Oxidized buckypaper for stir-disc solid phase extraction: evaluation for

Subscriber access provided by UNIVERSITY OF THE SUNSHINE COAST

Article

Oxidized buckypaper for stir-disc solid phase extraction: evaluation for several classes of environmental pollutants recovered from surface water samples Pierpaolo Tomai, Andrea Martinelli, Stefano Morosetti, Roberta Curini, Salvatore Fanali, and Alessandra Gentili Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b00927 • Publication Date (Web): 29 Apr 2018 Downloaded from http://pubs.acs.org on May 3, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Stir-disc solid phase extraction with oxidized buckypaper as sorbent membrane 84x47mm (300 x 300 DPI)

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 27

Oxidized buckypaper for stir-disc solid phase extraction: evaluation for several classes of environmental pollutants recovered from surface water samples Pierpaolo Tomai†, Andrea Martinelli†, Stefano Morosetti†, Roberta Curini†, Salvatore Fanali‡, Alessandra Gentili†*. †

Department of Chemistry, University of Rome “La Sapienza“, Piazzale Aldo Moro n°5, P.O. Box

34, Posta 62, 00185 Roma, Italy. ‡

PhD School in Natural Science and Engineering, University of Verona, 37129 Verona, Italy.

Istituto di Metodologie Chimiche, Consiglio Nazionale delle Ricerche (C.N.R.), Monterotondo, Roma, Italy.

*Corresponding author: Fax number: + 39-06-490631. E-mail address: [email protected] (A. Gentili)

Field Code Changed

1


Page 3 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Abstract This paper describes, for the first time, the use of oxidized buckypaper (BP) as a sorbent membrane of a stir-disc solid phase extraction (SD-SPE) module. The original device, consisting of a BP disc (d = 34 mm) enveloped in a polypropylene mesh pouch, was designed to extract organic micropollutants (OMPs) from environmental water samples in dynamic mode. High-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) was used to analyze the extracts. Several classes of pesticides and pharmaceuticals were chosen as model compounds to evaluate key parameters affecting the recovery rates. To this end, the effects of High-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) was used to analyze the extracts. Aadsorption time, desorption time, stirring speed, kind type and volume of solvent, and sample volume were the thoroughly studiedexamined parameters. After optimization, a novel and in-depth study was conducted to findA a correlation between physicochemical properties of the analytes and extraction yields was thoroughly investigated. Recoveries were mainly governed by a combination of logP and pKa values. Indicatively, hydrophilic compounds with a logP 1 exhibited recoveries ranging between 50% and 100% depending on their pKa, while compounds with pKa between 6 and 7.5 gave low yields irrespective of their logP. The analytical method was also validated and tested as large scale screening method of OMPs in surface waters. The analysis of real samples revealed the presence of some non-steroidal antiinflammatory drugs, sulfonamides and pesticides at low ng L−1 concentration levelss with relative standard deviations lower than 8%.

Keywords: buckypaper; solid phase extraction; sample preparation; membranes; carbon nanotubes; environmental samples; LC-MS.

2



Page 4 of 27

Among sample preparation techniques currently employed, solid phase extraction (SPE) is undoubtedly the most used one.1,2 There is a large variety of commercially available cartridges, typically packed with classical microparticles (50-60 µm) of silica, polymeric and carbonaceous materials. A careful choice of the sorbent, made according to the nature of analytes and matrix of interest, allows obtaining high recovery efficiencies and enrichment factors between 20 and 1000. Over the years, with the intent to preserve its basic advantages, various options have been studied to widen the SPE applicability and circumvent its limitations.3 In this context, efforts have been devoted to study new materials, formats and realization modes to avoid the drawbacks related to channeling phenomena, breakthrough and occlusion of the cartridge when particularly complex samples are processed (e.g. waste water, surface water, milk, etc.). As far as the format is concerned, disc SPE was the first solution devised to abate plugging phenomena by suspended particulate matter and to provide increased processing rates. Currently, at least three different disc formats are available: i) particle-loaded membranes, ii) particle-embedded glass-fiber discs, and iii) laminar discs. In the first two types, the sorbent microparticles are immobilized in polytetrafluoroethylene or glass fibers and the whole device is laid on a porous support. In the laminar discs, a bed of microparticle sorbent is packed between two filters. Discs are commonly used in a similar way as cartridges, i.e. letting the sample pass through the device in a manner analogous to filtration (flow-through mode). Compared to cartridges, discs offer some clear advantages that favor their use especially for environmental applications where large sample volumes (0.25-2 L) are processed to reach ng-pg L-1 limits of quantification (LOQs). The larger cross-sectional area and decreased pressure drop allow higher loading rates and, consequently, shorter processing time. Discs also have an increased contact surface per unit bed mass which facilitates interactions with analytes. The use of smaller particles (10-30 µm) and the greater mechanical stability minimize channeling which, on the contrary, can be a serious effect of cartridges due to their low packing density.

3


Page 5 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Sorbent-based sorptive microextraction (SBSM) techniques, such as membrane-stir extraction,4 stirbar sorptive extraction5 and stir fabric phase sorptive extraction,6 are other modes to realize SPE: they make use of a thin fiber covered by a small amount of sorbent and take advantage of magnetic stirring to speed up sorption and desorption of analytes.3,7,8 Well-known for their overwhelming advantages, SBSM techniques usually do not produce an exhaustive extraction since, unlike SPE, they are equilibrium-based approaches.9 Moreover, the limited sorbent loading capacity can lead to high detection limits not enough to detect ultra-traces of OMPs occurring in environmental samples.10 As far as sorbent materials are concerned, typical porous phases, such as silica-based gels (C18, C8, etc) and co-polymer resins (styrene-divinylbenzene, poly(divinylbenzene-N-vinylpyrrolidone, etc.), display surface areas ranging from 400 to 800 m2 g-1, while carbonaceous materials (e.g. nonporous graphitized carbon-black, porous graphitic carbon) exhibit the more modest values of 100210 m2 g-1.411 The chemical stability of bonded-phase silica is undermined at pH extremes (i.e., pH8), but that of polymer and carbon-based sorbents is preserved throughout the full pH range. All of these materials are easily available on the market at low-moderate costs. Over the last few years, the development of new sorbents has been object of unceasing research activities. The unique physicochemical, mechanical and morphological properties of carbon nanostructured materials make them ideal candidates for SPE operations.3,512-815 Carbon nanotubes (CNTs) have been the most investigated allotropic forms.916 A CNT can be regarded as a minute cylinder resulting by the rolling up of either an individual graphitic sheet (single-walled CNT, SWCNT) or concentric multiple sheets (multi-walled CNT, MWCNT). The distortion of the planar graphene sheet creates fluctuating induced dipole moments responsible for van der Waals and π-π stacking interactions which account for an excellent extraction capability of non polar, moderately polar and aromatic organic compounds.1017 Strong oxidizers are able to provide edges of CNTs and strained/defective sites on their sidewalls with hydroxyl, carbonyl and carboxyl groups mainly the edges of CNTs but also the strained or defective sites on their sidewalls. The introduced polar 4



functionalities

make

it

the

surface

available

for

further

Page 6 of 27

surface

chemical

functionalizationsfunctionalization with ester or amide groups.1017,1118 Another key advantage is the high specific surface area (SSA) even if its values are broaden on a very large range (1315-50 m2g1

) depending on number of walls and/or diameter of CNTs.1219 When CNTs are packed into a SPE

cartridge, the SSA is reduced to a large degree because of spontaneous aggregation of CNTs into bundles. During the sample loading and elution steps, the sorbent bed becomes more and more compact, causing high backpressure and long operational time. To overcome such undesired drawbacks,

CNTs

have

been

used

as

discs,1320

combined

with

magnetic

nanoparticles14nanoparticles21 or employed in microextraction devices.916,1522 The remarkable van der Waals forces of adhesion exerted by CNTs have been exploited to prepare CNT-based membranes which have opened up attractive prospects for sample preparation.1623 There are three main arrangements of CNTs in membranes: i) vertically aligned CNTs immobilized onto polymeric membranes, ii) bundles of CNTs on inert membranes and iii) self-supporting entangled assemblies of CNTs, the latter otherwise known as buckypaper (BP). Although such membranes can be prepared in formats of different shapes (e.g. disc, square) and size for SPE operations, the number of current applications has still been very limited. Niu et al.1320 prepared SWCNT discs and used them in dynamic flow-through mode to extract sulfonylurea herbicides from large-volume water samples (up to 3 L). Nevertheless, since the lab-made disc was mechanically fragile, a filter paper was used as support and only one side on the membrane was suitable for adsorption. BP in the strict sense was used for the first time by our group to assist a microparticle copolymeric sorbent in the SPE clean-up of cobalamins from milk.1724 Since BP (two discs of 12-mm), was packed into the cartridge, its absorption capability was not exploited to the full since the dynamic flow-through mode did not permit to achieve an adequate contact time with the analytes either in the loading or in the elution step, so its absorption capability was not completely exploited. In this work, a new stir-membrane deviceis proposed for capitalizing the full potential of BP as sorbent for SPE operations. Analogous sorbent-based sorptive microextraction (SBSM) techniques, 5


Page 7 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


such as membrane-stir extraction,18 stir-bar sorptive extraction19 and stir fabric phase sorptive extraction20, make use of a thin fiber covered by a small amount of sorbent.3,21,22 Well-known for their overwhelming advantages, SBSM techniques are similar to SPE with the difference that, being equilibrium-based approaches, they usually do not produce an exhaustive extraction.23 Moreover, the limited sorbent loading capacity can lead to high detection limits not enough to detect ultratraces of OMPs occurring in environmental samples.24The aim of this work is to present a novel SPE extraction device which integrates the main advantages of disc-SPE with the properties of CNTs and magnetic stirring. To this end, a dedicated extraction unit was developed, optimized, and quantitatively evaluated to capitalize the full potential of BP as sorbent. The here presented device has beenwas conceived to overcome the typical disadvantages of SPE cartridges and the abovementioned limitations of the analogous SBSM techniques. It relies on a nanoporous oxidized CNTbased membrane that is completely composed of sorbent material, displays the high-desirable nanomaterial properties and can be designed in several shapes, sizes and realization modes. In the present design, the device is particularly suitable for environmental applications. For this reason, it was tailored for extraction, clean-up and pre-concentration of OMPs in from environmental waters. Different classes of compounds including drugs, hormones and pesticides, covering a broad spectrum of polarities, were selected as model analytes to study the correlation between their physicochemical properties and adsorption capability of oxidized BP. As a result, no disc SPE device with such unique characteristics has been reported to date.

EXPERIMENTAL SECTION Chemicals,

Materials

and

Solutions.

Acifluorfen,

procimione,

17-β-estradiol,

17-α-

ethynylestradiol, 17-α-estradiol, diethylstilbestrol, estriol, equilin, progesterone, testosterone, methyl-testosterone, testosterone propionate, trenbolone, acetaminophen, acetylsalicylic acid, carprofen, diclofenac, etodolac, flunixin, ibuprofen, ketoprofen, meclofenamic acid, naproxen, 6



Page 8 of 27

nimesulide, phenylbutazone, salicylic acid, tolfenamic acid, β-zearalenol, sulfachloropiridazine, sulfadiazine, sulfaguanidine, sulfamerazine, sulfamethoxazole, sulfamonomethoxine, sulfanilamide, sulfaquinoxaline, sulfathiazole, tapazole, mercaptobenzimidazole, thiouracil, butocarboxim, carbaryl, methomyl, pirimicarb, propoxur, bensolfuron methyl, cinosolfuron, diuron, linuron, isoproturon, metobromuron, metsolfuron methyl, monolinuron, rimsolfuron, thifensulfuron methyl, 2,4-dichlorophenoxyacetic

acid,

2-methyl-4-chlorophenoxyacetic

acid,

carbendazim,

imazamethabenz-methyl, imazaquin, iprodion, 4-nitrophenol, atrazine, picloram, desmedipham, phenmedipham, dimethoate, chlorosolfuron, 2,4 dichlorophenol, 4-chloro-2-methylphenol were purchased

from

Aldrich–Fluka–Sigma

S.r.l.

(Milan,

Italy).

Mecoprop,

4-(4-chloro-2-

methylphenoxy)butanoic acid, malathion, fenthion, acephate and flamprop were purchased from Dr. S. Ehrenstorfer Promochem (Wesel, Germany). All standards had a purity grade greater than 93%. Methanol, dichloromethane and acetonitrile of RS grade (elevated purity grade) were purchased from Aldrich–Fluka–Sigma S.r.l. (Milan, Italy). Sodium chloride (NaCl), fFormic acid 50% and nitric acid 65% were obtained from Aldrich–Fluka–Sigma S.r.l. and Carlo Erba, respectively. Ultrapure water was obtained from a Milli-Q water generator (Millipore, Bedford, MA, USA). Structures, exact masses, logP and pKa of all 76 model compounds are shown in Table S-1 in the Supplementary Material. Commercial BP was purchased from Nanolab, Inc.(Nanolab,Waltham, USA). The stock solutions of each analyte were prepared dissolving weighed standard amounts (OHAUS DV215CD Discovery Semi-Micro and Analytical Balance 81 g/210 g capacity, 0.01 mg/0.1 mg readability) in methanol at a concentration of 1 mg mL-1. Working standard solutions were prepared from the individual stock solutions by dilution with methanol at concentrations of 10 ng µL-1 and 1 ng µL-1. All standards and solutions were stored in the darkness at 4 °C. The polytetrafluoroethylene (PTFE) filters (0.45 µm, 25 mm i.d.; 0.45 µm, 13 mm i.d.) were purchased from Alltech Associates Inc.

7


Page 9 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Environmental Samples. Natural water samples were collected from Lakes Vico, Bracciano and Martignano and from different locations along the River Tiber: Farfa natural area; Tiber Island in the center of Rome; Marconi Bridge in Rome’s suburbs; Fiumicino at the mouth of the River Tiber , Italy (Figure S-1). All samples were collected in 5-L dark glass bottles. Once in the laboratory, the samples were filtered through 1.2 µm Whatman glass microfiber filters (Whatman International Ltd, Maidstone, UK) and stored at 4 °C until the extraction. Samples from Lake Martignano Lake were used as blank matrices for the method validation. Design of the stir-BP device and BP activation. The BP used in this work is a porous flexible felt of 17.5 x 22.5 cm2 and about 0.15–0.25 mm thickness, composed of entangled unoriented oxidized MWCNTs. According to the supplier information, free-standing continuous nanotube sheet was prepared from purified CNTs, after hydrochloric and nitric acid-treatment. The oxidized CNTs were first suspended in water with a surfactant, then filtered and dried on a membrane and, finally, the layer of BP was peeled off the membrane. As evaluated by our group in previous works,1724,25 BP from Nanolab possesses a specific surface area (110.6 m2g-1), and a porosity of about 80 % and an O/C atomic ratio of 0.23.. The photograph of the extraction device is displayed schematically represented in Figure 1, while components, assembly and operational steps are . schematically represented in Figure 2. EachA BP disc (d = 34 mm) was cut by using scissor and inserted into a polypropylene mesh pouch whose edges were heat-welded to maintain the mechanical integrity of BP during stirring. The pouch was submerged in a nitric acid water solution (65%) for 2 h and, then, thoroughly rinsed with Milli-Q water to remove the acid and any residue of catalyst and surfactant from BP. Subsequently, methanol was used to wash off carbonaceous impurities (3 x 6 mL) and Milli-Q water to condition the membrane (3 x 6 mL). After these treatments, the BP disc was fixed on the top of a Teflon hollow cylinder (dint= 1 cm, dext = 1.9 cm, h = 2.1 cm), endowed with a hole for the insertion of a magnetic stir bar on the bottom.

8


Formatted: Font: Bold


Page 10 of 27

Stir-disc solid phase extraction. A 600-mL beaker was filled with 500 mL of surface water. The extraction unit was immersed into the aqueous sample and left under magnetic stirring overnight (about 16 hours) to favor analyte absorption on BP membrane. At the end of this period, the pouch with BP was freed from the assembly. A small magnetic stir bar was inserted into a narrow polypropylene tube fixed on a side of the pouch (see Figure 12), which was then placed on the bottom of a 75-mL weighing bottle. The analytes were eluted with 9-mL of CH2Cl2:CH3OH (50:50, v/v) under continuous stirring for 30 minutes. The solvent volume was enough to submerge the pouch. The organic fractionextract was collected in a 2-cm i.d. glass tube with a conical bottom and evaporated to dryness in a water bath at 40°C under N2 flux. The residue was dissolved in 500 µL of H2O: CH3CN (80:20, v/v) solution; after filtration, a volume of 40 µL was injected into the HPLC-MS/MS system. High-performance liquid chromatography-tandem mass spectrometry. Liquid chromatography was performed by using to a micro HPLC/ autosampler/ vacuum degasser system PE Series 200 (Perkin Elmer, Norwalk, CT). The seventy-six model compounds were separated on an Alltima C18 column (250 × 4.6 mm, 5 µm) (Alltech, Deerfield, IL, USA) equipped with a precolumn of the same type (7.5 × 4.6 mm, 5 µm). Mobile phases consisted of water (phase A) and acetonitrile (phase B), both of them containing 2 mM formic acid. The gradient profile was as follows (t in min): t0, A = 80%; t20, A = 100%; t20.5, A = 0%; t40, A = 0%. The flow rate of the mobile phase was 1 mL/min, but only 200 µL/min of the LC column effluent was diverted to the Turbo V source of the mass spectrometer. Analytes were detected and quantified by an API 4000 Qtrap® (ABSCIEX, Foster City, CA, USA) mass spectrometer equipped with an Electrospray Ionization (ESI) probe on the Turbo V source and an Electrospray Ionization (ESI) probe. The detection was performed in dual-polarity by setting the capillary voltage at 5500 V for the positive ionization and at -4500 V for the negative one. The source heather temperature was fixed at 450◦C to warm the drying gas. High-purity nitrogen was used as a curtain gas (5 L min-1) and collision gas (4 mTorr), while air as nebulizer gas (2 L min-1) 9


Page 11 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


and drying gas (20 L min-1). Unit mass resolution (0.6 ± 0.1 m/z full width at half maximum, FWHM) was set in each mass-resolving quadrupole. Quantitative analysis of the target analytes was carried out in scheduled multiple reaction monitoring (SMRM) mode, selecting one MRM transition per analyte. The Scheduled MRM™ algorithm was used with an MRM detection window of 90 s in the retention window characteristic of each analyte (tr ± 90 s) and a target scan time of 0.3 s. A settling time of 50 ms was used for polarity switching. A total of 25 transitions in positive mode and 51 transitions in negative mode were monitored with an MRM pause time of 2 ms. Table S-2 lists the HPLC-ES(±)-SMRM parameters employed for identification and quantification. HPLC–MS/MS data were acquired and elaborated by Analyst® 1.6 Software (AB Sciex).

RESULTS AND DISCUSSION The aim of this work is to present a novel SPE extraction device which integrates the main advantages of disc-SPE with the properties of CNTs and magnetic stirring. To this end, a dedicated extraction unit was developed, optimized, and quantitatively evaluated. Optimization of the extraction unit and procedure. The impact of some parameters on the extraction yields of this novel device was studied by planning a series of tests to assess their actual influence. Method Ooptimization experiments were carried owas doneut by using 500 mL of MilliQ water samples spiked at 4 µg L-1 with the 76 model compounds. In order to choose the format of the BP disc, we considered that a classical one-side 47-mm SPE disc has a surface of 1735 mm2. Since BP has two sides available for adsorption, a diameter of 34 mm was enough to obtain a disc with an effective surface of 1816 mm2 (908 mm2 per side). Sorption step. Sorption on BP is a complex process involving the analyte transfer from the aqueous sample solution to the porous solid phase and resulting in: i) infiltration of molecules into the sorbent bulk (absorption); ii) migration into the smaller pores (diffusion); iii) .interaction of molecules with the sorbent surface via Van der Waals, dipole-dipole, hydrogen-bonds and electrostatic interactions (adsorption). There are several parameters that can influence these events 10



Page 12 of 27

such as contact time, stirring rate, sorbent mass, pH in the case of acid or basic compounds, temperature, and sample volume. In order to simplify the study, all operations were carried out at ambient temperature, and sorbent mass, and sample volume were kept constant. For the latter, optimization will be discussed later (see section “Sample volume”). Moreover, all operations were carried out at ambient temperature. At this level, sample pH was still not taken into account because of the great variety of the model compounds’ pKa values. Therefore, attention was focused on , while sample volume will be evaluated later (see section “Sample volume”). Sstirring rate and adsorption time, parameters that can are crucial parameters affecting diffusive phenomena crucially. Evaluation of the optimal value of stirring rate was conducted in the range from 90 to 170 rpm, in order to avoid the central vortex that could precluding the complete contact sample-BP. Actually, since the extraction remained constant in this range, an intermediate value of 130 rpm was chosen for the following experiments. Sorption time was assessed in the range 0-40 h; desorption was carried out with two 10-mL fractions of a mixture methanol: dichloromethane (50:50, v/v); each point was in triplicate. Since an analogous trend was observed for all analytes, the average of their normalized chromatographic areas (Ā) was plotted against sorption time (Figure 2). As can be observed, Tthe resulting curve increases steeply from 0 to 3 h and gradually approaches its maximum value at 24 h. The experimental data indicates that, like with for other porous membrane-based techniques, the adsorption kinetics is slow because it is controlled by the analyte diffusion inside of the smaller pores of BP. In fact, in our previous studies, the morphological analysis showed the presence of macropores and mesopores, the last ones having a distribution that ranges from 35 to 12.5 Å.25 However, taking into consideration both sensitivity and sample throughput, a 16-h contact time was selected as the optimum value for practical reasons. In fact, in this way, the sorption can be carried out overnight, while the remaining operations (desorption and evaporation) can be conveniently and rapidly completed the next morning.

11


Page 13 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Desorption step. As far as the desorption solvent is concerned, dichloromethane and methanol were chosen as extractants, also considering their described effectiveness for graphitized carbon black SPE .26 Dichloromethane/methanol mixtures were prepared in proportions 80:20, 50:50 and 20:80 v:v. Three experiments were conducted in parallel, eluting the analytes with two 10-mL fractions of each mixture. Obviously, depending on the logP of the model compounds, the more hydrophobic of them were preferentially eluted when the fraction of dichloromethane was higher, their interaction occurring preferentially with the graphenic portion of BP. On the other hand, polar compounds, interacting more tightly with the surface oxidized groups of BP, were better eluted when a higher proportion percentage of methanol was used. However, considering athe multi-residue nature of the method such as thathere presented in this paper, the best result was globally obtained with the mixture dichloromethane:methanol (50:50, v:v). Solvent volume, stirring rate and desorption time were optimized together applying a 23 full factorial design. In such type of design, each of the 3 factors runs at 2 levels for a total of 23 = 8 combinations of the levels. The full model, Y = b0 + b1X1 + b2X2 + b3X3 + b12X1X2 + b13X1X3 + b23X2X3 + b123X1X2X3, accounts for three main effects, three two-factor interactions, one threefactor interactions, and allows estimating the 8 biJ coefficients. The complete tabular form of this design with random run order and two center point runs is presented in the Supplementary Material (Table S-36). The two selected levels were: 6 mL (-1) and 12 mL (+1) for solvent volume; 90 rpm (-1) and 170 rpm (+1) for stirring speed; 20 min (-1) and 30 min (+1) as desorption time. To evaluate the performance of desorption, the experimental areas were normalized to give the same weight to all analytes and, then, averaged out. Since desorption time was the only significant (tStudent at 95%) and positive factor, it was set at the high level. Times longer than 30 min were not tested as not interesting for a practical use of the extraction device. The other two factors were selected at the center points. Sample volume. Finally, under the optimized conditions, a study on the sample volume that can be processed by the developed stir-disc SPE device was carried out by extracting . To this end, 12



different volumes (0.25 L, 0.5 L, 1 L) of an unspiked sample of natural water (from the River Tiber at Fiumicino) were extracted in triplicates. The enrichment factor obtained by processing 0.25 L was not enough to detect all the occurring OMPs observed by extracting 0.5-L and 1-L samples. On the other hand, compared to 0.5-L aliquots, tThe extraction of 1-L samples provided analyte areasS/N ratios only slightly greater and, hence, did not allow the determination of a higher number of analytes. Therefore, the 0.5-L volume was identified as the best compromise solution between proper enrichment factor and sample size and was selected as the sample volume to conduct the study of validation and feasibility.

Analytical method validationand application to real environmental samples. The LC-scheduled MRM method was validated by studying recovery, precision, linearity, sensitivity, limit of detection (LOD), and limit of quantification (LOQ) (see Tables S-3-S-5 for the results). Internal standardization was not applied in this study.. Matrix-matched calibration curves. For calibration curves, six 500-mL -aliquots of surface water were spiked with increasing concentrations of the analytes. Concentration ranges spanned over 2 orders of magnitude: 0.05-1 µgL-1, 0.05 – 2.5 µg L-1or 0.5-10 µg L-1. Extraction and analysis were carried out according to the protocol and conditions described in the Experimental Section. The curves were interpolated by linear regression, reporting chromatographic peak area against level of fortification. Sensitivity, accounted by the curve slope, depended on response to ES detection and on analyte losses during extraction procedure. Coefficients of determination (R2) were greater than 0.90 (Table S-3). Recovery, precision, LODs and LOQs. Recovery and precision were estimated fortifying five 500 m-L -aliquots (five replicates) of natural water with the analytes at three levels of concentration: 1 µg L-1, 10 µg L-1 and 50µg L-1. To calculate extraction efficiency, another aliquot was extracted and fortified post-extraction with the same nominal concentrations of the analytes. The six aliquots were analyzed in the same analytical session to evaluate recovery and intra-day precision at each level of 13


Page 14 of 27

Page 15 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


concentration. The same series of samples were prepared and analyzed in two additional analytical sessions to estimate inter-day precision. Precision, expressed as relative standard deviation (RSD), was around 15% regardless the extraction yields (see Table S-4). As for recovery, it will be analyzedin the following section. LODs and LOQs were estimated by analyzing six replicates of fortified water samples. LODs and LOQs were calculated as the analyte concentration detectable and quantifiable with a S/N ratio of 3 and 10, respectively. LODs ranged between 0.1 ng L-1 (nimesulide, thifensulfuron-methyl) and 10 µg L-1 (17-α-estradiol), depending on ES-MS sensitivity and recovery yield. For example, testosterone, which was recovered between 73% and 99% (depending on the concentration level) and exhibited high sensitivity (y = 50 x – 34; R2 = 0.984), was characterized by a LOD of 0.0038 µg L-1. See Table S-4 for the detailed data.

Correlation between physicochemical properties of the model compounds and adsorption capability of BP. Besides the optimization study previously discussed, an original approach was applied to investigate the correlation between physicochemical properties of the model compounds and adsorption capability of BP. Since the 76 analyzed molecules have different physicochemical characteristics, the adsorption capability of BP was initially investigated considering influence of two general parameters, namely logP e pKa, whose values are easily available in the literature. When the two parameters were considered individually, no linear correlation with the experimental data was observed. The curves resulting from a fourth-order polynomial interpolation are shown in Figure 3. Data are quite scattered, although some general trends can be identified: greater recoveries can be observed when logP is in the range 2-4, while lower values can be found for intermediate pKas. For such reason, recovery dependence from this pair of parameters was investigated through the construction of level curves (the graph in Figure 4 was obtained by means of Mathcad Prime 3.0). 14



Page 16 of 27

Visually, it can be noticed that the yields decrease when logP moves away from a value around 2.5 and increase when pKa moves away from 7.5. As a consequence of such evidence, data obtained at the three different fortification levels were interpolated by using equation 1: , = ∗ − | − | + ∗ | − |

eq.1

The optimization of the parameters , , , were performed through multivariate regression by using KaleidaGraph 4.03. Results are reported in Table S-47. The linear correlation of Graphing calculated recoveries (eq. 1) vs experimental recoveries (50 µg L-1), the obtained correlation is presented displayed in Figure 5. With the aim of increasing the correlation coefficient, other molecular parameters, such as mass, volume, surface, number of aromatic bonds, dipole, polarizability, inertial moments, were investigated. None of them is able to increase the established correlation in a significant extent. Such an outcome can be interpreted as a consequence of the complex nature of BP, with which the establishes various molecules can establish interactions of different nature with various molecules. However, it is possible to deduce that hydrophilic compounds with a logP 9 and a log P comprised between 2 to 4 have a higher probability of being absorbed as a result of hydrophobic interaction with the graphenic portion of BP. The latter can be enforced by supplementary interactions (i.e. hydrogen bonds and electrostatic interactions) between the functionalities of an analyte and the polar surface groups of the oxidized BP.

Analysis of environmental water samples. The developed validated method was finally applied to analyze surface water samples of Lazio region in order to test its feasibility. The detected OMPs were some non-steroidal anti-inflammatory drugs, sulfonamides and pesticides. All results are summarized in Table S-51. It is possible to observe a certain distribution of these analytes according to the anthropic impact. For instance, ketoprofen and acetaminophen, two of the most used over-the-counter drugs, exhibited high concentrations in urban areas. Sulfamethoxazole and 15


Page 17 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


sulfadiazine, medicinal products for human use, were especially found in the stretch of the Tiber after Tiber Island, where there is also an important hospital structure. Sulphaquinoxaline is an antibacterial drug for veterinary use that occurred at low levels in urban areas (pet animals) and higher concentrations in rural areas where livestock farms are present. The situation of pesticides is less clear since, for some of them, there are not differences between rural and urban areas. Eventually, diuron, forbidden in Italy since 2007, and atrazine, banned for about twenty years, were within the permitted limits.

CONCLUSIONS In the present article, a novel device which combines the advantages of a nano-sized sorbent membrane (high porosity, high specific adsorption capacity, two sides available for sorption) and positive effects of stirring is described. The extraction unit has been suitably designed to perform a more advantageous SPE mode, alternative to the flow-through one that is typical of SPE cartridges or discs. Stirring favors transference of analytes both in the sorption and desorption step, speeding up all operations. The device is quite manageable and simplifies all SPE operations allowing treating many samples simultaneously. The proposed configuration was methodically systematically optimized studying the behavior of 76 model compounds under different experimental conditions. Since the study has revealed that the two main parameters governing analyte sorption are logP and pKa, a contour map representing the recovery as a function of these parameters was constructed. Besides giving an overall view on BP performances, the map is usable to predict recovery of any other compound of interest whose values of logP and pKa are known. Last but not least, BP offers other two specific advantages over other sorbents: i) after its use, BP has the potential to be regenerated and recycled, but this aspect requires a careful further study to verify its surface modification; ii) BP is safer than free CNTs, since aggregation of nanotubes in a fabric avoids their potential dispersion into the atmosphere due to the dusty nature of CNTs. Finally, considering that the studied material can be easily functionalized as reported in an extensive literature,27,28 future 16



Page 18 of 27

research will be carried out to introduce polar groups or chiral selectors to extend the BP applicability to compounds with logP

Oxidized buckypaper for stir-disc solid phase extraction: evaluation for

Recommend Documents