknowledgments

The authors are grateful to the Universidad Nacional de RosarioCONICET (PIP 0163) y ANPCyT (PICT 2013-0136) and Junta deExtremadura and European FEDER Funds (Consolidation Project of

Research Group FQM003 Project GR10033) for 1047297nancial support M

C Hurtado-Saacutenchez thanks to Consejeriacutea de Economiacutea Comercio eInnovacioacuten of Junta de Extremadura and European Social Fund for afellowship (DOE 04012011)

Appendix A Supporting information

Supplementary data associated with this article can be found inthe online version at httpdxdoiorg101016jtalanta201411022

References

[1] C Mahugo-Santana Z Sosa-Ferrera ME Torres-Padroacuten JJ Santana-Rodriacuteguez TrAC Trends Anal Chem 30 (2011) 731ndash748

[2] SD Richardson TA Ternes Anal Chem 86 (2014) 2813ndash

2848[3] M Stuart D Lapworth E Crane A Hart Sci Total Environ 416 (2012) 1ndash21[4] KH Wammer AR Korte RA Lundeen JE Sundberg K McNeill WA Arnold

Water Res 47 (2013) 439ndash448

[5] R Meffe I de Bustamante Sci Total Environ 481 (2014) 280ndash295[6] Q Bu B Wang J Huang S Deng G Yu J Hazard Mater 262 (2013) 189 ndash211[7] E Vulliet C Cren-Oliveacute Environ Pollut 159 (2011) 2929ndash2934[8] M Clara B Strenn N Kreuzinger Water Res 38 (2004) 947ndash954[9] E Esteacutevez MC Cabrera A Molina-Diacuteaz J Robles-Molina MP Palacios-Diacuteaz

Sci Total Environ 433 (2012) 538ndash546[10] B Du AE Price WC Scott LA Kristofco AJ Ramirez CK Chambliss

JC Yelderman BW Brooks Sci Total Environ 466ndash467 (2014) 976ndash984[11] C Postigo SD Richardson J Hazard Mater 279 (2014) 461ndash475[12] D Fatta A Achilleos A Nikolaou S Mericcedil TrAC Trends Anal Chem 26 (2007)

515ndash533[13] J Robles-Molina FJ Lara-Ortega B Gilbert-Loacutepez JF Garciacutea-Reyes A Molina-

Diacuteaz J Chromatogr A 1350 (2014) 30ndash43[14] Z Migaszewski Gał uszka J Namieśnik TrAC Trends Anal Chem 50 (2013)

78ndash84[15] VA Lozano GM Escandar Anal Chim Acta 782 (2013) 37ndash45[16] AC Olivieri Anal Chem 80 (2008) 5713ndash5720[17] R Tauler Chemom Intell Lab Syst 30 (1995) 133ndash146[18] R Bro Chemom Intell Lab Syst 38 (1997) 149ndash171[19] J oumlhman P Geladi S Wold J Chemom 4 (1990) 135ndash146[20] AC Olivieri J Chemom 19 (2005) 253ndash265[21] MATLAB 710 2010 The MathWorks Inc Natick MA USA[22] AC Olivieri HL Wu RQ Yu Chemom Intell Lab Syst 96 (2009) 246 ndash251[23] langhttpwwwiquir-conicetgovardescargasmvc2rar rang accessed September

2014[24] GM Escandar Analyst 124 (1999) 587ndash591[25] HM Abdel-Wadood Bull Pharm Sci 31 (2008) 169ndash181[26] AC Olivieri GM Escandar A Muntildeoz de la Pentildea TrAC Trends Anal Chem 30

(2011) 607ndash617

[27] AC Olivieri GM Escandar Practical Three-way Calibration Elsevier Wal-tham USA 2014

[28] AV Gonzaacutelez MA Herrador AG Asuero Talanta 48 (1999) 729ndash736[29] F Allegrini AC Olivieri Anal Chem 86 (2014) 7858ndash7866[30] M Petrović B Škrbić J Ž ivančev L Ferrando Climent D Barceloacute Sci Total

Environ 468ndash469 (2014) 415ndash428[31] M Gros S Rodriacuteguez-Mozaz D Barceloacute J Chromatogr A 1248 (2012)

104ndash121[32] R Andreozzi M Raffaele P Nicklas Chemosphere 50 (2003) 1319ndash1330[33] P Punyapalakul T Sitthisorn World Acad Sci Eng Technol 69 (2010)

546ndash550[34] E Hapeshi A Achilleos MI Vasquez C Michael NP Xekoukoulotakis

D Mantzavinos D Kassinos Water Res 44 (2010) 1737ndash1746[35] GM Escandar AJ Bystol AD Campiglia Anal Chim Acta 466 (2002)

275ndash283[36] WP Gardiner Statistical Analysis Methods for Chemists A Software-Based

Approach The Royal Society of Chemistry Cambridge UK (1997) 64 [37] AYC Lin TH Yu CF Lin Chemosphere 74 (2008) 131ndash141

[38] Y Xiao H Chang A Jia J Hu J Chromatogr A 1214 (2008) 100ndash

108[39] JM Conley SJ Symes SA Kindelberger SM Richards J Chromatogr A 1185

(2008) 206ndash215[40] L Vera-Candioti MD Gil Garciacutea M Martiacutenez Galera HC Goicoechea

J Chromatogr A 1211 (2008) 22ndash32[41] MD Gil Garciacutea F Cantildeada Cantildeada MJ Culzoni L Vera-Candioti GG Siano

HC Goicoechea M Martiacutenez Galera J Chromatogr A 1216 (2009) 5489ndash5496[42] M Ibaacutentildeez C Guerrero JV Sancho F Hernaacutendez J Chromatogr A 1216 (2009)

2529ndash2539[43] F Tamtam F Mercier J Eurin M Chevreuil B Le Bot Anal Bioanal Chem 393

(2009) 1709ndash1718[44] PA Segura Garcia-Ac L Viglino A Fuumlrtoumls C Gagnon M Preacutevost S Sauveacute

J Chromatogr A 1216 (2009) 8518ndash8527[45] M Lombardo-Aguumliacute L Gaacutemiz-Gracia AM Garciacutea-Campantildea C Cruces-Blanco

Anal Bioanal Chem 396 (2010) 1551ndash1557[46] OP Togunde E Cudjoe KD Oakes FS Mirnaghi MR Servos J Pawliszyn

J Chromatogr A 1262 (2012) 34ndash42[47] S Xu C Jiang Y Lin L Jia Microchim Acta 179 (2012) 257ndash264

[48] TH Lim L Hu C Yang C He HK Lee J Chromatogr A 1316 (2013) 8ndash

14[49] N Stamatis V Triantafyllidis D Hela I Konstantinou Int J Environ AnalChem 93 (2013) 1602ndash1619


7172019 1-s20-S003991401400914X-main


In the present work three emerging pollutants representative of

different groups of therapeutic drugs were investigated the antic-onvulsant carbamazepine (CBZ) the antibacterial 1047298uoroquinoloneo1047298oxacin (OFL) and the non-steroidal anti-in1047298ammatory piroxicam(PX) (Scheme 1) because they are frequently found in environmen-

tal waters They display photo-induced 1047298uorescence (PIF) upon UV irradiation which could allow their quanti1047297cation Relatively fewmolecules are 1047298uorescent and 1047298uorescent photoproducts are even

fewer or parent compounds are photodegradated after UV irradia-tion This may lead to the erroneous conclusion that PIF-based

methods are free from interferences However as presently demon-strated in multicomponent systems the probability of the occur-

rence of interferences signi1047297cantly increases and in principle clean-up and separation procedures are almost unavoidable

Recently our research group quanti1047297ed CBZ as a single analyte inenvironmental waters using the PIF signals after UV irradiation of

acidic solutions in a simple laboratory-constructed reactor [15] Thelack of selectivity was overcome by the second-order advantage of multi-way calibration [16] and pollutant was quantitated in thepresence of unknown sample constituents Second-order data were

obtained as excitationndashemission photoinduced 1047298uorescence matrices(EEPIFMs) and processed by different algorithms although successfulresults were obtained with multivariate curve resolution-alternatingleast-squares (MCR-ALS) [17]

The critical difference of the present report with the earlier workis that the simultaneous resolution of three usual emerging con-taminants which strongly overlap their PIF spectra is presentlyintended with the concomitant change in both data analysis and

results interpretation Further the determinations are performed insolutions containing the analytes and additional pharmaceuticalssuch as ibuprofen (IBU) diclofenac (DICLO) salicylic acid (SAL) and

1047298ufenamic acid (FLU) (Scheme 1) The latter are profusely employed

in our geographical region and may thus be present in real watersand showed 1047298uorescence signals (either in native form or from theirphotoproducts) which signi1047297cantly overlap those of the analytes

Three chemometric algorithms achieving the second-order advan-tage ie parallel factor analysis (PARAFAC) [18] MCR-ALS and

unfolded partial least-squaresresidual bilinearization (U-PLSRBL)

[1920] were applied to process the EEPIFMs Noticeable differences

in the prediction capabilities of the employed algorithms were found

and discussedTo the best of our knowledge it is the 1047297rst time that the selectivity

offered by the chemometric analysis is evaluated for the simulta-neous determination of several analytes using EEPIFMs in very

interfering media The feasibility of determining the three emergentpollutants in real water samples using sustainable resources isdemonstrated

2 Experimental

21 Reagents and solutions

CBZ OFL and PX were purchased from Sigma (St Louis MOUSA) Methanol (MeOH) formic acid and hydrochloric acid (HCl)were obtained from Merck (Darmstadt Germany) IBU DICLO SAL and FLU were of analytical grade and were used as received Stock

standard solutions of individual analytes (4040 mg mL 1 CBZ4200 mg mL 1 PX and 5100 mg mL 1 OFL) were prepared bydissolving an appropriate amount of each compound in methanol

and stored at 4 1C Working analyte solutions of 20 mg mL 1 weredaily prepared by dilution of stock standard solutions in ultrapurewater Ultra pure Milli-Q water was used throughout the work

22 Instrumentation

Fluorescence measurements were performed on an Aminco

Bowman (Rochester NY USA) Series 2 luminescence spectro-photometer equipped with a 150 W xenon lamp EEPIFMs weremeasured in the ranges 246ndash333 nm (each 3 nm excitation) and380ndash480 nm (each 1 nm emission) leading to 29100 matrices

Excitation and emission slit widths were of 8 nm using 100 cmquartz cells The photomultiplier tube sensitivity was 600 V andthe cell temperature was regulated at 20 1C using a thermostaticbath (Cole-Parmer IL USA) EEPIFMs were saved and transferred

to a PC for subsequent chemometric analysisFor the reference chromatographic analysis see Supplementary

material

Scheme 1 Structures of carbamazepine (CBZ) o1047298oxacin (OFL) piroxicam (PX) ibuprofen (IBU) diclofenac (DICLO) salicylic acid (SAL) and 1047298ufenamic acid (FLU)


7172019 1-s20-S003991401400914X-main
















25 Water samples












graphic system

















calibration

Table 1



1 300 00 3002 300 600 300

3 300 300 00

4 300 300 600

5 00 300 300

6 600 300 3007 122 122 122

8 122 478 122

9 122 122 478

10 122 478 478

11 478 122 12212 478 478 122

13 478 122 478

14 478 478 478

15 300 300 300



7172019 1-s20-S003991401400914X-main



























7172019 1-s20-S003991401400914X-main
















322 Test samples

















ranging from 2103

to 016 ng mL 1





7172019 1-s20-S003991401400914X-main







7172019 1-s20-S003991401400914X-main



















Table 2



CBZ OFL PX

Validation samples


RMSEP 2 3 2REP 7 10 7

Test samplesa


RMSEP 2 3 2

REP 7 9 7




RMSEP 003 007 003




RMSEP 004 004 003

REP 9 10 6




RMSEP 003 004 005

REP 8 9 12




REP 8 9 6









article)


7172019 1-s20-S003991401400914X-main









Table 3


Sample CBZ OFL PX


Tap waterb 023 024 (001) 020 (001) 029 029 (001) 028 (001) 017 016 (004) 017 (001)

[104] [87] [100] [97] [94] [100]

049 054 (002) 060 (001) 072 081 (004) 079 (004) 042 038 (001) 056 (002)


[85] [115] [113] [119] [115] [92]

034 039 (001) 035 (001) 036 039 (006) 038 (002) 034 036 (002) 027 (001)

[115] [103] [108] [106] [106] [79]


[100] [125] [100] [100] [91] [93]029 024 (008) 028 (001) 092 09 (01) 100 (001) 063 053 (003) 059 (002)

[83] [97] [98] [109] [84] [94]

River watere 137 14 (2) 136 (01) 614 61 (03) 60 (03) 105 118 (23) 101 (04)[102] [99] [99] [98] [112] [96]

068 075 (009) 078 (001) 102 10 (02) 11 (01) 084 10 (02) 083 (003)

[110] [115] [98] [108] [119] [99]




Table 4




PX

Horm

Pharm



66103



RECfrac14993




LCndash

DADndash

MCR-ALS


20 RECfrac1472ndash

1193 RW WW [41]




142

RW [43]


column)




UW TW [45]



MSMS

CBZ OFL

PX


WWE WWI

[31]

RECfrac1440ndash124





MSMS

CBZ OFL

PX


24103


124

DW RW WE [30]


31103


128

RW WW [13]





7172019 1-s20-S003991401400914X-main






4 Conclusions






Acknowledgments






References












(2011) 607ndash617






















7172019 1-s20-S003991401400914X-main
















25 Water samples












graphic system

















calibration

Table 1



1 300 00 3002 300 600 300

3 300 300 00

4 300 300 600

5 00 300 300

6 600 300 3007 122 122 122

8 122 478 122

9 122 122 478

10 122 478 478

11 478 122 12212 478 478 122

13 478 122 478

14 478 478 478

15 300 300 300



7172019 1-s20-S003991401400914X-main



























7172019 1-s20-S003991401400914X-main
















322 Test samples

















ranging from 2103

to 016 ng mL 1





7172019 1-s20-S003991401400914X-main







7172019 1-s20-S003991401400914X-main



















Table 2



CBZ OFL PX

Validation samples


RMSEP 2 3 2REP 7 10 7

Test samplesa


RMSEP 2 3 2

REP 7 9 7




RMSEP 003 007 003




RMSEP 004 004 003

REP 9 10 6




RMSEP 003 004 005

REP 8 9 12




REP 8 9 6









article)


7172019 1-s20-S003991401400914X-main









Table 3


Sample CBZ OFL PX


Tap waterb 023 024 (001) 020 (001) 029 029 (001) 028 (001) 017 016 (004) 017 (001)

[104] [87] [100] [97] [94] [100]

049 054 (002) 060 (001) 072 081 (004) 079 (004) 042 038 (001) 056 (002)


[85] [115] [113] [119] [115] [92]

034 039 (001) 035 (001) 036 039 (006) 038 (002) 034 036 (002) 027 (001)

[115] [103] [108] [106] [106] [79]


[100] [125] [100] [100] [91] [93]029 024 (008) 028 (001) 092 09 (01) 100 (001) 063 053 (003) 059 (002)

[83] [97] [98] [109] [84] [94]

River watere 137 14 (2) 136 (01) 614 61 (03) 60 (03) 105 118 (23) 101 (04)[102] [99] [99] [98] [112] [96]

068 075 (009) 078 (001) 102 10 (02) 11 (01) 084 10 (02) 083 (003)

[110] [115] [98] [108] [119] [99]




Table 4




PX

Horm

Pharm



66103



RECfrac14993




LCndash

DADndash

MCR-ALS


20 RECfrac1472ndash

1193 RW WW [41]




142

RW [43]


column)




UW TW [45]



MSMS

CBZ OFL

PX


WWE WWI

[31]

RECfrac1440ndash124





MSMS

CBZ OFL

PX


24103


124

DW RW WE [30]


31103


128

RW WW [13]





7172019 1-s20-S003991401400914X-main






4 Conclusions






Acknowledgments






References












(2011) 607ndash617






















7172019 1-s20-S003991401400914X-main



























7172019 1-s20-S003991401400914X-main
















322 Test samples

















ranging from 2103

to 016 ng mL 1





7172019 1-s20-S003991401400914X-main







7172019 1-s20-S003991401400914X-main



















Table 2



CBZ OFL PX

Validation samples


RMSEP 2 3 2REP 7 10 7

Test samplesa


RMSEP 2 3 2

REP 7 9 7




RMSEP 003 007 003




RMSEP 004 004 003

REP 9 10 6




RMSEP 003 004 005

REP 8 9 12




REP 8 9 6









article)


7172019 1-s20-S003991401400914X-main









Table 3


Sample CBZ OFL PX


Tap waterb 023 024 (001) 020 (001) 029 029 (001) 028 (001) 017 016 (004) 017 (001)

[104] [87] [100] [97] [94] [100]

049 054 (002) 060 (001) 072 081 (004) 079 (004) 042 038 (001) 056 (002)


[85] [115] [113] [119] [115] [92]

034 039 (001) 035 (001) 036 039 (006) 038 (002) 034 036 (002) 027 (001)

[115] [103] [108] [106] [106] [79]


[100] [125] [100] [100] [91] [93]029 024 (008) 028 (001) 092 09 (01) 100 (001) 063 053 (003) 059 (002)

[83] [97] [98] [109] [84] [94]

River watere 137 14 (2) 136 (01) 614 61 (03) 60 (03) 105 118 (23) 101 (04)[102] [99] [99] [98] [112] [96]

068 075 (009) 078 (001) 102 10 (02) 11 (01) 084 10 (02) 083 (003)

[110] [115] [98] [108] [119] [99]




Table 4




PX

Horm

Pharm



66103



RECfrac14993




LCndash

DADndash

MCR-ALS


20 RECfrac1472ndash

1193 RW WW [41]




142

RW [43]


column)




UW TW [45]



MSMS

CBZ OFL

PX


WWE WWI

[31]

RECfrac1440ndash124





MSMS

CBZ OFL

PX


24103


124

DW RW WE [30]


31103


128

RW WW [13]





7172019 1-s20-S003991401400914X-main






4 Conclusions






Acknowledgments






References












(2011) 607ndash617






















7172019 1-s20-S003991401400914X-main
















322 Test samples

















ranging from 2103

to 016 ng mL 1





7172019 1-s20-S003991401400914X-main







7172019 1-s20-S003991401400914X-main



















Table 2



CBZ OFL PX

Validation samples


RMSEP 2 3 2REP 7 10 7

Test samplesa


RMSEP 2 3 2

REP 7 9 7




RMSEP 003 007 003




RMSEP 004 004 003

REP 9 10 6




RMSEP 003 004 005

REP 8 9 12




REP 8 9 6









article)


7172019 1-s20-S003991401400914X-main









Table 3


Sample CBZ OFL PX


Tap waterb 023 024 (001) 020 (001) 029 029 (001) 028 (001) 017 016 (004) 017 (001)

[104] [87] [100] [97] [94] [100]

049 054 (002) 060 (001) 072 081 (004) 079 (004) 042 038 (001) 056 (002)


[85] [115] [113] [119] [115] [92]

034 039 (001) 035 (001) 036 039 (006) 038 (002) 034 036 (002) 027 (001)

[115] [103] [108] [106] [106] [79]


[100] [125] [100] [100] [91] [93]029 024 (008) 028 (001) 092 09 (01) 100 (001) 063 053 (003) 059 (002)

[83] [97] [98] [109] [84] [94]

River watere 137 14 (2) 136 (01) 614 61 (03) 60 (03) 105 118 (23) 101 (04)[102] [99] [99] [98] [112] [96]

068 075 (009) 078 (001) 102 10 (02) 11 (01) 084 10 (02) 083 (003)

[110] [115] [98] [108] [119] [99]




Table 4




PX

Horm

Pharm



66103



RECfrac14993




LCndash

DADndash

MCR-ALS


20 RECfrac1472ndash

1193 RW WW [41]




142

RW [43]


column)




UW TW [45]



MSMS

CBZ OFL

PX


WWE WWI

[31]

RECfrac1440ndash124





MSMS

CBZ OFL

PX


24103


124

DW RW WE [30]


31103


128

RW WW [13]





7172019 1-s20-S003991401400914X-main






4 Conclusions






Acknowledgments






References












(2011) 607ndash617






















7172019 1-s20-S003991401400914X-main







7172019 1-s20-S003991401400914X-main



















Table 2



CBZ OFL PX

Validation samples


RMSEP 2 3 2REP 7 10 7

Test samplesa


RMSEP 2 3 2

REP 7 9 7




RMSEP 003 007 003




RMSEP 004 004 003

REP 9 10 6




RMSEP 003 004 005

REP 8 9 12




REP 8 9 6









article)


7172019 1-s20-S003991401400914X-main









Table 3


Sample CBZ OFL PX


Tap waterb 023 024 (001) 020 (001) 029 029 (001) 028 (001) 017 016 (004) 017 (001)

[104] [87] [100] [97] [94] [100]

049 054 (002) 060 (001) 072 081 (004) 079 (004) 042 038 (001) 056 (002)


[85] [115] [113] [119] [115] [92]

034 039 (001) 035 (001) 036 039 (006) 038 (002) 034 036 (002) 027 (001)

[115] [103] [108] [106] [106] [79]


[100] [125] [100] [100] [91] [93]029 024 (008) 028 (001) 092 09 (01) 100 (001) 063 053 (003) 059 (002)

[83] [97] [98] [109] [84] [94]

River watere 137 14 (2) 136 (01) 614 61 (03) 60 (03) 105 118 (23) 101 (04)[102] [99] [99] [98] [112] [96]

068 075 (009) 078 (001) 102 10 (02) 11 (01) 084 10 (02) 083 (003)

[110] [115] [98] [108] [119] [99]




Table 4




PX

Horm

Pharm



66103



RECfrac14993




LCndash

DADndash

MCR-ALS


20 RECfrac1472ndash

1193 RW WW [41]




142

RW [43]


column)




UW TW [45]



MSMS

CBZ OFL

PX


WWE WWI

[31]

RECfrac1440ndash124





MSMS

CBZ OFL

PX


24103


124

DW RW WE [30]


31103


128

RW WW [13]





7172019 1-s20-S003991401400914X-main






4 Conclusions






Acknowledgments






References












(2011) 607ndash617






















7172019 1-s20-S003991401400914X-main



















Table 2



CBZ OFL PX

Validation samples


RMSEP 2 3 2REP 7 10 7

Test samplesa


RMSEP 2 3 2

REP 7 9 7




RMSEP 003 007 003




RMSEP 004 004 003

REP 9 10 6




RMSEP 003 004 005

REP 8 9 12




REP 8 9 6









article)


7172019 1-s20-S003991401400914X-main









Table 3


Sample CBZ OFL PX


Tap waterb 023 024 (001) 020 (001) 029 029 (001) 028 (001) 017 016 (004) 017 (001)

[104] [87] [100] [97] [94] [100]

049 054 (002) 060 (001) 072 081 (004) 079 (004) 042 038 (001) 056 (002)


[85] [115] [113] [119] [115] [92]

034 039 (001) 035 (001) 036 039 (006) 038 (002) 034 036 (002) 027 (001)

[115] [103] [108] [106] [106] [79]


[100] [125] [100] [100] [91] [93]029 024 (008) 028 (001) 092 09 (01) 100 (001) 063 053 (003) 059 (002)

[83] [97] [98] [109] [84] [94]

River watere 137 14 (2) 136 (01) 614 61 (03) 60 (03) 105 118 (23) 101 (04)[102] [99] [99] [98] [112] [96]

068 075 (009) 078 (001) 102 10 (02) 11 (01) 084 10 (02) 083 (003)

[110] [115] [98] [108] [119] [99]




Table 4




PX

Horm

Pharm



66103



RECfrac14993




LCndash

DADndash

MCR-ALS


20 RECfrac1472ndash

1193 RW WW [41]




142

RW [43]


column)




UW TW [45]



MSMS

CBZ OFL

PX


WWE WWI

[31]

RECfrac1440ndash124





MSMS

CBZ OFL

PX


24103


124

DW RW WE [30]


31103


128

RW WW [13]





7172019 1-s20-S003991401400914X-main






4 Conclusions






Acknowledgments






References












(2011) 607ndash617






















7172019 1-s20-S003991401400914X-main









Table 3


Sample CBZ OFL PX


Tap waterb 023 024 (001) 020 (001) 029 029 (001) 028 (001) 017 016 (004) 017 (001)

[104] [87] [100] [97] [94] [100]

049 054 (002) 060 (001) 072 081 (004) 079 (004) 042 038 (001) 056 (002)


[85] [115] [113] [119] [115] [92]

034 039 (001) 035 (001) 036 039 (006) 038 (002) 034 036 (002) 027 (001)

[115] [103] [108] [106] [106] [79]


[100] [125] [100] [100] [91] [93]029 024 (008) 028 (001) 092 09 (01) 100 (001) 063 053 (003) 059 (002)

[83] [97] [98] [109] [84] [94]

River watere 137 14 (2) 136 (01) 614 61 (03) 60 (03) 105 118 (23) 101 (04)[102] [99] [99] [98] [112] [96]

068 075 (009) 078 (001) 102 10 (02) 11 (01) 084 10 (02) 083 (003)

[110] [115] [98] [108] [119] [99]




Table 4




PX

Horm

Pharm



66103



RECfrac14993




LCndash

DADndash

MCR-ALS


20 RECfrac1472ndash

1193 RW WW [41]




142

RW [43]


column)




UW TW [45]



MSMS

CBZ OFL

PX


WWE WWI

[31]

RECfrac1440ndash124





MSMS

CBZ OFL

PX


24103


124

DW RW WE [30]


31103


128

RW WW [13]





7172019 1-s20-S003991401400914X-main






4 Conclusions






Acknowledgments






References












(2011) 607ndash617






















7172019 1-s20-S003991401400914X-main






4 Conclusions






Acknowledgments






References












(2011) 607ndash617






















1-s2.0-S003991401400914X-main

Documents

Transcript of 1-s2.0-S003991401400914X-main