One-Step Fabrication of a Microfluidic Device with an Integrated

A microfluidic application of this was demonstrated by Begolo et al., who fabricated ... chemicals across chambers that were only fluidically connecte...
0 downloads 0 Views 1MB Size
Subscriber access provided by University of Newcastle, Australia

Article

One step fabrication of a microfluidic device with an integrated membrane and embedded reagents by multimaterial 3D printing Feng Li, Petr Smejkal, Niall P. Macdonald, Rosanne M Guijt, and Michael C. Breadmore Anal. Chem., Just Accepted Manuscript • Publication Date (Web): 21 Mar 2017 Downloaded from http://pubs.acs.org on March 21, 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 19

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

One step fabrication of a microfluidic device with an integrated membrane and embedded reagents by multi-material 3D printing

Feng Li1,2,3, Petr Smejkal1, Niall P. Macdonald1,3, Rosanne M. Guijt2,3, Michael C. Breadmore1,3* 1

Australian Centre for Research on Separation Science, School of Chemistry, University of Tasmania, Private Bag 75, Hobart, Tasmania 7001, Australia.

2

School of Medicine and Australian Centre for Research on Separation Science, , University of Tasmania, Private Bag 26, Hobart, Tasmania 7001, Australia 3

ARC Centre of Excellence for Electromaterials Science (ACES), School of Chemistry, University of Tasmania, Hobart, 7001, TAS, Australia

Abstract One of the largest impediments in the development of microfluidic-based smart sensing systems is the manufacturability of integrated, complex devices. Here, we propose multi-material 3D printing for the fabrication of such devices in a single step. A microfluidic device containing an integrated porous membrane and embedded liquid reagents was made by 3D printing and applied for the analysis of nitrate in soil. The manufacture of the integrated, sealed device was realised as a single print within 30 min. The body of the device was printed in transparent acrylonitrile butadiene styrene (ABS), and contained a 400 µm wide structure printed from a commercially available composite filament. The composite filament can be turned into a porous material through dissolution of a water-soluble material. *

Email: [email protected]

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

Liquid reagents were integrated by briefly pausing the printing before resuming for sealing the device. The devices were evaluated by the determination of nitrate in a soil slurry containing zinc particles for the reduction of nitrate to nitrite using the Griess reagent. Using a consumer digital camera, the linear range of the detector response ranged from 0-60 ppm, covering the normal range of nitrate in soil. To ensure the sealing of the reagent chamber is maintained, aqueous reagents should be avoided. When using the non-aqueous reagent, the multi-material device containing the Griess reagent could be stored for over 4 days but increased the detection range to 100-500 ppm. Multi-material 3D printing is a potentially new approach for the manufacture of microfluidic devices with multiple integrated functional components.

Introduction One of the main impediments to the uptake of microfluidic technologies beyond the academic setting has been the difficulty in translating academic research into commercially viable products due to the difficulty in the manufacture of the microfluidic devices1. A major contributing factor to this is the widespread use of polydimethylsiloxane (PDMS) devices in early stage research.2 The material properties make PDMS an unattractive choice for many applications: it adsorbs small hydrophobic molecules from solution,3 has serious absorption of proteins,4 and is gas permeable.5 Casting is also unattractive for large volume manufacturing. The attraction of PDMS for early stage research is the simplicity and hence accessibility to different generations of devices in the design phase. The production of a new single layer PDMS microfluidic device takes about a day and costs approximately $215.6 3D printing provides a new approach that offers similar simplicity similar but provides access to materials that may be more compatible with commercial translation. Significant improvement in the resolution and the reduction in price of 3D printers over the past 5 years has enabled the production of microfluidic devices in hours for a few dollars of material.7 The 3D printing technologies most widely adopted in microfluidics8,7,9 are stereolithography (SL) 10-13, fused deposition modelling (FDM) 14,15, and photopolymer inkjet printing. 16,17 To move from academically interesting research towards microfluidic systems capable of handling real world applications, complex, integrated microfluidic systems that typically comprise more than one material are desirable. Multi-material 3D printing, enabling the integration of functional parts with different material properties in a single run, is attractive for the manufacture of these devices. Chio and co-workers developed a multi-material stereolithography (MMSL) machine

ACS Paragon Plus Environment

Page 2 of 19

Page 3 of 19

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

containing 4 resins. MMSL can produce unique multi-material complex parts that are functional and visually illustrative.18 This approach however, is expensive and slow because the printed structure must be moved between the different resin vats during the print process for each material in each layer. Multi-material-inkjet based systems are much quicker as the different materials can be printed in the same print pass for each layer.19,20 For example, the Stratasys Objet Connex printers can print up to 3 different materials, and various combinations of these, in a single run. A microfluidic application of this was demonstrated by Begolo et al. who fabricated a pumping lid by printing two materials with different mechanical properties (rigid and elastic).21 Amorn and co-workers developed their own high resolution, low cost inkjet multi-material 3D printer which can simultaneously print up to 10 different materials at a cost of $700022. While impressive, this printing approach also suffers from one of the major limitations of inkjet printers for microfluidics: the difficulty in removing the support material. FDM printers do not require a support material and multi-material printing is easier to realise than for SL, but offers poorer resolution than SL and inkjet. FDM uses thermoplastics, providing access to a much greater range of materials than the other printing technologies. Filaments including acrylonitrile butadiene styrene (ABS), polylactic acid (PLA), polycarbonate, polyamide, and polystyrene provide a wide range of physicochemical and mechanical properties properties, and hence a choice of colour, optical transparency, chemical and biocompatibility. Cronin et al. demonstrated pausing the FDM print process to embed solid particles and liquid reagents during the print job, providing a next level of integration.23 The liquid reagents were used for chemical synthesis by inverting the device, allowing for mixing of different chemicals across chambers that were only fluidically connected on the top.14,23 However, there was still only one material used for the fabrication of the device. FDM printers with multiple print heads allow for the extrusion of different materials during a print run. This has been used to create advanced functional ceramic actuators and sensors from to four different materials24, but there are no reports of multi-material FDM printing in microfluidics. In this paper, we examine multi-material FDM printing for microfluidics through the integration of a porous membrane between two microfluidic chambers. The membrane was made from a composite filament and integrated into an ABS microfluidic device. After exposure to water, a water-soluble compound was removed from the composite leaving a porous membrane. The device, with the Griess reagent embedded in one of the chambers, was used for the colorimetric determination of nitrate in soil. The integration of materials with different physicochemical properties provides a new way to integrate functionalities in a microfluidic device, and provides opportunities for the creation of complex

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

integrated devices in a way that is compatible with rapid prototyping and high volume manufacturing.

Experimental section Materials and chemicals Three different FDM filaments (1.75 mm PLA, clear ABS and Lay-Felt™) were used for fabrication, all purchased from MatterHackers Inc. (Orange County, California, USA). Ferric chloride (FeCl3), Zinc dust, naphthylethylenediamine dihydrochloride, calcium sulfate (CaS04) and sulphanilamide were purchased from Sigma-Aldrich (St. Louis, USA), potassium thiocyanate (KSCN), potassium nitrate (KNO3),and phosphoric acid were purchased from AJAX Chemicals (Sydney, Australia). All solutions were prepared in mili-Q water obtained from a Millipore (North Ryde, Australia) purification system.

Device design and fabrication The microfluidic device with an integrated membrane was printed using a FELIX 3D FDM printer with dual extruders (FELIXprinters, IJsselstein, The Netherlands). The construction of this device included 4 steps: (i) computer-aided design (CAD); (ii) export the STL file; (iii) digital slicing into multiple 2D layers; (iv) export for printing as a gcode file. Steps (i) and (ii) were accomplished using “AutoCAD 2016 (Student version)” software (Autodesk Inventor, San Rafael, California). Kisslicer PRO was used to perform steps (iii) and (iv), this software is available free or as pro version at www.kisslicer.com. The print time of each device was approximately 30 min, and the printed chip and cross-section image are shown in Figure 1 a) and b). The dimensions of the microfluidic channels in ABS were 20 x 0.7 x 1 mm to avoid blockage from particulate samples, and the Lay Felt™ membrane was printed to be 29 x 1.6 x 0.4 mm. Characterisation of the membrane was performed by pre-washing the membrane with mili-Q water using a Harvard 2000 syringe pump (Harvard Apparatus, Holliston, The USA) to dissolve the soluble template component from the extruded filament leaving behind the permeable membrane. To embed the Griess reagent in the integrated devices used for the Griess reaction, the printer was paused to pipette 15 µL of the liquid reagent before sealing the device, and the printing process was schematically demonstrated in Figure 2. These devices were not washed with water. Each device was only used once to minimize adsorption of analytes onto the channel walls and fouling of the membrane.

Membrane evaluation

ACS Paragon Plus Environment

Page 4 of 19

Page 5 of 19

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

The surface area and microporosity were assessed using a Tristar II analyser for the nitrogen adsorption/desorption isotherm at 77 K (Particle and Surface Science, Gosford, Australia). Experiments to determine the minimum printable membrane thickness were performed using solutions of 25 mM FeCl3 and 2.5 M KSCN. These were stored in a refrigerator at 4⁰C when not used. Calibration curves for nitrate were constructed from KNO3 (10, 20, 40, 60 ppm NO3-). Griess reagent was freshly prepared by adding 0.02 g naphthylethylenediamine dihydrochloride, 0.2 g sulphanilamide and 0.5 g phosphoric acid to 10 mL mili-Q water. A soil surface sample was collected from University of Tasmania grounds, air dried and homogenized with a mortar, and 1 g was placed into a plastic tube mixed with 0.05 g CaS04, 0.01 g zinc dust and 3 mL milli-Q water, shaken for 10 minutes to make a soil slurry.25

Detection and data analysis Colorimetric detection was performed by image processing using image J software photographs taken by a Canon 60D DSLR (Canon Inc., Tokyo, Japan) with a 70 mm F2.8 EX DG Macro lens (Sigma Corp., Kawasaki, Japan). To determine intensity, images were converted to greyscale before determining the mean grey value after background subtraction.

Results and discussion Membranes can be integrated into microfluidic devices in a number of ways. The simplest is to sandwich an existing membrane between two chip halves, positioning it on top of one substrate and sealing it in place by bonding with the second substrate 26,27. This approach was the first approach to integrating membranes and has successfully been demonstrated for a diversity of membrane materials. Unfortunately, it requires precise alignment of the membrane relative to the fluidic channels, complicating the microfabrication process. Sealing the membranes in a leakage-tight manner is also challenging 28. An alternative approach to fabricating membranes in microfluidic devices is to create the membrane in situ after bonding. Nafion™ membranes and photopolymerised hydrogels have been created in situ, both using laborious and technically challenging methods again introducing a number of processing steps post-fabrication 29,30. Multi-material FDM printing offers the ability to integrate multiple materials into a device during fabrication, eliminating additional processing steps and maintaining an automated fabrication pathway. Additionally, FDM printing allows for the embedding of liquid reagents

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

during fabrication, a unique capability that is difficult to realise with the same ease using traditional approaches.

Membrane thickness Transport through the membrane, as well as dissolution of the template to reveal its porous properties, are expected to be diffusion-based processes, and hence it was expected that the membrane thickness would correlate inversely with analysis time. To establish the minimum feature width that could be printed in a fluidically sealed ABS device, PLA was used as a non-porous substitute for the composite filament and structures separating the sample and reagent channels with a thickness of 200, 300, 400 and 700 μm were printed inside the ABS device. Fluidic leaks around the membrane were visualised by the formation of the red iron thiocyanate complex using 20 mM FeCl3 and 2.5 M KSCN at either side of the structure. In devices containing 200 and 300 μm wide structures, the red colour was observed indicating fluidic contact between the channels. The absence of the red colour when using structures thicker than 400 μm indicated the two channels were fluidically sealed. According to the manufacturers’ product specification, the porous filament contains two different components, the insoluble porous elastomer and a water-soluble template (polyvinyl alcohol analogue). The soluble template component can be washed away using water, leaving behind the flexible porous structure.31 Scanning electron microscope (SEM) images comparing washed and unwashed composite are shown in Figure 1 c) and d), showing a more uneven structure for the washed membrane. Imaging at higher magnification did not reveal any pores. The membranes were examined by Brunauer–Emmett–Teller (BET), with results revealing an average pore diameter of 12 nm in the washed filament while no pores detected in the unwashed membrane. To demonstrate the ability of the printed composite filament to function as a membrane, its ability to transport small inorganic ions was investigated. Again using iron thiocyanate, the formation of the red colour was studied in devices containing washed and unwashed 400 µm wide composite structures, and also with PLA as a control. Twenty mM FeCl3 and 2.5 M KSCN were placed on different sides of the membrane, and photographs were taken every 5 min for 5 hours. For devices containing the washed composite material, SCN- rapidly diffused across the porous structure into the chamber containing Fe3+, as shown by development of the red colour within 3 min. For the unwashed composite material, 3 hr was required to first observe the red colour. This longer time includes the time required to dissolve the PVA-like template molecule from the filament to create the porous structure. With the dissolution of the template being a time-dependent process, this property may

ACS Paragon Plus Environment

Page 6 of 19

Page 7 of 19

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

present future opportunities for time-controlled processes in simple microfluidic devices, such as the sequential addition of reagents for chemical reactions. The influence of the width of the porous structure on the diffusion time was studied using washed composite structures of four different thickness (400, 500, 600, and 700 μm). Again, aqueous solution of 20 mM FeCl3 and 2.5 M KSCN were placed on each side of the porous structure and photographs of the microdevice recorded every 5 min for 1 hr. The results shown in Figure 3 indicate that the thinner the structure, the quicker the ions permeate through it, with the same colour intensity observed for the 400 µm wide structure after 5 min taking 20 min to reach with the 700 μm wide structure. For the 400 µm wide porous structure, the colour intensity stabilised after 15 minutes, indicating equilibrium of transport of SCN- across the membrane had been established; for the 700 µm wide structure it took 55 min to reach equilibrium. The intensity increased linearly with thickness over the first 20 minutes for the thinner structures (R2 = 0.99 for the 400 μm and 500 μm wide), but showed some non-linearity for the thicker structures as evidenced by the slight curvature in the intensity vs time plot for the 700 μm wide structure. Based on the time to reach equilibrium, all further experiments were conducted with 400 µm. To assess the repeatability of the proposed chips, 5 different devices with same membrane thickness (400 μm) were tested using same reagent (20 mM FeCl3 and 2.5 M KSCN). Measurement of the colour intensity revealed a relative standard deviation (RSD) of 9.0%.

Colourimetric determination of nitrate The colourimetric determination of nitrate in soil in microfluidic devices is challenging due to the known difficulty in processing samples containing particulates. Nitrate is an essential nutrient in soil, and the content of nitrate in soil is used as an index for the nitrogen nutrition of crops. Fertilisers are used to ensure sufficient nitrate levels in the soil to ensure crop production32, but overuse has a detrimental effect on the environment, causing eutrophication of lakes and seas and consequent algae blooms threatening the ecosystem. Additionally, a large amount of the unabsorbed nitrogen fertilizer volatizes to produce nitrous oxide, a powerful greenhouse gas. It is desirable to have an in-field test to establish the nitrate content in the soil and adjust the fertilizer amount to not over-fertilize. Nitrate in soil is typically determined using a colourimetric test based on the Griess reagent. The Griess reagent reacts with NO2-, hence zinc dust is typically used to reduce the NO3to NO2-. Before the colorimetric measurement, particulate matter from the soil and zinc dust need to be removed by filtration or centrifugation25,33, adding an unwanted processing step that complicates the implementation of on-site testing at agricultural locations. Here, a 3D printed device is presented for the direct

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

determination of nitrate from a soil slurry, requiring only the addition of some zinc dust for the reduction reaction. The 3D printed chip was assessed for the quantitative determination of nitrate by mixing KNO3 solutions (0, 10, 20, 40, 60 ppm NO3-) with zinc dust and then inserting this solution into the left fluidic chamber. The zinc dust reduces the NO3- to NO2which diffuses through the porous structure to react with Griess reagent to form a pink complex, while the porous structure prevents the zinc dust and soil particulates from entering the reagent chamber, allowing for direct photometric determination. Photographs of the device were taken every 60 s using a digital camera and processed using ImageJ software to determine the colour intensity. As illustrated in Figure 4b, the colour intensity increased at rates proportional to the initial NO3concentration, to stabilise at a colour intensity once equilibrium had been established. The equilibrium intensity increased linearly with concentration for 0-60 ppm NO3- (y=0.50x+1.7, intercept=1.7, R2 = 0.99). The LOD of 5 ppm from our method is higher than the 1 ppm reported for a μPAD34 and is likely due to optical scattering from the rough extruded surface of the microchip, but could be addressed by increasing the optical path length. When fast answers are required, the rate at which the colour intensity develops over the first 5 min can be used as an alternative (linear for 0-60 ppm NO3-, R2 = 0.99) but sensitivity (LOD = 40 ppm) is compromised due to insufficient intensity for the lower concentrations in this interval.

Printed devices with embedded liquid reagents For in-field use, devices with integrated reagents are preferred. Using concepts introduced by Cronin et al. in fluidically interconnected channels14,23, we embedded the liquid Griess reagent in the device by pausing the print process. In contrast to previous work, our channels are only fluidically connected when the template molecule in the composite is dissolved to make the structure porous. Initially, aqueous Griess reagent was embedded into the device and used immediately or stored for 1, 2, 3 and 4 days. The performance was assessed by placing 30 ppm nitrite in the sample chamber and monitoring the intensity of Griess reagent on both sides of the membrane using imageJ software. The data in figure 5 a) shows that as the reagent is stored in the device, it dissolves the membrane such that after 4 days the reaction occurs almost immediately on the sample side. However, for the analysis of dirty particulate samples, it is desirable to monitor the reagent side where there will be no interference from the particles. Colour formation on the reagent side (figure 5 b) did not follow a predictable trend. Increasing after one day of storage, then decreasing after two, before stabilizing after 3-4 days to be approximately the same as using the device immediately after fabrication. This is presumably related to the amount of composite that has been dissolved at each time and the diffusion of the nitrate, Griess reagent and/or product through the

ACS Paragon Plus Environment

Page 8 of 19

Page 9 of 19

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

material. There is also an appreciable increase in the time required to reach equilibrium intensity when the composite is not washed out before (compare figure 4b with figure 5b). To try and remove the variability observed with aqueous storage, a solvent which does not dissolve the composite was evaluated. The Griess reagent was prepared in absolute ethanol and stored for 0, 1, 2, 3 and 4 days before 200 ppm nitrite was placed in the sample chamber. As shown in Figure 5c, all the chips showed formation of the pink color first in the reagent side with similar results obtained irrespective of the storage time, with no colour formed on the sample side (figure 5d). This suggests that reagents could be stored for several days when prepared in ethanol. However, almost 20 hrs was required for the color intensity to reach equilibrium and the intensity of 100 ppm KNO3 was very low due to the reaction being performed in ethanol (reduced intensity confirmed in off-chip experiments). While a linear response was obtained (100-500 ppm KNO3, R2=0.98), the sensitivity was lower and the LOD (100 ppm) higher than the level required to detect nitrate in soil, making it not practically useful without further reagent or solvent development. To validate the potential of the device with embedded reagents, a sample containing dry soil was mixed with zinc dust and water and placed in the left chamber, with the aqueous Griess reagent on the right. The colour intensity was measured in the reagent side (Figure 6) and compared to a standard calibration curve created using standard nitrate in 6 different microfluidic devices. The soil slurry was determined to contain 15.2 ppm NO3-, providing a soil concentration of 45.6 ppm, which is in the normal range of nitrate level in soil extract,25 and demonstrates the potential of these printed devices for soil analyses once the sensitivity and time issues can be addressed. Further integration of additional chambers for standard addition and/or printed filament for automatic calibration and white balance is needed to make the devices more robust and reliable for use by untrained users.

Conclusions 3D printing shows much potential for microfluidics, particularly for the fabrication of complex, integrated devices. FDM printers with multiple extruders provide a new approach for the fabrication of multi-material devices with unprecedented simplicity. As material properties can be translated to microfluidic functionalities, multi-material printing may ultimately lead to the fabrication of highly complex, integrated devices using a simple, automated and single-step manufacturing process. Here, we demonstrate the integration of a nanoporous membrane in a microfluidic device in a single print, in a fabrication process allowing for liquid

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

reagent integration. The analytical use of the integrated device was demonstrated by its use for the colorimetric determination of nitrate from soil using the Griess reaction, eliminating the need for filtration/centrifugation. Continued development of materials and printer capability will see this approach established as a low-cost flexible method for the fabrication of complex, integrated microfluidic devices.

Acknowledgements FL would like to acknowledge the University of Tasmania for the provision of a scholarship. MCB wish to acknowledge an Australian Research Council Future Fellowship award (FT130100101). RMG would like to acknowledge the Alexander von Humboldt Foundation for the award of a fellowship for Experienced Researchers. Support from the ARC Centre of Excellence for electromaterials Science (ACES) (Grant CE140100012) for funding and the UTAS Central Science Laboratory for SEM images is also acknowledged. The authors would also like to acknowledge Dr. Min Zhang for discussions.

ACS Paragon Plus Environment

Page 10 of 19

Page 11 of 19

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

References (1) Whitesides, G. M. Nature 2006, 442, 368-373. (2) McDonald, J. C.; Duffy, D. C.; Anderson, J. R.; Chiu, D. T.; Wu, H.; Schueller, O. J. A.; Whitesides, G. M. ELECTROPHORESIS 2000, 21, 27-40. (3) Toepke, M. W.; Beebe, D. J. Lab Chip 2006, 6, 1484-1486. (4) Mukhopadhyay, R. Analytical chemistry 2007, 79, 3248-3253. (5) Heo, Y. S.; Cabrera, L. M.; Song, J. W.; Futai, N.; Tung, Y.-C.; Smith, G. D.; Takayama, S. Analytical chemistry 2007, 79, 1126-1134. (6) Au, A. K.; Lee, W.; Folch, A. Lab on a Chip 2014, 14, 1294-1301. (7) Waheed, S.; Cabot, J. M.; Macdonald, N. P.; Lewis, T.; Guijt, R. M.; Paull, B.; Breadmore, M. C. Lab on a Chip 2016, 16, 1993-2013. (8) Au, A. K.; Huynh, W.; Horowitz, L. F.; Folch, A. Angewandte Chemie International Edition 2016, 55, 3862-3881. (9) Macdonald, N.P., Cabot, J. M., Smejkal, P. , Guijt, R., Paull, B., Breadmore, M.C.. Anal. Chem. Just accepted, 8 March, 2017. (10) Bhargava, K. C.; Thompson, B.; Malmstadt, N. Proceedings of the National Academy of Sciences 2014, 111, 15013-15018. (11) Au, A. K.; Bhattacharjee, N.; Horowitz, L. F.; Chang, T. C.; Folch, A. Lab on a Chip 2015, 15, 1934-1941. (12) Shallan, A. I.; Smejkal, P.; Corban, M.; Guijt, R. M.; Breadmore, M. C. Analytical chemistry 2014, 86, 3124-3130. (13) Cabot, J. M.; Fuguet, E.; Rosés, M.; Smejkal, P.; Breadmore, M. C. Analytical Chemistry 2015, 87, 6165-6172. (14) Kitson, P. J.; Rosnes, M. H.; Sans, V.; Dragone, V.; Cronin, L. Lab on a Chip 2012, 12, 3267-3271. (15) Krejcova, L.; Nejdl, L.; Rodrigo, M. A. M.; Zurek, M.; Matousek, M.; Hynek, D.; Zitka, O.; Kopel, P.; Adam, V.; Kizek, R. Biosensors and Bioelectronics 2014, 54, 421-427. (16) Erkal, J. L.; Selimovic, A.; Gross, B. C.; Lockwood, S. Y.; Walton, E. L.; McNamara, S.; Martin, R. S.; Spence, D. M. Lab on a Chip 2014, 14, 2023-2032. (17) Anderson, K. B.; Lockwood, S. Y.; Martin, R. S.; Spence, D. M. Analytical chemistry 2013, 85, 5622-5626. (18) Choi, J.-W.; Kim, H.-C.; Wicker, R. Journal of Materials Processing Technology 2011, 211, 318-328. (19) Visser, J.; Peters, B.; Burger, T. J.; Boomstra, J.; Dhert, W. J.; Melchels, F. P.; Malda, J. Biofabrication 2013, 5, 035007. (20) Murphy, S. V.; Atala, A. Nature biotechnology 2014, 32, 773-785. (21) Begolo, S.; Zhukov, D. V.; Selck, D. A.; Li, L.; Ismagilov, R. F. Lab on a Chip 2014, 14, 46164628. (22) Sitthi-Amorn, P.; Ramos, J. E.; Wangy, Y.; Kwan, J.; Lan, J.; Wang, W.; Matusik, W. ACM Transactions on Graphics (TOG) 2015, 34, 129. (23) Kitson, P. J.; Symes, M. D.; Dragone, V.; Cronin, L. Chemical Science 2013, 4, 3099-3103. (24) Jafari, M.; Han, W.; Mohammadi, F.; Safari, A.; Danforth, S.; Langrana, N. Rapid Prototyping Journal 2000, 6, 161-175. (25) Singh, J. Plant and Soil 1988, 110, 137-139. (26) Li, F.; Guijt, R. M.; Breadmore, M. C. Analytical Chemistry 2016, 88, 8257-8263. (27) Cannon, D. M.; Kuo, T.-C.; Bohn, P. W.; Sweedler, J. V. Analytical chemistry 2003, 75, 2224-2230. (28) De Jong, J.; Lammertink, R.; Wessling, M. Lab on a Chip 2006, 6, 1125-1139. (29) Haywood, D. G.; Saha-Shah, A.; Baker, L. A.; Jacobson, S. C. Analytical chemistry 2014, 87, 172-187.

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

(30) Kim, S. J.; Han, J. Analytical chemistry 2008, 80, 3507-3511. (31) 2013. (32) Sah, R. Communications in Soil Science & Plant Analysis 1994, 25, 2841-2869. (33) Nelson, J.; Kurtz, L. T.; Bray, R. Analytical Chemistry 1954, 26, 1081-1082. (34) Jayawardane, B. M.; Wei, S.; McKelvie, I. D.; Kolev, S. D. Analytical chemistry 2014, 86, 7274-7279.

Figure legend Figure 1. a) 3D printed chip integrated with membrane, b) Cross section image of the chip, c) SEM image of unwashed Lay-Felt, d) SEM image of washed lay-felt with water.

Figure 2. Schematic of embedding Griess reagent during 3D printing process. a) chip prinitng with ABS, b) membrane prinitng with Lay-felt, c) embedding Griess reagent while pausing the printing process, d) continueing printing to seal the reservoir.

Figure 3. a) Comparison of chips with membrane of washed layfelt (left), unwashed layfelt (middle) and PLA (right) in transporting small inorganic ions, 20 mM FeCl3 and 2.5 M KSCN were used. b) Color intensity vs Lay-Felt membrane thickness, 20 mM FeCl3 and 2.5 M KSCN were used.

Figure 4. a) 3D printed chips with 0.4 mm washed Lay-Felt membrane. Griess reagent was injected into left channel, and different concentrations of NO3- (left to right: 0, 10, 20, 40, 60 ppm) with zinc dust was injected to the right channel. b) Color intensity vs time for different concentrations of KNO3. Figure 5. Color intensity vs time for chips with embedded Griess reagent. a) reagent side and b) sample side with aqueous Griess, 30 ppm KNO2 was injected into the sample side. c) reagent side and d) sample side with ethanolic Griess, 200 ppm KNO2 was injected into the sample channel.

Figure 6. Direct determination of NO3 in soil. Soil slurry with CaSO4 and Zinc dust were mixed and injected into the left channel, and Griess reagent was injected into the right channel.

ACS Paragon Plus Environment

Page 12 of 19

Figure 1

Page 13 of 19

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

Analytical Chemistry

a)

b)

1 mm

c)

d)

0.2 mm

ACS Paragon Plus Environment

0.2 mm

Analytical Chemistry

Figure 2 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

a)

Page 14 of 19

b)

Layfelt ABS Heated nozzle Chip

membrane

Build plate

c)

d) Griess reagent

ACS Paragon Plus Environment

Figure 3

Page 15 of 19

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

Analytical Chemistry

a) T=0 b)

T=3 min

T=3 hrs

ACS Paragon Plus Environment

Analytical Chemistry

Figure 4 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

0 ppm

Page 16 of 19

10 ppm 20 ppm 40 ppm 60 ppm

a) b)

T=0

T=70 min

ACS Paragon Plus Environment

Page 17 of 19

Figure 5

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

Analytical Chemistry

a)

b)

c)

d)

ACS Paragon Plus Environment

Figure 6 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

Analytical Chemistry

T=0 min

Page 18 of 19

T=55 min

ACS Paragon Plus Environment

Page 19 of 19

ToC Entry

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

Analytical Chemistry

ACS Paragon Plus Environment