3.1 Amount of NaCl on the µPAD

Initially, we optimized the volume of NaCl deposited on the µPAD. A volume of 6 µL was considered optimal as it ensured the rectangular channel was fully filled when the solution entered the opposite end, which guaranteed adequate coating of the section of the channel in contact with the water sensor.

Then, the concentration of the NaCl solution was varied within a range of 0.05–2 M, as shown in Fig. 2. Large errors were observed at concentrations of 0.05 and 0.25 M, and the measured time was longer. This indicated that insufficient NaCl had been deposited, which caused deionized water to flow beyond the second electrode of the water sensor without stopping the timer. Consequently, electric contacts were added to the third and/or fourth electrodes to stop the timer. The sensor stabilized its response when the concentration surpassed 0.5 M. However, a further increase in the concentration of NaCl slightly increased the stop time. This could be attributed to the change in flow rate, which depended on the amount of NaCl deposited in the channel. The sample solution flows into the channel at a rate that matches the rate at which NaCl dissolves. Consequently, as the concentration of NaCl increases, the flow speed decreases due to the solid NaCl impeding the flow and creating resistance. Therefore, at concentrations of 1 and 2 M, it took a longer time to stop the timer. Furthermore, NaCl dissolved more slowly in samples with high viscosity. Therefore, a lower concentration of NaCl would be preferable for producing consistent and accurate results in a shorter time. These results suggested that the most suitable NaCl concentration would be 0.5 M.

Fig. 2The alternative text for this image may have been generated using AI.

Effect of NaCl concentration on flow time. The channel of the µPAD consists of a 10-mm detection distance and a 4-mm width; 50 µL of deionized water was introduced into the µPAD. The paper substrate vehicle is Grade 1 CHR chromatography paper

It should be noted that salts dissolved in samples showed no influence on the detection time. This was confirmed by conducting a preliminary study in which solutions with different concentrations of NaCl were flowed through a channel that did not contain solid NaCl. Therefore, the salts in the samples exerted negligible influence on the detection time in the present detection system.

3.2 Optimizing the µPAD Design and Paper

Initially, Grade 1 CHR chromatography paper was employed as the substrate to investigate the effect of channel width and length on the µPADs. Channel widths that varied from 4 to 7 mm were tested, with flow times measured at 10, 20, 30, and 40 mm intervals. Regardless of the channel width, an increase in travel distance reduced the flow speed. This is consistent with the Lucas-Washburn equation, which states that the time required for liquid flow is proportional to the square of the travel distance. Therefore, the time required to flow a specific distance increases quadratically with an increase in distance. An increase in flow time led to significantly greater error and indicated that a shorter length was preferable for achieving reproducible, short flow times. Consequently, a length of 10 mm provided a shorter measurement time and higher reproducibility.

Conversely, the channel width had a smaller impact on flow time than the effect of the traveling length. However, elongating the traveling length results in a longer flow time for a narrower channel. While the Lucas-Washburn equation does not account for the effect of channel width on flow speed, narrower channels would exhibit greater friction between the liquid and the hydrophobic boundary within the µPADs. This would result in slower speeds [31]. Nevertheless, narrower channels led to smaller errors in flow time owing to a more uniform fluid front (i.e., a flat front) within the channel. Hence, a channel width of 4 mm minimized the error in the flow time.

Additionally, we compared the use of Grade 1 CHR chromatography paper with that of No. 2 filter paper. Water flowed faster through No. 2 filter paper than Grade 1 CHR chromatography paper. Generally, pore size significantly influences flow speed. Therefore, the No. 2 filter paper (pore size, 5 μm) had a faster flow rate than the Grade 1 CHR chromatography paper (pore size, 11 μm). Despite having different thicknesses (0.18 mm for Grade 1 CHR chromatography paper and 0.26 mm for No. 2 filter paper), thickness appeared to have little effect.

Although a shorter flow time is preferable for enhancing reproducibility, there was no significant difference in standard deviation when using either a 4-mm or a 10-mm channel with the two paper substrate vehicles. Consequently, the relative standard deviation of No. 2 filter paper was higher than that of Grade 1 CHR chromatography paper, which was likely due to the shorter average flow time when using the No. 2 filter paper. Given the short analysis time and high reproducibility, µPADs were fabricated with a 4-mm channel width and a 10-mm length, using Grade 1 CHR chromatography paper as the substrate vehicle.

3.3 Reproducibility of Detectors

The detection system is equipped with ten time detectors, which enable rapid successive measurements of up to ten samples. To evaluate the reproducibility of flow times, two experimental setups were compared: repeated measurements of water flow were conducted ten times using a single channel, and single measurements were conducted in rapid succession across ten channels. In the first setup, ten consecutive measurements were performed using a single detector; in the second, water samples were introduced to ten detectors in rapid succession for one-time measurements. The results demonstrated flow times of 38.92 ± 1.57 s and 37.45 ± 1.76 s for the ten consecutive measurements using one channel and for the single measurements using ten channels, respectively. Both methods exhibited comparable reproducibility, with relative standard deviations of less than 5%. These findings confirm that both repeated single-channel measurements and rapid, successive multi-channel measurements yield reliable and reproducible results.

3.4 Effect of Sample Volume

To elucidate the effect of an introduced volume on the measurement, volumes of deionized water ranging from 40 to 70 µL were added to the µPAD. The results indicated that flow time was independent of sample volume, which is consistent with the Lucas-Washburn equation. Consequently, the proposed system is user-friendly as it eliminates the need for a volumetric tool, such as a micropipette, to introduce the sample. This makes the system suitable for use in non-laboratory environments.

3.5 Relative Viscosity of BSA and Glucose

The solutes dissolved in the solution could affect the relationship between flow time and viscosity, because flow time depends on factors such as surface tension and contact angle. Therefore, in the present study, we investigated the relationships between viscosity and flow time for BSA and glucose solutions. The viscosities of the solutions were measured using an Ostwald viscometer, and their corresponding flow times were plotted against them, as shown in Fig. 3. Both BSA and glucose solutions exhibited a linear relationship, whereby flow time increased with increasing solution viscosity. According to Eq. 2, viscosity directly correlates with flow time provided that other factors, such as surface tension, contact angle, travel length, and the channel radius, remain constant, because these correspond to the paper’s pore size. Therefore, the results listed in Fig. 3 successfully express the relationship in Eq. 2.

However, the results in Fig. 3 also highlight a limitation. While the proposed method could be used to evaluate the viscosity of solutions containing similar chemicals, it cannot be relied on to compare flow times when the constituents differ greatly. This means that the flow time only directly reflects the relative viscosity of solutions containing similar solutes when α in Eq. 2 is constant. Therefore, measuring a series of related samples, such as biological fluids, will not pose a problem due to the expected similarity of their constituents. It is evident that the proposed system would be useful for evaluating the relative viscosities of solutions.

Fig. 3The alternative text for this image may have been generated using AI.

Relationships between viscosity and flow time of BSA and glucose. Appropriate amounts of either BSA or glucose were dissolved in deionized water, and the viscosities were measured using an Ostwald viscometer. Other conditions are described in the text

3.6 Protease-Activity Assay

The linear relationships between viscosity and flow time, as shown in Fig. 3, indicate that viscosity measurement will be used in a protease-activity assay. Proteases are enzymes that digest proteins. While protein solutions are viscous, their viscosity is reduced in the presence of a protease due to an enzymatic reaction that cleaves peptide bonds into small peptides and amino acids. Therefore, we employed the proposed system to measure viscosity after protein digestion to determine protease activity.

In the protease-activity assay based on viscosity measurement, gelatin acted as a substrate due to the high viscosity of its solutions. Gelatin solutions at concentrations of 1%, 3%, and 5%, as well as a bromelain solution at 0.06 mg/mL, were prepared in MES buffer at a pH of 6.0. These solutions were then mixed together at a volume ratio of 1:9 between bromelain and gelatin. When a 5% gelatin solution was introduced into the µPAD without bromelain, it did not flow into the channel, independent of the channel and pore sizes. Therefore, the protease assay did not examine a 5% gelatin solution. The flow times for 1% and 3% gelatin solutions in the absence of bromelain were 77 and 585 s, respectively. Adding 0.06 mg/mL of bromelain to the gelatin solutions and incubating the mixtures for 1 h at room temperature decreased the flow time to 58 s for 1% gelatin and 306 s for 3% gelatin. Using 1% gelatin resulted in a decrease in flow time of only 19 s, which was insufficient to obtain reproducible results. However, the change in flow time was 279 s for a 3% gelatin solution, which would allow for reproducible and distinguishable results. Therefore, a 3% gelatin solution was employed for the protease-activity assay.

The proposed protease-activity assay was used to determine the activities of trypsin, bromelain, and papain. Figure 4 illustrates the relationships between protease concentration and flow time. All proteases showed a decrease in flow time as the concentration increased, which indicated the digestion of gelatin. Linear relationships were obtained between concentration and flow time, with the slope depending on the type of protease. The slope reflects the activity of the individual protease: a large slope indicates high activity in the digestion of gelatin. The order of activity was bromelain > papain > trypsin, which is consistent with our previous findings. This order was also consistent with that obtained by using fluorescent casein as a substrate, as well as with a previously reported PADs, as shown in the previous study [32]. Thus, the present method could be successfully used to evaluate the enzyme activity of proteases.

Fig. 4The alternative text for this image may have been generated using AI.

Calibration curves for trypsin, papain, and bromelain. The appropriate amount of each enzyme was mixed with a 3% gelatin solution at a volume ratio of 1:9 in 50 mM of MES buffer (pH 6.0). The mixture was incubated for 1 h before introduction into the µPAD. Other conditions are described in the text

3.7 Saliva Viscosity

We also investigated how the viscosity of saliva samples changed over time after collection (Fig. 5a). While the initial viscosities of the three samples were significantly different, they decreased rapidly within the first few hours and eventually converged to a similar level. Rogers reported a similar decrease in viscosity and explained that mucin, the primary component responsible for the high viscosity of saliva, is metabolized and degraded by bacteria present in the saliva [33]. Therefore, the viscosity of saliva should be measured as soon as possible after sampling.

Fig. 5The alternative text for this image may have been generated using AI.

Viscosity measurement of saliva samples. a Changes in viscosity after sampling; b the relationship between salivary viscosity and exercise duration. Other conditions are described in the text

In addition, the change in the viscosity of saliva was determined after exercise, as shown in Fig. 5b. After 30 min of exercise, the saliva viscosity of the three volunteers generally increased. This aligns with the subjective input reported by the participants. As they were permitted to rehydrate during exercise, the occasional decrease in viscosity observed in volunteers 3 and 2 at 10 and 20 min, respectively, could have been caused by saliva dilution. However, the results clearly showed that the viscosity of saliva had increased after 30 min of exercise, even when the participants consumed water during the 5-minute rest intervals. Increased viscosity after exercise has been reported in a study where each subject cycled for 15 min and showed a 4-fold increase in viscosity [12]. Studies have shown that the salivary secretion rate decreases after exercise due to dehydration, which leads to an increase in protein concentration and secretion of salivary amylase and peroxidase [34]. Ligtenberg et al. [35] also observed an increase in mucin secretion after exercise. The results of previous studies support those obtained in the present study.

The present method has a significant advantage over conventional methods for measuring saliva samples, as it requires only 50 µL of sample volume. By contrast, conventional methods—such as rotational, capillary, and falling-ball viscometry—typically require several to hundreds of milliliters of a sample, which makes them unsuitable for saliva analysis due to the difficulty of collecting such large volumes. Ligtenberg et al. studied the effect that exercise exerts on salivary viscosity using 1 mL of saliva with a viscoelastometer [12]. Although instruments capable of measuring volumes of several microliters are available, they are generally expensive. Therefore, the present method is superior to conventional viscometric techniques and viscoelastometers in terms of required sample volume and cost-effectiveness. Furthermore, as shown in Fig. 5, the viscosity of saliva changes after collection. Therefore, the multichannel detection capability of the proposed system is advantageous because it enables rapid measurement immediately after sampling, which is crucial for obtaining accurate results. Saliva samples contain similar matrices for individuals, and the viscosity could be measured directly. However, it will be difficult to directly compare the viscosity of samples with different matrices. In such cases, a multi-channel detection system is advantageous because it enables rapid, successive measurements of viscosity for samples and references. For example, when measuring the viscosity of biological samples, the results could be normalized by including a reference solution in rapid succession, and the proposed detection system enables correction of the flow rate.