DSpace at VNU: Screening determination of pharmaceutical pollutants in different water matrices using dual-channel capillary electrophoresis coupled with contactless conductivity detection

Screening determination of pharmaceuticalpollutants in different water matrices using dual-channel capillary electrophoresis coupled with contactless conductivity detection Minh Duc Le,

Screening determination of pharmaceutical

pollutants in different water matrices using

dual-channel capillary electrophoresis coupled with

contactless conductivity detection

Minh Duc Le, Hong Anh Duong, Manh Huy

Nguyen, Jorge Sáiz, Hung Viet Pham, Thanh Duc

Mai

Received date: 8 May 2016

Accepted date: 12 July 2016

Cite this article as: Minh Duc Le, Hong Anh Duong, Manh Huy Nguyen, Jorge Sáiz, Hung Viet Pham and Thanh Duc Mai, Screening determination of pharmaceutical pollutants in different water matrices using dual-channel capillary

http://dx.doi.org/10.1016/j.talanta.2016.07.032

This is a PDF file of an unedited manuscript that has been accepted for publication As a service to our customers we are providing this early version of the manuscript The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

www.elsevier.com/locate/talanta

Screening determination of pharmaceutical pollutants in different water matrices using channel capillary electrophoresis coupled with contactless conductivity detection

dual-Minh Duc Le1, Hong Anh Duong1, Manh Huy Nguyen 1, Jorge Sáiz2, Hung Viet Pham1*, Thanh Duc Mai1 *

Keywords: capacitively coupled contactless conductivity detection (C4D); capillary

electrophoresis (CE); dual-channel CE; water analysis,;pharmaceutically active compounds (PhACs); water samples

Abstract

In this study, the employment of purpose-made dual-channel compact capillary

electrophoresis (CE) instrument with capacitively coupled contactless conductivity detection (C4D) as a simple and inexpensive solution for screening determination of various

pharmaceutical pollutants frequently occurring in surface water and hospital wastewater in Hanoi, Vietnam is reported Five negatively charged pharmaceutically active compounds, namely ibuprofen, diclofenac, bezafibrate, ketoprofen and mefenamic acid were determined using the first channel whereas three positively charged ones, namely diphenhydramine, metoprolol and atenolol were determined with the second channel of the CE-C4D instrument Two different background electrolytes (BGEs) were used in these two CE channels

independently The best detection limits achieved were in the range of 0.2 - 0.8 mg/L without sample pre-concentration Enrichment factors up to 200 were obtainable with the inclusion of

a solid phase extraction step Good agreement between results obtained from CE-C4D and those with the standard confirmation method (HPLC-DAD) was achieved, with correlation coefficients higher than 0.98

1 Introduction

Many pharmaceutically active compounds are classified as environmental contaminants due

to their low biodegradability and their potential to cause undesirable ecological and human health effects [1-5] While having positive effects on the treatment of various pathologies and diseases, the use of pharmaceutical products results in the contamination of the aquatic

environment, mainly through municipal and hospital effluents [6-8] The occurrence of a large number of pharmaceutical contaminants in the environment has been reported for different developed countries (see some recent examples in [7, 9-12]) In emerging countries, although the situation is often worse, only limited information is available as the regulations for

environmental protection are not well established In Vietnam, the first study on

pharmaceutically active compounds contamination was recently reported by Tran et al., focusing on hospital wastewater and surface water in Hanoi [13] Water contamination in Hanoi caused by direct discharge of domestic and industrial wastewater to rivers and lakes poses a big threat to public health This leads to an urgent need for regular control of these contaminants in different water sources

So far, various analytical methods for determination of pharmaceutical contaminants in

different environmental matrices have been proposed [14, 15] Among all techniques, high performance liquid chromatography (HPLC) is the most common due to its high degree of confidence, reliability and reproducibility The utilization of HPLC nevertheless requires long analysis time, a large amount of costly HPLC-grade solvents as mobile phases as well as high setup and maintenance costs (especially for the high pressure pumps and accessories) In Vietnam, only a few central environmental monitoring agencies or companies with abundant funding and sufficient expertise can afford the installation, long-term operation and

maintenance of such instruments On the other hand, capillary electrophoresis (CE), relying

on a high voltage rather than high pressure for driving the analytes in a micro separation channel, offers a more economic and higher-throughput alternative The determination of pharmaceutical residues in water with CE using UV or mass-spectrometric detection has been reported repeatedly (see refs [16-22] for example)

Three additional positive features of CE that make it even more suitable for screening analysis purposes are portability for mobile deployment (see [23-26] and references listed therein), customer-oriented CE configuration for adaptation to different financial and expertise

situations [27, 28] and the possibility of using dual-channel setup for the concurrent

determination of positively and negatively charged analytes [29-31] The portability and the use of more than one channel were enabled when CE was used together with capacitively coupled contactless conductivity detectors (C4D) In these detectors, the difference between the conductivities of the analytes and that of the background electrolyte (BGE) can be

measured by employing a pair of tubular electrodes fitted around the capillary wall Some notable advantageous features of C4D include high versatility, ease in construction and

operation, low power consumption and the potential of miniaturization (for more details see [32, 33] and the references listed therein) CE - C4D has been repeatedly used by Richter et al for determinations of active ingredients together with their counter ions or degradation

products in pharmaceutical formulations [34-39] Simultaneous determination of cations and anions in these pharmaceutical applications were implemented through the use of BGEs with

pH above 7 to produce high magnitudes of electro-osmotic flow (EOF) in fused silica

capillaries for sweeping the anions with low and opposite electrophoretic mobilities towards the detector.In environmental applications, the employment of CE-C4D has been

communicated for separation of negatively and positively charged pharmaceutical

contaminants in standard solutions [40] and for tracing of some negatively charged ones in a

hospital wastewater sample [41] Only one single CE channel was used and no cross check with another well established analytical method was realized in both cases

Herein we report a straightforward and cost-effective method based on purpose-made channel compact CE-C4D for concurrent screening determination of different positively charged pharmaceutical residues (diphenhydramine, metoprolol and atenolol) and negatively charged ones (ibuprofen, diclofenac, bezafibrate, ketoprofen and mefenamic acid) in various water matrices in Hanoi, Vietnam Compared to the recently reported dual-channel CE

dual-instruments that share a common BGE for both channels [31, 42-44], the system reported herein allows the employment of two different BGEs, allowing the separations of analytes belonging to different categories

2 Experimental

2.1 Chemicals and Materials

All chemicals were of analytical or reagent grade and purchased from Fluka (Buchs,

Switzerland) or Merck (Darmstadt, Germany) Stock solutions (1 mM) of diphenhydramine, metoprolol, atenolol, ibuprofen, diclofenac, ketoprofen, bezafibrate and mefenamic were used for the daily preparation of the standard solutions Chemicals used for the preparation of BGEs included: Acetic acid (Ace), histidine (His), 2-(N-morpholino)ethanesulfonic acid (MES), lactic acid (Lac), tris(hydroxymethyl)aminomethane (Tris), 3-(N-

morpholino)propanesulfonic acid (MOPS) and hydroxypropyl-beta-cyclodextrin (HP-β-CD) Fused silica capillaries of 50 µm ID and 365 µm OD were obtained from Polymicro

Technologies (Phoenix, AZ, USA) Prior to their use, the capillaries were pre-conditioned with 1 M NaOH for 10 min and de-ionized water for 10 min, followed by flushing with the BGE The capillaries were then used continuously for successive separations De-ionized

water, purified using a water purification system from Millipore - model Simplicity UV (Bedford, MA, USA), was used for the preparation of all standard solutions and for sample dilution if required Commercial solid phase extraction (SPE) cartridges, including i)

LiChrolut RP-18 with the cartridge volume of 3 mL containing 500 mg sorbents (Merck) and ii) hydrophilic modified, styrene-based polymer (hydrophilic lipophilic balance, HLB) SPE cartridges (200 mg sorbents per cartridge, 30 µm particle size, Waters Corporation), as well as the SPE vacuum manifold (Visiprep 5-7030, Sulpelco) were used for sample treatment and pre-concentration Cross-checking was carried out using an HPLC instrument (LC-20AB) equipped with a UV-VIS-based diode array detector (DAD) from Shimadzu Corp (Japan)

2.2 Instrumentation

All experiments were performed on a purpose-made dual-channel CE instrument The high voltage (HV) modules (DX250 series) capable of providing up to 25 kV were obtained from EMCO (Sutter Creek, CA) The HV end of the capillary was isolated with a safety cage made from Perspex, which was equipped with a microswitch to interrupt the HV on opening The miniature membrane pumps (NF-5-DCB) for sample aspiration were purchased from KNF (Balterswil, Switzerland) Micro-graduated needle valves were obtained from IDEX (P-470, Oak Harbor, WA) and solenoid valves from NResearch (product nos 116T021 and 116T031, West Caldwell, NJ) All fluidic connections were made with 0.02" I.D and 1/16" O.D Teflon tubing and with polyether ether ketone (PEEK) flangeless nuts and ferrules 10-32 or

¼-28 UNF (IDEX) The injection interface that accommodates the grounded end of the

capillary and the ground electrode was machined from a Perspex block (2 cm × 2 cm × 3 cm) Detection was carried out with in-house built miniature HV - C4D according to a design reported previously [43, 45].The resulting signals were recorded with an ADC-20 data

acquisition system (Pico Technology, St Neots, UK) connected to the USB-port of a personal

computer A lithium battery pack of 14.8 V (CGR 18650CG 4S3P, Contrel, Hünenberg, Switzerland) and a separate pair of smaller Li-ion batteries (CGR 18659CG 4S1P, Contrel) fitted with 12 V regulators of appropriate polarities were used for powering the CE-C4D system Mains power can be utilized whenever available

2.3 Field sampling

The sampling sites are located in Hanoi - the capital of Vietnam (see details in Fig S1)

Surface water samples were collected from the Nhuệ river (SN1, SN2 and SN3), Tô Lịch river (TL1 - TL5), Lừ river (SL1), Sét river (SS1), Kim Ngưu river (KN1), the lake near Hanoi Medical University (YHN) and West Lake (HT2) The distances between the sampling sites from the same river or lake were at least 500 m The samples from these rivers and lakes were collected near the municipal discharges at the distance of 1 m from the borders and at the depth of 20 - 30 cm below the surface Untreated and treated wastewater samples were

collected from the influents and effluents of wastewater treatment plants of the central

pharmaceutical manufactory (TW1 and TW2, respectively) and the VCP pharmaceutical manufactory (VCP1 and VCP2, respectively) Water samples from discharges to the receiving points in residential zones were also collected These samples came from a wastewater

discharge (NCT2), Vietnam sports hospital (TT1) and Vietnam national hospital of pediatrics (NH) Totally 20 samples representative for different surface water bodies in Hanoi that may

be contaminated by pharmaceutical pollutants were collected for analyses Water samples were firstly filtered with 0.45 µm membrane filters (Sartorius, Göttingen, Germany), then collected in amber glass bottles and stored at 4 °C (up to one week) The collected water

samples, especially wastewater samples may contain microorganisms whose activity can lead

to modification of the concentrations of pharmaceutical pollutants via microbial degradation

of pharmaceutical compounds [46] The samples therefore were stored at 4°C rather than at

room temperature in order to inhibit / minimize the activity of microorganisms Operations with CE-C4D and HPLC were carried out immediately upon conclusion of the sampling campaign

2.4 Analytical procedures

For pre-concentration of negatively charged pharmaceutical pollutants (i.e ibuprofen,

diclofenac, bezafibrate, ketoprofen and mefenamic acid), the surface water and wastewater samples were passed through RP-C18 extraction cartridges at a flow rate of 1 - 2 mL / min Before sample loading, SPE cartridges were conditioned with 5 mL acetonitrile, 5 mL

deionized water, 5 mL buffer composed of 9 mM Tris adjusted to pH 8 with lactic acid and finally 5 ml water Samples or standard solutions were first acidified to pH 2 - 3 in order to preserve the target pharmaceutical compounds in the protonated forms and then loaded

through the SPE cartridges The loading volumes were adjusted from 200 to 900 mL

depending on the sample matrices Upon completion of the extraction, the SPE cartridges were flushed with 5 mL buffer of 9mM Tris / Lactic acid (pH 8) and vacuum dried for 30 min Elution was then implemented with 6 mL eluent composed of 9 mM Tris / lactic acid, pH 8 (40 % by volume) and acetonitrile (60 % by volume) The pre-concentrated samples were subsequently analyzed with CE-C4D and with HPLC- DAD for cross-checking

For pre-concentration of positively charged pharmaceutical pollutants (i.e diphenhydramine, metoprolol and atenolol), the samples or standard solutions were acidified to pH 4 with

hydrochloric acid (8 M) The acidified sample solution (200 mL) was then passed at a flow rate of 1drop / second through a HLB SPE column that had been conditioned with 6 mL of methanol and 6 mL of deionized water The HLB SPE column was then flushed with

deionized water and air-dried for 5 min before elution was carried out with 10 mL of formic

acid 0.1 % (v/v) in methanol The eluent was subsequently dried with nitrogen and filled to 1

mL with a 2.5 mM His / 2.5 mM MOPS (pH 6.5) solution Shaking was needed for complete dissolution of the eluted analytes in this His / MOPS medium

For CE-C4D separation of negatively charged pharmaceuticals, a high voltage of 15 kV was applied over a capillary with 75 cm total length (Lt) and 65 cm effective length (Leff), The optimal BGE composition found was composed of 36 mM Tris and 0.5 mM HP-β-CD

adjusted to pH 8 with lactic acid As for the positively charged species, the CE-C4D analyses were carried out using the optimized BGE composed of 25mM His, 25 mM MOPS and 1mM HP-β-CD (pH 6.5) A high voltage of 15 kV was applied over a capillary with 80 cm Lt and

70 cm Leff The HPLC - DAD procedure (detection at 260 nm) for cross-checking was

developed according to a protocol reported elsewhere [47]

3 Results and Discussion

3.1 Dual-channel CE setup using individual BGEs for analytical throughput improvement

As numerous medicinal compounds can be present in the aqueous environment [10, 12, 13], it

is desirable to develop a screening method that allows simultaneous determination of a wide range of analytes in one single run In the first work on determination of pharmaceutical contaminants with CE-C4D, 13 pharmaceutical compounds in both anionic and cationic forms

in standard solutions (artificial samples) were successfully separated [40] The satisfactory separation resolution for these standards, with the cationic species being incompletely

separated, however could not be achieved with real samples having complex matrices, at least

in our hands To our experience, separating many species having close electrophoretic

mobilities in a single CE profile is not always the best approach for increasing analysis

throughput because fluctuation of migration times and unidentified (unwanted) peaks from the sample matrices make it harder to identify and quantify the target species In addition, finding

a single BGE that is optimal for all pharmaceutical compounds is not trivial Instead, the use

of dual-channel CE setup with individual BGEs allows independent optimizations of

separation conditions for positively and negatively charged species Concurrent determination

of cations and anions in a single run is an approach for improvement of analytical throughput

A photograph of our compact CE instrument with two independent separation channels is demonstrated in Fig 1 Each channel functions independently as an automated compact CE system The fluidic arrangement and electronic boards were designed to allow

accommodation of two separated CE modules in an aluminum briefcase with dimensions of

45 (w) × 35 (d) × 15 cm (h) (see Fig 1) The fluidic parts of the two CE modules are arranged

on the left The Perplex cage to the right with 2 chambers contains the high voltage electrodes The small metal boxes sitting on the top of this cage are the C4D detectors To the far right of the briefcase are the pneumatic parts for pressurization of the buffer reservoirs The control electronics are located underneath the supporting black Perplex panel and some manual switches and connectors are mounted on the top (at the far upper left corner of the system) A schematic drawing of the system is shown in Fig 2 Precise propulsion of fluids through the two CE channels is carried out by pressurizing two reservoirs of different background

electrolytes with compressed air The samples are loaded into two sample loops each of which

is situated between two 3-way valves Subsequently the sample plugs are moved to the

injector blocks by switching the three-way valves V1-1 and V2-1 to allow the individual background electrolytes to flow from the pressurized reservoirs (see Fig 2) In each CE channel, a fraction of the sample plug is injected into the capillary hydrodynamically by applying a back-pressure for a determined period of time when the plug is located at the front end of the capillary The back-pressures are set by adjustment of the needle valves each of which splits the flow into two paths; and applied for the desired duration by closure of gate valves V1-3 and V2-3 (while V1-4 and V2-4 stay open) Flushing the interfaces and the

manifolds ahead of the interfaces, as well as the capillaries, is possible by either opening or closing V1-3 and V1-4 at the same time for the channel 1, and V2-3 and V2-4 for the channel

2 Compared to previous dual-channel CE setups that share one common buffer for both separation channels [31, 42-44], this instrument has been the first dual-channel configuration that allows utilizing two individual running buffers Concurrent determination of analytes belonging to different categories in a single run is therefore possible High throughput

analyses were therefore implemented without having to sacrifice the separation performance

3.2 Optimization of the CE-C4D and SPE conditions for concurrent separation and detection

of pharmaceuticals in water

The structures of the concerned pharmaceutically active compounds and their respective pKavalues are showed in Table S1 The determination of ibuprofen, diclofenac, bezafibrate, ketoprofen and mefenamic acid by CE is possible when the carboxylic groups are in the deprotonated anionic form The analytical procedure for these compounds was optimized using a Tris/lactic acid (Tris/Lac) buffer as (i) the eluent together with acetonitrile to release the pharmaceutical compounds retained on the sorbents after the SPE preconcentration process and (ii) the CE-C4D running buffer to assure de-protonation and thus anionic

ionization of the target compounds Optimal pH of the Tris/Lac BGE was investigated from 7

to 9 by varying the concentration of lactic acid As for CE-C4D a dependence of the peak heights on the buffer concentration is observed [48], the Tris concentration in the Tris/Lac BGE was varied from 9 mM to 45 mM for signal height optimization It was found that a buffer of 36 mM Tris adjusted to pH 8 with lactic acid offered the best signal-to-noise ratios (data not shown) In order to obtain baseline separation of the five target pharmaceutical compounds, recourse to HP-β-CD, which is a complexing reagent, is needed Like other cyclodextrins, HP-β-CD has a toroidal shape with hydrophilic external surface and relatively

hydrophobic cavity [40, 49] Different pharmaceutical compounds have different interactions

of their hydrophobic groups with the cavity of HP-β-CD As HP-β-CD is neutral, the to-size ratios of the analyte ions are reduced after complexation In addition, the hydroxyl groups of HP-β-CD could form hydrogen bonds with the analytes These in turn lead to better selectivity in electrophoretic separations and achievement of baseline separation of the peaks The effect of varying HP-β-CD concentrations in the Tris/Lac BGE on the separation

charge-performance is shown in Fig 3 The electrophoretic mobilities of mefenamic and ketoprofen were most affected by the addition of this complexing reagent The best separation was achieved with HP-β-CD concentration of 0.5 mM As HP-β-CD does not possess any charge, its presence does not result in any change in conductivity of the BGE Using the optimized BGE composed of 36 mM Tris and 0.5 mM HP-β-CD adjusted to pH 8 with lactic acid, the migration time of EOF to the detector was determined to be 890 s The EOF arrived at the detector much earlier than the tested negatively charged pharmaceutical pollutants and

therefore did not affect the electrophoretic separation of these compounds

For CE-C4D determination of diphenhydramine, metoprolol and atenolol, their amino groups need to be protonated to render them positively charged The following BGEs that have pH values lower than the analytes' pKa values were tested: i) 9 mM Tris / 5 mM Lac (pH 8), ii)

50 mM Tris / 50 mM MOPS (pH 7.7) iii) 9 mM Tris adjusted to pH 5.4 with acetic acid, iv)

25 mM His / 25 mM MOPS (pH 6.5), v) 50 mM His adjusted to pH 4 with lactic acid and vi)

10 mM His / 10 mM MES (pH 6) The electropherograms obtained with these BGEs are shown in Fig 4 As can be seen, the best performance in terms of high peaks and baseline separation were achieved with the BGE containing 25 mM His, 25 mM MOPS and 1mM HP-

β -CD The migration time of the EOF to the detector was determined to be 1300 s with the optimized BGE The EOF was well separated from the concerned positively charged

pharmaceutical pollutants and therefore did not affect the electrophoretic separation of these compounds Compared to the previous work by Quek et al [40], improved separation

performance was obtained with our optimized conditions with which diphenhydramine, metoprolol and atenolol were completely separated

The salient performance data for the direct determinations without pre-concentration of the target pharmaceutical contaminants with the optimal CE-C4D conditions are shown in Table

1 The best detection limit achieved for the conditions is 0.2 mg/L and calibration curves were acquired up to 50 mg/L For higher concentrations peak overlap would occur The

reproducibilities of the measurements of peak areas and migration times were found to be better than 6 % and 1 %, respectively In order to minimize the risk of peak overlap due to the possible presence of unidentified compound(s) when working with real-world environmental samples, the CE-C4D optimizations were implemented towards achievement of good

separation resolutions rather than very fast separations for a high analytical throughput This therefore came with a compromise of the separation time (up to 20 min) The throughput in our case was improved with the use of dual separation channels instead of conventional single channel setups

The SPE procedure for matrix removal and pre-concentration of ibuprofen, diclofenac,

bezafibrate, ketoprofen and mefenamic acid was adapted from our previous work for

ibuprofen, diclofenac, bezafibrate and naproxen [41] Different parameters, i.e loading

volumes (200 - 900 mL), elution volumes (2 - 10 mL), eluent composition

(Tris/Lac/acetonitrile) were re-optimized to adapt to different matrices from various sample sources The SPE cartridges could be tolerant of a loading volume up to 600 mL for ‘dirty’ samples from Tô Lịch and Lừ rivers until they were completely blocked For ‘cleaner’

samples, higher loading volumes were possible A volume larger than 900 mL was

nevertheless avoided as otherwise it would require too long preconcentration time at the loading speed of 1 - 2 mL/min Elution volumes of more than 6 mL were found to give

satisfied elution efficiency with the recovery being higher than 94% (data not shown) With an elution volume of 6 mL, enrichment gains of 30 - 150 (based on the volume ratio calculation with the assumed SPE recovery of 100 %) were achieved after this sample treatment process These theoretical enrichment factors would decrease with a reduction of the SPE efficiency

To evaluate the SPE recovery, the optimized SPE procedure was then tested with different matrices containing these negatively charged pharmaceutical compounds spiked at 40 µg/L

As shown in Table S2, the highest recoveries (more than 97%) were achieved with the matrix prepared from de-ionized water An enrichment factor of 150 at the SPE recovery of 100 % accordingly decreased to 145 when the recovery was reduced to 97 % The recoveries

decreased to around 90 % when SPE was carried out with a waste water matrix Competitive adsorption of other interfering compounds from the waste water matrix onto the sorbents may

be a reason for this reduced SPE performance In all cases, satisfied recoveries were always achieved, regardless of the water matrix tested

For preconcentration of cationic pharmaceutical pollutants (i.e diphenhydramine, metoprolol and atenolol), C18 and HLB SPE cartridges were tested; and their performance were

compared The alkalization of the samples to pH ≥ 10, which is needed to render these

compound deprotonated to be retainable on a hydrophobic C18 column, nevertheless results in possible precipitation of different cationic species in the sample matrix This would lead to potential performance loss due to i) possible adsorption of pharmaceutical pollutants on the newly formed precipitates and ii) (partial) blocking of the SPE cartridge due to these

precipitates The other SPE option, using HLB columns that favor the retaining of the analytes

in the hydrophilic (protonated) states, was therefore implemented Optimization of the SPE conditions for diphenhydramine, metoprolol and atenolol using HLB columns was

implemented according to the previous work by Maldaner and Jardim [50] Sample

acidification prior to SPE and elution at pH 6.4 were needed to achieve a good SPE

performance [50] Conveniently, the optimized BGE for separation of diphenhydramine, metoprolol and atenolol with CE-C4D (i.e His / MOPS, pH 6.5) could be used directly as the eluent to desorb these compounds from the HLB sorbents In order to maintain the stacking effect in electrophoresis to achieve high and sharp peaks, the conductivity of the eluent that serves also as the sample matrix to be injected into the CE-C4D system should be significantly lower than that of the BGE [41] A 10-fold diluted BGE (i.e 2.5 mM His / 2.5 mM MOPS,

pH 6.5) was therefore employed as the eluent With a loading volume of 200 mL and a final volume after elution of 1 mL, an enrichment factor of 200 was obtained based on volume calculation at the theoretical SPE recovery of 100 % The recoveries for diphenhydramine, metoprolol and atenolol spiked in de-ionized water obtained with the optimized

preconcentration conditions were 108 %, 99.4 % and 26.9 % respectively (see table S2) These values decreased to 88.3 %, 95.6 % and 24.6 % when a waste water matrix (TW1) was tested The respective enrichment factors obtained with these reduced SPE recoveries were

176, 191 and 49 The poor recovery obtained for atenolol was possibly due to its stronger retaining on the HLB cartridge induced from the two amino groups in its structure A similar phenomenon was also observed by Maldaner and Jardim when carrying out SPE

preconcentration of atenolol using HLB sorbents [50] Metoprolol and diphenhydramine on the other hand possess only one amino group in their structures, which in turn may render desorption of these compounds from the HLB particles more facile Efficient stripping of atenolol from a HLB cartridge therefore would require a more severe eluting condition Note that offline SPE preconcentration was implemented instead of on-line coupling of the SPE

module to the CE setup By doing so, a high SPE throughput could be achieved by carrying out preconcentration of many samples using several SPE cartridges at the same time Several preconcentrated samples thus could be collected simultaneously and made ready for the subsequent CE-C4D analyses The CE-C4D process therefore is not affected or delayed by the long extraction time in the SPE preconcentration step

3.3 Monitoring of pharmaceutical pollutants in different water matrices

The purpose-made dual-channel system employing the optimized SPE-CE-C4D conditions was then used to analyze different target pharmaceutical pollutants in different water bodies in Hanoi (Vietnam) whose population is approximately 6.5 million and water consumption is estimated about 150 L per capita per day [51] Hanoi is the place where most of the hospitals

in the North of Vietnam, including 32 big hospitals, more than 200 medical centers and 2000 private clinics, are located For this reason, among the municipal and hospital wastewater discharges in Hanoi (approximately 800,000 m3/day), the latter contributes considerably to the total discharge [51] Water from this discharge is not treated and is utilized for different activities, such as fish farming, vegetable watering and household use, which may lead to direct and / or indirect intakes of pharmaceutical pollutants CE-C4D profiles for the analyses

of negatively charged pharmaceutical pollutants in various samples taken from different surface and wastewater sources after the enrichment process are shown in Fig 5 The cross-checking results with HPLC-DAD are shown in Table 2 The deviation of less than 10% (see Table 2) and coefficients of determination (r2) for the two pairs of data of more than 0.98 were achieved, which proved very good agreement between the results obtained from two methods Ibuprofen was found in many samples with the concentrations determined with CE-C4D of 8 -

40 µg/L Other negatively charged pharmaceutical pollutants were also detected in some water samples Tô Lịch river and the water outlet of Vietnam sports hospital were

contaminated with more than one pharmaceutical pollutant, i.e ibuprofen, mefenamic and bezafibrate Some unidentified peaks also appear in the CE-C4D profiles of these samples It

is thus speculated that some other pharmaceutical contaminants may also be present in surface water and wastewater in Hanoi Further effort to identify these species however was not made

in the scope of this work Diphenhydramine, metoprolol and atenolol were not detected in all tested samples, with an exemplary electropherogram obtained with the waste water TW1 illustrated in Fig S2 Therefore no data for these compounds were included in the table 2 In order to demonstrate that the developed methodology works well will real-world samples, trace amounts of diphenhydramine, metoprolol and atenolol were spiked into the sample TW1 and the CE-C4D procedure was repeated As can be seen in Fig S2, the signals of these spiked cationic species were well detected Note that the developed SPE-CE-C4D methodology is meant for screening operation to preliminarily indentify the contamination sources Other well-established (sophisticated) separation techniques, such as HPLC-MS/DAD or GC-MS are always recommended when available for further confirmation of the presence of

pharmaceutical contaminants in the water sources showing positive results with SPE-CE-C4D before any legal reporting and further treatment action will be correspondingly implemented

4 Conclusions

A purpose-made automated CE-C4D instrument with the option of dual-channel setup for high throughput operation was successfully applied for the concurrent screening determination of different pharmaceutical pollutants in various surface water and wastewater matrices in Hanoi (Vietnam) where many hospitals and clinic centers are located Cross check with the reference method (HPLC-DAD) proved the reliability of the results obtained with CE-C4D With the simplicity of the operation, the cost-effectiveness of the instrument and chemicals used, CE-

C4D can be seen as a low-cost screening method for quick identification of pharmaceutical