Inkjet printing of paraffin on paper allows low-cost point-of-care diagnostics for pathogenic fungi

We present a high resolution, ultra-frugal printing of paper microfluidic devices using in-house paraffin formulation on a simple filter paper. The patterns printed using an office inkjet printer formed a selective hydrophobic barrier of 4 ± 1 µm thickness with a hydrophilic channel width of 275 µm. These printed patterns effectively confine common aqueous solutions and solvents, which was verified by solvent compatibility studies. SEM analysis reveals that the solvent confinement is due to pore blockage in the filter paper. The fabricated paper-based device was validated for qualitative assessment of Candida albicans (pathogenic fungi) by using a combination of L-proline β-naphthylamide as the substrate and cinnamaldehyde as an indicator. Our studies reveal that the pathogenic fungi can be detected within 10 min with the limit of detection (LOD) of 0.86 × 106 cfu/mL. Owing to its simplicity, this facile method shows high potential and can be scaled up for developing robust paper-based devices for biomarker detection in resource-limited settings.


Introduction
Most often, the presence of microbial pathogens (including bacteria, fungi, viruses and parasites) inside and around humans has been shown to have detrimental effects on their health, which ranges from mild infections all the way up to lethal diseases (Savary et al. 2019;Chow et al. 2010). As a result, pathogen detection is more imperative than ever and is readily employed in wide-spread areas like food testing (Priyanka et al. 2016), environmental monitoring (Rajapaksha et al. 2019) and clinical diagnosis (Lazcka et al. 2007). Recent times have seen a steady increase in antimicrobial resistance and delayed disease detection which has become a bludgeoning problem among patient care and hospital settings (Prestinaci et al. 2015). Therefore, the development of efficient diagnostic technique is now of paramount importance. Conventional and frugal ways of detecting causative pathogens mainly rely on culture-based methods, where the samples are isolated and grown in a suitable media. Later, the cultured samples are subjected to microscopic examination, biochemical testing and chromogenic media for detection (Wormser and Ryan 2003;Law et al. 2014;Prestinaci et al. 2015;Váradi et al. 2017). Though the said methods are cheap, robust and well-implemented for clinical samples, they are less sensitive and requires frequent media preparation, expert intervention and longer assay time. This, eventually results in delayed detection as well as targeted therapy (Wang and Salazar 2016;Goluch 2017).
As a result, the last decade has seen rapid progress in the field of pathogen detection due to advancements in optical biosensors (Lazcka et al. 2007) and microfluidics (Kou et al. 2016). Microfluidics technology in particular has made significant contributions in chemical assays (Yeo et al. 2011), single-molecule manipulation (Mani et al. 2013), biomedical sensing (Whitesides 2006) and more importantly, point-ofcare diagnostics (Nasseri et al. 2018). In the interest of developing frugal point-of-care diagnostic devices 'Paper' has been shown to hold great promise as a viable material as it is affordable, abundant, biodegradable and highly porous. Additionally, surface of paper possesses hydrophilic groups, which facilitate sample movement via capillary flow without necessitating external pumps (Songok et al. 2014). Owing to its myriads of benefits, paper is considered as the better candidate for diagnostic device development when compared to its counterparts, namely polymers and glass Tang et al. 2020). As a result of its versatility, paper platforms have also been combined with electrochemical, colorimetric and fluorescence detection systems for sensing various analytes, biomarkers and other applications (Ahmed et al. 2016;Chatterjee et al. 2018;Lee et al. 2018;Rosati et al. 2019;Xiao et al. 2019;Mani et al. 2020;Lin et al. 2020;e Silva et al. 2020). A significant step in transforming paper into a paper-based device mainly relies on the patterning of hydrophobic barriers. This can be achieved by using conventional and low-throughput techniques like photolithography (OuYang et al. 2014;He et al. 2013), vapour phase deposition (Haller et al. 2011;Kwong and Gupta 2012), screen printing Sameenoi et al. 2014), flexography printing (Olkkonen et al. 2010), plasma treatment (Li et al. 2008), wax dipping (Songjaroen et al. 2011) and correction pens (Mani et al. 2019). Alternatively, highthroughput wax printing (Lu et al. 2009;Carrilho et al. 2009;Nilghaz et al. 2019) and inkjet printing (Abe et al. 2008;Li et al. 2010;Elsharkawy et al. 2014;Yamada et al. 2015;Su et al. 2016;Matsuda et al. 2017;Punpattanakul et al. 2018) are also used to make paper-based microfluidic devices for chemical sensing and diagnostic purposes (Li et al. 2012;Yang et al. 2017).
The main drawbacks of commercial wax printers are their high cost ([ 150$), poor compatibility to organic solvents ) and unavailability in developing countries (Lin et al. 2020), primarily due to constraints in affordability and service rendering. In addition to that, the exact ink composition of commercially available Xerox Phaser and Colorqube (solid wax printer) is a trade secret (Potter et al. 2019), therefore hindering their reproducibility even more. Hence, there is a growing interest towards the development of alternative facile fabrication methods. In this work, we have developed a high-throughput, robust, ultra-frugal inkjet printing set-up (\ 25$) with an in-house formulation (based on ''paraffin wax'', heptane, colorant) which we have used to achieve selective hydrophobization on a paper surface (Fig. 1). This hydrophobic formulation is inert, opaque and compatible with major aqueous and organic solvents. Additionally, we have achieved a hydrophilic channel resolution of 275 lm. As a novel application, we have leveraged the printed paraffin-based paper devices for the rapid (\ 10 min) and qualitative detection of pathogenic fungi Candida albicans, which is prevalent among immunocompromised patients and neonates.

Materials
Whatman (R) filter paper (Grade 1) of 180 lm thickness and 11 lm pore size was purchased from GE Life Sciences. Paraffin was purchased from Merck, India. n-Heptane was obtained from SRL, India and colorants (red and green) were procured from local candle manufacturer. L-proline b-naphthylamide (PRO) was procured from Sigma Aldrich, India and p-dimethylaminocinnamaldehyde (DCA) was obtained from Loba Chemie, India. Other chemicals (analytical grade) were purchased from Merck, SRL, Loba Chemie and Himedia. Candida albicans (ATCC 24433) strain was obtained from the culture collection at Mycology Laboratory, Department of Microbiology, Kasturba Medical College, Manipal.

Preparation of paraffin formulation
The paraffin formulation was prepared using a mixture of Paraffin, n-Heptane and colorant.
The paraffin was initially prepared in n-Heptane solvent at 6.5% (w/v) solution. 20 lL of the colorant was further added to the prepared solution and vortexed gently. The surface tension of prepared paraffin formulation was found to be 20.18 mN/m.

Inkjet printing of paper-based devices
A HP Deskjet printer (1112) was used to print the hydrophobic formulation by replacing the original ink in the cartridge with the in-house formulation. The ink formulation was filled in the printer cartridge using a syringe. The cartridge was cleaned rigorously using n-Heptane and Isopropyl alcohol before filling the ink formulation and loading it into the printer. The pattern of the hydrophobic barrier for the paper devices were designed using CorelDraw X6 software and printed 15 times on single side of Whatman (R) filter paper (Grade 1). The cartridge was frequently refilled using the syringe after depletion. The printed sheet of paper was placed in a hot air oven set at 100°C for 15 min to allow the paraffin to penetrate into the paper pores. The minimum hydrophobic barrier width achieved was 3 mm. The thickness of the uncoated and coated filter paper was measured using a Mitutoyo Digital Micrometer (293-831). As per the manufacturer's guidelines, the volume of droplet deposited by HP Deskjet 1112 printer cartridge is 22 picolitre. (https:// store.hp.com/in-en/default/hp-803-2-pack-economyblack-ink-cartridges-3yp94aa.html).

Surface characterization
Surface analysis of coated and uncoated paper was performed using a JEOL Scanning Electron Microscope. The water penetration rate (lateral) for plain paper and the hydrophobic barrier device was determined using a measuring scale and a timer.

Paper-based assay for Candida albicans
Initially, an assay for Candida albicans was conducted in a vial by taking L-proline b-naphthylamide (PRO) as a substrate and Cinnamaldehyde as a colorimetric indicator. Here, 50 lL of the PRO substrate solution (2 mg/mL) was taken in a vial followed by the addition of 200 lL sample of Candida albicans (ATCC 24433) in water. This suspension was incubated at 37°C for 15 min. Later, 80 lL of the colorimetric indicator was added and was checked for pink colour formation. This study was extended to the inkjet-printed paper-based devices. Here, 50 lL of the PRO substrate solution (2 mg/mL) was dropcasted into the circular inkjet-printed paper device and allowed to dry for about 15 min. A sample of 200 lL of Candida albicans (different colony forming units) was added to the paper device and incubated at 37°C for 15 min. Later, 80 lL of the colorimetric indicator DCA was added all over the sample zone and observed for pink colour formation for 30 min.
Proteinase K treatment for confirmation of colour change 10 lL of 20 mg/mL Proteinase K was added to the 200 lL suspension of C. albicans and heated at 56°C for 30 min in a dry bath. This Proteinase K treated cell suspension was added to the circular inkjet-printed paper device having the drop casted PRO substrate solution and incubated at 37°C for 15 min. Later, 80 lL of the colorimetric indicator DCA was added all over the sample zone and observed for colour change.

Image acquisition and analysis
Images of the device for barrier compatibility and colorimetric test were captured using Canon Eos 3000D DSLR camera at a fixed distance and ambient lighting. The images for the colorimetric test were captured from 10 min up to 30 min after the addition of the DCA indicator, and the average G channel intensity value was measured using FIJI software. First, the images of the colorimetric test including the controls were split into red, green and blue channels in the FIJI software. Then the entire circular area of the test zone in the green channel image was selected, inverted and the mean intensity of the zone was measured.

Results and discussion
Inkjet printing of paper-based devices Inkjet printed regions contain 'paraffin', a hydrophobic material which has penetrated and blocked the perforated pores of the paper due to heating. This has been confirmed by adding water in the patterned institute logo, where we observed no leakage or disruption (Fig. 2a). The surface tension of the hydrophobic formulation was found to be 20.18 mN/ m, which is concordant with the literature (Yamada et al. 2015). Using our frugal inkjet printing, we have printed two hydrophobic line patterns of 3 mm width and 10 mm length. The non-printed region between the two printed hydrophobic lines serves as a hydrophilic channel. Upon introduction, water wicks up into the finer hydrophilic channels. Using a microscope, we assessed the hydrophilic channel's resolution. Interestingly, the channel's width is found to be 275 lm (Fig. 2b).
Characterization of paper-based devices Whatman (R) filter paper (Grade 1) is highly porous with an average pore size of 11 lm. The inkjet printer precisely deposited small droplets (Pico litres) of the hydrophobic formulation on paper. After the heating step (100°C for 15 min), paraffin molecules penetrate and block the paper pores, which in turn gives the paper the property to confine solvents. The Scanning Electron Microscopy image (Fig. 3a) reveals a stark difference between the patterned and plain surface. This result clearly attributes to the pore blockage on paper by the paraffin molecules. Using a Mitutoyo Digital Micrometer, we have also measured the hydrophobic barrier thickness, which was found to be 4 ± 1 lm. From various reported studies (Schilling et al. 2012;Younas et al. 2019), it is evident that printing count as well as heating could contribute to penetration of the ink. We have optimized the number of printing or printing count (15 times) based on the water confining capability of the printed hydrophobic barriers. To assess the ink penetration in paper substrate, we have calculated grey scale intensity (using FIJI software by inverting the image) of plain as well as printed filter paper (front & rear side). The grey scale intensity values are found to be 117 ± 5, 154 ± 4, 145 ± 3 for Plain paper, Printed paper (Front) and Rear side respectively (Supplementary Figure 1). Interestingly, comparable or similar grey scale intensity values of printed and rear side corroborates to the complete penetration of paraffin-based ink from one side to other side of the paper.
A growing body of literature suggest the importance of measuring hydrohead values for plain porous substrates and coated hydrophobic substrates (Mates et al. 2014;Sen et al. 2018;Chatterjee et al. 2018). We have measured hydrohead values for the plain Whatman filter paper (Grade 1) and printed paper device (hydrophobic coated) by fixing 5 mm diameter circular area of the substrates at the end of a glass tube and measuring the height of water added till the water penetrated the substrates completely. The penetration pressure and the hydrohead values for the uncoated Fig. 2 a Inkjet-printed paper device (hydrophobic barrier of 3 mm) with water confinement depicted by an arrow (Design: Institute logo) b Geometry of printed hydrophobic barriers and microscopic image of hydrophilic channel (wicked-up water) in transmission illumination mode (right) and coated filter paper are 74.91 N/m 2 (7.67 ± 1.15 mm) and 289.87 N/m 2 (29.67 ± 2.08 mm) respectively. Additionally, we have measured the penetration length of water (lateral direction) in the paper channels using a measuring scale and stopwatch (Supplementary Figure. 2). In an interesting study by Malekghasemi et al., the authors have compared the water penetration rate in printed filter paper-based channels and unmodified filter paper with the theoretical Washburn equation (Malekghasemi et al. 2016). In their seminal work, Hong et al., has reported that Washburn's law is not applicable to very low resolution channels formed by wax boundaries since the flow speed is dependent on the channel width, where the surface tension acts at the barrier in the inverse direction to the flow, hence lowering the flow speed (Hong and Kim 2015). We have plotted a graph comparing the water penetration rates of plain and printed filter paper with the Washburn equation (Fig. 3b). The curve for the plain paper or without barrier follows the Washburn law. However, the channels printed with paraffin-based formulation deviated from Washburn's law which is in concordance with Hong et al.
A crucial step in converting a paper into a sensing device relies on the hydrophobic barrier's compatibility to various solvents. Several researchers have examined the barrier's compatibility in paper-based devices (Dornelas et al. 2015;Wang et al. 2014;Mani et al. 2019). With this goal in mind, we have subjected the printed paraffin-based circular patterns (Fig. 4) to different solvents. We have drop-casted 30 lL of solvent in circular device to assess the barrier's compatibility. Ten circular devices were taken for each solvent and further observed for any leakage of the hydrophobic barrier. Interestingly, 9 out of 10 devices exhibited ubiquitous resistance (i.e., no leakage) to all aqueous and certain polar solvents. This experiment confirms that paraffin-based barriers can be deployed for myriads of sensing applications.

Colorimetric detection of Candida albicans
To authenticate the functionality of the printed paperbased device for point-of-care diagnostics, we focused on leveraging the device for the detection of Candida albicans, a major opportunistic fungal pathogen causing invasive candidiasis among immunocompromised hosts (Mayer et al. 2013;Kullberg and Arendrup 2015;Clancy and Nguyen 2018). The traditional way of diagnosing Candida albicans includes culturebased methods and nucleic-acid testing, which are laborious, expensive and demand expert intervention. To circumvent these challenges, we have utilized L-proline b-naphthylamide (PRO) substrate targeting the enzyme L-proline aminopeptidase, secreted by Candida albicans (Perry et al. 1990) as a means of detection as its results can be interpreted without a skilled personnel (Supplementary Figure 3). Firstly, L-proline b-naphthylamide (PRO) substrate was imbibed on the printed paper device. Upon addition of pathogenic Candida albicans, L-proline aminopeptidase hydrolyses the L-proline b-naphthylamide (PRO) substrate, resulting in a visible pink colour within 10 min, after adding the indicator (Fig. 5a), which is a significant and easy way to interpret the result. When compared to existing microbial identification methods like BacT/ALERT, Culture & VITEK system which cost between 10 and 15$ per sample. This approach greatly reduces cost (\ 0.20$), can be used as a presumptive identification test for Candida albicans. This method can also be extended further to other pathogen detection assays as well.
To confirm that this distinct colour change (pink) is due to the secretory enzyme from Candida albicans and not due to artefacts, we have carried out two confirmatory studies. In the first study, we treated the C. albicans cell suspension as well as the supernatant with Proteinase K solution by heating it at 56°C for 30 min. Then the proteinase K treated inoculum suspension and the supernatant were added to the paper device (with PRO substrate) and incubated at 37°C for 15 min. Upon addition of the cinnamaldehyde indicator, no colour change to pink was observed in the paper device (Fig. 5b). In the second study, we have performed the same qualitative colorimetric assay for other fungi like S. cerevisiae and Candida tropicalis. Similarly, we didn't observe amenable colour change (Fig. 5c). Moreover, our results have been concordant with existing literature which confirms that L-proline aminopeptidase is not secreted by these organisms (Rawlings and Salvesen 2013). Both these tests prove that the colour change is due to the presence of a secretory enzyme (L-proline aminopeptidase) from the organism, which when disrupted, doesn't give the colour change to pink.
Additionally, we have measured the G channel intensity values of the control and test sample images. Figure 6 depicts the sudden increase in the G channel intensity value. This change is attributed to the hydrolysis of L-proline b-naphthylamide (PRO) by the enzyme L-proline aminopeptidase secreted by Candida albicans resulting in a visible colour change within 10 min, which can pave the way for frugal diagnostics. To analyse the limit of detection (LOD) for the developed paper-based colorimetric test, we used different Candia albicans suspensions at varying concentrations (0.86-8.35 9 10 6 cfu/mL). Interestingly, cell samples less than 0.86 9 10 6 cfu/mL induced no amenable colour change to pink, and the G channel intensity of the samples was similar to the control values with little higher intensity value (may be due to the addition of colorimetric DCA indicator). Figure 7 shows an increasing trend in the G channel intensity for an increase in the concentration of the cells. Strikingly, samples of 8.35 9 10 6 cfu/mL and above showed significant visible pink colour with higher G intensity values.
Finally, we have assessed shelf-life of the device by drop-casting L-proline b-naphthylamide (PRO) in the paper-based device. A similar pink colour was observed for the device stored at 4°C for 1 month (Data not shown). Moreover, no leakage from the paper-device was observed during this experiment. An outline of cost analysis, based on (i) fabrication of paper-based devices (using ColorQube), (ii) conventional diagnostic method for fungi with our frugal method is provided in the supplementary material. It is clear from the results that the paraffin-based fabrication method can potentially replace cumbersome bulk experiments. Our method boasted fabrication of devices with adequate shelf life for long-term usage, which would be especially useful in the case of resource-constrained settings for point-of-care detection.

Conclusion
To recapitulate, in this work, we have described the development of a cost-effective (\ 0.20 $) paperbased device leveraged for point-of-care detection of pathological fungi. The paper-based device was fabricated by inkjet printing an in-house formulated paraffin ink. The printed designs exhibited hydrophobic property due to the penetration of the paraffin into the paper pores which rendered them capable of confining major aqueous solutions and other solvents. The functionality of the fabricated paper device was validated through a colorimetric test developed for Candida albicans (a common fungal pathogen). Our study provides the framework for converting a simple desktop printer into a high throughput paper-device fabrication unit. In our opinion, the results represent an excellent initial step towards the development of frugal and robust fabrication methods for point-of-care sensing or diagnostics with the scope of extending it to smart-phone based studies.