Low-Cost Passive Sampling Device with Integrated Porous Membrane

In Vivo Stable-Isotope Labeling and Mass-Spectrometry-Based Metabolic Profiling of a Potent Tobacco-Specific Carcinogen in Rats. Analytical Chemistry...
0 downloads 0 Views 7MB Size
Article pubs.acs.org/ac

Cite This: Anal. Chem. 2018, 90, 12081−12089

Low-Cost Passive Sampling Device with Integrated Porous Membrane Produced Using Multimaterial 3D Printing Umme Kalsoom,†,‡ Chowdhury Kamrul Hasan,†,§,∥ Laura Tedone,† Christopher Desire,† Feng Li,†,‡ Michael C. Breadmore,†,‡,§ Pavel N. Nesterenko,†,‡,§ and Brett Paull*,†,‡,§

Downloaded via UNIV OF SUNDERLAND on October 23, 2018 at 12:57:02 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.



Australian Centre for Research on Separation Science (ACROSS), School of Natural Sciences, University of Tasmania, Private Bag 75, Hobart, Tasmania 7001, Australia ‡ ARC Centre of Excellence for Electromaterials Science (ACES), School of Natural Sciences, University of Tasmania, Sandy Bay, Hobart, Tasmania 7001, Australia § ARC Training Centre for Portable Analytical Separation Technologies (ASTech), School of Natural Sciences, University of Tasmania, Private Bag 75, Hobart, Tasmania 7001, Australia ∥ Department of Environmental Science, School of Environmental Science and Management, Independent University, Bangladesh, Dhaka, 1229, Bangladesh S Supporting Information *

ABSTRACT: Multimaterial 3D printing facilitates the rapid production of complex devices with integrated materials of varying properties and functionality. Herein, multimaterial fused deposition modeling (MM-FDM) 3D printing was applied to the fabrication of low-cost passive sampler devices with integrated porous membranes. Using MM-FDM 3D printing, the device body was produced using black polylactic acid, with Poro-Lay Lay-Felt filament used for the printing of the integrated porous membranes (rubber-elastomeric polymer, porous after removal of a water-soluble poly(vinyl alcohol) component). The resulting device consisted of two interlocking circular frames, each containing the integrated membrane, which could be efficiently sealed together without the need for additional O-rings, and prevented loss of enclosed microparticulate sorbent. Scanning electron microscopy (SEM) analysis of the purified composite filament confirmed the porous properties of the material, an average pore size of ∼30 nm. The printed passive samplers with various membrane thicknesses, including 0.5, 1.0, and 1.5 mm, were evaluated for their ability to facilitate the extraction of atrazine as the model solute onto the internal sorbent, under standard conditions. Gas chromatography−mass spectrometry was used to determine the uptake of atrazine by the device from standard water samples and also to evaluate any chemical leaching from the printed materials. The sampler with 0.5 mm thick membrane showed the best performance with 87% depletion and a sampling rate of 0.19 Ld−1 (n = 3, % RSD = 0.59). The results obtained using these printed sampling devices with integrated membranes were in close agreement to devices fitted with a standard poly(ether sulfone) membrane. screening techniques (discrete site and time specific “grab sampling” and mechanical sampling) are both challenging and costly. Due to their presence at ultratrace concentrations and the inability of grab sampling approaches to provide timeweighted average (TWA) concentrations (average concen-

E

nvironmental pollutants arising from industrial waste, agricultural runoff, and chemical spills pose a serious risk to the aquatic ecosystem and represent an increasing human health concern.1 Strict government policies, increasing initiatives to improve water quality, and growth of industries, e.g., pharmaceutical, food and beverages, and agricultural, increase the demand for monitoring and analysis of water based pollutants.2,3 Long-term monitoring and exposure assessments for many of these pollutants through conventional © 2018 American Chemical Society

Received: June 26, 2018 Accepted: September 17, 2018 Published: September 17, 2018 12081

DOI: 10.1021/acs.analchem.8b02893 Anal. Chem. 2018, 90, 12081−12089

Article

Analytical Chemistry

difficulty in removing the support materials in the case of the latter. Fused deposition modeling (FDM) 3D printing on the other hand provides an inexpensive readily accessible alternative to the above approaches and does not require a support material to realize the printing process. A wide range of thermoplastic materials are available for FDM 3D printers, including acrylonitrile butadiene styrene, polylactic acid (PLA), polyamide, and polycarbonate,17 together with an increasing variety of functional and composite materials.13 Using this variety of starting materials, multimaterial fused deposition modeling (MM-FDM) printers can now be used to produce devices containing components or regions of greatly differing strength, flexibility, porosity, electrical conductivity, etc. One area of considerable interest is the ability to print materials of controlled or selective porosity, including active and passive integrated membranes.25−27 For example, Li et al. recently reported the printing of an ABS based microfluidic device containing an integrated porous membrane, the latter simultaneously printed using a composite print material, namely, Poro-Lay Lay-Felt filament, which contains a watersoluble poly(vinyl alcohol) (PVA) component, which when removed leaves the porous structure in place.28 A similar material print was simultaneously being used by Belka et al., namely, Lay-Fomm 60, to produce a tube shaped porous sorbent insert for use within Eppendorf tubes, and applied to the centrifugal solid phase microextraction of glimepiride from water.29 Here, we present for the first time the use of MM-FDM 3D printing for the rapid low-cost fabrication of passive sampling devices with integrated porous membranes. The devices consist of interlocking top and bottom plates, each containing a central 60 mm diameter self-supported integrated porous membrane produced using the Poro-Lay Lay-Felt filament. The integrated device can be loaded with sorbent and closed to provide a leak-free seal, without additional O-rings, sealants, or compression bands, and obviously needs no additional fragile membrane materials, as is the case with most commercial passive water samplers available. To demonstrate the potential utility of the printed device, herein we demonstrate its performance for the extraction and uptake of the herbicide atrazine from water samples. The performance of the printed membrane is also compared to that of the commonly applied commercial counterpart.

tration of analytes over the sampling period), for such target contaminants, a great deal of uncertainty exists in attempting to establish continuous and accurate analytical data for dynamic systems.4 Passive sampling is an alternative sampling approach which is now seeing widespread application. Passive sampling sees the continuous exposure of an extractive sorbent to the dynamic water system and provides the means to obtain the TWA concentration of the target solutes over the deployment period.5 TWA concentrations of the water contaminants can be determined by measuring the sampling rate (Rs), the volume of water extracted by the passive sampling device per day (Ld−1).6 Additionally, for ultratrace analysis, minimum solvent consumption, low cost, and the ability to provide equilibrium pollutant concentrations7 make this an attractive approach to widespread sampling. Passive sampling devices, commercially available in several formats, are commonly used for monitoring aquatic systems for low level organic contaminants, e.g., pharmaceuticals and personal care products (PPCPs), surfactants, pesticides, herbicides, etc.4,8 Popular examples include the “Chemcatcher” configuration (TelLab, T.E. Laboratories, Carlow, Ireland) and the polar organic chemical integrative sampler (POCIS) (Environmental Sampling Technologies Inc., St. Joseph, MO, USA). The latter format typically consists of a membrane− sorbent−membrane sandwich housed between two metallic compression rings to provide adequate sealing and restrict the loss of the internal sorbent.7,9 Various types and/or combination of sorbent can be used in either format, with selectivity tailored to particular target solutes; e.g., commercial variants of the POCIS include the “POCISpest” for pesticides and the “POCISpharm” for PPCPs.10 Passive samplers are commonly used to monitor herbicides, including atrazine, 2-chloro-4-ethylamino-6-isopropyilamino-striazine, which is one of the most extensively investigated herbicides, due to its prevalence in the aquatic systems11 and potential adverse health effects.12 Commercial passive sampling devices are available in limited sizes and are produced via traditional engineering and subtractive manufacturing techniques. These techniques are obviously rather labor intensive, and iterative design improvements are both costly and rather time-consuming. However, over the past decade, 3D printing has gained a very strong foothold in the area of design and rapid prototyping, due to its ability to readily create customized complex objects from an increasing variety of materials,13,14 across a huge range of application areas. This has been of particular interest to the analytical science community for applications in sensors,15 microfluidics,16,17 and separation sciences.18−21 In 3D printing, complex three-dimensional objects, designed by freely available computer aided design (CAD) drawing software packages, are produced in a layer-bylayer process, until the whole object is fully printed.22 Recently, the availability of multimaterial 3D printers has expanded the potential of 3D printing for “active” device prototyping considerably. Modern multimaterial 3D printing allows sequential printing using differing materials (polymers), thus providing potential to produce integrated objects with regions/sections of varying functionality. Multimaterial stereolithography (MMSL) based 3D printers containing up to four resins have been developed,23 together with multimaterial inkjet based printers.24 However, currently, both of the above are expensive and time-consuming, with issues of moving the printed body between the various resins associated with the former and



EXPERIMENTAL SECTION Materials. NaCl and atrazine were obtained from SigmaAldrich (Sydney, Australia). KNO3 and CuSO4 were obtained from AJAX Chemicals (Sydney, Australia), while HPLC-grade methanol was obtained from Scharlab, S.L. (Barcelona, Spain). Deionized water (Millipore, Melbourne, Australia) was used for the preparation of all solutions. The hyper-cross-linked polystyrene (HCPS) resin (MN-150) (particle size: 75−125 μm; surface area: 800 m2 g−1; bimodal pore size distribution with micropores of diameter 1.4 nm and macropores of diameter 46 nm; pore volume: 0.6 mL g−1) was obtained from Purolite International Limited (Pontyclun, UK). Supor poly(ether sulfone) (PES) membrane disc filters (pore size: 0.1 μm; thickness: 101.6−157.5 μm) were purchased from Pall Corporation (Michigan, USA). Stainless steel wing nuts and bolts were obtained from Nuts & Bolts, Moonah (Hobart, Australia). Device Design and 3D Printing. The CAD designs for the device were created using Inventor Pro (Autodesk, San 12082

DOI: 10.1021/acs.analchem.8b02893 Anal. Chem. 2018, 90, 12081−12089

Article

Analytical Chemistry

Figure 1. (a) CAD design and (b) 3D printed image of the passive sampling device; (c) homemade device for diffusion experiment consisting of two reservoirs with the 3D printed membrane in the middle.

filament (Matter Hackers, CA, USA) was used for the simultaneous printing of the membrane. The extrusion temperature was set at 230 °C for Lay-Felt and 200 °C for the PLA; the nozzles were kept warm at 180 °C when not being used to minimize material leakage. The layer thickness of 150 μm and extrusion width of 300 μm were used for printing the device. The printing bed temperature was 70 °C. The freshly printed devices were washed by soaking in water for a minimum of 48 h (replacing the water every 24 h) or until the PVA was fully removed from the membrane, at which point the wash solution appeared clear. The devices were then further soaked in methanol for 3 h, before being stored in polyethylene bags when not in use. Each device was then immersed in deionized water for at least 3 h prior to use. Membrane Characterization. Scanning electron microscopy (SEM) was carried out using a Hitachi SU70 instrument (Hitachi High Technologies America, USA). Lay-Felt filament samples were prepared by placing a small section of the printed membrane onto carbon tape on an Al SEM stub. Samples were platinum sputter coated (approximately 4 nm layer) at 1.5 kV on a Bal-Tec SCD 050 Sputter Coater (Bal-Tec AG, Balzers, Liechtenstein). The average pore size was determined by nitrogen adsorption with a Micromeritics Tristar II automated gas sorption−desorption instrument (Micromeritics, Norcross, GA, USA) using the Brunauer−Emmett−Teller (BET)

Rafael, USA), and the CAD files, exported as .stl files, were converted into G-code using KISSlicer PRO (freely available at www.kisslicer.com). The print time of each device was approximately 2.5 h, and the final CAD design and 3D printed passive sampler are shown in Figure 1a,b, respectively. The representative images of the top and bottom part of the final design are provided within the Supporting Information. The passive sampling device with integrated porous membrane was fabricated using a FELIX 3D FDM printer consisting of dual extruders (FELIXprinters, IJsselstein, The Netherlands). PLA and acrylonitrile butadiene styrene (ABS), with recommended printing temperatures of 200 °C and 225− 235 °C, respectively, are both compatible with the Lay-Felt material (printing temperature, 225−235 °C). In this current study, PLA was chosen for printing the device body due to its eventual biodegradability. However, biodegradability of PLA is a very slow process, with a reported degradation of ∼10 nm thickness over a period of over 2 weeks (aqueous, pH = 7).30 Previous research has reported up to 30 days were required to detect any decrease in dry matter for PLA disks immersed in water.31 With this information, no measurable degradation to the device body was expected during the planned deployment testing periods (14 days at neutral pH). The main circular housing structure of the sampler was printed with black PLA (Matter Hackers, CA, USA), and Poro-Lay Lay-Felt porous 12083

DOI: 10.1021/acs.analchem.8b02893 Anal. Chem. 2018, 90, 12081−12089

Article

Analytical Chemistry

Figure 2. Low resolution SEM images of the (a) 3D printed membrane before washing, (b) 3D printed membrane after washing, and (c) Lay-felt filament after washing; high resolution SEM images of the (d) 3D printed membrane before washing, (e) 3D printed membrane after washing, and (f) Lay-Felt filament after washing.

method. Prior to analysis, ∼200 mg of each sample in triplicate was dried in a Micromeritics VacPrep-061 at 80 °C for up to 48 h. Preliminary diffusion experiments were conducted by employing a simple dual reservoir (500 mL) setup, separated by the 1 mm thick 3D printed porous membrane (Figure 1c). For initial diffusion studies, absorbing test probes (0.025 M KNO3, 0.1 M CuSO4, and 50 mg L−1 atrazine) were added to solution in the first reservoir and deionized water was used to fill the second reservoir. 1.5 mL of samples was taken out at regular intervals to determine the transport rates for various probes across the membrane using UV/vis absorbance spectroscopy (Metertech SP-8001 UV/Visible Spectrophotometer, Metertech Inc., Taipei, Taiwan), measuring the absorbance of the deionized water over regular intervals until no further increase in absorbance was observed and an equilibrium was achieved (nitrate, copper, and atrazine at 200, 645, and 280 nm, respectively). For leaching experiments, approximately 0.5 g of each of the Lay-Felt filament, PLA filament, and commercial PES membrane was soaked in 20 mL of 100% methanol. Samples of methanol were collected every hour for analysis and replaced with fresh methanol, for a period of 3 h. The methanol extracts were analyzed by GC/MS for the presence and identification of volatile leachables. Atrazine Extraction and GC Analysis. For atrazine sampling, the 3D printed sampler devices, containing the integrated 3D printed membrane and commercial PES membrane, were loaded with 200 mg of the hyper-crosslinked polystyrene (Purolite, MN-150) sorbent and deployed in a tank containing 3 L of artificially created seawater (3.5% NaCl) spiked with 10 μg L−1 of atrazine. All experiments were conducted in the dark (setup wrapped in aluminum foil) at room temperature with continuous magnetic stirring. Aliquots (9 mL) of the seawater solution were collected in triplicate every 2 days over a total period of 14 days. PES membranes and HCPS resin were washed prior to use. PES was prepared by soaking in 20%, 50%, and 100% methanol, each for 24 h.

HCPS resin was prepared by a 6 h Soxhlet extraction in 100% methanol followed by 30 min of drying in an oven at 70 °C. A GC2010 system equipped with a GC/MS QP2010 plus (Shimadzu Corp., Kyoto, Japan), a CombiPAL autosampler, and a GC capillary column, 5% phenyl polysilphenylenesiloxane (BP5MS) (30 m × 0.25 mm × 0.25 μm) (Supelco Inc., Pennsylvania, USA), was used for the analysis of the artificial seawater solutions. Solid phase microextraction (SPME) was performed for the extraction of atrazine from extracted water samples for GC/MS analysis. A polydimethylsiloxane/divinylbenzene (PDMS/DVB) SPME fiber assembly (part no.: 57311 Supelco) was obtained from SigmaAldrich (Sydney, Australia). The PDMS/DVB SPME fiber (fiber coating thickness (df) of 65 μm; needle size: 24 G) was initially conditioned following the guidelines of the manufacturer prior to application. The GC/MS autosampler consisted of two heating chambers: (1) incubator and (2) SPME fiber conditioning station. The preincubation of the SPME vial with samples was carried out in the incubator at room temperature for 5 min. The SPME/DVB fiber was then immersed in the sample, and the extraction was performed at room temperature in the incubator for 30 min. Following extraction, the SPME/DVB fiber was desorbed in the GC injector at 250 °C for 3 min. Helium gas (flow rate = 1.4 mL min−1) and splitless injection (at 250 °C) were used for analysis. The temperature program included: initial oven temperature of 50 °C for 1 min and final oven temperature of 250 °C for 5 min (at 15 °C min−1), MS ion source of 220 °C, and transfer line of 250 °C. Mass spectra were obtained in the mass range of 50−350 m/z at a scan speed of 3333 μ s−1. The chromatographic peak area was integrated using GC/MS solution software (Shimadzu Corp., Kyoto, Japan). A calibration curve of peak area against concentration (1−10 μg L−1) was obtained for quantitative analysis of the atrazine in water samples (n = 3). Rs (Ld−1) was calculated from the following equation 12084

DOI: 10.1021/acs.analchem.8b02893 Anal. Chem. 2018, 90, 12081−12089

Article

Analytical Chemistry Rs =

Ci − C t V × Ci t

from the printing process were superficial only. This was further confirmed by the BET analysis of the printed Lay-Felt membranes, which revealed an average pore diameter of 34 nm, further confirming the larger diameter pores/cavities (≥1 μm) from the printing process are shallow surface features only. It should be noted here that average pore diameter (34 nm) obtained in this study was slightly larger than the 12 nm average previously reported for this Lay-Felt material,28 which is a likely reflection of the extended washing time in this study to ensure complete removal of PVA. Previous studies have used SEM for an estimate of various surface properties35,36 including pore size, shape, and pore size distribution.35 The pore size distribution and the average pore size were determined by measuring the diameter of 500 pores (using ImageJ software), and the histograms for these data sets were obtained using 22 bins, where the bin width = range of values/no. of bins (Figure 3). The pore size distribution

(1)

where Ci and Ct are the concentrations of atrazine at the beginning of the experiment and end of exposure (μg L−1), V is the volume of water sample, and t represents the number of deployment days/exposure time.32



RESULTS AND DISCUSSION Device Design and 3D Printing. The passive sampler device was designed considering three key requirements: (1) having an integrated porous membrane, (2) providing adequate sealing to ensure no loss of loaded microparticulate sorbent, and (3) offering very simple assembly. An initial design consisting of a top and bottom ring-shaped holder (I.D. = 60 mm, O.D. = 90 mm, thickness = 4 mm) with an integrated membrane (diameter = 60 mm; thickness = 1 mm) was evaluated. However, these preliminary devices saw detachment of the membrane due to poor adhesion of LayFelt with the supporting PLA material, which resulted in the loss of sorbent enclosed within the device. To improve sealing and membrane attachment to the device, a membrane of 70 mm diameter was utilized with 10 mm embedded in the body of the PLA supporting ring structure (Figure 1a), still leaving a 60 mm span of the porous membrane exposed (see inset of Figure 1b). Additionally, a 3 mm deep groove in the bottom section and a 3 mm raised ring in the top section were introduced in the design to effectively attach and seal the two parts of the device, without the need for an O-ring, and inhibit any sorbent loss. Finally, the design included 3 mm openings for marine grade nuts and 2 mm openings at the boundaries of the samplers for attachment during deployment (Figure 1a,b). Membrane Characterization. Lay-Felt material used for printing the membrane consists of two major components, a rubber-elastomeric polymer and a water-soluble PVA component, which can be dissolved in water after the printing process is complete. The resulting porous membrane is relatively elastic in comparison to the unwashed membrane, which is rigid. SEM was used for characterization of the pore size distribution within the membranes. Figure 2a−c shows the obtained SEM images of the unwashed printed membrane, washed printed membrane, and washed Lay-Felt filament, respectively. It can clearly be seen that both washed and unwashed membranes have striations from the printing process; however, the unwashed membrane has a relatively smooth surface (Figure 2a) in comparison to the washed membrane, which appeared to be rather rough (Figure 2b). Imaging at higher magnification confirmed the presence of pores in the nanometer range and several cavities (up to 1 μm) on the surface of the washed 3D printed membrane (Figure 2e). These cavities/larger size pores (≥1 μm) were likely to originate from the printing process itself, as they were found to be absent in the washed filament surface, which consisted of only smaller pores with an average pore size of ∼30 nm (Figure 2c,f). The formation of cavities/air pockets due to heterogeneity between the layers and filament strands is a common issue when using FDM printing.33,34 To confirm if these air pockets due to printing process were interconnected and could allow unwanted liquid transport, a small cuvette (H × W × D: 50 mm × 15 mm × 15 mm) with 0.5 mm wall thickness (minimum thickness that could be fabricated with mechanical stability) was printed using PLA. No leakage of water through the walls confirmed that these cavities arising

Figure 3. Histogram of pore size distribution (from the SEM image) of washed Lay-Felt filament, obtained by measuring 500 pores.

obtained by SEM for the Lay-Felt filament was found to be in the range of 15−240 nm with most of the pores located in the 30−80 nm range (average pore size ∼30 nm corresponding with BET results). Preliminary experiments were conducted to explore the selectivity of the 3D printed membrane. As passive sampling is typically applied to monitoring of low level persistent organic pollutants, atrazine (a neutral molecule at environmental pH of between 5 and 8, with pKa1 = 1.60 and pKa2 = 1.95)37 was chosen as a suitably representative solute (e.g., for triazines and triazoles),38 and significant previous experimental data for extraction using passive samplers exist for it (see Table 1). Furthermore, to confirm the charge selectivity, if any, of the printed membrane, the transport of small inorganic anions (NO3−) and cations (Cu2+) across the membrane was also determined, using the simple setup shown in Figure 1c. Transport of the above test solutes across the membrane was determined by measuring an increase in the absorbance of the reservoir containing deionized water at regular intervals until apparent equilibrium was achieved. Figure 4 shows the absorbance vs time plot for NO3− and atrazine, reflecting the selectivity of the membrane for small inorganic anion species and neutral organic compounds. However, the transfer of Cu2+ across the membrane was not observed, and no change in the absorbance was observed for the reservoir containing 12085

DOI: 10.1021/acs.analchem.8b02893 Anal. Chem. 2018, 90, 12081−12089

Article

Analytical Chemistry

Table 1. Comparison of Atrazine Extraction Using the 3D Printed Passive Sampling Device with Integrated Membrane and Other Passive Sampling Devices41 device

membrane

sorbent

sample

3D printed sampler

Lay-Felt

hyper-cross-linked polystyrene (Purolite, MN150)

distilled water

POCIS

PES

Oasis HLB

tap water

POCIS Chemcatcher Chemcatcher POCIS

PES PES none PES

Oasis HLB Oasis HLB SDB-XC Empore disk 80:20 (weight/weight) Isolute ENV1/Ambersorb 1500 dispersedon S-X3 Bio Beads

river water river water tap water distilled water

POCIS POCIS

PES PES

Chemcatcher

PES

POCIS

PES

(i) polystyrene divinylbenzene polymer, (ii) anion exchange sorbent poly(styrene-co-divinylbenzene) reversed phase sulfonated 80:20 (weight/weight) Isolute ENV1/Ambersorb 1500 dispersedon S-X3 Bio Beads

seawater drinking water, river water treated sewage water seawater

calibration type static (continuous stirring) static (turbulent) laboratory laboratory continuous flow static (continuous stirring) laboratory static (turbulent) flow through static (continuous stirring)

atrazine spiking (μg/ L)

sampling days

Rs (d L−1)

10

14

0.19

5

21

0.2346

0.4 0.4 100 5

26 26 14 56

0.57, 0.974 0.03, 0.104 0.2849 0.0550

24 21

0.0451 0.10, 0.19, 0.2052 0.12−0.5253

0.17 10

3−24 0.5

7

0.2154

Lay-Felt and PLA materials were also investigated for the leaching of any unreacted polymer, plasticizers, or solvents, into the sample solution, as these could potentially interfere with the extraction or subsequent detection of solutes of interest. The leaching of material from the Lay-Felt material, PLA, and a commercial PES membrane was determined. The methanol wash of each material was examined for volatile leachable using GC/MS. Figure 5a shows the chromatograms for time studies of the leaching experiment for PLA, PES, and Lay-Felt. The methanol extracts from PLA and PES contained peaks of dodecamethylcyclohexasiloxane and tetrasiloxane possibly from unreacted materials or byproducts. The methanol extract from Lay-Felt material showed minimum leaching of material, with the chromatogram consisting of several peaks but at negligible concentrations, potentially related to the washing step with deionized water for more than 48 h prior to the experiment. These results suggest the materials used for printing the sampler can be employed for extraction after simple rinsing with water and methanol, with minimum amount of leachable material, which could otherwise potentially contaminate the extraction sorbent. Figure 5b shows a comparison of chromatograms from the SPME of the artificially created seawater samples after 2 and 10 days of extraction, using the 3D printed passive sampler with the integrated Lay-Felt membrane (solution sampled using SPME), revealing the concentration of remaining atrazine. The sampler, consisting of a 0.5 mm thick integrated membrane, was deployed for extraction after first rinsing with deionized water for 48 h, followed by 3 h of washing with 100% methanol. The chromatogram for the 10 day depletion experiment showed a 77% decrease in peak area for atrazine in comparison to the 2 day depletion chromatogram (the calibration curve for peak area vs concentration using the SPME method is also shown in Figure 5b). However, in addition to the peak for atrazine, the chromatograms also revealed the presence of several organic acid species (lauric and palmitic acid), which are known to be present within both PLA and thermoplastic materials,39,40 and had likely leached from the sampler body. However, neither species interfered with the determination of atrazine.

Figure 4. Time vs absorbance graph for diffusion of atrazine (50 mg L−1) and nitrate (0.025 M KNO3) across the membrane (n = 3 ± SD).

deionized water (assuming SO42−, like NO3−, was free to transport across the membrane, any Cu2+ in the receiving reservoir should have given rise to an increase in absorbance, here not seen). These results point to the 3D printed membrane being selective for anions and neutral solutes only and not allowing the transfer of cationic species. This is in agreement with previously reported work using Lay-Felt membranes, where thiocyanate was shown to migrate across the membrane into a chamber containing Fe3+, represented by development of red color on only one side of the membrane, showing no transport of the cation in the reverse direction.28 12086

DOI: 10.1021/acs.analchem.8b02893 Anal. Chem. 2018, 90, 12081−12089

Article

Analytical Chemistry

deployed in artificial seawater spiked with atrazine. The artificial seawater was analyzed using SPME-GC/MS at 2 day intervals up to a total of 14 days for % depletion of atrazine and to calculate the Rs. Figure 6a,b shows atrazine depletion (%) in

Figure 5. (a) GC/MS chromatograms for leaching extracts of PLA filament, PES membranes, and Lay-Felt filament in 100% methanol collected after every hour up to 3 h. (b) GC/MS chromatograms for the concentration of atrazine remaining in the artificial seawater samples collected after 2 and 10 days of extraction (concentration at 0 day = 10 μg L−1), using the passive sampling device with integrated membrane of 0.5 mm thickness. Peak identification: (1) 2,3dimethylbutyric acid methyl ester; (2) dodecamethylcyclohexasiloxane; (3) tetrasiloxane; (4) diphenyl sulphone; (5) lauric acid; (6) 1,3,5-triazine-2,4-diamine; (7) palmitic acid.

Figure 6. (a) % depletion vs days and (b) sampling rates (L d−1) vs days plots for atrazine using the commercial membrane and 3D printed POCIS device with integrated membranes of 0.5, 1.0, and 1.5 mm thickness.

artificial seawater and Rs for the printed devices with integrated membranes of various thicknesses, together with a 3D printed device fitted with a commercial membrane. These results showed that the performance of the device with the 0.5 mm thickness integrated membrane, with depletion values of 87% and a Rs value of 0.19 L d−1 (average from n = 3), was very comparable with the depletion (92%) and the Rs values (0.20 L d−1) (at 14 days) recorded using the sampling device fitted with the commercial PES membrane. The transfer across the 1.5 mm thick membrane was negligible, with % depletion and sampling rate of 5.1% and 0.011 L d−1 (n = 3). These results establish that the transfer through the membrane is a diffusion based process and the performance of the passive sampling device depends on the membrane thickness, which could also be the cause of the slightly better performance of the commercial PES material (thickness >0.2 mm) in comparison to the 0.5 mm thick 3D printed membrane. The Rs for atrazine obtained using this 3D printed sampling device with the integrated membrane is in close agreement with previously reported values from laboratory based experiments, which confirms the potential of the printed device for real applications. A comparison of atrazine extraction using the 3D printed passive sampler with previously reported devices is shown in Table 1.

Membrane Thickness and Atrazine Depletion Experiment. Transfer of solutes through the Lay-Felt membrane is a diffusion based process, and the performance of sampling devices in terms of extraction efficiency and Rs is expected to be highly dependent on membrane thickness. Therefore, devices with various membrane thicknesses, including 0.5, 1, and 1.5 mm, were investigated. A sampling device with a membrane thickness of less than 0.5 mm was not examined as it was not structurally robust and resulted in mechanical failure of the membrane during handling, resulting in leakage of the resin powder into the water reservoir. Laboratory based deployment and calibration were used in this study to evaluate the performance of newly developed 3D printed passive sampling devices, in an attempt to deliver reliable Rs for the test solute (atrazine) under controlled sampling and sample conditions.41 Previous studies have used similar approaches to evaluate the performance of sampling devices for providing TWA concentration of solutes under investigation.32,42−46 HCPS MN-150 was selected as adsorbent for this set of experiments, as this resin has been successfully applied previously for the quantitative removal of neutral pesticides (e.g., methomyl) from aqueous solutions.47 The passive sampling devices loaded with the adsorbent were 12087

DOI: 10.1021/acs.analchem.8b02893 Anal. Chem. 2018, 90, 12081−12089

Analytical Chemistry



CONCLUSION



ASSOCIATED CONTENT



In this work, MM-FDM 3D printing was used for the very first time to fabricate a passive sampling device body with an integrated porous circular membrane. As the membrane circumference is embedded in the structure of the sampler body, it does not require additional O-rings or other assembly parts to prevent the encased sorbent leakage, therefore providing very simple and quick assembly. Using a dual nozzle FDM printer, cost-effective passive sampling devices tailored to the customer demands can be produced, which are otherwise difficult to achieve using traditional subtractive manufacturing techniques. The focus of this study was device development and characterization of porous Lay-Felt membrane, and atrazine was chosen as the representative solute to demonstrate the potential application of the multimaterial 3D printed sampler. This work provides proof-of-concept that such 3D printable membranes on this relatively large scale can perform similarly to commercial polymer membranes and be printed simultaneously to supporting structures, as here in the case of these passive samplers. The developed sampler device is a lowcost alternative to current sampling products and can be produced in-house, in customized formats, by researchers with access to low-cost printers. With continuous development in porous printing materials48 and an increasing number of commercially available printable materials with various porosities, 3D printing shows a great potential for the manufacturing of the highly selective sampling devices by selecting membrane materials with desired properties (e.g., pore size and surface charge) for a wide range of environmental applications.

* Supporting Information The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.8b02893. Representative images of the top and bottom part of the final design of the passive sampling device (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Michael C. Breadmore: 0000-0001-5591-4326 Pavel N. Nesterenko: 0000-0002-9997-0650 Brett Paull: 0000-0001-6373-6582 Notes

The authors declare no competing financial interest.



REFERENCES

(1) Eichelberger, J. W.; Lichtenberg, J. Environ. Sci. Technol. 1971, 5, 541−544. (2) Spies, B.; Dandy, G. Sustainable water management, Securing Australia’s future in a green economy. A Report of a study by the Australian academy of technological sciences and engineering (ATSE); Australian Academy of Technological Sciences and Engineering: Melbourne, Australia, 2012. (3) https://en.unesco.org/waterquality-IIWQ/activities-projects/ monitoring-water-quality-wastewater. Access date 17−05−2018. (4) Ahrens, L.; Daneshvar, A.; Lau, A. E.; Kreuger, J. J. Chromatogr. A 2015, 1405, 1−11. (5) Hale, S. E.; Oen, A. M. P.; Cornelissen, G.; Jonker, M. T. O.; Waarum, I.-K.; Eek, E. Mar. Pollut. Bull. 2016, 111, 33−40. (6) Dalton, R. L.; Pick, F. R.; Boutin, C.; Saleem, A. Environ. Pollut. 2014, 189, 134−142. (7) Seethapathy, S.; Górecki, T.; Li, X. J. Chromatogr. A 2008, 1184, 234−253. (8) Miege, C.; Budzinski, H.; Jacquet, R.; Soulier, C.; Pelte, T.; Coquery, M. J. Environ. Monit. 2012, 14, 626−635. (9) Alvarez, D. A.; Huckins, J. N.; Petty, J. D.; Jones-Lepp, T.; StuerLauridsen, F.; Getting, D. T.; Goddard, J. P.; Gravell, A. In Comprehensive Analytical Chemistry; Greenwood, R.; Mills, G.; Vrana, B., Eds.; Elsevier: New York, 2007; pp 171−197. (10) Cernoch, I.; Franek, M.; Diblikova, I.; Hilscherova, K.; Randak, T.; Ocelka, T.; Blaha, L. J. Environ. Monit. 2011, 13, 2582−2587. (11) Hua, W.; Bennett, E. R.; Letcher, R. J. Water Res. 2006, 40, 2259−2266. (12) Villanueva, C.; Durand, G.; Coutte, M.; Chevrier, C.; Cordier, S. Occup. Environ. Med. 2005, 62, 400−405. (13) Kalsoom, U.; Nesterenko, P. N.; Paull, B. RSC Adv. 2016, 6, 60355−60371. (14) Kalsoom, U.; Peristyy, A.; Nesterenko, P. N.; Paull, B. RSC Adv. 2016, 6, 38140−38147. (15) Cecil, F.; Zhang, M.; Guijt, R. M.; Henderson, A.; Nesterenko, P. N.; Paull, B.; Breadmore, M. C.; Macka, M. Anal. Chim. Acta 2017, 965, 131−136. (16) Waheed, S.; Cabot, J. M.; Macdonald, N. P.; Lewis, T.; Guijt, R. M.; Paull, B.; Breadmore, M. C. Lab Chip 2016, 16, 1993−2013. (17) Salentijn, G. I. J.; Oomen, P. E.; Grajewski, M.; Verpoorte, E. Anal. Chem. 2017, 89, 7053−7061. (18) Gupta, V.; Talebi, M.; Deverell, J.; Sandron, S.; Nesterenko, P. N.; Heery, B.; Thompson, F.; Beirne, S.; Wallace, G. G.; Paull, B. Anal. Chim. Acta 2016, 910, 84−94. (19) Sandron, S.; Heery, B.; Gupta, V.; Collins, D. A.; Nesterenko, E. P.; Nesterenko, P. N.; Talebi, M.; Beirne, S.; Thompson, F.; Wallace, G. G.; Brabazon, D.; Regan, F.; Paull, B. Analyst 2014, 139, 6343−6347. (20) Gupta, V.; Mahbub, P.; Nesterenko, P. N.; Paull, B. Anal. Chim. Acta 2018, 1005, 81−92. (21) Kalsoom, U.; Nesterenko, P. N.; Paull, B. TrAC, Trends Anal. Chem. 2018, 105, 492−502. (22) Wong, K. V.; Hernandez, A. ISRN Mech. Eng. 2012, 2012, 208760. (23) Choi, J.-W.; Kim, H.-C.; Wicker, R. J. Mater. Process. Technol. 2011, 211, 318−328. (24) Murphy, S. V.; Atala, A. Nat. Biotechnol. 2014, 32, 773−785. (25) Low, Z.-X.; Chua, Y. T.; Ray, B. M.; Mattia, D.; Metcalfe, I. S.; Patterson, D. A. J. Membr. Sci. 2017, 523, 596−613. (26) Pinger, C. W.; Heller, A. A.; Spence, D. M. Anal. Chem. 2017, 89, 7302−7306. (27) Yuan, S.; Strobbe, D.; Kruth, J.-P.; Van Puyvelde, P.; Van der Bruggen, B. J. Mater. Chem. A 2017, 5, 25401−25409. (28) Li, F.; Smejkal, P.; Macdonald, N. P.; Guijt, R. M.; Breadmore, M. C. Anal. Chem. 2017, 89, 4701−4707. (29) Belka, M.; Ulenberg, S.; Bączek, T. Anal. Chem. 2017, 89, 4373−4376. (30) Xu, L.; Crawford, K.; Gorman, C. B. Macromolecules 2011, 44, 4777−4782.

S



Article

ACKNOWLEDGMENTS

The authors acknowledge the Central Science Laboratory for access to SEM and Dr. Petr Smejkal for technical assistance. C.K. Hasan acknowledges a postgraduate scholarship from the ARC Training Centre for Portable Analytical Separation technologies (ASTech, Grant No. IC140100022). Financial support for this research was also provided by the ARC Centre of Excellence for Electromaterials Science (ACES, Grant No. CE140100012). 12088

DOI: 10.1021/acs.analchem.8b02893 Anal. Chem. 2018, 90, 12081−12089

Article

Analytical Chemistry (31) Rocca-Smith, J. R.; Whyte, O.; Brachais, C.-H.; Champion, D.; Piasente, F.; Marcuzzo, E.; Sensidoni, A.; Debeaufort, F.; Karbowiak, T. ACS Sustainable Chem. Eng. 2017, 5, 2751−2762. (32) Rujiralai, T.; Bull, I. D.; Llewellyn, N.; Evershed, R. P. J. Environ. Monit. 2011, 13, 1427−1434. (33) Tothill, A. M.; Partridge, M.; James, S. W.; Tatam, R. P. J. Micromech. Microeng. 2017, 27, 035018. (34) Cook, K.; Canning, J.; Leon-Saval, S.; Reid, Z.; Hossain, M. A.; Comatti, J.-E.; Luo, Y.; Peng, G.-D. Opt. Lett. 2015, 40, 3966−3969. (35) Abdullah, S. Z.; Bérubé, P. R.; Horne, D. J. J. Membr. Sci. 2014, 463, 113−125. (36) Wang, X.; Bai, H.; Shi, G. J. Am. Chem. Soc. 2011, 133, 6338− 6342. (37) Colombini, M. P.; Fuoco, R.; Giannarelli, S.; Pospíšil, L.; Trsková, R. Microchem. J. 1998, 59, 239−245. (38) Fontecha-Cámara, M. A.; López-Ramón, M. V.; Á lvarezMerino, M. A.; Moreno-Castilla, C. Langmuir 2007, 23, 1242−1247. (39) Inkinen, S.; Hakkarainen, M.; Albertsson, A.-C.; SödergÅrd, A. Biomacromolecules 2011, 12, 523−532. (40) Ryan, K.; Lupton, K. E.; Pape, P.; John, V. J. Vinyl Addit. Technol. 2000, 6, 7−19. (41) Morin, N.; Miège, C.; Coquery, M.; Randon, J. TrAC, Trends Anal. Chem. 2012, 36, 144−175. (42) Ibrahim, I.; Togola, A.; Gonzalez, C. Environ. Sci. Pollut. Res. 2013, 20, 3679−3687. (43) MacLeod, S. L.; McClure, E. L.; Wong, C. S. Environ. Toxicol. Chem. 2007, 26, 2517−2529. (44) Arditsoglou, A.; Voutsa, D. Environ. Pollut. 2008, 156, 316− 324. (45) Jones-Lepp, T. L.; Alvarez, D. A.; Petty, J. D.; Huckins, J. N. Arch. Environ. Contam. Toxicol. 2004, 47, 427−439. (46) Mazzella, N.; Dubernet, J.-F.; Delmas, F. J. Chromatogr. A 2007, 1154, 42−51. (47) Chang, C.-F.; Chang, C.-Y.; Hsu, K.-E.; Lee, S.-C.; Höll, W. J. Hazard. Mater. 2008, 155, 295−304. (48) Mu, X.; Bertron, T.; Dunn, C.; Qiao, H.; Wu, J.; Zhao, Z.; Saldana, C.; Qi, H. Mater. Horiz. 2017, 4, 442−449. (49) Gunold, R.; Schäfer, R. B.; Paschke, A.; Schüürmann, G.; Liess, M. Environ. Pollut. 2008, 155, 52−60. (50) Alvarez, D. A.; Petty, J. D.; Huckins, J. N.; Jones-Lepp, T. L.; Getting, D. T.; Goddard, J. P.; Manahan, S. E. Environ. Toxicol. Chem. 2004, 23, 1640−1648. (51) Hernando, M. D.; Martínez-Bueno, M. J.; Fernandez-Alba, A. R. Bol. Inst. Esp. Oceanogr. 2005, 21, 37−46. (52) Fauvelle, V.; Mazzella, N.; Delmas, F.; Madarassou, K.; Eon, M.; Budzinski, H. Environ. Sci. Technol. 2012, 46, 13344−13353. (53) Vermeirssen, E. L. M.; Bramaz, N.; Hollender, J.; Singer, H.; Escher, B. I. Water Res. 2009, 43, 903−914. (54) Martínez Bueno, M. J.; Hernando, M. D.; Agüera, A.; Fernández-Alba, A. R. Talanta 2009, 77, 1518−1527.

12089

DOI: 10.1021/acs.analchem.8b02893 Anal. Chem. 2018, 90, 12081−12089