Paper-based microfluidic sensing devices fabricated by inkjet printing technology

Paper-based microfluidic sensing devices fabricated by inkjet printing technology Bernat Guirao Expósito Master’s Thesis Advisor: Daniel Citterio Analytical Chemistry laboratory Department of Applied Chemistry Keio University Chemical Engineering Master’s Thesis Validated by ETSEIB (UPC) 1 Index 1. Abstract 2 2. Introduction 2 2.1. Paper patterning technologies 2 2.1.1. Photolithography 3 2.1.2. Wax printing 3 2.1.3. Inkjet printing technologies 3 2.1.4. Other paper patterning technologies 4 2.1.5. Paper patterning technology election 4 2.2. Substrate: paper 4 2.3. Quantitative detection of analytes 5 2.4. Previous research 5 3. Experimental section 7 3.1. Preparation of the patterns 7 3.2. Preparation of the hydrophobic ink 8 3.3. Preparation of the pH sensor 8 3.4. Preparation of the pH buffers 9 3.5. Geometry of the patterns 9 3.6. Printers 9 4. Results and discussion 10 4.1. Simple test patterns 10 4.2. Two path patterns 11 4.2.1. Patterns C_s3w03ch15 and C_s3cir3w03ch15 12 4.2.2. shortpath+ pattern series 12 4.3. Four path patterns 16 4.3.1. Patterns 16 to 21 16 4.3.2. Pattern C_s3s4w1ch24_cr_ch04_18_ shortpath4 17 4.3.3. Pattern C_s45s4w1ch24_cr_ch04_18_shortpath4 18 4.3.4. Pattern C_s45s4w1ch24_cr_ch04_18_shortpath4_sext9 22 4.4. Simple symmetry patterns 23 4.4.1. Pattern Simple symmetry 1 24 4.4.2. Pattern Simple symmetry 2 and Simple symmetry 3 24 4.4.3. Pattern Simple symmetry 4_square 27 4.4.4. Pattern Simple symmetry 4 and Simple symmetry 5 28 4.4.5. Pattern Simple symmetry 6 30 4.4.5.1. Test using EPSON and Dimatix 30 4.4.5.2. Test using only EPSON 32 5. Conclusions 33 6. Future research 33 7. Acknowledgement 34 8. Bibliography 34 2 1. Abstract The detection of some analytes, for example heavy metal ions, is crucial in locations such as developing countries or humanitarian emergencies, where no qualified professionals or sophisticated equipment is available. In this project, a detection system using inkjet printing technology is developed. The system relies on printing of hydrophobic UV-curable inks to define microfluidic channels in paper, and to print a sensor inside this channel. After that, a solution sample can be dropped into the channel, resulting in a colour change of the sensor. By digitally measuring the colour change, the analyte quantity can be determined. Actually, the main focus of the project is the testing of different channel geometries in order to improve their performance. The aim of this project is not only to develop a detection system, since nowadays a lot of them are already available. The system proposed here requires neither qualified professionals nor very sophisticated equipment for operation. Actually, the goal is to detect with as simple equipment as possible. Only a computer, a commercially available printer and a colour scanner would be sufficient, making it possible to use them in areas such as developing countries or in humanitarian emergencies. Besides, this detection system has been designed to be as environmentally friendly as possible. 2. Introduction The system studied in this project, which is explained in 3. Experimental section in more detail, consists basically in printing a hydrophobic microfluidic channel in a piece of paper, creating a barrier that does not let hydrophilic liquids to escape from the channel created. Then, the sensor reagent is printed inside the channel, immobilizing the sensor reagent and requiring small quantities of it. Once the system has been created, analyte is dropped into the channel. When analyte sample reaches the sensor reagent, colour change is detected. It can be measured with a scanner and a simple colour detection program. From the colour obtained, analyte parameters such as concentration can be deducted. The system proposed is simple and low price. It only requires using a printer to print the reagent sensor and an availably commercial scanner to scan the paper and detect the colour change. Actually, a person with only basic computer skills is able to use this system. In developing countries there is a lack of qualified professionals, so it is crucial to design an easy to use system. Otherwise, industrial fabrication of the system is cheap, since inkjet printing technology and UV light are available technologies in the industry. Moreover, the system is made of paper, which is an available and cheap material, and the other reagents, basically the ink and sensor ones, are expensive, but only small quantities are used. This system presents an incredible amount of applications. Actually, there are as much applications as sensor reagents that are able to be printed into the channels. Besides, the system is cheap and easy to use, so no qualified professionals or large amounts of money are required to use it, which increases the applications. Some examples of applications are on site detection or analyte properties quantification in humanitarian emergencies or in difficult access zones (applications in exploration and military fields) as well as household applications. It must be taken into account that more research must be conducted in this project field. So, in the nearly future it may substitute some current applications of well established and commonly used technologies such as photolithography. 2.1. Paper patterning technologies Printing technologies have experienced a great development during the last decades, and they are no more exclusive to the publishing sector, since applications in fields like chemistry and biology, for example, have been found. 19 3 Actually, paper-based sensor fabrication is one of the research fields where printing technologies are being applied. The idea of making sensors using paper-based technologies is very attractive, since they are easy to use and may provide valuable information. However, a lot of research is needed to make them cheaper and more precise, since they often require so expensive technologies and the information provided is almost qualitative and not as exact as desired. Most of paper-based sensors do not require external equipment or reagents. So, this kind of sensing devices are adequately inexpensive to be commonly used, including in developing countries, in emergency first aid situations, and in home health-care settings, in contrast to conventional medical diagnostic technologies that were designed for air conditioned laboratories, highly trained personnel, a constant supply of reagents, and considerable volumes of sample. Besides, low sample and sensor volumes are required, which is important in the case of available clinic samples or high cost sensing reagents. Among the paper patterning technologies available, inkjet printing, photolithography or wax printing, among others, may be useful for sensor fabrication. In this project inkjet printing technology was selected as the most appropriate one. However, important advances are taking place on the other technologies, too. 2.1.1. Photolithography In the photolithography field, a lot of research has been conducted, trying to apply it to paperbased sensors. 1 Photolithography consists in using a light to transfer a geometric pattern from a photo mask to a light-sensitive chemical. One of the main problems it presents are its high cost, since sophisticated equipment and qualified professionals are needed, making it more expensive to implant in mass scale production than other printing methods, and being almost impossible to be applied in developing countries. Besides, the substrate used must be chemically treated, including the hydrophilic part, implying a substrate contamination risk. However, important improvements and research are being conducted to solve these problems, trying to make it a more feasible method. 2 2.1.2. Wax printing Another printing technology used regularly is wax printing. Actually, wax printing is a promising technology in the fabrication of microfluidic paper-based analytical devices, since printing velocity is extremely high and it is feasible to use it for mass scale production. For example, wax printing takes less than 3 min to prepare four multizone plates, while photolithographic methods require about 20 min to prepare a single plate. 3 Moreover, wax printing is a flexible technology, so it is possible to change the fabrication process easily, like the design of the patterns. The system consists basically in printing wax on the paper and heat it in a hot plate, penetrating the molten wax into the paper, creating a hydrophobic barrier. It must be taken into account molten wax penetration into substrate makes the channels wider than in the original design, although some models exist to predict channel widening. 4 Otherwise, the same phenomenon occurs in inkjet printing technology. In terms of producing industrially paper based microfluidics, wax printing could be a feasible technology, due to its velocity, as it was told before, and simplicity and nature of the process itself. However, another technology should be used to print the sensor reagent, like inkjet printing technology. And creating a hydrophobic barrier could be complicated due to wax solvent compatibility problems. 2.1.3. Inkjet printing technologies Inkjet printing has become a useful and feasible technology in a lot of industrial fields for accurately depositing very small quantities (tens of picoliters) of materials at defined spots on the surface of a wide variety of substrates. Moreover, it is a single step process which does not require highly qualified professionals or an important amount of facilities, in contrast to photolithography. And it is a suitable technology for scale industry mass production, too. So, its 4 time saving, low cost and flexibility will probably establish inkjet printing technology as one of the most suitable printing technologies in the future. Furthermore, modern inkjet printing devices have multiple print heads which are able to print different kinds of ink simultaneously, being able to print both the hydrophobic barrier of the pattern and the sensor reagent, and, in contrast with other printing technologies like plotting, the place, time and quantity of liquid deposition can be controlled precisely (“on demand” inkjet printing technology). In spite of all the advantages mentioned, inkjet printing technology has been mostly applied in plastic electronics and for the manufacture of polymer light-emitting diodes 5-8 , and in the analytical chemistry field only in the fabrication of electrochemical sensing devices. 9-10 Inkjet printing technology is also a feasible technology for the fabrication of microfluidic sensing devices incorporating various chemical functions. 11-12 2.1.4. Other paper patterning technologies Although other technologies for patterning paper exist, they are not as important as the previous ones. Some examples may be cutting and plotting. In the case of cutting, the knife cost is representative and sample and sensor are not isolated in a zone of the paper, getting wet all the paper when making analysis, and consequently being able to contaminate people and objects around. 13 Regarding plotting technology, it is not readily adoptable to the dispensing of the reagents required for chemical sensing. Moreover, plotting technique is a continuous line drawing method, with the amount of dispensed liquid being more difficult to control and, therefore, limiting the resolution of plotted structures to 1 mm, approximately. Besides, lately felt tip-based x,y-plotters are being replaced by inkjet printing devices, being gradually more difficult to find in the market 14-15 . 2.1.5. Paper patterning technology election Definitely, in order to achieve properly this project goals, inkjet printing is the most suitable technology, since it is simple, cheap and flexible and it can be used in emergency and first aid situations. Moreover, it is feasible to apply it in mass production scale and multiple print heads can be used to print both the hydrophobic barrier of the pattern and the sensor reagent. 2.2. Substrate: paper Different substrates can be used in microfluidic sensing devices, such as glass or paper. In the case of this kind of sensing devices, the substrate must let the analyte to flow through it, so it would have a porous structure or some external equipment, such as external pumps, would be used. Paper has a porous structure, since it is normally made of cellulose fibers. But paper also has more advantages, since it is abundant, inexpensive (6$/m 2 even for high-quality chromatography paper), sustainable, easily and environmentally friendly disposable, easy to use, store, and transport and easy to modify chemically. Paper is also a familiar material and everybody is used to it. This is an important factor, since one of the final goals of this research branch is making sensing devices to be used in developing countries or emergency situations, so no qualified professionals must be able to use them easily. But perhaps its most important advantage is that an enormous variety of inkjet printing techniques, concretely inkjet printing technology, is available for its functionalization. Moreover, paper has been used in the analytical chemical field for so many years, and a lot of experience working with this material has been acquired, so a lot of advantage can be taken from existing analytical techniques. Paper physical properties make it very useful for chemical analysis, too. It is thin, lightweight (10 mg/cm 2 ) and available in a wide range of thicknesses (0.07-1 mm). Moreover, it is easy to stack, store, and transport. Its composition and structure makes it feasible to work with biological samples, since it is made of cellulose or cellulose-polymer blends, and it can be modified chemically to incorporate a wide variety of functional groups that can be covalently bound to proteins, DNA, or small molecules. Another advantage is the fibrous nature of paper, which 5 provides to the system intrinsic capillary pumping 18 , whence no external pumps or similar sophisticated equipment is not required. Moreover, nowadays printing techniques are able to make sophisticated microdevices which incorporate more functionality, for example metal electrodes, without the need of clean rooms or other kinds of expensive infrastructure. Paper colour is white (it scatters light), so it provides contrast with a coloured substrate, being an available substrate for colorimetric tests, and it is available in a wide range of highly engineered forms with a very wide range of properties. For example, filter papers with welldefined pore sizes can separate suspended solids from samples before an assay or remove erythrocytes from blood. From an industrial point of view, paper can be easily manufactured on a large scale by the wellestablished coating and printing techniques 17 , which can further lower the cost of the final products. Moreover, little initial investment would be required. 2.3. Quantitative detection of analytes Quantitative colorimetric detection of analytes is possible by reflectance detection when the intensity of the colour that develops in the test zones is a function of the concentration, or another parameter, of the analyte. In other words, the colour change produced in the sensor due to the contact with the analyte can be detected, and with this information and a calibration curve, the analyte desired parameter can be calculated. As it was told before, the system designed must be easy to use, so that no qualified people can use it. To see the colour change with the naked eye may be impossible in some cases, so simple systems like scanners and colour registering software may be used. 2.4. Previous research Whitesides and coworkers introduced the concept of using patterned paper substrate as a microfluidic platform for multiplex analyte detection. In their studies, hydrophobic polymers were photolithographically patterned 19 or printed using a desktop plotter 20 onto hydrophilic paper so that the millimeter-sized fluidic channels can be well-defined in the paper matrix. Owing to the natural capillary action of paper, fluids can flow along these channels without the external pumps that are required in the conventional microfluidic devices. The use of these patterned paper platforms minimizes the use of sample volume and makes multiple analysis possible. Whitesides and coworkers were able to detect glucose and protein simultaneously on a single patterned paper. In a more recent work 21 , an imaging device (camera phone or portable scanner) was added to the system in order to make quantitative analysis possible. Actually, inkjet printing can be used to deposit a wide variety of materials, such as inorganic particles or biological species, on paper. For example, Pelton et al. coated biomoleculeattached microgel inks on paper 22 and Abe et al. printed antibodies to analyze biological parameters 11 . The concept of the system used in this project is similar to the system used by Abe et al. 12 . That system consisted in soaking filter paper in a 1.0% wt solution of polystyrene in toluene for 2h and then drying at room temperature for 15 minutes. This turned the paper, which is hydrophilic, in hydrophobic. Then, a hydrophilic pattern was etched by the ejection of toluene droplets, which dissolved the hydrophobic polymer material. The dissolved excess polymer was accumulated at the outside of the liquid droplets according to the coffee ring effect upon toluene evaporation and taken up by the surrounding paper. Doing this, the paper became hydrophilic again, almost exactly as at the beginning. Once the hydrophilic channel had been created, sensor reagents could be printed using inkjet printing technology, too. Using this method, multiple analyte sensors can be done. However, this system has been criticized due to the risk that some hydrophobic polymer remained in the hydrophilic pattern, since at the beginning all the paper was soaked. 6 In figure 1, a schematic representation of the system used by Abe et al. to make multianalyte sensors is shown. Figure 1. Schematic representation of the system used by Abe et al. to make multianalyte sensors. Source: Abe et al., 2008 12 The main problem of this system was the coffee ring effect that took place when the analyte reached the sensor. The analyte pushed the sensing reagents to the sensing area limits, resulting in a non homogeneous colour, which made difficult to take out valuable information of the experiments. In figure 2, an outline describing the coffee ring effect consequences can be observed. Figure 2. Outline showing the coffee ring effect consequences In figure 3, the sensor reagent can be observed before and after application of the analyte. The effects of coffee ring effect can be observed, too. Figure 3. Sensor reagent in different sensing areas, before (patterns numbered with “1”) and after (patterns numbered with “2”) analyte application. Source: Abe et al., 2008 12 7 3. Experimental section 3.1. Preparation of the patterns The system proposed to make patterns proposed by Abe et al. 12 used toluene, which is an environment and health affecting volatile organic solvent (VOC) that requires some security measures when it is used. Moreover, it had the problem that the analyte flowed for a path where previously a polymer had been printed. Despite polystyrene was removed using toluene, which vaporizes easily, some contamination may occur, leading to incorrect results. So, a new method of creating patterns was used, which consisted basically in printing the hydrophobic barrier only in the parts required to be hydrophobic. Doing this, the hydrophilic part where the analyte had to flow remained unchanged, no risk of contamination existed, and toluene was not used. The reagents used to create the hydrophobic barrier were changed, too. One of the reagents used previously was toluene, but low vaporization temperature solvents are not desired reagents in the industry field, since they are considered dangerous. So, instead of the toluene and polystyrene solution, a solution of Octadecyl Acrylate, 1,10-Bis(acryloyloxy)decane and 2,2Dimethoxy-2-phenylacetophenone, which was used as initiator, were used. Once the solution was printed, it was polymerized using ultraviolet light (Hamamatsu Lightning Cure 6686). In all the experiments 600 mW/cm 2 were used. It must be taken into account that once this solution was printed, it spread very fast, so UV light must be applied immediately after printing. If some time was awaited between printing and polymerization, spread of the ink would be produced, and when it was polymerized, the pattern would get a different form from the original, causing symmetry and flow problems. It must be taken into account that the hydrophobic barrier printing procedure consisted in two steps, since the front part and the back part of the substrate were polymerized independently. The back part of the pattern was totally covered by the hydrophobic barrier, to prevent the sensor reagent and the analyte from leaking outside the pattern. Once the hydrophobic pattern had been printed, the sensor reagent could be printed in the desired zone of the pattern. Once both the hydrophobic barrier and the sensor reagent were printed, analyte was dropped into the pattern and some time later colour change was analysed using a colour detection software (Digital Color Meter software Ver. 3.6.1 provided with the Mac OS X Ver. 10.5 operating system). The substrate used to create the patterns was filter paper (Advantec 5C 185mm). The substrate used must be absorbent and resistant to UV light. 8 In Figure 4, a schematic representation of the preparation of the patterns is shown. Figure 4. Representation of the preparation of the patterns 3.2. Preparation of the hydrophobic ink As it was told before, the solution used to create the hydrophobic barrier consisted of Octadecyl Acrylate (ODA, provided by WAKO), 1,10-Bis(acryloyloxy)decane (DDA, provided by Tokyo Chemical Industries) and 2,2-Dimethoxy-2-phenylacetophenone (provided by WAKO), which was used as initiator. To prepare it, a solution of ODA and DDA (ODA-DDA, 70-30 %wt) was prepared, and later initiator was added to the solution (ODA:DDA-initiator, 85-15 %wt). Then, some stirring time was applied. This solution showed good performances at short time term, but after some days the pattern performances clearly worsened. In order to improve the pattern durability, the stabilizer of ODA and DDA can be removed. To remove the ODA stabilizer, it was mixed in an ethanol and 5% wt NaOH solution, and it was submitted to a separation process. After the separation process, distiller and a vacuum bomb were used to evaporate the remaining ethanol. In the case of DDA, the process was the same, but no ethanol was used. 3.3. Preparation of the pH sensor Two pH sensors were used in this project, one made of particles and another one without them. In the first experiments made with pH sensors, the particle based sensing reagent was used, since the capacity of particles to get stuck to the paper made it more resistant to the coffee ring effect. However, it has been demonstrated yet that coffee ring effect can be avoided using pH based particle sensing reagent even with the simplest patterns (the ones with only an analyte entrance, such as the “Simple test” series patterns). Preparation method of this reagent is explained by Abe et al. 11 Non particle pH sensing reagent, which is more easily affected by the coffee ring effect than the previous one, was used to demonstrate that multiple path channels can avoid it. Preparation method of this sensing reagent is explained by Abe et al. 12 Back pattern printing UV light polymerization 15 Figure 11. Outline of ink spread and its resistance depending on the path width It also seems logic that the ink resistance is the main producer of the analyte flow asymmetry, so in wider paths, less asymmetry existed. Although it is true that paper does not possess an homogeneous structure and it probably causes flow asymmetry, it is hydrophilic and the analyte can flow through it, while monomer ink is hydrophobic and the parts of the path invaded by it cannot be crossed by the analyte anymore, producing an important effect in terms of flow asymmetry. Looking at table 4, this fact is checked, since the wider the path is, the less delay between flows occurred. Regarding the polymerization time, despite time differences could be noticed on the same pattern depending on the polymerization time applied, no relation between the polymerization time and effects in flow velocity and symmetry could be established. It must also be taken into account that experiments reproducibility was very low and standard deviation values confirm it. This poor reproducibility makes difficult to establish empirical relations. Obviously, an important factor that affects seriously the reproducibility of the experiments is the printer used. EPSON is a low cost commercially available printer, and the hydrophobic ink used probably produces partial clogging of the nozzles, leading to differences in the dispensed amounts of ink. No experiments involving sensor reagents were done, since the performances of the patterns designed was not satisfactory enough. To check the patterns performance durability made with monomer ink with stabilizer, that needs more polymerization time but has a lower average life than without stabilizer, some patterns were tested again an hour later. In table 5, patterns flow timings observed after polymerization and an hour later can be observed. Pattern Name-date Polymerization time Leaki ng t 1 (s) t 2 (s) t delay (s) t sensor area (s) 1-2/5/2011 1 min/pp No 65 87 22 160 1'-2/5/2011 1 min/pp No 20 38 18 90 C_s3w03ch15_ shortpath+ (after polymerization) 3'-2/5/2011 2 min/pp No 83 83 0 220 1-2/5/2011 1 min/pp No 1'-2/5/2011 1 min/pp No 28 38 10 70 C_s3w03ch15_ shortpath+ (an hour later) 3'-2/5/2011 2 min/pp No 40 55 15 165 2-2/5/2011 1 min/pp No 80 310 230 330 C_s3cir3w03ch15_ shortpath+ (after polymerization) 4-2/5/2011 2 min/pp No 150 210 60 270 2-2/5/2011 1 min/pp No 80 240 C_s3cir3w03ch15_ shortpath+ (an hour later) 4-2/5/2011 2 min/pp No 79 170 91 330 Table 5. Effect of time on patterns performance made of monomer ink with stabilizer comparing their flow behaviours just after polymerizing and an hour later 16 Comparing the five patterns tested it can be observed that in all cases the results changed significantly, partly due to bad reproducibility, and in most of them, patterns performance worsened, becoming the paths blocked in some cases. So, with these experiments it can be concluded that patterns durability using monomer ink with stabilizer is low. This was taken into account in following experiments when the patterns testing was not produced just after polymerization, using in these cases monomer ink without stabilizer. 4.3. Four path patterns In this section, results obtained using patterns from number 16 to 25 are analysed. Amost all these patterns geometry consisted of two areas connected by two symmetrical paths that bifurcated into two more paths, resulting in four symmetrical paths that entered the sensor reagent area. The idea of designing four paths is to surround the sensor reagent area more effectively than with the two paths patterns, reaching a better flow cancellation and consequently, reducing the coffee ring effect. Obviously, these patterns were designed taking into account the conclusions of the previous section. 4.3.1. Patterns 16 to 21 The first patterns tested was pattern number 16 and 17 (C_s3cir3w03ch15_shortpath4 and C_s3w03ch15_shortpath4), which were printed using EPSON printer and whose geometry can be observed in figure 12. This pattern consisted of a 3 mm side dropping square and 3 mm diameter sensor reagent area. Channel width was 1.5 mm, and the bifurcated paths width was 0.75 mm. Figure 12. C_s3cir3w03ch15_shortpath4 pattern geometry Deionized water could not flow through the bifurcated paths, since their width (0.75 mm) was too narrow to let water pass, and water became blocked or it flowed extremely slowly. Next patterns designed were patterns 18, 19 and 20. The basic difference with the previous ones was a wider path (18 or 30 mm) for both the main paths and the bifurcated ones. The paths geometry was changed to a curved shape, expecting to get better flowing performances than with angular shapes. However, serious leaking and blocking problems appeared, being almost impossible for the flow to reach the sensor reagent area, independently of the deionized water volume applied (10 to 60 µl) or the polymerization time (1 to 4 minutes for every part of the paper). The cause of these problems was the polymerization method used. The patterns of this section were so wide that the UV light could not surround all the pattern properly, causing an insufficient and inhomogeneous polymerization. So, to polymerize all the next experiments of this section, the UV light was focused alternatively on different parts of the pattern in a symmetric way, in order to assure that all parts of the pattern got polymerized in the same way. An outline of the different ways to polymerize the patterns can be observed in figure 13. 17 Figure 13. Outline of the different methods to polymerize wide patterns 4.3.2. Pattern C_s3s4w1ch24_cr_ch04_18_ shortpath4 To check this polymerization method, pattern number 21 (C_s3s4w1ch24_cr_ch04_18_ shortpath4) was printed and tested with 20 µl of both deionized water and red dye. Polymerization time for every part of the pattern was 3 min, although every part of the pattern was not polymerized for 3 minutes, since as it was told before the UV light focus position was changed alternatively. EPSON printer was used. In order to have a better and faster polymerization, from this time on in the experiments of this section, monomer ink used had been purified, removing the stabilizer according to the method explained in 3.2.Preparation of the hydrophobic ink. In table 6, C_s3s4w1ch24_cr_ch04_18_ shortpath4 experiments results can be observed. In figure 14, C_s3s4w1ch24_cr_ch04_18_ shortpath4 pattern geometry and the different pattern zones nomenclature are shown. Name-date Test substance Leaking tA (s) tA1 (s) tA2 (s) tB (s) tB1 (s) tB2 (s) tC (s) 25-3/6/2011 Dye No 10 25 25 19 25'-6/6/2011 Dye No 2 13 14 42 25''-6/6/2011 Dye No 16 34 53 97 26-3/6/2011 Deionized water No 164 390 390 86 312 312 420 26'-6/6/2011 Deionized water No 20 48 48 96 99 26''-6/6/2011 Deionized water No 45 153 150 83 390 Table 6. C_s3s4w1ch24_cr_ch04_18_ shortpath4 patterns test results Figure 14. C_s3s4w1ch24_cr_ch04_18_ shortpath4 pattern geometry and its zones nomenclature 18 Comparing C_s3s4w1ch24_cr_ch04_18_ shortpath4 and the previous four path patterns polymerized with the inhomogeneous method, it can be observed that pattern performances were clearly better. To begin with, no leaking problems were detected, and blocking and inefficient flow transport problems were reduced considerably. However, in two of the paths analysed, the analyte flowed so slow that did not reach the bifurcation, and in four cases it did not arrive to the sensor reagent area, so this pattern is still very inefficient in terms of analyte flowing. It is remarkable that in patterns which dye was used as analyte, the flow was considerably faster than the ones which used deionized water. This result is interesting, since it means that some substance properties, such as surface tension, may affect the flowing properties of patterns. So, it would be interesting to do some future research in this field. In figure 15, pattern 25-3/6/2011 can be observed. Figure 15. C_s3s4w1ch24_cr_ch04_18_ shortpath4 pattern tested with red dye. Not real dimensions 4.3.3. Pattern C_s45s4w1ch24_cr_ch04_18_shortpath4 This pattern design is similar to the previous one, but the dropping square is bigger (4.5 mm), allowing to drop more than 20 µl of analyte. In the next series of experiments, particle based pH sensor reagent, whose synthesizing method is explained at Abe et al. 13 , was printed into the center of the sensor reagent area. EPSON printer was used to print the patterns and analyte used was a pH 12 NaOH solution. Polymerization time was 3 min. The sensor reagent was printed with Dimatix. Cartridge setting used was pH_indicator_paticle_cartridge. Sensor reagent geometry was 3x3 mm 2 square, and different printing layers were tested. 19 In table 7, C_s45s4w1ch24_cr_ch04_18_ shortpath4 experiments results can be observed. Name-date Volume of analyte (µl) Printing layers tA (s) tA1 (s) tA2 (s) tB (s) tB1 (s) tB2 (s) tC (s) Test Observations 2913/6 /201 1 30 1 15 21 21 10 15 15 26 No colour remained. Weak colour 3013/6 /201 1 20 1 15 22 22 35 Reagent was spread to the right side and bifurcated paths. Weak colour 3113/6 /201 1 20 1 49 110 79 48 114 96 Sensor shape was modified. Weak colour 3214/6 /201 1 20 10 42 17 29 29 Reagent was spread to the left. Inhomogeneous colour 3314/6 /201 1 20 15 24 48 48 40 86 100 Colour change only in some part of the sensor square 3414/6 /201 1 30 15 12 19 19 Reagent was spread to the right. Inhomogeneous colour 3514/6 /201 1 30 15 15 21 21 23 32 32 Inhomogeneous colour Table 7. C_s45s4w1ch24_cr_ch04_18_ shortpath4 patterns test results Figure 16. C_s45s4w1ch24_cr_ch04_18_ shortpath4 pattern geometry and its zones nomenclature Analysing the results, it can be observed that only one printing layer is such a little amount of sensor reagent, and consequently the colour change obtained was so weak that may not be detected by colour detection software or the quantity used was so little that it was washed by the analyte and disappeared. Otherwise, and independently of the printing layers printed, inhomogeneous colour and sensor reagent spreading occurred, probably due to flow asymmetry. So, these experiments proved that flow asymmetry must be eradicated in order to avoid coffee ring effect and obtain a homogeneous colour change. 20 Similarly to experiments performed in table 6, pH 12 NaOH solution flowed faster than deionized water solution. Causes of this phenomena may be the same as for dye solution. Moreover, in this case basicity may be an important parameter, too. No leaking was observed in any of the patterns. In order to improve patterns performances and avoid the flow asymmetry, more C_s45s4w1ch24_cr_ch04_18_ shortpath4 were printed using EPSON printer, and different polymerization procedures were applied. In previous experiments, it was noticed that patterns deformed a bit when they became wet, and this deformation may enhance flow asymmetry. In order to solve this problem, the zone around the patterns was sticked with adhesive tape to the surface where the experiments were being conducted. However, patterns still bent a bit. To print the patterns, all of them were printed one by one and polymerized, and then tests of the patterns were conducted. So, it must be taken into account that there were time differences between polymerization and tests of the patterns, although they were not significative (no more than 30 minutes). Deionized water was used to perform the tests. 4 polymerization methods were tested: 1) Each side of the paper was polymerized for 3 minutes with 600 mW UV light. On each side of the paper, polymerization was conducted in two zones, changing from one zone to the other every 5s during the first 20s, and changing every 10s from then. 2) Each side of the paper was polymerized for 3 minutes with 600 mW UV light. On each side of the paper, polymerization was conducted in two zones, changing from one zone to the other every 5s. 3) Each side of the paper was polymerized for 6 minutes with 300 mW UV light. 4) Each side of the paper was polymerized for 3 minutes with 300 mW UV light. 21 In table 8, experiments performed with each method can be observed. Experiments in orange were made and tested with a day difference. C_s45s4w1ch24_cr_ch04_18_ shortpath4 pattern geometry and its zones nomenclature can be observed at figure 16. Polymerization method Name-date Volume of analyte (µl) tA (s) tA1 (s) tA2 (s) tB (s) tB1 (s) tB2 (s) tC (s) PT116/6/2011 30 308 1080 1080 310 848 780 1200 PT1'- 16//2011 30 440 680 380 820 900 PT1d121/6/2011 30; after 25' 30µl more 400 1 PT1'd121/6/2011 30; after 25' 30µl more 1950 1860 PT1''- 16/6/2011 30 33 42 42 69 PT1'''- 16/6/2011 30 270 270 PT1''d121/6/2011 30; after 25' 30µl more 590 1680 1320 810 1740 1740 1890 2 PT1'''d121/6/2011 30; after 25' 30µl more 1680 PT216/6/2011 30 PT2'- 16//2011 30 160 400 510 230 PT2d121/6/2011 30; after 25' 30µl more 3 PT2'd121/6/2011 30; after 25' 30µl more PT316/6/2011 30 150 360 360 150 360 360 460 PT3'- 16//2011 30 80 120 120 210 PT3d121/6/2011 30; after 25' 30µl more 420 1860 960 4 PT3'd121/6/2011 30; after 25' 30µl more 420 1500 1080 1020 2040 1980 2220 Table 8. C_s45s4w1ch24_cr_ch04_18_ shortpath4 patterns test results using different polymerization methods The worst results were obtained with method 3), so it may be concluded that an excessive time of polymerization will lead to unsatisfactory performances. The other polymerization methods showed so dispersed results that proper conclusions could not be made. Results were not reproducible and they showed in some cases extremely asymmetric speed flow results. The same patterns were made again, and a day between polymerization and test with deionized water was left. Results (which are marked with orange in table 8) showed worse performances and reproducibility than the previous ones. Actually, results were so unsatisfactory that more deionized water had to be dropped after 25 minutes of test time due to the low performances obtained. Although it is possible that time influences pattern performances, these results could be a consequence of the printing conditions, since EPSON printer is a comercially available printer and probably the amount of ink dispensed was changing. 22 In conclusion, patterns previously designed could not provide symmetric flows, so it is not possible to analyse any analyte parameter with these patterns, since coffee ring effect cannot be avoided. In order to study more in detail the effect of basic solutions on the patterns, the following experiment was conducted. Patterns PT1d1 and PT2d1, which had been tested with deionized water, were tested with pH 12 NaOH solution once deionized water had been vaporized. And once NaOH solution vaporized, deionized water was applied again. Results can be observed in table 9. Polymerization method Name-date Analyte Volume of analyte (µl) tA (s) tA1 (s) tA2 (s) tB (s) tB1 (s) tB2 (s) tC (s) Deionized water 30; after 25' 30µl more 400 NaOH (pH 12) 30 24 34 34 57 50 2 PT1d121/6/201 1 Deionized water 30 26 34 34 41 52 Deionized water 30; after 25' 30µl more NaOH (pH 12) 30 30 60 70 70 95 97 3 PT2d121/6/201 1 Deionized water 30 86 98 98 91 111 111 115 Table 9. C_s45s4w1ch24_cr_ch04_18_ shortpath4 patterns test results using different analytes Observing the results obtained, it can be observed that pH 12 NaOH solution flowed extremely faster than the deionized water one. However, when deionized water was dropped again, it flowed as fast as NaOH solution. This may mean that basic solutions can dilute polymer ink contamination that locates in the path or reduce somehow the path resistance. However, it exists the possibility that some NaOH contamination remained in the paper. Future research could be focused on this phenomena understanding. 4.3.4. Pattern C_s45s4w1ch24_cr_ch04_18_shortpath4_sext9 Finally, pattern C_s45s4w1ch24_cr_ch04_18_shortpath4_sext9 was designed. This pattern is similar to the previous one, but an additional 9x9 mm 2 dropping square was made and connected with the previous dropping water square. The basic idea was, on one hand, to create an area to drop a large quantity of analyte in order to surpass the path resistance and on the other hand to drop the analyte in a zone which was not the zone where paths get divided. If the dropping square is the same square where paths begin, it is possible that flow does not distribute homogeneously in it depending on the exact dropping position, which is clearly influenced by human error, and consequently flowing different water quantities on each path, which will produce flow asymmetry. However, if analyte is dropped in another area, when analyte arrives at the bifurcation area, analyte distribution will be more symmetric. Figure 17 has been created to explain it in more detail. 23 Figure 17. Outline representing analyte distribution depending on pattern geometry Some patterns C_s45s4w1ch24_cr_ch04_18_shortpath4_sext9 were tested, but no satisfactory results were achieved. Although a lot of water was dropped (until 400 µl), it remained in the big dropping square and only a little part flowed, so the results obtained were the same as C_s45s4w1ch24_cr_ch04_18_ shortpath4 results, where blocking, flow asymmetry and low flowing performances occured. However, the idea of having a dropping square and a separate path bifurcation zone was used in future patterns. In figure 18, a tested C_s45s4w1ch24_cr_ch04_18_shortpath4_sext9 pattern can be observed. Figure 18. C_s3s4w1ch24_cr_ch04_18_ shortpath4 pattern tested with red dye. Not real dimensions 4.4. Simple symmetry patterns In order to solve the asymmetry problems, a new series of patterns were made. All the geometry tips learned from the previous sections were applied. For example, the dropping area and the path bifurcation area were located in different zones. To enhance flow symmetry, patterns designed could be polymerized without changing the UV spotlight position. In other words, all the patterns designed could be located in a 20 mm diameter circumference, which is the UV spotlight diameter. It must be taken into account that changing the UV spotlight position produces polymerization inhomogeneity, since central parts of the pattern get more polymerized than outer parts, and it is not possible to move the spotlight manually locating it in the same position all time. Apart from reducing the patterns size to be polymerized at once, paths were made as big as possible, in order to reduce the ink spread effects in the flow. 24 This section experiments were performed using monomer ink without stabilizer. 4.4.1. Pattern Simple symmetry 1 To begin with, Simple symmetry 1 pattern was designed. This pattern design is the simplest one, and it is very similar to Simple test pattern series, but both dropping, sensor reagent squares and paths dimensions were bigger. In figure 19, Simple symmetry 1 geometry is shown. Figure 19. Simple symmetry 1 pattern geometry In figure 20 a tested Simple symmetry 1 pattern can be observed. Figure 20. Simple symmetry 1 pattern tested with red dye. Not real dimensions The reason to do this pattern was to check that a pattern with the polymerization conditions (without moving the spotlight) and path dimensions commented before would be able to reach satisfactory flow symmetry performances. These patterns were printed with EPSON printer and polymerized for three minutes. 20 µl of deionized water were used to do the tests. Results can be observed at table 10, where time required for the analyte to reach the opposite square is shown. Name-date Time to reach the sensor area (s) Name-date Time to reach the sensor area (s) 3923/6/2011 3 4423/6/2011 135 4023/6/2011 2 4523/6/2011 4 4123/6/2011 9 4623/6/2011 8 4223/6/2011 9 4723/6/2011 69 4323/6/2011 4 Table 10. Simple symmetry 1 pattern test result Although two abnormal results were obtained, the rest were quite similar and reproducible. Once checked this pattern flow performance was satisfactory, two path patterns were designed in order to check the flow symmetry. 4.4.2. Pattern Simple symmetry 2 and Simple symmetry 3 Next pattern designed was Simple symmetry 2. This pattern geometry consisted of a dropping square and two sensor reagent squares. The reason to make this pattern was not to print two sensor reagents, but checking the pattern flow symmetry of this kind of pattern. This pattern test results are shown at table 11. In the test 20 µl of deionized water were used and a polymerization time of 3 minutes was applied. 31 to measure analytes properties with the system studied in this project. The procedure to create the different pH buffer solutions used are specified at section 3.4 (Preparation of the pH buffers). The analyte amount used in the patterns was 12 µl, except in pattern 106, that needed 23µl. Before and after doing the patterns tests, they were scanned in order to detect the sensor reagent colour digitally (Digital Color Meter software Ver. 3.6.1 provided with the Mac OS X Ver. 10.5). The colour detection system used was the L*a*b* system, and the parameter used to analyse the colour difference was a*, since the most important colour change occurred in the red tonality. Moreover, in order to check the relation between colour change and time, patterns were scanned before testing and at different times after testing. In Annex III. Simple symmetry 6 (made with EPSON and Dimatix) patterns using non particle pH based sensor reagent. Colour and image data patterns colour data, as well as their images, can be observed. Observing all patterns colour data before being tested, it can be noticed that the printing reproducibility was so high, and the initial a* values were very similar, being the average value 15.782 and the standard deviation 0.671. So, in order to analyse the colour change, only the testing a* value must be taken into account, since the initial a* value was approximately the same for all the patterns. Looking at the results, it can also be observed that the colour values depended on time. The most reproducible and coherent results were obtained, approximately, between 10 and 25 minutes after the test. It is logical that some time is required to let the flow reach all the analyte layers and that after some time the pH sensor reagent may be affected by other factors that may interfere in the colour change, such as its dilution into the analyte. Moreover, depending on the pH buffer, different times were required to change the colour, so, although the optimal colour data was obtained between 10 and 25 minutes, depending on the case these time lapse may be slightly modified. In figure 29, a calibration curve relating the colour and the pH can be observed. As it was commented previously, the values used to do it were taken between 10 and 25 minutes after the test. -5 0 5 10 15 20 56789 pH a* Figure 29. Calibration curve correlating the colour data obtained using pH buffer solutions in Simple symmetry 6 patterns A good calibration curve with so little error could be created. The error bars indicate the samples standard deviation values (4 samples were performed for every pH buffer). However, it must be taken into account that only two of the four experiments made with pH 7 buffer were used, since the rest showed very different values. Regarding coffee ring effect, it could be avoided in a great quantity of the patterns tested. However, some of them showed some, since the flow symmetry obtained was not perfect, pushing away the sensor reagent in some cases. This is not a serious problem, since the coffee ring effect produced in these cases was not pronounced at all and it did not affect the results. 32 In conclusion, the creation of this calibration curve demonstrates that two paths patterns are a feasible method to detect analyte properties, avoiding to a large degree the coffee ring effect and providing so exact results. 4.4.5.2. Test using only EPSON Once it was demonstrated that two path channels are a feasible and useful tool, the same experiment was done again, in the same conditions, but this time using only EPSON printer. In other words, the sensor reagent was printed using EPSON printer. The sensing reagent zone dimensions were the same than in the previous experiment, as well as the interlayer printing time. However, the number of layers printed was changed to 10 layers, instead of 15. As it can be observed in Annex IV. Simple symmetry 6 (made only with EPSON) patterns using non particle pH based sensor reagent. Colour and image data , EPSON results showed a good reproducibility. For example, the standard deviation value of the sensor reagent colour data (a*) before being tested was 0.97, which is similar to the result obtained when Dimatix was used in the previous experiment (0.67). Moreover, colour data values were almost similar, being 15.78 when Dimatix was used and 15.09 when EPSON was used. In figure 30, the calibration curve obtained with the different pH values can be observed. Figure 30. Calibration curve correlating the colour data obtained using pH buffer solutions in Simple symmetry 6 patterns. Only EPSON printer was used. As it can be observed, results were so reproducible, and standard deviation values were acceptable in all cases, although in pH 7 and 8 the standard deviation values were higher than in the previous experiment. It must be taken into account that in some cases coffee ring effect appeared, as it can be observed in figure AIV6 of Annex IV. Simple symmetry 6 (made only with EPSON) patterns using non particle pH based sensor reagent. Colour and image data, and in some of these cases the colour detection zone was moved from the original sensing reagent zone, in order to achieve more reliable results. 33 It must be taken into account that for this experiment new non particle based pH sensor reagent was done, so colour data obtained for the different pHs varied from the ones obtained in the previous experiment. In conclusion, in this last experiment it was demonstrated that it is possible to detect quantitatively a substance property, like pH, using only a commercially available printer to print the reagents, making more feasible the possibility of using and creating this kind of sensing system in the same place where it is required. Although in this case the sensor reagent used was a pH sensor reagent, it could be used with other kinds of sensor reagents, too. 5. Conclusions The main goal of this project was to develop a sensing system that was able to detect analytes or analyte properties without the necessity of either qualified professionals or sophisticated equipment. This goal has been achieved, since a sensing device was created using a commercially available printer, an UV light, a scanner and a simple colour detection software. The system designed was proved using a particle and a non particle based pH sensing reagent, and for both of them a calibration curve relating pH values and sensor reagent colour data were perfomed. Moreover, the system designed is not only limited to pH sensor reagents, and more sensor reagents may be used, since the system is versatile, accepting even the use of aqueous (non particle based) reagents, too. A better comprehension of the ink spread effect and its consequences was obtained. Ink spread was the main cause of assymetry in the patterns, since hydrophobic ink invades the hydrophilic path in an irregular way, producing irregular flow resistance, which is increased enormously when the path is lengthened. So, interesting pattern design tips, such as the most suitable polymerization conditions, were learned in order to avoid assymetry and path resistance. It was also demonstrated that two path and four path channels are feasible methods to reduce flow asymmetry, and consequently the coffee ring effect produced when the analyte washed away the sensor reagent. 6. Future research In this project a sensing system was presented, but it was only tested with pH sensing reagent. However, one of the project aims was to create a sensing device as universal as possible, so more sensing reagents should be tested. Patterns design could be improved in order to improve patterns performances. For example, a four path channel may be designed using the Simple symmetry patterns design criteria. This new design may be able to reduce the coffee ring effect. It was observed that depending on the analyte used, flow velocity changed. It was also observed that the more basic or acid the analyte pH was, the faster it flowed, while the more neutral, the slower it flowed. It would be interesting to check which are the factors that produce this phenomena, as well as checking if the paper structure was modified. Finally, it would be interesting to modify the patterns created in order to create multianalyte sensing devices, such the one showed in figure 31. 34 Figure 31. Multianalyte sensing pattern designed using the Simple symmetry patterns design criteria 7. Acknowledgement I would like to thank Daniel Citterio (Master’s Thesis advisor), Imoto Shota and Maejima Kento for the assistance and advice provided, and to Suzuki Koji for letting me to do the project in the laboratory he directs. 8. Bibliography [1] Martinez, A.W. et al. Diagnostics for the developing world: microfluidic paper-based analytical devices. Analytical Chemistry 82(1), 2010. [2] Martinez A.W. et al. Flash: a rapid method for prototyping paper-based microfluidic devices. Lab on a Chip [www.rsc.org/loc]. [3] Martinez A.W. et al. Three-dimensioanl microfluidic devices fabicated in layered paper and tape. Proc. Natl. Sci. U.S.A. 105, 2008. [4] Carrilho E. et al. Understanding wax printing: a simple micropatterning process for paperbased microfluidics. Analytical Chemistry 81(16), 2009. [5] Sele, C. W. et al. Self-aligned inkjet printing with sub-hundred nanometer resolution. Adv. Mater. 17, 2005. [6] Sirringhaus, H. et al. High-resolution ink-jet printing of all-polymer transistor circuits. Science 290, 2000. [7] de Gans, B. et al. Inkjet printing of polymers: State of the art and future developments. Advanced Materials, 16(3), 2004. [8] Yoshioka Y. et al. Simple modification of sheet resistivity of conducting polymeric anodes via combinatorial ink-jet printing techniques. Macromolecular Rapid Communications, 26(4), 2005. [9] Setti L. et al. An amperometric glucose biosensor prototype fabricated by thermal inkjet printing. Biosensors and Bioelectronics, 20(10), 2005. [10] Setti L. et al. An HRP-based amperometric biosensor fabricated by thermal ink-jet printing. Sensors and Actuators, 126(1), 2007. [11] Abe K. et al. Inkjet-printed paperfluidic immuno-chemical sensing device. Analytical and bioanalytical chemistry, 398(2), 2010. 35 [12] Abe K. et al. Inkjet-Printed Microfluidic Multianalyte Chemical Sensing Paper. Analytical Chemistry, 80(18), 2008. [13] Fenton E. M. et al. Multiplex lateral-flow test strips fabricated by two-dimensional shaping. Applied Materials & Interfaces. 1(1), 2009. [14] Bruzewicz D. A. et al. Low-cost printing of poly(dimethylsiloxane) barriers to define microchannels in paper. Analytical Chemistry, 80(9), 2008. [15] Martinez A.W. et al. Diagnostics for the developing world: microfluidic paper-based analytical devices. Analytical Chemistry, 82(1), 2010. [16] Lab on paper. Lab on a Chip, 8, 2008. [17] Aikio S. et al. Bioactive Papers and Fiber Products: Patent and Literatura Survey. VTT Working Papers, 51, 2006. [18] Zimmermann M. et al. Capillary pumps for autonomous capillary systems. Lab on a Chip, 7(1), 2007. [19] Martinez A.W. et al. Patterned paper as a platform for inexpensive, low-volume, portable bioassays. Andgewandte Chemie International Edition, 46(8), 2007. [20] Bruzewicz D.A.et al. Low-cost printing of poly(dimethylsiloxane) barriers to define microchannels in paper. Analytical Chemistry, 80(9), 2008. [21] Martinez A.W. et al. Simple telemedicine for developing regions: camera phones and paper-based microfluidic devices for realtime, off-site diagnosis. Analytical Chemistry, 80(10), 2008. [22] Su S.et al. Microgel-based inks for paper-supported biosensing applications. Biomacromolecules, 9(3), 2008. ANNEX Annex. Index I. Pattern glossary 2 II. Printer settings 8 III. Simple symmetry 6 (made with EPSON and Dimatix) patterns using non particle pH based sensor reagent. Colour and image data 10 IV. Simple symmetry 6 (made only with EPSON) patterns using non particle pH based sensor reagent. Colour and image data 13 2 I. Pattern glossary In this section a glossary with all the patterns designed for the project has been done. All patterns shapes can be seen observed together with their respective names and important measures. In most cases, pattern names provide useful information about them. For example, “Simple test” pattern series refer to the patterns made at the beginning of the project to establish some basic geometry measures, and “Simple symmetry” patterns refer to the last patterns done, which, were very simple in order to solve their symmetry problems. However, most of the pattern names provide information about their geometry and measures. The standard name may be: C_sX1sX2cirX3wX4chX5 (1) • sX1 indicates the dropping square side length • sX2 indicates the sensor square side length, in case it would be different from the dropping square one • cirX3 indicates the sensor zone diameter, in case its shape would be a circumference • wX4 indicates the pattern wall width • chX5 indicates the channel width It must be taken into account that all measures are given in mm (decimal points were omitted). Pattern names endings provide information, too. For example, “_shortpath+” refers to the patterns whose path was made shorter in order to enhance their performance. “_shortpath4” refers to patterns whose channels are divide into two. In the case of “_shortpath4” patterns, their name may be lengthened in order to provide more useful information. A standard name may be: C_sX1sX2cirX3wX4chX5_cr_ch04_X6_shortpath4 (2) • _cr indicates that the path between the dropping square and the sensor zone has a bent shape. If it is not the case, _cr does not appear • _ch04_ X6 indicates path width of the bifurcated channels For example, C_s45s4w1ch24_cr_ch04_18_shortpath4 indicates that the pattern is composed by two squares, the dropping one has a side of 4.5mm and the sensor one of 4 mm. The channel width is 1 mm and the path width is 2.4 mm. The paths are bifurcated in two, and each has a width of 1.8 mm. The path between the two squares has a bent shape. Other: Wall width: 0.24 mmSensor square: 3.2x3.2 mm 2 Channel width: 1.12 mmDropping square: 3.2x3.2 mm 2 Name: Simple test 1 Other: Wall width: 0.3 mmSensor square: 3.2x3.2 mm 2 Channel width: 1.5 mmDropping square: 3.2x3.2 mm 2 Name: Simple test 2 Other: Wall width: 0.3 mmSensor square: 3x3 mm 2 Channel width: 1.5 mmDropping square: 3x3 mm 2 Name: Simple test 3 Other: Wall width: 0.3 mmSensor square: 2.5x2.5 mm 2 Channel width: 1.5 mmDropping square: 2.5x2.5 mm 2 Name: Simple test 4 Other: Wall width: 0.3 mmSensor square: 2x2 mm 2 Channel width: 1.5 mmDropping square: 2x2 mm 2 Name: Simple test 5 Other: Wall width: 0.3 mmSensor square: 1.5x1.5 mm 2 Channel width: 1.5 mmDropping square: 1.5x1.5 mm 2 Name: Simple test 6 Other: Wall width: 0.3 mmSensor square: 1.2x1.2 mm 2 Channel width: 1.5 mmDropping square: 1.2x1.2 mm 2 Name: Simple test 7 Other: Wall width: 0.3 mmSensor area: 3 ϕmm Channel width: 1.5 mmDropping square: 3x3 mm 2 Name: C_s3cir3w03ch15 1 2 3 4 5 6 7 8 3 Other: Wall width: 0.3 mmSensor square: 3x3 mm 2 Channel width: 1.5 mmDropping square: 3x3 mm 2 Name: C_s3w03ch15 Other: Wall width: 0.3 mmSensor square: 3x3 mm 2 Channel width: 1.2 mmDropping square: 3x3 mm 2 Name: C_s3w03ch12_shortpath+ Other: Wall width: 0.3 mmSensor square: 3x3 mm 2 Channel width: 1.5 mmDropping square: 3x3 mm 2 Name: C_s3w03ch15_shortpath+ Other: Wall width: 0.3 mmSensor square: 3x3 mm 2 Channel width: 1.8 mmDropping square: 3x3 mm 2 Name: C_s3w03ch18_shortpath+ Other: Wall width: 0.3 mmSensor area: 3 ϕmm Channel width: 1.2 mmDropping square: 3x3 mm 2 Name: C_s3cir3w03ch12_shortpath+ Other: Wall width: 0.3 mmSensor area: 3 ϕmm Channel width: 1.5 mmDropping square: 3x3 mm 2 Name: C_s3cir3w03ch15_shortpath+ Other: Wall width: 0.3 mmSensor area: 3 ϕmm Channel width: 1.8 mmDropping square: 3x3 mm 2 Name: C_s3cir3w03ch18_shortpath+ Other: bifurcated paths were approximately 0.75 mm Wall width: 0.3 mmSensor area: 3 ϕmm Channel width: 1.5 mmDropping square: 3x3 mm 2 Name: C_s3cir3w03ch15_shortpath4 14 15 16 13 12 11 10 9 4 11 pH 9 23 min after test Pattern L* a* b* 107 65.982 -1.996 -5.785 108 58.820 -2.539 -6.605 109 110 Table AIII1. Colour data of the Simple symmetry 6 patterns tested with pH buffers III.2. Image data Before test After test After test Pattern pH buffer used Image Time (min) Image Time (min) Image 93 94 5 7 25 97 98 6 7 20 100 102 7 17 104 20 105 8 5 15 109 110 9 5 14 Table AIII2. Image data of the Simple symmetry 6 patterns tested with pH buffers Some images of the patterns are shown, too. As it can be observed, some of them were slightly affected by coffee ring effect, although this was not a serious problem. Figure AIII1. Pattern 100 (pH 7), Figure AIII2. Pattern 101 (pH 7), 17 minutes after the test 17 minutes after the test Figure AIII3. Pattern 96 (pH 6), Figure AIII4. Pattern 106 (pH 8), 20 minutes after the test 10 minutes after the test 12 Figure AIII5. Pattern 108 (pH 9), Figure AIII6. Pattern 110 (pH 9), 23 minutes after the test 10 minutes after the test 13 IV. Simple symmetry 6 (made only with EPSON) patterns using non particle pH based sensor reagent. Colour and image data IV.1. Colour data pH 5 Before test 14 min after test 20 min after test Pattern L* a* b* L* a* b* L* a* b* 129 77.263 15.742 12.688 68.659 17.605 16.195 70.747 16.594 16.695 130 75.860 17.098 13.828 71.662 18.930 19.402 73.983 16.539 16.664 131 77.292 14.730 12.562 72.374 18.551 18.949 74.369 16.273 18.793 132 77.694 16.379 16.121 74.786 18.227 18.242 75.798 15.523 18.066 pH 5 23 min after test Pattern L* a* b* 129 71.458 15.094 15.621 130 75.602 15.430 14.453 131 75.882 14.910 16.246 132 76.703 14.176 15.945 pH 6 Before test 13 min after test 23 min after test Pattern L* a* b* L* a* b* L* a* b* 113 77.101 14.383 10.656 76.976 14.910 19.691 76.828 12.867 14.758 114 77.635 14.543 10.629 77.916 13.945 21.977 77.508 11.961 18.922 115 76.655 14.398 10.887 76.036 14.949 21.469 74.495 14.668 18.234 116 77.867 14.742 9.793 77.824 13.664 20.957 76.532 12.227 16.941 pH 7 Before test 10 min after test 16 min after test Pattern L* a* b* L* a* b* L* a* b* 117 76.544 14.938 11.941 78.906 7.582 25.652 80.490 7.387 21.031 118 76.791 15.004 12.723 81.032 5.008 26.832 82.036 5.176 22.703 119 77.642 14.488 12.672 79.791 5.410 27.227 82.040 4.594 21.492 120 78.665 13.613 12.004 81.492 3.973 24.215 82.196 4.055 22.016 Before test 10 min after test 15 min after test 133 74.438 15.531 15.848 81.556 6.309 29.762 78.676 7.281 27.527 pH 7 22 min after test Pattern L* a* b* 117 80.378 7.812 18.949 118 82.681 5.090 19.391 119 81.993 5.281 19.438 120 82.904 4.504 19.242 20 min after test 133 79.355 7.102 24.66 14 pH 8 Before test 11 min after test 16 min after test Pattern L* a* b* L* a* b* L* a* b* 121 79.237 12.816 11.859 84.149 -1.617 24.250 85.858 -0.273 20.695 122 76.880 14.609 12.652 81.694 1.902 25.527 83.084 2.074 23.445 123 76.159 14.500 12.688 80.609 -2.238 26.672 82.960 -1.121 25.984 124 76.600 15.422 12.277 81.433 -2.871 26.555 84.215 -0.910 24.438 Before test 10 min after test 15 min after test 134 74.937 15.992 15.418 89.868 -3.223 13.754 87.597 -2.578 15.039 135 76.242 15.754 18.145 88.862 -3.297 21.203 86.011 -3.340 22.941 pH 8 23 min after test Pattern L* a* b* 121 85.890 -0.098 20.059 122 82.835 1.785 23.074 123 84.228 -0.184 22.395 124 85.629 -0.059 21.262 20 min after test 134 87.700 -2.402 14.789 135 86.220 -2.086 22.160 pH 9 Before test 9 min after test 17 min after test Pattern L* a* b* L* a* b* L* a* b* 125 76.771 14.367 12.461 73.883 -7.449 0.746 79.451 -4.691 -0.242 126 76.652 14.641 12.660 70.494 -7.320 0.051 75.404 -4.285 -1.270 Before test 7 min after test 13 min after test 127 74.861 15.852 15.805 77.807 -6.363 2.293 81.416 -5.051 1.613 128 73.892 16.363 16.332 70.582 -7.691 1.242 78.843 -4.656 0.848 136 76.144 16.297 18.031 76.184 -7.047 1.188 81.079 -5.105 0.852 pH 9 24 min after test Pattern L* a* b* 125 78.744 -4.723 -0.949 126 75.053 -4.691 -2.348 20 min after test 127 128 77.957 -4.426 -0.32 136 Table AIV1. Colour data of the Simple symmetry 6 patterns tested with pH buffers using only EPSON printer 15 IV.2. Image data Before test After test After test After test Pattern pH buffer used Image Time (min) Image Time (min) Image Time (min) Image 130 131 5 14 20 23 115 116 6 13 23 117 118 7 10 16 22 122 123 8 11 16 23 125 126 9 9 17 24 Table AIV2. Image data of the Simple symmetry 6 patterns tested with pH buffers Some images of the patterns are shown, too. As it can be observed, some of them were strongly affected by coffee ring effect. Figure AIV1. Pattern 130 (pH 5), Figure AIV2. Pattern 131 (pH 5), 20 minutes after the test 20 minutes after the test Figure AIV3. Pattern 117 (pH 7), Figure AIV4. Pattern 118 (pH 7), 16 minutes after the test 16 minutes after the test 16 Figure AIV5. Pattern 125 (pH 9), Figure AIV6. Pattern 126 (pH 9), 17 minutes after the test 17 minutes after the test