3D-Printed Microflow Injection Analysis Platform for Online Magnetic

Oct 17, 2017 - In this work, the concept of 3D-printed microflow injection (3D-μFI) embodying a dedicated multifunctional 3D-printed stator onto a ro...
2 downloads 0 Views 632KB Size
Subscriber access provided by University of Florida | Smathers Libraries

Article

3D printed micro-flow injection analysis platform for on-line magnetic nanoparticle sorptive extraction of antimicrobials in biological specimens as a front end to liquid chromatographic assays Han Wang, David J Cocovi-Solberg, Bin Hu, and Manuel Miró Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b03767 • Publication Date (Web): 17 Oct 2017 Downloaded from http://pubs.acs.org on October 18, 2017

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 free 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 accessible to all readers and 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.

Analytical Chemistry 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 23

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

3D printed micro-flow injection analysis platform for on-line magnetic nanoparticle sorptive extraction of antimicrobials in biological specimens as a front end to liquid chromatographic assays Han Wanga, David J. Cocovi-Solbergb, Bin Hua, Manuel Mirób

*

a) Key Laboratory of Analytical Chemistry for Biology and Medicine, Department of Chemistry, Wuhan University, Wuhan 430072, PR China b) FI-TRACE group, Department of Chemistry, University of the Balearic Islands, Carretera de Valldemossa, km. 7.5, Palma de Mallorca, Spain

Abstract In this work, the concept of 3D printed microflow injection (3D-µFI) embodying a dedicated multi-functional 3D-printed stator onto a rotary microvalve along with a mesofluidic sample preparation platform is proposed for the first time. A transparent 3D-printed stereolithographic mesofluidic chip device accommodating polyaniline (PANI) decorated magnetic nanoparticles (32.5±3.8 mg) is harnessed to in-line sorptive microextraction as a front end to liquid chromatography with peak focusing. As a proof of concept application, the 3D-µFI assembly was resorted to matrix clean-up and automatic programmable-flow determination of organic emerging contaminants (4-hydroxybenzoate analogs and triclosan as antimicrobial model analytes) in human saliva and urine samples. By using a sample volume of 1.0 mL with a loading flow rate of 200 µL min-1, an eluent volume of 120 µL at 80 µL min-1, and on-line HPLC injection of 300 µL of the mixture of eluate and Milli-Q water (in a 1:2 ratio) to prevent band broadening effects of the most polar analytes, the limits of detection (3σ criterion) ranged from 1.1-4.5 ng mL-1 for methylparaben (MP), ethylparaben (EP), propylparaben (PrP), phenylparaben (PhP), butylparaben (BP) and triclosan (TCS). Enhancement factors of 16-25 were obtained for the target analytes. Spike recoveries ranged from 84 to 117% for both saliva and urine samples. The on-line 3D-µFI hyphenated method is synchronized with the chromatographic separation and features a chip lifetime of more than 20 injections with minimal losses of moderately nonpolar compounds on the walls of the mesofluidic device.

*

Corresponding author. E-mail: [email protected]. Tel: +34-971172746 1

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 23

Alkyl esters of 4-hydroxybenzoic acid (commonly known as parabens) are widely used as antimicrobial agents in cosmetics, pharmaceuticals, foods and beverages. Compared to other antimicrobial agents, alkylated analogs of 4-hydroxybenzoate are widely used as a result of their low cost, broad spectrum of antimicrobial activity and good chemical and thermal stability1. Traditionally, paraben compounds have been considered relatively safe, but studies within the last decade suggested that they might play a role as endocrine disruptors, possessing weak estrogenic activity but lead to a potential increase in breast cancer incidence, whereby they are regarded as emerging contaminants1-3. Hereto, the European Union amended EU regulation 1223/2009 on cosmetic products for preservatives with compliance from 30 October of 2014 (Commission Regulation (EU) No. 358/2014)4 and 16 April of 2015 (Commission Regulation (EU) No. 1004/2014)5 in which analogs of 4-hydroxybenzoate esters are restricted to a maximum concentration of 0.4% but butylparaben (BP) down to 0.14% and maximum concentration of 0.8% (w/w) for total parabens. Triclosan (TCS), or 5-chloro-2-(2,4dichlorophenoxy) phenol has also been used in a variety of consumer products as antimicrobial and preservative agent6. It is also found to be a potential endocrine disruptor, specifically regarding disruption of thyroid hormone homeostasis. Therefore, the content of triclosan should not exceed 0.2% (w/w) in mouthwash and 0.3% (w/w) in the remaining cosmetic products, which are regulated by the European Community Cosmetic Directive (Commission Regulation (EU) No. 358/2014 and Regulation (EC) No. 1223/2009)4,5 and the US Food and Drug Agency (USFDA))7 in Europe and US, respectively. Therefore, the development of simple, rapid, selective and/or sensitive and cost-effective analytical methods for parabens and triclosan aimed at high-throughput analysis of cosmetic and biological samples are highly demanded. Chromatographic methods, such as high performance liquid chromatography (HPLC)8-11, gas chromatography (GC)12-14 and capillary electrophoresis (CE)15,16 have been reported over the past few years for the determination of parabens and triclosan in matrices of varying complexity. Compared to GC and CE, HPLC based methods need no derivatization steps and allow greater sample volumes, which make HPLC well suited as a routine separation technique for anti-microbial agents with a broad spectrum of polarity. However, the concentration of parabens and triclosan in biological specimens is usually at the trace or ultra-trace level concentrations in normally troublesome matrixes, thereby suitable sample pretreatment methods before instrumental analysis are commonly called for. Up to now, several sample preparation methods with distinct underlying principles including solid phase extraction (SPE)17-20, solid phase microextraction (SPME)21,22, liquid phase microextraction (LPME)23-25, and stir bar sorptive extraction (SBSE)11,26 have been harnessed to the quantitation of analogs of phydroxybenzoate esters in biological samples (e.g., saliva, urine, serum and plasma). SPE/SPME approaches feature a great variety of sorptive materials and formats commercially available, along with excellent matrix tolerance for parabens assays in biological samples17-20. 2

ACS Paragon Plus Environment

Page 3 of 23

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

Comparatively, the capability of enduring matrix ingredients by the LLE/LPME counterparts is rather limited, and sometimes the extraction phase is not amenable to direct on-line/in-line injection into separation techniques such as HPLC. Meso/microfluidic-based

analytical

platforms

for

the

miniaturization

of

sample

handling/pretreatment with subsequent on-line automatic detection has evolved as functional tool-sets for chemical analysis so as to cope with most of the twelve principles of green chemistry27,28. For example, Bodor et al.29 fabricated a poly(methylmethacrylate) chip with two separation

channels

in

a

column-coupling

arrangement

for

isotachophoresis-zone

electrophoresis separation of parabens with on-column conductivity detection, yet standards at a concentration level as high as the mg L-1 level were analyzed and no real sample analysis were reported. De Malsche et al.30 designed a pillar array chip modified with C8 hydrophobic coating that was directly coupled to a commercial micro-LC so as to replace normally used capillary columns for the separation of parabens. The developed system was able to isolate MP, EP, PrP and BP, but no quantitation results or real sample analysis were performed as well. Hartwell et al.31 developed a method of low-pressure LC separation based on a miniature monolithic analytical column for the determination of parabens as active ingredients in skin lotions and other cosmetic products, yet at levels again occurring at the mg L-1 level. Therefore, the above publications are merely focused on on-chip separation principles and detection of major sample constituents. To the best of our knowledge, there is no report in the literature concerning onchip sample pretreatment for parabens and triclosan aiming at trace level analysis. Recent efforts in the microfluidics field have been directed to exploiting sorptive microextraction approaches for removal of matrix ingredients or amelioration of detection capabilities32-36. While the capabilities of microfluidic platforms for sample preparation have made great strides over the years, there remain important features that are still lacking including the facility for custom-designed miniaturized platforms that can be fabricated rapidly and at low cost. Three-dimensional (3D) printing involving customized additive manufacturing has emerged as a versatile platform to overcome barriers created by the skills and equipment required for production of conventional lab-on-a-chip devices while facilitating rapid prototyping at will using affordable desktop printers37-42. Over the past few years, 3D printed fluidic scaffolds have been already resorted to a variety of research fields (e.g., environmental assays43, cell culture assays44 and drug analysis45) with advanced fluid controlling46. Besides, 3D printing technologies are also amenable to the fabrication of liquid drivers and liquid-diverting valves used for fluidic control47,48 along with flow-through module structures for on-chip analyte preconcentration.44,49 With an automation suite, operation of the device might be also advantageously accomplished in unattended mode 24/7. Custom-grade 3D printers are well suited for fabrication of mesofluidic/millifluidic structures with ≥ 500 µm cross-section features, yet improvements in printing materials and types of printers for fabrication of truly 3

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 4 of 23

microfluidic platforms with channel size regimes down to 100 µm are called for.50 This work is aimed at developing a novel concept of micro-flow injection (3D-µFI) analysis based on 3D stereolithographic (SLA) printing. Though the surface roughness of the printed conduits is not comparable to conventional PTFE/PFE FI-tubing, recent studies confirmed that SLA printers produce the smoothest surfaces against fuse deposition modelling or Polyjet counterparts with good laminar flow features51. Automatic fluidic control via pressure driven flow is herein conducted by a 3D-printed stator onto the head valve of a commercial microsyringe pump. Polyaniline (PANI)-modified magnetite nanoparticles (Fe3O4@SiO2@PANI, MNPs) packed on-chip were used as a proof-of-concept sorptive nanoentities for sample preparation by micro-solid phase extraction (µSPE) of alkyl benzoate analogs (including MP, EP, PrP, BP, PhP and TCS) as model analytes in saliva and urine samples prior to on-line liquid chromatographic separations. To the best of our knowledge, this is the first report of a versatile 3D printed flow injection device with mesofluidic/sub-millifluidic features, equipped with a dedicated 3D printed multi-functional valve stator, which integrates automatic sample handling as a front end to on-line LC separations. EXPERIMENTAL Apparatus, reagents, Fe3O4@SiO2@PANI MNP synthesis, chip loading, and sample information are available as Supporting Information (SI). Transparent 3D-printed mesofluidic device configurations were fabricated with the aid of a consumer-grade stereolithographic 3D printer (Form 2, Formlabs, Somerville, USA) loaded with colourless Clear Resin (FLGPCL02, Formlabs). A KA-9180 16W UV-oven equipped with two 8W low pressure mercury lamps was used for curing the 3D printed parts (PSKY, China). Fluidic setup The micro-flow analysis platform consisted of a 30 mm stroke Cavro Xcalibur micro-syringe pump (Tecan, Männedorf, Switzerland) equipped on top with a dedicated multi-port 3D-printed valve, described below, replacing the original head valve and operating simultaneously as a multi-position and diverting valve. The pump is furnished with a 1.0 mL syringe glass barreled syringe. A schematic illustration of the flow setup permitting automatic programmable flow (i.e., forward; backward and stopped flow at will) is depicted in Fig. 1. The port connecting the syringe bypassing the holding coil (port #6, see Fig. 1) was used for aspirating doubly distilled water as a carrier. The other ports of the valve connecting the syringe through the holding coil were assigned as Fe3O4@SiO2@PANI–loaded magnetic µSPE chip (port #4), autosampler (port #2), sampling cup for receiving the eluate and performing at-line dilution (port #3), HPLC transfer line (port #5) and waste (port #1). In regards to the autosampler, the eluent (2% acetic acid in acetonitrile) is contained in rack position 1, the TRIS buffer in position 2, and the water

4

ACS Paragon Plus Environment

Page 5 of 23

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

reservoir for cleaning the autosampler’s tip between successive samples in position 3. The remaining positions were used for standards and samples. When the tip is moved to the ‘up’ position, that is, outside of vials, the flow system is programmed to aspirating air. All of the manifold tubing for connecting the modules is made of polytetrafluoroethylene with 0.78 mm ID, except the 1.5-mL holding coil with an ID of 1.5 mm. The freeware automation suite Cocosoft 4.452, and the extensions thereof for Cavro XCalibur micro-syringe pump and AIM3200 autosampler are utilized in this work for user-friendly control of the syringe pump displacement and flow rate, position of the AIM3200 autosampler, selection of ports from the 3D printed head valve and synchronization with the HPLC system for on-line analysis. The communications were performed through USB-RS232 adapters (Parallax). The automatic method employed to control all of the fluidic setup is described in the SI (Table S1). Hyphenation of the mesofluidic platforms to the HPLC equipment, controlled by its own manufacturer’s software, viz., Jasco’s ChromNAV 2.0, was performed by exploiting the HPLC autosampler’s own capabilities, that is, the ‘injection delay’, and the ‘injection signal’ output on the rear panel. The HPLC operational procedure for separation and determination of target species is listed in Table S2. Design and configuration of 3D printed mesofluidic devices The head valve of a microsyringe pump (Cavro XCalibur, 30 mm stroke, Tecan, Männedorf, Switzerland) was substituted by a dedicated 3D-printed stator with an intricate channel design in order to fabricate a complete micro-flow injection assembly (namely, a Sequential Injection manifold53) over a single head valve, that is, allowing the carrier stream (water) to flow into the syringe, but preventing other fluids to do so by placing a holding coil (HC) between the syringe and the samples/reagents/unit operations, which are nested to the external ports of the 3D printed stator. 3D printed channels of the stator were of 0.8 mm ID, similar to the other fluidic connections (1/32’’ ID). In order to connect the stator system to other fluidic components, each channel was ended with an 8 mm long cylinder (5.5 mm OD) for thread connection via flangeless fitting using Delrin® nuts (Sigma-Aldrich) for tubing 1/16’’ OD following tapping screw threads with ¼ 28’’ taper and bottoming taps. The overall shape of the 3D-printed stator was identical to that of the syringe manufacturer’s, and had the same overall port counts than the sum of the original head valve (3 ports, but only port #6, is used here, shown in Fig. 1) and a conventional 6-port multi-position valve (only ports #1-5, are used in the present design, see Fig. 1). The ports #1-5 communicate with the syringe pump through the holding coil (HC, see Fig. 1). With this new 3D-printed design, the original XCalibur head valve could be easily replaced without further mechanical or firmware modification. The 3D printed mesofluidic chip platform for further MNP packing comprised a single parallelepiped channel with dimensions of 0.8 × 3.0 × 50 mm, along with ended channels of 5.5 5

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 6 of 23

mm OD for tapered thread coupling to the other fluidic hardware (see Fig. S1). The fluidic structure was designed to nest two magnets to the external sample walls, separated by 1.5 mm, that is, the wall thickness of the 3D print was a mere 0.35 mm. The 3D-printed sample preparation platform was designed in a millifluidic (rather than microfluidic) format with thin walls in order to increase the amount of sorbent material available for analyte retention with the two magnets not too distant from each other for the sake of a good retention of the magnetic nanosorbent. The magnetic field in the lumen was estimated to be ca. 550 ± 50 mT. A diagrammatic description of the hyphenated 3D printed micro-flow injection analysis platform for on-line extraction prior to HPLC assays of parabens and triclosan in biological samples is illustrated in Fig. 1. A close-up of the miniaturized fluidic system that showcases the compactness of the setup exploiting the new 3D-printed valve design is available as Fig. S2. Fabrication of 3D printed mesofluidic devices Modelling of all fluidic scaffolds (microvalve stator and µSPE extraction chip, see above) were performed via the 123D Design freeware (Autodesk). The as-created models were loaded into the 3D printer via the manufacturer’s PreForm software (Formlabs) used as slicer to create the object with appropriate orientation onto the built platform. The µSPE extraction chips were printed at 100 µm nominal resolution and tilted 10º along their major axis against the built platform in order to draw the liquid resin from the inner channel during the printing process. The supports for holding the device in place during the printing process were created automatically by the PreForm software with the default parameters (density = 1, point size = 0.60 mm, flat spacing = 5 mm, slope multiplier = 1.00, base thickness = 2.00 mm, height above base = 5.00 mm). Four of the µSPE devices could be printed simultaneously in 274 layers, which lasted in total about 2 h. A single mesofluidic platform is built with 4.2 mL of FLGPCL02 clear resin with a final cost per piece of ca. 1.32 € (0.71 € resin + 0.61 € power). The tailor-made stator mesofluidic channel system was printed at 50 µm nominal resolution as the trade-off between appropriate resolution and rapid prototyping. Early prints at 100 µm resolution yielded occluded channels due to poor resolution. Prints at 25 µm afforded occluded channels as well inasmuch as the printing process lasted ca. 4 h which in turn lead to the solidification of the liquid resin in the conduits of the scaffold. The stator was one-step printed with the external face lying flat on the building platform without supports. This allowed a faster print with flat and smooth faces easy to sand, and a quick removal of the liquid resin from the interior of the fluidic channels in the course of the print. Under the selected printing conditions, the stator was sliced in 340 layers, and prototyped in 139 min using 8.93 mL of resin with a final cost of 2.24 € per device (1.52 € resin + 0.72 € power).

6

ACS Paragon Plus Environment

Page 7 of 23

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

Following the 3D printing process, the fluidic platforms were taken apart from the built platform and soaked in 2-propanol for ca 10 min in order to remove all of the non-polymerized resin from the final print. To this end, the devices were manually shaken in a 2-propanol bath and the 3D printed interior mesochannels and fluidic structures were perfused with 2-propanol with the aid of a 5 mL two-body polypropylene/polyethylene syringe. After cleansing, the modules were irradiated under a 16W low pressure mercury lamps for 24h for curing of the resin and ameliorating the mechanical properties of the prints. The valve stator was finally sanded consecutively with 800, 1200 and 4000 grain sandpaper on both faces in order to fit it tightly to the valve rotor and improve its transparency for naked-eye QC/QA assessment. To make the stator fully transparent, after sanding, a minute amount of clear resin was applied to the external face with an optical grade cellulose pad, which was homogenized by rubbing and cured overnight under 16W UV light. Automated analytical procedure The automatic analytical method starts with filling the holding coil with water (port #6) and emptying the micro-syringe pump barrel. The custom-built 3D printed multi-way valve was turned to port #2 (See Fig. 1) and the autosampler was moved to the vial of the first sample (position 4). Next, 1.0 mL of sample solution was drawn into the holding coil followed by pumping it via flow reversal toward the mesofluidic channel containing Fe3O4@SiO2@PANI at 200 µL min-1 for analyte concentration and matrix removal. The sample matrix flows into the mixer (port #3), and from there to the waste (port #1) by backward-forward flow. A metered volume of 120 µL of eluent (98:2 (v/v) ACN:HAcO) was aspirated from position 1 in the autoampler and delivered by forward flow to the mesofluidic channel at 80 µL min-1. Realization of HPLC peak focusing was done by prior mixing of the eluate within the sampling cup (port #3) with 240 µL of water (used as carrier) followed by at-line gentle perfusing with 700 µL of air (autosampler’s port with the tip up) for homogenization of the eluate and water plugs. On-line analysis of the eluate was performed by hear-cut injection of 300 µL of homogenate into the LC system. Prior to further analysis, the mesofluidic MNP-laden scaffold was rinsed with TRIS buffer (pH 8.7) aspirated from an autosampler vial (position 2). The operational steps and the workflow of the analytical method are compiled in Table S1 and Fig. 2, respectively. A detailed description of the synchronization of the on-line sample preparation/microextraction approach with HPLC operation for on-line handling and analysis of the eluate is available in the SI. RESULTS AND DISCUSSION Sorptive materials Preliminary experiments were undertaken so as to evaluate the sorptive and elution behavior of

7

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 8 of 23

PANI against two conventional commercially available sorbents for non-polar or moderately polar organic species, namely, C18-chemically modified silica gel and divinylbenzene-co-Nvinylpyrrolidone copolymer (Oasis HLB®). Experimental results discussed in SI and summarized in Figures S3 and S4 demonstrated the improved analytical performance of PANI against C18-chemically modified silica gel for alkyl esters of p-hydroxybenzoic acid in terms of sorption efficiency and superior feasibility for on-line µSPE against Oasis HLB in terms of elution rates. Therefore, Fe3O4@SiO2@PANI was selected as the magnetic sorptive material for on-chip µSPE/sample clean-up in this work.

Effect of pH Sample pH plays a crucial role on the adsorption of parabens and TCS onto Fe3O4@SiO2@PANI MNPs because electrostatic interactions beside hydrophobic and π-π stacking might account to the binding of the target analytes to PANI54,55. The effect of sample pH on the on-line retention efficiency of target analytes onto the MNP-laden mesofluidic chip was investigated within the pH range of 4 to 10 (see Fig. S5). The average adsorption efficiencies of the most polar compounds, namely, MP and EP, decreased by 13% and 23% for MP, and by 10% and 14% for EP, at pH 4 and pH 10, respectively, compared to the adsorption efficiencies at pH 8. Within the range of pH 5-9, the adsorption efficiency of all the target analytes remained virtually invariable (Fig. S5). This signals that the major interactions between the target analytes and Fe3O4@SiO2@PANI NPs are of π-π type and hydrophobic nature, because electrostatic interactions with PANI should have been ameliorated at alkaline pH. A value of pH 8.7 (buffered with TRIS), for which RSDs were below 10% (slightly better than those of neutral pH), was chosen as the loading pH for the following experiments, yet the proposed method admits a broad range of pH with excellent analytical performance.

Effect of sample flow rate and sample volume The effect of sample flow rate on the stability of the sorptive MNPs and on-line adsorption efficiency of the target species was explored within the range of 0-220 µL min-1. For sample flow rates ≤ 200 µL min-1, the adsorption efficiencies were in all cases higher than 85% (see Fig. S6) for standard solutions with negligible washout of MNPs from the 3D printed mesofluidic platform. Because the higher the loading flow rate across the on-line microextraction platform the better the sample throughput, a nominal flow rate of 200 µL min-1 was chosen for the remainder of the work, which in fact is much higher than those used in microfluidic MNPs-based extraction microsystems (10-15 µL min-1) 32,36 and on-chip MNPsbased enzymatic reaction microsystems (5 µL min-1)56. The dependence of the extraction yield upon the sample volume was studied in the range of 0.52.5 mL for a given amount of every individual analyte, namely 125 ng, so as to study potential 8

ACS Paragon Plus Environment

Page 9 of 23

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

breakthrough effects. Sample volumes ≤ 2.0 mL afforded average adsorption efficiencies for the overall target analytes within the range of 85-105 % (see Fig. S7) for standard solutions. In consideration of the potential sample throughput demands, 1.0 mL was chosen for further studies.

Effect of the eluent nature Experimental data of the role of the sample pH on the adsorption efficiency let us to conclude that the main interaction between target analytes and sorptive NPs is of hydrophobic nature, and that low/high pH might in fact foster desorption of the target analytes from Fe3O4@SiO2@PANI MNPs. Therefore, MeOH, ACN, ACN with 2% HAcO and ACN with 0.1 M NaOH were investigated as eluents. Compared with MeOH, ACN provided better elution efficiencies for the majority of parabens and TCS for volume of eluent of 200 µL (See Fig. 3). With alkaline ACN, the retention times of target analytes slightly changed and an interfering peak (most likely from the leaching of the polymer fluidic mesodevice) jeopardizes the accurate quantification of BP. The extraction and elution efficiencies of all of target analytes using 100% ACN or acidic ACN (2% HAC in ACN) as eluent are statistically identical. Therefore, a deeper investigation was conducted by studying the elution profile in four consecutive eluent fractions of 50 µL each (see Fig. 4). For the most hydrophobic compounds, namely, PhP and TCS, acidic ACN afforded improved extraction efficiencies by 14% and 7%, respectively, for the first 100 µL eluent against those obtained by raw ACN. Therefore, 2% HAcO in ACN was selected as the most appropriate eluent for on-line elution of target analytes and realization of HPLC heart-cut injection.

Effect of the elution flow rate and eluent volume The minimum available flow rate of the Cavro XCalibur microsyringe pump furnished with 1.0mL gastight syringe for repeatable handling of solutions across the flow setup is 50 µL min-1, whereby the effect of the eluent flow rate on stripping of the concentrated analytes from the MNP was studied from 50 to 100 µL min-1. Experimental results revealed that the extraction efficiency of PhP, BP and TCS decreased sharply by 22-28% with flow rates above 80 µL min-1 (see Fig. S8). An elution flow rate of 80 µL min-1 was therefore chosen as the trade-off between throughput and absolute recoveries of the on-line microextraction method. The greater the eluent volume handled in the automatic 3D-µFI setup and on-line injected into the chromatograph, the better the method’s sensitivity is expected. In fact, absolute recoveries of 72%, 74%, 77% 74%, 77%, 91% of MP, EP, PrP, PhP, BP and TCS, respectively, were obtained from standard solutions with an eluent volume of 100 µL. On-line injection of volumes of eluate (98:2 (v/v) ACN:HAcO) above 20 µL was however proven inappropriate as a result of band broadening effects for the most polar parabens, namely MP and EP. To tackle this issue, 9

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 10 of 23

the 3D-µFI platform was designed to dilute the eluate with water in a 1:2 ratio using the at-line mixing arrangement described in the Experimental section and SI. Experimental results showed that with a heart-cut injection volume of 300 µL (100 µL eluate + 200 µL water) in airsegmented mode (see Table S1), the chromatographic peaks displayed satisfactory resolution (R>1.5). To make sure that no air plug was to be injected into the HPLC column, more than 300 µL of fluid (i.e., loop capacity) were proven necessary. Through several attempts, 360 µL was found to be the smallest liquid volume to be delivered toward the HPLC valve, that is, two segments of 30 µL eluate would remain at both ends of the HPLC loop. The eluent volume was finally set to 120 µL, which is in turn mixed with 240 µL of water at-line for appropriate eluate dilution and peak focusing onto the chromatographic column.

Effect of the ionic strength In biological fluids, the occurrence of electrolytes is expected to potentially affect the sorptive efficiency of nanomaterials. As a result, the effect of the ionic strength on the µSPE was studied by adding 0-25% NaCl (m/v) to the buffered standards (pH=8.7) analyzed on-line by the 3DµFI setup as a front end to LC. The extraction efficiencies remain statistically identical with NaCl < 5% (m/v), but strikingly lower for the most polar species (viz., MP and EP) with further increase of the NaCl concentration (see Fig. S9). This phenomenon is most likely a consequence of the anion-exchange electrostatic competition of the analytes at pH 8.7 for the PANI surface against concomitant matrix electrolytes. However, the salt concentration in saliva and urine is usually < 2%57, which in turn does not jeopardize the extraction of parabens and TCS in our mesofluidic system. Therefore, the ionic strength of the samples was not adjusted prior to analysis.

Analytical performance Under the selected experimental conditions, the analytical performance of the proposed on-line 3D-µFI method was evaluated (see Table 1). The limits of detection (LODs) were calculated according to the IUPAC guidelines (three times the standard deviation of the background for 10 runs divided by the slope of the calibration curve) and ranged from 1.1 to 4.5 ng mL-1. The linearity of the calibration graph was tested by employing mixed standard solutions (in TRIS buffer, pH 8.7) with increasing concentrations of MP, EP, PrP, PhP, BP and TCS from about the limit of quantification (3.3 × LOD) to 2000 ng mL-1. The calibration curves showed a linear response with determination coefficients > 0.9956. The MNP-laden mesofluidic chip could be reused for more than 20 times with RSD <15%. The on-line sample processing method lasted about 20 min, which is in turn well synchronized with the chromatographic separation of the preceding sample. Compared with the traditional silicon microfluidic chips, the 3D printed mesofluidic platforms 10

ACS Paragon Plus Environment

Page 11 of 23

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

are inexpensive and more suitable for resource-limited settings: There is no need for clean room facilities, photolithographic machinery, spin coater, hot plates, photoresist and monocrystalline silicon pieces, which overall cost tens of thousands of dollars. The desktop 3D printer employed in this work just costs 3800 €; and 1 L of the polymeric resin has a market price of 136 €, whereby a single 3D printed mesofluidic chip costs less than 2 € including power expenses. 3D printing also reduces the tedious manual operation of lithographic procedures and is amenable to both flow injection practitioners and novices in the field. The analytical performance of the on-line miniaturized 3D printed mesofluidic channel configuration for antimicrobials was compared with those of previous works using microextraction approaches in combination to LC separation with UV detection10,11,58-60 (see Table 2). To the best of our knowledge, our 3D-µFI setup is the only one affording fully automatic µSPE with in-line sample loading and on-line elution as a front end to the separation column. Further, the proposed mesofluidic method is applied to a broader range of organic species with a wide spectrum of polarity; consumes much lower sample volume as compared to batchwise stir-bar microextraction counterparts10,11,59, which in turn, are deemed inappropriate for the handling of micro-volumes of biological samples, such as saliva; and features better LOD against a previous sorptive microextraction approach using identical sample volume, that is, 1.0 mL58.

Sample analysis The 3D-µFI system embodying the integrated microvalve and mesofluidic sample preparation platform was harnessed to the automatic on-line MNP-based microextraction and determination of parabens and triclosan in saliva and urine samples (see Table 2, Table 3 and Table S3). The samples were collected from volunteers who had used parabens and triclosan containing personal care and hygiene products as indicated in SI. For the first saliva sample (see Fig. S10a and Fig. S10c), MP and PrP could only be found just after using the toothpaste and mouth rinse solution, yet TCS was encountered in all sampling times at different concentration levels. For the second saliva sample (see Fig. S10b and Fig. S10d), only MP and TCS could be detected right after using the antimicrobial-laden personal care product. For assessment of method trueness, aliquots of saliva sample #1 and sample # 2 at t=0 h were spiked with MP, PrP and TCS at the 300, 300, 600 ng mL-1 levels and 400, 0, 60 ng mL-1 levels, respectively, with relative recoveries spanning from 103-114% and 91-111% for the saliva samples, respectively, using external calibration (see Table 3) with the possibility of resorting to PhP as internal standard. Paraben/triclosan-free saliva samples were also doped with MP, EP, PrP, PhP, BP and TCS at three concentration levels within the range of 10-1000 ng mL-1 for every single target analyte. Relative recoveries (see Table S3) ranged from 84% to 117%. For further method validation, saliva #1 (t=0 h) was also determined by LC-MS as a reference method after in-line 11

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 12 of 23

matrix clean-up and preconcentration by the developed µFI system (see supporting information). The concentrations of MP, PrP and TCS found in the sample after 50-fold eluate dilution were 155±12, 85±4 and 451±17 ng mL-1, respectively, which are in good agreement with experimental results obtained by our proposed mesofluidic method (P 4, such as TCS, because of potential absorption onto the polymeric parts of the magnetic bar and the dialytic module itself61. Experimental results and spike recovery data for the human urine sample as obtained by the developed automatic online 3D-µFI platform are compiled in Table S4. In humans, parabens might be hydrolysed to p-hydroxybenzoic acid and then excreted in urine as glycine, glucuronide and sulphate conjugates, but also the unchanged parabens can be excreted as glucuronide and sulphate conjugates62. Hereto, an enzymatic hydrolysis as described in SI was harnessed in this work prior to sample handling by the mesofluidic platform for the determination of the total amount of parabens. Spike recoveries at the 10, 100 and 1000 ng mL-1 for method trueness exploration afforded values ranging from 84% to 117% using external calibration. However, the recoveries of PhP were in all instances < 80%. To figure this out, the urine sample was spiked with the two most hydrophobic analogs of 4-hydroxybenzoate in our work, namely, PhP and BP at the 1000 ng mL-1 level, and then 1 mL of spiked urine sample was subjected to dialysis for 12 h. The dialysate results as obtained by LC-MS revealed that the free concentration of PhP was a mere one-fifth of that of BP. This indicates that PhP is bound to large-molecular weight, undialyzed compounds in urine, most likely by π-π interactions, which in turn made it impossible to recover the nominal (spiked) concentration by the on-line µSPE method. The set of validation data in this section indicates that our 3D-µFI method is free from multiplicative matrix interferences in the determination of the target anti-microbials in urine and saliva with no need of matrix matching calibration nor the method of the standards addition. Data reported in the literature63 signaled that the concentrations of MP, PrP and TCS in human urine are greatly variable, viz, δ= 3230 ng mL-1 for MP, δ= 1530 ng mL-1 for PrP and δ= 1630 ng mL-1 for TCS. Previous reports also indicated that the concentrations of MP, PrP and TCS in human saliva samples collected after using parabens and triclosan containing personal care and hygiene products are at mg L-1 level (at 0 h, 1.6 mg L-1 of TCS59, 71.6 mg L-1 of MP64 and 16.7 mg L-1 of PrP64), thus demonstrating the feasibility of our method for automatic monitoring of free parabens and triclosan species in human urine and saliva. 12

ACS Paragon Plus Environment

Page 13 of 23

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

Conclusion In this paper, a new concept for microflow injection analysis based on a custom-built 3D printed multi-channel layered valve structure assembled on top of a microsyringe pump alongside a 3D printed single-channel mesofluidic platform for handling PANI-decorated magnetic nanoparticles on-chip for on-line sample microextraction is herein presented for the first time. The 3D printed multi-channel module replaces conventional switching and multiposition/selection valves in standard flow systems and admits manifold tubing (including holding coil). In combination with the bi-directional operation of the microsyringe pump the hyphenated setup fosters automatic flow programming and implementation of flow-through unit operations. The analytical features and performance of the proposed microflow setup were explored for on-line biological sample clean-up and concentration of anti-microbial agents (viz., triclosan and alkyl p-benzoate analogs) as model analytes. An asset of the sample preparation module is the possibility of interfacing to LC via heart-cut injection for automatic handling of the eluent/eluate solutions including at-line modulation of solvent composition for further chromatographic peak focusing. The 3D-µFI mesofluidic method showcases good matrix tolerance ability and could be mechanized for on-line 24/7 analysis. The LODs of the developed mesofluidic method ranged from 1.1-4.5 ng mL-1 with enhancement factors of 16-25. Further work is underway in our lab to explore opportunities and resilience of 3D printed mesofluidics for scale-up applicability including the clean-up of matrix co-extractives in troublesome samples containing body fluids as obtained from oral bioaccessibility testing of organic emerging contaminants in environmental solids and foodstuff. Efforts are also directed to minimize polymer leaching in the course of the analytical procedures by amelioration of the resin curing throughout the postprocessing of the 3D print by resorting to a high-pressure, 150W-mercury vapour lamp.

Acknowledgements M. Miró and D.J. Cocovi-Solberg acknowledge financial support from the Spanish State Research Agency (AEI) through project CTM2014-61553-EXP (AEI/FEDER, UE). M. Miró extends his appreciation to AEI for supporting project CTM2014-56628-C3-3-R (AEI/FEDER, UE). H. Wang acknowledges financial support from the Research Plan for PhD Short-time Mobility Program, Wuhan University.

13

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 14 of 23

REFERENCES (1) Ocaña-Gonzalez, J. A.; Villar-Navarro, M.; Ramos-Payan, M.; Fernandez-Torres, R.; BelloLopez, M. A. Anal. Chim. Acta, 2015, 858, 1-15. (2) Bledzka, D.; Gromadzinska, J.; Wasowicz, W. Environ. Int., 2014, 67, 27-42. (3) Giulivo, M.; de Alda, M. L.; Capri, E.; Barcelo, D. Environ. Res., 2016, 151, 251-264. (4) Official Journal of the European Union, Commission Regulation (EU) No. 358/2014 of 9 April 2014 amending Annexes II and V to Regulation (EC) No. 1223/2009 of the European Parliament and of the Council on cosmetic products, L 107/5−9, Brussels, 2014. (5) Official Journal of the European Union, Commission Regulation (EU) No 1004/2014 of 18 September 2014 amending Annex V to Regulation (EC) No 1223/2009 of the European Parliament and of the Council on cosmetic products, L 282/5−8, Brussels, 2014. (6) Montaseri, H.; Forbes, P. B. C. Trac-Trends Anal. Chem., 2016, 85, 221-231. (7) US Food and Drug Agency, section 505 (b) of the Federal Food, Drug and Cosmetic Act (FDCA), Rockville, 2012. (8) Zhang, Q. L.; Lian, M.; Liu, L. J.; Cui, H. Anal. Chim. Acta, 2005, 537, 31-39. (9) Ye, X.; Tao, L. J.; Needham, L. L.; Calafat, A. M. Talanta, 2008, 76, 865-871. (10) Almeida, C.; Nogueira, J. M. F. J. Chromatogr. A, 2014, 1348, 17-26. (11) Xu, J.; Chen, B. B.; He, M.; Hu, B. J. Chromatogr. A, 2013, 1278, 8-15. (12) Lopez-Darias, J.; Pino, V.; Meng, Y. J.; Anderson, J. L.; Afonso, A. M. J. Chromatogr. A, 2010, 1217, 7189-7197. (13) Ferreira, A. M. C.; Moder, M.; Laespada, M. E. F. J. Chromatogr. A, 2011, 1218, 3837-3844. (14) Albero, B.; Sanchez-Brunete, C.; Miguel, E.; Perez, R. A.; Tadeo, J. L. J. Chromatogr. A, 2012, 1248, 9-17. (15) Huang, H. Y.; Lai, Y. C.; Chiu, C. W.; Yeh, J. M. J. Chromatogr. A, 2003, 993, 153-164. (16) Blanco, E.; Casais, M. D.; Mejuto, M. D.; Cela, R. Anal. Chim. Acta, 2009, 647, 104-111. (17) Tahmasebi, E.; Yamini, Y.; Mehdinia, A.; Rouhi, F. J. Sep. Sci., 2012, 35, 2256-2265. (18) Han, F.; He, Y. Z.; Yu, C. Z. Talanta, 2008, 74, 1371-1377. (19) Chen, Z. F.; Ying, G. G.; Lai, H. J.; Chen, F.; Su, H. C.; Liu, Y. S.; Peng, F. Q.; Zhao, J. L. Anal. Bioanal. Chem., 2012, 404, 3175-3188. (20) Shen, H. Y.; Jiang, H. L.; Mao, H. L.; Pan, G.; Zhou, L.; Cao, Y. F. J. Sep. Sci. 2007, 30, 48-54. (21) Yang, T.-J.; Tsai, F.-J.; Chen, C.-Y.; Yang, T. C.-C.; Lee, M.-R. Anal. Chim. Acta, 2010, 668, 188-194. (22) Gonzalez-Marino, Quintana, J.B; Rodriguez, I.; Schrader, S.; Moeder, M. Anal. Chim. Acta, 2011, 684, 59-66. (23) Saraji, M.; Mirmahdieh, S. J. Sep. Sci. 2009, 32, 988-995. (24) Prichodko, A.; Jonusaite, K.; Vickackaite, V. Cent. Eur. J. Chem., 2009, 7, 285-290. (25) Jain, R.; Mudiam, M. K. R.; Chauhan, A.; Ch, R.; Murthy, R. C.; Khan, H. A. Food Chem., 2013, 141, 436-443. (26) Ramirez, N.; Marce, R. M.; Borrull, F. J. Chromatogr. A, 2011, 1218, 6226-6231. (27) Armenta, S.; Garrigues, S.; de la Guardia, M. Trac-Trends Anal. Chem., 2015, 71, 2-8. (28) Gałuszka, A.; Migaszewski, Z.; Namieśnik, J. Trac-Trends Anal. Chem., 2013, 50, 78-84. (29) Bodor, R.; Zuborova, M.; Olvecka, E.; Madajova, V.; Masar, M.; Kaniansky, D.; Stanislawski, B. J. Sep. Sci., 2001, 24, 802-809. (30) De Malsche, W.; De Bruyne, S.; De Beek, J. O.; Sandra, P.; Gardeniers, H.; Desmet, G.; Lynen, F. J. Chromatogr. A, 2012, 1230, 41-47. (31) Hartwell, S. K.; Kehling, A.; Lapanantnoppakhun, S. Chromatographia, 2014, 77, 883-891. (32) Chen, B. B.; Heng, S. J.; Peng, H. Y.; Hu, B.; Yu, X.; Zhang, Z. L.; Pang, D. W.; Yue, X.; Zhu, Y. J. Anal. Atom. Spectrom., 2010, 25, 1931-1938. (33) Giordano, B. C.; Burgi, D. S.; Hart, S. J.; Terray, A. Anal. Chim. Acta, 2012, 718, 11-24. (34) Park, M.; Ha, H. D.; Kim, Y. T.; Jung, J. H.; Kim, S.-H.; Kim, D. H.; Seo, T. S. Anal. Chem. 2015, 87, 10969-10975. (35) Gasilova, N.; Srzentić, K.; Qiao, L.; Liu, B.; Beck, A.; Tsybin, Y. O.; Girault, H. H. Anal. Chem. 2016, 88, 1775-1784. (36) Wang, H.; Chen, B.; Zhu, S.; Yu, X.; He, M.; Hu, B. Anal. Chem. 2016, 88, 796-802. (37) Kataoka, É. M.; Murer, R. C.; Santos, J. M.; Carvalho, R. M.; Eberlin, M. N.; Augusto, F.; Poppi, R. J.; Gobbi, A. L.; Hantao, L. W. Anal. Chem. 2017, 89, 3460-3467. (38) Morgan, A. J. L.; Hidalgo San Jose, L.; Jamieson, W. D.; Wymant, J. M.; Song, B.; Stephens, 14

ACS Paragon Plus Environment

Page 15 of 23

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

P.; Barrow, D. A.; Castell, O. K. PLOS ONE, 2016, 11, e0152023. (39) Zhang, Y.; Ge, S.; Yu, J. Trac-Trends Anal. Chem., 2016, 85, Part C, 166-180. (40) Waheed, S.; Cabot, J. M.; Macdonald, N. P.; Lewis, T.; Guijt, R. M.; Paull, B.; Breadmore, M. C. Lab Chip 2016, 16, 1993-2013. (41) Bhattacharjee, N.; Urrios, A.; Kang, S.; Folch, A. Lab Chip 2016, 16, 1720-1742. (42) Lockwood, S. Y.; Meisel, J. E.; Monsma, F. J.; Spence, D. M. Anal. Chem. 2016, 88, 18641870. (43) Au, A. K.; Bhattacharjee, N.; Horowitz, L. F.; Chang, T. C.; Folch, A. Lab Chip 2015, 15, 1934-1941. (44) Su, C. K.; Peng, P. J.; Sun, Y. C. Anal. Chem. 2015, 87, 6945-6950. (45) Gross, B. C.; Anderson, K. B.; Meisel, J. E.; McNitt, M. I.; Spence, D. M. Anal. Chem. 2015, 87, 6335-6341. (46) Gross, B.; Lockwood, S. Y.; Spence, D. M. Anal. Chem. 2017, 89, 57-70. (47) Su, C.-K.; Hsia, S.-C.; Sun, Y.-C. Anal. Chim. Acta 2014, 838, 58-63. (48) Gong, H.; Woolley, A.T.; Nordin, G.P. Lab Chip 2016, 16, 2450-2458. (49) Kataoka, E.M.; Murer, R.C.; Santos, J.M.; Carvalho, R.M.; Eberlin, M.N.; Augusto, F.; Poppi, R.J.; Gobbi, A.L; Hantao, L.W. Anal. Chem., 2017, 89, 3460–3467 (50) Beauchamp, M.J.; Nordin, G.P.; Woolley, A.T. Anal. Bioanal. Chem. 2017, 409, 4311–4319. (51) Macdonald, N. P.; Cabot, J. M.; Smejkal, P.; Guijt, R. M.; Paull, B.; Breadmore, M. C. Anal. Chem. 2017, 89, 3858-3866. (52) Cocovi-Solberg, D. J.; Miro, M. Anal. Bioanal. Chem. 2015, 407, 6227-6233. (53) Hansen, E. H.; Miró, M. TrAC Trend. Anal. Chem. 2007, 26, 18-26. (54) Mehdinia, A.; Esfandiarnejad, R.; Jabbari, A. J. Sep. Sci., 2015, 38, 141-147. (55) Jiang, X. L.; Cheng, J.; Zhou, H. B.; Li, F.; Wu, W. L.; Ding, K. R. Talanta 2015, 141, 239246. (56) Liu, J. Y.; Lin, S.; Qi, D. W.; Deng, C. H.; Yang, P. Y.; Zhang, X. M. J. Chromatogr. A, 2007, 1176, 169-177. (57) Caroli, S.; Alimonti, A.; Coni, E.; Petrucci, F.; Senofonte, O.; Violante, N. Crit. Rev. Anal. Chem., 1994, 24, 363-398. (58) Liu, D.; Song, N.; Cheng, Y. C.; Chen, D. X.; Jia, Q.; Yang, Y. W. RSC Advances 2014, 4, 49153-49160. (59) Silva, A. R. M.; Nogueira, J. M. F. Talanta 2008, 74, 1498-1504. (60) Kim, D.; Han, J.; Choi, Y. Anal. Bioanal. Chem., 2013, 405, 377-387. (61) Koc, A.; Orhon, K. B.; Ogutverici, A.; Yilmaz, L.; Furi, L.; Oggioni, M. R.; Dilek, F. B.; Yetis, U. Desalin. Water Treat., 2014, 52, 7101-7107. (62) Janjua, N. R.; Frederiksen, H.; Skakkebaek, N. E.; Wulf, H. C.; Andersson, A. M. Int. J. Androl., 2008, 31, 118-129. (63) Moos, R. K.; Anger, J.; Wittsiepe, J.; Wilhelm, M.; Bruning, T.; Koch, H. M. Int. J. Hyg. Environ. Health. 2014, 217, 845-853. (64) Park, Y. D.; Jang, J. H.; Park, J. E.; Kim, J. H.; Kim, E. C.; Song, Y. J.; Kwon, H. J. Biomed. Chromatogr., 2014, 28, 1692-1700.

15

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 16 of 23

Figure legends:

Figure 1. Diagrammatic representation of the proposed micro-flow injection system illustrating the mesofluidic channel configuration containing PANI-MNPs, the sampling cup, and the transfer line for on-line HPLC. The 3D printed stator concomitantly operates as the head valve of the syringe and a multi-port selection valve. HC: Holding Coil. Note: The blue lines in the figure indicate the communication channels between the syringe pump and the external ports.

Figure 2. Flow diagram of the analytical method and synchronization of the automatic sample preparation procedure with on-line HPLC analysis

Figure 3. Effect of the eluent nature on the extraction efficiency of target analytes. Error bars represent the standard deviation of three replicate experiments. (CMP, EP, PrP, PhP, BP, TCS=250 ng mL-1, sample flow rate: 200 µL min-1, sample volume: 1.0 mL, elution flow rate: 50 µL min-1, elution volume: 200 µL)

Figure 4. Effect of the eluent composition on the extraction efficiency of target analytes. Error bars represent the standard deviation of three replicate experiments. (CMP, EP, PrP, PhP, BP, TCS=250 ng mL-1, sample flow rate: 200 µL min-1, sample volume: 1.0 mL, elution flow rate: 50 µL min1

, elution volume: 50 µL in four consecutive fractions)

16

ACS Paragon Plus Environment

Page 17 of 23

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

Figure 1

17

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 18 of 23

Figure 2

18

ACS Paragon Plus Environment

Page 19 of 23

MP EP PrP PhP BP TCS

100

Extraction efficiency (%)

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

80 60 40 20 0 MeOH

ACN

2% HAcO in ACN

0.1M NaOH in ACN

Figure 3

19

ACS Paragon Plus Environment

Analytical Chemistry

MP EP PrP PhP BP TCS

60

Extraction efficiency (%)

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 20 of 23

50 40 30 20 10 0 1st 2nd

3th 4th

1st 2nd

ACN

3th

4th

2% HAcO in ACN

Figure 4

20

ACS Paragon Plus Environment

Page 21 of 23

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

Table 1. Analytical performance of the on-line 3D-µFI microextraction procedure using Fe3O4@SiO2@PANI NPs as the sorptive material Linear Intra-day Inter-day Determination Linear equation LODs range RSDsb RSDsc coefficient EFsa -1 -1 (ng mL ) (ng mL ) -1 2 (ng mL ) (n=7, %) (n=7, %) (R ) MP 5-2000 y=1572x+1049 0.9996 1.4 19 5.1 11 EP 10-2000 y=1642x+4598 0.9991 3.4 22 9.0 9.8 PrP 5-2000 y=1635x+2559 0.9984 1.1 23 6.9 8.9 PhP 5-2000 y=1576x+1263 0.9956 1.3 24 6.0 7.8 BP 5-2000 y=1662x+1113 0.9994 1.1 25 8.2 10 TCS 20-2000 y=1805x+5259 0.9993 4.5 16 7.9 11 a: EFs: enhancement factors, EFs is obtained by the ratio of the linear slope of the on-line method against that of direct injection of 20 µL into HPLC without preconcentration. b: Relative Standard Deviation using a single chip platform, CMP, EP, PrP, PhP, BP =25 ng mL-1, CTCS=50 ng mL-1 c: Relative Standard Deviation in seven consecutive days with different chip platforms, CMP, EP, PrP, -1 -1 PhP, BP=25 ng mL , CTCS=50 ng mL

Analytes

21

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

Page 22 of 23

Table 2. Comparison of the analytical performance of our 3D-µFIA setup system with that of previous reports in the literature using sorptive microextraction of alkyl hydroxyl benzoate analogs and triclosan as a front end to LC-UV assays Analytes

Pretreatment

Detector

LODs (ng mL-1)

Linear range (ng mL-1)

Online or offline detection

Sample throughput (h-1)

Sample volume (mL)

Real samples

Ref.

MP, EP, PrP, BP

Bar adsorptive microextraction

HPLCDAD

0.1

0.5-28.0

offline

-

25

Environmental water, urine, body lotion and hand cream

10

MP, EP, PrP, BP

Stir bar sorptive microextraction

HPLC-UV

0.6-2.7

5-1000 for BP and PP, 10-1000 for EP, 10-2000 for MP

offline

-

10

Beverage samples including cola, orange juice and herbal tea

11

Polymer monolith solid phase microextraction

HPLC-UV

8-33

50-5000

offline

-

1.0

Red wine sample

58

Stir bar sorptive microextraction

HPLCDAD

0.1

0.4-108

offline

-

25

Toothpaste, saliva and urban wastewater

59

20

10

River water and wastewater

60

3

1.0

Urine and saliva samples

This work

MP, EP, PrP, BP TCS

TCS

Solid phase microextraction

HPLC-UV

0.005

0.01-168

Semiautomatic method (on-line detection)

MP, EP, PrP, PhP, BP and TCS

Chip-based magnetic solid phase microextraction

HPLC-PDA

1.1-4.5

5-2000 for MP, PrP, PhP and BP, 10-2000 for EP, 20-2000 for TCS

online

22

ACS Paragon Plus Environment

Page 23 of 23

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

Table 3. Analytical results (Mean ±SD, n=3) for alkyl esters of 4-hydroxybenzoic acid and triclosan in saliva using the proposed 3D-µFI platform following on-line PANI-MNP sorptive microextraction

Found (ng mL-1)

Target analytes

Saliva sample #1

MP PrP TCS

0 h*

1h

8h

134±2 79±1 506±48

ND ND 89±5

ND ND 54±4

Added** (ng mL-1)

Found (after spike) (ng mL-1)

Recovery (%)

300 300 600

449±16 432±11 1164±72

103 114 111

Saliva MP 221±8 ND ND 400 568±18 91 sample # 2 TCS 27±3 ND ND 60 97±6 111 ND: not detected *Sample was 50 fold diluted **Compounds were added to the saliva sample collected right after the application of personal care and hygiene products (t=0 h)

23

ACS Paragon Plus Environment