URINE GLUCOSE DETECTION VIA GOLD NANOPARTICLE FORMATION USING 3D-CONNECTOR MICROFLUIDIC PAPER-BASED ANALYTICAL DEVICES

A metabolic disorder that has experienced a significant increase in the world is diabetes mellitus. Diabetes is caused by two main factors: the first is damage to pancreatic beta cells, which prevents insulin from being produced, and the second is impaired insulin secretion and function. Chronic diabetes, if not treated properly, can lead to acute complications including eye, kidney, lung, nerve, and even death. Diabetes can be diagnosed through blood and urine. In general, glucose detection is carried out using invasive methods that use blood samples, which can cause pain and discomfort for users. Current research is developing non-invasive glucose detection using urine samples. This research aims to develop non-invasive glucose detection technology using 3D-connector μPADs (Microfluidic Paper Based Analytical Devices) which have the advantages of being safe, easy, and simple. The three-dimensional connector on the device functions as a connector to facilitate the coordination of fluid flows in the sample zone and detection zone. The glucose detection method uses gold (III) chloride as a gold nanoparticle (AuNPs) precursor, an aqueous extract of Acalypha indica Linn as a stabilizing agent, sodium hydroxide (NaOH) as a catalyst, and glucose in artificial urine as a sample. Method validation results using imageJ software indicated linearity with a coefficient of determination value (R 2 ) of 0.9714, precision with a %RSD value (Relative Standard Deviation) of 2.69, and an accuracy level ranging from 92.22-99.23%.


INTRODUCTION
Diabetes mellitus is a complex metabolic condition characterized by high blood sugar levels (hyperglycemia) caused by insulin production or its effectiveness [1].Insulin is crucial in regulating blood glucose levels, glucose assimilation, and utilization.Insulin resistance, experienced by individuals with diabetes mellitus lead body cells to reject insulin hormone, accumulating glucose in the bloodstream [2].Diabetes is divided into three primary types: type 1 diabetes, type 2 diabetes, and gestational diabetes induced by pregnancy [3].According to the International Diabetes Federation (IDF), In 2017, there were 451 million adults living with diabetes worldwide.Diabetes is estimated to impact 693 million individuals by 2045 if not properly managed and prevented, despite many efforts to increase life expectancy [4].
Excessive consumption of sugary foods and an unhealthy lifestyle are contributing factors to diabetes, leading to an increased Body Mass Index (BMI).Individuals with a higher BMI are more likely to develop type 2 diabetes [5].Besides BMI, the characteristics of diabetes patients include having blood glucose levels that reach ≥ 126 mg/dL [6].According to the American Diabetes Association, individuals with diabetes mellitus need to preserve blood glucose by strictly controlling blood glucose levels below 10 mmol/L (180 mg/dL).Failure to comply can lead to serious complications including vision impairment, kidney damage, stroke, cardiovascular issues, liver problems, coma, and death.Regular blood glucose level monitoring is essential for medical diagnosis and preventing complications due to hyperglycemia, especially in diabetic individuals [7][8].Early detection of glucose levels in individuals with diabetes mellitus could be performed using both invasive and non-invasive methods.The invasive method requires blood samples, which can cause pain due to tissue damage, while the non-invasive method employs samples other than blood, such as interstitial fluid, urine, sweat, saliva, and tears, thus eliminating pain [9].
Microfluidic Paper-Based Analytical Devices (μPADs) hold an immense potential for the development of Point-of-Care Testing (POCT), the first concept was proposed by Whitesides in 2007 [10][11].µPADs as glucose detection devices fulfill the WHO (World Health Organization) "ASSURED" criteria, namely Affordable, Sensitive, Specific, User-friendly, Robust and rapid, Equipment-free, and Deliverable to individuals with diabetes melitus [12].The fabrication of µPADs has undergone significant advancements, including their potential as visually based detection devices (colorimetry) [13][14], electrochemistry [14], chemiluminescence [15], fluorescence [16], and Raman spectroscopy [17].µPADs as glucose detection devices could be categorized into two types: the colorimetric method and the distance-based method.The colorimetric method utilizes µPADs designs by observing the formed color without determining the distance of reaction, while the distance-based method involves measuring the distance based on the length of the generated color band.Influencing factors of µPADs analysis include color intensity calculation and lighting effects, which affect sensitivity and reproducibility [18].
The detection system used in µPADs is based on the formation of nanoparticles such as gold and silver.These nanoparticles play a role in inducing color changes that occur on the µPADs [19].Gold nanoparticles (AuNPs) are widely employed as catalysts due to their excellent biocompatibility, high stability, and low cytotoxicity [21], drug delivery [22], biosensors [23], cancer photothermal therapy [24], bioelectric devices [25], dan x-ray imaging [26].The presence of glucose in diabetic mellitus patients can be detected using gold nanoparticles [27].Gold nanoparticles are employed as sensing components due to their unique physicochemical and optical properties, ranging from their shape and surface characteristics to their size, which falls within the range of 1-100 nm [20].In light of the rising prevalence of diabetes mellitus, particularly in underdeveloped and developing countries, the development of 3D-Connector µPADs based on the distance-based reaction method holds promise as a reliable breakthrough for rapid, cost-effective, and safe glucose detection.

Instrumentation
The instrumentation used in this research included a Xerox ColorQube 8580 DN printer, solid ink color for fabricating 3D-μPADs, an oven for wax penetration into the paper substrate, and CorelDRAW Graphics Suite X7 software for designing the 3D-μPADs.Image capture was performed using a black box equipped with LED lights measuring 10x10x10 cm, LED lights, and a 12MP mobile phone camera with 2x magnification.

Fabrication of 3D-μPADs
Fabrication of 3D-Microfluidic Paper-Based Analytical Devices (μPADs) involved designing the μPADs using CorelDRAW Graphics Suite X7 and printing them onto Whatman Number 1 paper using Xerox ColorQube 8580 DN printer.After printing, wax penetration was performed on the paper substrate by heating it in an oven at 120°C for 3 minutes to maximize the formation of a hydrophobic barrier on the 3D-μPADs.Subsequently, the paper was cut according to the design using a paper trimmer and scissor.The fabrication process of the 3D-μPADs is illustrated in Figure 1.

In-situ Synthesis of AuNPs on 3D-μPADs
In-situ synthesis of AuNPs on 3D-μPADs is depicted in Figure 2. 10 μL solution of 100 mM gold (III) chloride (HAuCl4) and Acalypha indica Linn aqueous extract 0.25% (5:1 volume ratio) was evenly dropped onto the detection zone from bottom to top then dried for 15 minutes at room temperature.A 0.25% extract of Acalypha indica Linn was prepared by mixing 5 g of Acalypha indica Linn powder with 200 g of distilled water, then stirred and heated at 60°C for 20 minutes.Subsequently, the 3D-connector was folded and sealed with transparent tape and clamped with acrylic to maximize the reaction distance efficiency.After folding, a 20 μL sample containing 100 mg/dL glucose in artificial urine and 10 M NaOH was dropped onto the detection zone and dried for 15 minutes at room temperature.The color change and length of the reaction distance were observed for 15 minutes.Data were collected through analysis using ImageJ software by calculating the slope value (y/x) based on Equation 1 and Figure 3.
A clear color boundary is indicated by a high slope value.The blue color intensity is chosen because the slope value of the blue color intensity is higher than those of the red and green colors in the ImageJ analysis.

Validation Method
Validation method was performed after obtaining the optimal conditions from the optimization of reagent volume, sample volume, drying time, reagent concentration, and reaction time.Method validation included linearity, precision, and accuracy.Linearity measurement is performed using artificial urine with the composition shown in Table 1 with standard glucose analyte concentrations of 0, 20, 40, 60, 80, and 100 mg/dL.Then the obtained data are processed to obtain SD (Standard Deviation) and %RSD (Relative Standard Deviation).Precision determination is carried out by measuring the concentration of standard glucose analyte 100 mg/dL for six times.Accuracy is determined by measuring the concentration of standard glucose analyte in artificial urine 15, 25, 50, and 75 mg/dL, then compared to the previous standard glucose analyte at equivalent concentrations.

Effect of Wax Penetration On 3D-Μpads
Wax penetration into the 3D-μPADs serves to create a hydrophobic barrier on the paper substrate, thereby optimizing the solution permeation process.During the wax heating process, there is a narrowing of the detection zone, sample zone, and 3D connector as shown in Table 2. Based on the results of wax penetration measurements, there is the narrowing of the detection zone to (0.89 x 0.69) mm, the sample zone to (0.77 x 0.76) mm, and the 3D connector zone to (0.86 x 0.8) mm.

In Situ Synthesis of Gold Nanoparticles (AuNPs)
In-situ synthesis of AuNPs is carried out using HAuCl4 precursor, Acalypha indica Linn aqueous extract as a stabilizer, NaOH as a catalyst, and glucose as a reductant, which takes place insitu on 3D-μPADs.3D-connectors in μPADs control fluid flow by connecting and separating reagent and sample fluids, thereby enhancing the detection accuracy and performance of 3D μPADs.The purpose of folding the 3D connector part is to prevent air contamination and accelerate the reaction distance of the sample solution.The order of dropping is determined based on the detectable color change and length of the reaction distance.Glucose, which plays a role in reducing Au 3+ to Au 0 , binds to the surface of gold nanoparticles through hydrogen bonding of hydroxyl groups (OH -).AuNPs cannot form under acidic conditions, hence NaOH is required to create a basic environment [28].Detection of glucose through the formulation of AuNPs is as follows [29].extract.The results of the reagent volume optimization can be seen in Figure 4. Visually, there is an increase in color sharpness from 6 μL to 8 μL reagent volume, followed by a decrease at 10 μL volume.The optimum reagent volume is 8 μL.The observation results are consistent with the ImageJ analysis, which shows the color sharpness threshold.The slope value measurement is performed three times using ImageJ software.The color formed at 8 μL volume is sharper than at 10 μL volume.At 10 μL volume, the color appears uneven, which is caused by the increasing amount of fluid accommodated in the detection zone, leading to a decrease in the formation of gold nanoparticles (AuNPs) and affecting the color sharpness.Sample volume determination is performed to identify the optimal sample volume from the mixture of glucose solution in artificial urine and NaOH.The results of the sample volume optimization can be seen in Figure 5.The observations reveal that the length of the reaction distance increases as the sample volume rises.The optimum sample volume is 18 μL.The color sharpness threshold starts to decrease when the sample volume is 20 μL and 22 μL.The decrease in color sharpness is attributed to the increased fluid volume in the sample zone, which prolongs the reaction time for gold nanoparticle (AuNPs) formation, leading to an indistinct color change threshold.sharpness produced at a reagent drying time of 10 minutes is fainter compared to 15 minutes.Color sharpness increases with increasing drying time from 10, 15, to 25 minutes.The drier the detection zone, the more AuNPs will form, as indicated by a sharper color threshold.At 15 and 25 minutes of drying time, the resulting color appears similar, but the color sharpness threshold at 25 minutes appears more even and sharper.This color sharpness enhancement ceases at 35 minutes when the resulting color is fainter.The determination of gold (III) chloride (HAuCl4) precursor concentration aims to identify the amount of AuNPs formed.The results of the concentration of HAuCl4 optimization are shown in Figure 7.The optimum HAuCl4 precursor concentration is 75 mM.The color observed at 25 mM HAuCl4 concentration is fainter compared to concentration of 50 mM.This indicates that at low concentrations, gold nanoparticles are not fully formed, resulting in an invisible color.As the HAuCl4 concentration increases, the color sharpness threshold or slope also increases.At HAuCl4 concentration of 25 mM, there is a significant increase to 50 mM, demonstrating that HAuCl4 plays a significant role in AuNPs formation and consequently influences color sharpness.The color sharpness threshold increases up to 75 mM concentration and then decreases at 100 mM.concentration decreases, AuNP formation diminishes due to the hindered glucose reduction reaction in converting Au 3+ to Au 0 .AuNP formation is more effective in alkaline conditions [29].The optimum NaOH concentration is 10 M. The optimal concentration of NaOH is 10 M due to the solubility of NaOH in water being 1 g/L at 25°C.At 4 M NaOH concentration, the length of the reaction distance is high but decreases at 6, 8, and 10 M NaOH concentrations.As the NaOH concentration increases, the resulting color becomes sharper and more distinct.The determination of reaction time aims to identify the optimal reaction duration for the detection of artificial urine using a mixture of glucose solution and NaOH applied to the sample zone.The results of the reaction time optimization can be seen in Figure 9.The optimum reaction time is 15 minutes.At a reaction time of 5 minutes, the resulting color appears faint, indicating that AuNPs are not fully formed.Significant color enhancement is observed at reaction times of 10 and 15 minutes.Prolonged reaction times promote AuNPs formation, consequently influencing the sharpness of the resulting color.

Linearity
Linearity is performed using the previously tested optimization on 3D-μPADs with glucose concentration variations of 0, 20, 40, 60, 80, and 100 mg/dL.Data analysis was carried out using ImageJ software and then processed in Microsoft Excel to measure the length of the reaction distance, resulting in a graph between log concentration (x) and the length of the reaction distance (y).The results of the linearity determination are shown in Figure 10.At a glucose concentration of 0 mg/dL, no color is formed because without glucose, the AuNPs formation reaction cannot occur.The color change becomes visible at glucose concentrations of 20-100 mg/dL, whereas at 20 mg/dL, the color is quite faint.This indicates that the AuNPs formed are also few, but the color sharpness is still clearly visible.The higher the glucose concentration used, the sharper and more distinct the resulting color.The linear regression correlation obtained with ImageJ software results in a coefficient of determination (R 2 ) value close to one, which is 0.9714.

Precision
Precision is a method of validation to assess the quality and consistency of measurements performed under identical conditions repeatedly.Precision was performed six times using a glucose concentration of 100 mg/dL in artificial urine.The analysis of the length of the reaction distance was carried out using ImageJ software.The precision test results in a %RSD of 2.69 and a standard deviation (SD) of 0.57.Accuracy is a method validation to determine the closeness of the measurement to the true value.Accuracy is performed using standard glucose solutions with concentrations of 15, 25, 50, and 75 mg/dL.Data analysis was performed using ImageJ software to obtain the length of the reaction distance.The results are then processed in Ms. Excel to obtain %error and %accuracy.The accuracy test results with ImageJ software produce a higher percentage of accuracy, namely in the range of 92.22-99.23%.The results of the accuracy determination are shown in Figure 11.

CONCLUSION
Non-invasive glucose detection can be achieved using 3D-μPADs through AuNPs formation based on color sharpness and reaction distance.AuNPs are one of metal nanoparticles with high biocompatibility and environmental friendliness.The optimization results of 3D-μPADs parameters include a reagent volume of 8 μL, sample volume of 18 μL, drying time of 25 minutes, gold (III) chloride concentration of 75 mM, NaOH concentration of 10 M, and reaction time of 15 minutes.After optimization, method validation for linearity yielded a coefficient of determination (R 2 ) of 0.9714, precision with %RSD of 2.59, and accuracy in the range of 92.22-99.23%.

Figure 4 .
Figure 4. (a) Graph of the relationship between the average slope value (y/x) and the reagent volume, (b) 3D-μPADs image for reagent volume optimization.

Figure 5 .
Figure 5. (a) Graph of the relationship between the average slope value (y/x) and the sample volume, (b) 3D-μPADs image for sample volume optimization.

Figure 6 .
Figure 6.(a) Graph of the relationship between the average slope value (y/x) and the reagent drying time, (b) 3D-μPADs image for reagent drying time optimization.The determination of reagent drying time is employed to identify the optimal drying time for the mixture of HAuCl4 solution and Acalypha indica Linn aqueous extract applied to the detection zone with equal reagent and sample volumes.The results of the reagent drying time optimization are shown in Figure 6.The optimum reagent drying time is 25 minutes.The color

Figure 7 .
Figure 7. (a) Graph of the relationship between the average slope value (y/x) and the concentration of gold (III) chloride, (b) 3D-μPADs image for gold (III) chloride concentration.

Figure 8 .
Figure 8.(a) Graph of the relationship between the average slope value (y/x) and the concentration of NaOH, (b) 3D-μPADs image for NaOH concentration.

Figure 9 .
Figure 9. (a) Graph of the relationship between the average slope value (y/x) and the reaction time, (b) 3D-μPADs image for reaction time.

Figure 10 .
Figure 10.(a) Graph of the relationship between glucose concentration and length of the reaction distance using imageJ, (b) 3D-μPADs image for linearity.

Table 1 .
Composition of artificial urine solutions.