Fabrication of 3D microstructure array on chip for rapid pathogen detection

Abstract An advanced process of free angle photolithography (FAPL) is used for making 3D supercritical angle fluorescence (SAF) structures and transfer them to the polymeric chip by injection molding for low-cost microfluidic devices with the embedded optical sensing. The FAPL was performed via a motorized stage to control the angle of incidence of light and achieve the desired shape with dimension from 50 μm to 150 μm and slope required for the 3D optical structure. These 3D structures are used for enhancement of fluorescent signal through the unique properties of SAF. The presented SAF structure has a reduced active area (50 μm) that allows enrichment of the fluorescence efficiency and reduces the amount of sample required for detection on the polymeric microfluidic chip. Herein, we are presenting reduced dimension of SAF structures, fabricated by FAPL process and increases the number of SAF per mm2 area. This also reduced the volume of sample required per test. Improvement in the limit of detections (LOD) is observed when using the small dimensions of SAF. Solid phase polymerase chain reaction (SP-PCR) on these SAF structures permits for on-chip pathogen detection. These 3D structures have the potential to be widely used in microfluidic chips as a tool for signal enrichment and low-cost point of care systems for optical detection.

A C C E P T E D M A N U S C R I P T 2 potential to be widely used in microfluidic chips as a tool for signal enrichment and lowcost point of care systems for optical detection.
 The intensity of the signal vs. probe concentration measured from the rear face for different size of SAF structures and defined the LOD (0.05 nM) of the smallest SAF structure. The use of such small size of SAF structure will enable us to use less volume (220 pL) and save expensive primer and molecular probes in an experiment. The presented smaller SAF have great potential to be used for the multiple detections in a small area with high sensitivity and large field of view which can help to use such chip without a sophisticated optical system. This will allow us to make a cost-effective experimental design with a low volume of the sample and optical readout system. In future development, a lower detection limit could be achieved by reducing the background signal and by using higher wavelength fluorophores. The combination of SP-PCR and L-SAF array would be widely used as a high-throughput biosensor to analyze food, clinical and environmental samples.

Abstract
An advanced process of free angle photolithography (FAPL) is used for making 3D supercritical angle fluorescence (SAF) structures and transfer them to the polymeric chip by injection molding for low-cost microfluidic devices with the embedded optical sensing. The FAPL was performed via a motorized stage to control the angle of incidence of light and achieve the A C C E P T E D M A N U S C R I P T 3 desired shape with dimension from 50 µm to 150 µm and slope required for the 3D optical structure. These 3D structures are used for enhancement of fluorescent signal through the unique properties of SAF. The presented SAF structure has a reduced active area (50 m) that allows enrichment of the fluorescence efficiency and reduces the amount of sample required for detection on the polymeric microfluidic chip. Herein, we are presenting reduced dimension of SAF structures, fabricated by FAPL process and increases the number of SAF per mm 2 area.
This also reduced the volume of sample required per test. Improvement in the limit of detections (LOD) is observed when using the small dimensions of SAF. Solid phase polymerase chain reaction (SP-PCR) on these SAF structures permits for on-chip pathogen detection. These 3D structures have the potential to be widely used in microfluidic chips as a tool for signal enrichment and low-cost point of care systems for optical detection.

Mohammad-Ali Shahbazi
A C C E P T E D M A N U S C R I P T

Introduction
In recent years various polymer lab-on-a-chip systems have been produced and commercialized with the final goal of obtaining systems that are disposable and compatible with large-scale production [1]. Cyclic olefin copolymers (COCs) are an extremely attractive for large-scale production [2,3]. COC have high optical transparency, chemical inertness, and high heat resistance [4][5][6][7]. Before making a disposable chip mold, one should first think of making a suitable quality (smoother surface and easy to replicate) mold for fabrication of chip. Methods such as micro milling, 3D laser cutting and laser milling are developed to fabricate molds for a costeffective polymeric chip. Among these technologies, micro milling and laser milling are widely used to make the molds, and injection molding is used to make the positives replica of those molds. Role of the master mold has strongly been highlighted to achieve defect-free polymer chips in injection molding process [8,9]. Past years, studies took place to make a better mold with low surface roughness and defects in the molds to achieve better, smoother and reproducible polymer chip [10]. The micro milling is conventionally used to make a rapid mold for injection molding. However, it cannot achieve lower surface roughness as can be achieved by photolithography. Though, molds made by conventional photolithography process also having some defects or changes of microstructures due to the imperfections in the photolithography process and photoresist [11,12]. These defects in structures cause problems during the molding and demolding of the polymer chips. To reduce the surface defects, which are created during micro milling, and defects during the conventional photolithography process, we demonstrated the use of FAPL process to produce 3D structures with smoother surface and fewer defects. Using FAPL process allows us to precisely transfer microstructures to Ni master mold and then to the polymeric chip.
In our previous work, we presented the fabrication of supercritical angle fluorescence (SAF) microstructures in polymer by means of injection molding with a micro-milled stamp (here referred as shim) using micro-milling method [13]. The advantage of using injection molding is two-fold: the reproducibility of structures for Lab-on-chip applications with dimensions of tens of nanometers, together with a technique that is ideal for production of large number of samples [14]. The advantage of using SAF structures can be explained if we consider a fluorophore molecule available on the front of a generic surface of a microstructure: for a generic structure, most of the emitted light onto the material is refracted outside the structure and so a very little amount of light passes through to give a signal [15,16]. Collection of this unused part of light by a structure that exhibits a supercritical angle of reflection delivers an enhancement in the signal intensity up to 46 folds [13]. Usually this refracted fluorescent light is lost or not been gathered when using a flat surface (microscopic polymer slide) for the signal capturing. Moreover, the SAF structures also offer a great field of view for more efficient signal collection.
In this paper, we introduce the FAPL fabrication process and demonstrate the possibility of achieving lithography made SAF (L-SAF) structures with lower surface A C C E P T E D M A N U S C R I P T 6 roughness and higher optical efficience respect to the our previously described micromilled SAF (M-SAF) [13]. Particularly, we present a FAPL fabrication procedure ( Figure 1A) for L-SAF structures of different sizes (50 µm, 100 µm, 150 µm) and demonstrate their advantages over the SAF structures obtained by injection molding of micro-milled shims. Furthermore, reduction in the dimensions (50 m) of SAF both improves the number of available detection site per area and reduces the required sample volume. We also consider defining the efficiency of the fluorescent signal against the number of molecules. The SP-PCR was performed using DNA probes on the L-SAF array [17,18] ( Figure 1B) for the purpose of on-chip pathogen detection. Our experimental outcomes confirmed that by using such kind of SAF arrays, the limit of detection (LOD) can be improved up to 0.05 nM corresponding to 6.62×10 3 molecules.
The LOD is calculated by data obtained from a high-end fluorescent scanner. The presented work is strategic in order to have effective multiple on-chip detection sites per mm 2 area. Furthermore, fabrication and incorporation of these L-SAF arrays on the chip can easily be done by injection molding for mass production. This also adds-on in the direction of the portable, high-throughput bio-sensing system that can be ideal for onsite diagnosis.

Fabrication of SAF structures
To achieve L-SAF structures by FAPL method, the initial fabrication process was  [19,20]. Since SU-8 presented the issue of removal from shim and long processing times, a new negative tone photoresist THB 151N from JSR Co. was used in the fabrication of Ni shim for injection molding of SAF structures. The main advantage of THB 151N is that the fabrication processes does not require a postexposure bake and that it easy to remove after the Ni electroplating step. This makes THB 151N preferable over SU-8 for integration in production-like processes [21].
At first, the photoresist was spin-coated at 1200 rpm on a Si wafer and soft baked for 3-5 min to obtain a resist layer of about 50 m thickness. To achieve the higher thickness for 100 and 150 m SAF structures the spin coating step was repeated 2 and 3 times. A final resist thickness of 100, 150 ± 5 m was achieved. We chose such thickness since the goal was to obtain structures with an aspect ratio of about 1 which is ideal for injection molding [22]. The complete fabrication process is described schematically in

Fabrication of master mold and transfer of L-SAF array to polymeric chip
As a result, the Ni layer was electroplated to a thickness of around 340 m over the

Spotting on SAF structures
The precise spotting on each SAF structure in an array was performed by a sciFLEXARRAYER S5 (Scienion, Germany). A microtiter plate was used to keep the serial dilution (1 µM to 1 pM) while a glass tip (type 1) was taking the solution and washing tip each time before and after changing the sample for spotting. This process is automated and controlled by the software developed by Scienion. The BioAnalyzer 4F/4S scanner

DNA preparation
Streptococcus pneumonia was from culture collection of National Food Institute, Technical University of Denmark (DTU-Food). S. pneumonia genomic DNA was isolated using DNeasy Blood and Tissue kit (Qiagen, Germany) as instruction from the supplier. The DNA concentration was determined by Nano drop 1000 (Thermo Scientific, USA) and the DNA preparation was stored at minus 20 o C before use.

Primers
To make a model to confirm the performance of SP-PCR on FAPL-generated SAF structures for pathogen detection, we selected a set of primers to amplify pneumolysin gene for Streptococcus pneumonia detection [23]. Forward and Cy3-labelled reverse primers were used for the liquid phase amplification. The surface primer was designed as an internal primer and modified at the 5' end with a poly T (10) and C (10) tail to immobilize on the front of SAF. All PCR primers were synthesized and purchased by DNA technology (Aarhus, Denmark) and the sequences are listed in Table S1 (Supplementary data).

Immobilization of surface primers
The 50 µM of surface primers were prepared in a solution containing 5× SSC and 0.004% Triton X, and spotted on front of SAF structure inside the chamber of a polymeric (COC) by sciFLEXARRAYER. After spotting, the microchip was dried and treated with UV irradiation at the wavelength of 254 nm with power of 3 mW/cm 2 for 10 min (Stratalinker 2400, Stragtagene, CA, USA) to directly immobilize the poly (T) poly (C)tagged DNA oligonucleotide on plastic surface without any surface modification [24].
Before conducting the SP-PCR, the chip was washed with 0.1× SSC for 5 min and rinsed with Milli-Q water and dried in incubator 37°C. The chip was treated with BSA 2.5 mg/mL for 30 min then rinsed with Milli-Q water and dried in an incubator.

Solid phase-PCR
After immobilization of surface primers, the microchip was bonded and fixed with a gene frame (Thermo Fisher Scientific) to create a 25-μL reaction chamber surrounding the solid support primer immobilized SAF array. The PCR master mix was loaded by

Data analysis
The microchip was scanned using a BioAnalyzer 4F/4S scanner with 200-ms shutter time (LaVisionBioTec GmbH, Bielefeld, Germany). Fluorescence intensity was quantified using ImageJ software [25]. A circle was adjusted to the size of the SAF, and the mean value of grey levels of the pixels inside a fluorescent spot was calculated. A square was drawn surrounding the circle, and the mean signal was taken as background 1. To determine the signal to noise ratio (SNR), the mean signal fluorescent spot with the sample without probe target (NC, negative control) was used as background 2. SNR in this study was defined as the mean signal intensity of the feature subtracted from the mean background 2 and divided by the variation of the background 2. The limit of detection was determined at SNR of 3. All the experiments were performed in triplicate.

Structural characterization of 3D structures
The design of the chip is based on the calculation of the distribution of fluorescence intensity emitted into the polymer for fluorophores on air/water/polymer interfaces as described elsewhere [13]. To select a suitable substrate for fabrication of the disposable chips, one should consider the proportion of the light emitted into the chip part as a function of its refractive index for two interfaces: air/substrate and water/substrate. Moreover, the higher the refractive index of the substrate, the higher the portion of the fluorescent light is emitted into it. Theoretically, these SAF structures can collect light at emission angles of up to 79.2° and 63.2° for air/polymer and water/polymer interfaces, respectively (the maximum angle for total internal reflection). So the slope of the SAF structure defines the intensities of the signal generated. The fluorescence intensities collected by using these SAF structures are about 98% (air/Polymer) and 78% (water/Polymer) is calculated in comparison to the percentage acquired using a perfect parabolic lens for signal intensity collection [26]. It has been discussed in our previous work that a lower signal is collected if the front face of the SAF structure is used because the light is emitted in all the directions that make the collection of light dependent on the numerical aperture of the optical device ( Figure 1A). However, use of rear face of the SAF structure allows us to collect a large portion of the light that was lost in case of collection from the front side. So using the same size of the numerical aperture of the optical system, we can collect more signal, which is required for various low-cost optical detection systems. A graphic sketch of FAPL is shown in Figure 2B In Figure 4, the SEM images for all the three sizes are shown. The SAF structures exhibit uniformity and low surface roughness after the lithography as well as after the injection molding. The optical characterization is presented in the Figure 4A

Comparisons between micro-milled SAF (M-SAF) and lithography made (L-SAF) array
The SEM images show the apparent difference in between the surface roughness of SAF structures fabricated by FAPL and micro-milling process. As discussed in our previous report, the surface roughness at the sidewall of the SAF structures will give light scattering effects [13]. Since where the fluorescent light is collected using total internal reflection at the surface, roughness will cause a loss of signal.  Figure 6B). The LOD for L-SAF lies between the 0.05 to 0.5 nM, corresponding to 6.62×10 3 and 6.62×10 4 molecules, respectively. While the LOD for M-SAF is 10 nM (1.32×10 6 molecules). Therefore, it was clearly observed that L-SAF provided better performance over the M-SAF and fabricated smaller size of L-SAF should be tested and compared with the other size of SAF structures in their signal performance. concentration ranges from 0.05 nM to 1 µM then signal starts to get saturated due to a high exposure time. It is also important to mention here that the small volume of the sample is sufficient to cover the all front surface of the smaller size of the L-SAF surface to give the better signal. However, if the sample volume is increased up to the volume when it is sufficient to completely cover the front face surface of SAF with bigger size (100 and 150 µm), the LOD for these size can be further improved with higher volume.

SP-PCR on fabricated SAF structures
As mentioned in our previous reports [18,27,28], the SP-PCR has become increasingly popular for molecular diagnosis integrated with lab-on-a-chip devices. In this study, we addressed a number of advantages of the small size of the SAF structure (50 um) with fabricated by free angle lithography process compared to the bigger size as well as SAF structure from the conventional process. To test the applicability of these 3D SAF structure for further biomedical application, we carried out the SP-PCR on this SAF structure embedded in a microchip. The surface primers were spotted on the front of SAF array as indicated in Figure 8. The DNA target gene (pneumolysin gene) from Streptococcus pneumonia was specifically amplified using SP-PCR on the front of the L-SAFs on chip. Where were modified with surface primers for DNA target gene. As showed in the figure 8b, it shows the high fluorescent signal and there are no any fluorescent signals observed containing non-specific primers. Since there was not any specific primer available so no amplification took place during SP-PCR reaction as presented in Figure 8b.

Conclusions
In this paper, we have presented a novel free angle lithography method to make a disposable chip with a miniaturized (50 µm) L-SAF array, with the successful SP-PCR for selective and sensitive detection of pathogen. The intensity of the signal vs. probe concentration measured from the rear face for different size of SAF structures and defined the LOD (0.05 nM) of the smallest SAF structure. The use of such small size of SAF structure will enable us to use less volume (220 pL) and save expensive primer and molecular probes in an experiment. The fabrication of SAF structure is important to achieve the small size and higher density of the SAF on a chip. Despite the complication of the FAPL process for fabrication of SAF structures it still embraces advantages of lower surface roughness and background signal noise ( Figure 6A). The presented smaller SAF have great potential to be used for the multiple detections in a small area with high sensitivity and large field of view which can help to use such chip without a sophisticated optical system. This will allow us to make a cost-effective experimental design with a low volume of the sample and optical readout system. In future development, a lower detection limit could be achieved by reducing the background signal and by using higher wavelength fluorophores. The combination of SP-PCR and L-SAF array would be widely used as a high-throughput biosensor to analyze food, clinical and environmental samples.