INTRODUCTION
Electrochemical glucose biosensors remain important analytical platforms because they combine rapid electrical readout, low sample-volume requirements, and compatibility with compact point-of-care devices [1,2]. Carbon nanomaterials are particularly attractive for screen-printed electrodes because their conductivity and high surface area can strengthen interfacial electron transport while retaining the low cost and disposability of the printed platform [3,4]. Sensor performance nevertheless depends strongly on the interfacial material used to immobilize the enzyme and transduce the enzymatic reaction.
In first-generation glucose biosensors, glucose oxidase (GOx) catalyzes glucose oxidation using molecular oxygen as the electron acceptor, producing hydrogen peroxide (H₂O₂) as an electroactive product [5]. H₂O₂ can then be detected electrochemically, linking enzymatic recognition to the measured current. Recent nanostructured GOx electrodes have demonstrated that carbon nanotubes and high-surface-area platinum architectures can improve enzyme loading, electron transport, and peroxide transduction [4,6,7]. PtNPs are effective electrocatalysts for H₂O₂, while carbon nanotubes provide conductive pathways and a high-surface-area support [8,9] 
PAMAM dendrimers provide a complementary interfacial function. Their highly branched architecture and numerous terminal groups can organize metal nanoparticles, interact with functionalized carbon nanomaterials, and support biomolecule immobilization [10,11]. The stepwise synthesis and characteristic generation structure of PAMAM were originally established by Tomalia and co-workers [12]. In the present study, G5-PAMAM, PtNPs, and acid-functionalized MWCNTs were integrated on a screen-printed carbon electrode. The composite was first evaluated for H₂O₂ transduction and was subsequently coupled to immobilized GOx for enzymatic glucose sensing. The study examines structure-property relationships, coating-layer optimization, scan-rate behavior, calibration, and same-electrode repeatability.

MATERIALS AND METHODS
Materials and electrode platform
MWCNTs (outer diameter 13-18 nm, length 1-12 µm, purity >99%), ethylenediamine, methyl acrylate, chloroplatinic acid (H₂PtCl₆), sodium borohydride, polyvinylpyrrolidone, GOx, chitosan, glutaraldehyde, ethanol, methanol, nitric acid, sulfuric acid, potassium ferricyanide, potassium chloride, H₂O₂, D-glucose, and phosphate-buffer constituents were used as received. Commercial HY-PC100 screen-printed carbon electrodes (ICHY, Shenzhen Haoyang Technology Co., Ltd., China) contained a 3-mm-diameter carbon working electrode (geometric area 0.0707 cm²), a carbon counter electrode, and an integrated Ag/AgCl reference electrode.

Synthesis of G5-PAMAM
G5-PAMAM was synthesized by alternating Michael addition and amidation steps [12]. Ethylenediamine (10 mL) was diluted in ethanol (40 mL), cooled in an ice bath under nitrogen, and reacted with methyl acrylate (112 mL) for 1-2 h to form the ester-terminated G0.5 intermediate. Volatile components were removed by rotary evaporation at 65 °C. For G1 formation, 16 g of G0.5 was reacted with ethylenediamine (54 mL) in ethanol (40 mL). Alternating half- and full-generation reactions were continued to G5, with solvent removal and purification between steps. Reaction completion was monitored using the ninhydrin assay, and the final product was purified by dialysis.

Acid functionalization of MWCNTs
Pristine MWCNTs (100 mg) were dispersed in a mixed HNO₃/H₂SO₄ oxidizing solution and maintained at 90 °C for 4 h to introduce oxygen-containing surface groups [13]. After cooling, the material was filtered and repeatedly washed with deionized water until neutral, then dried to obtain f-MWCNTs.

Synthesis of PtNPs and ternary nanocomposite
For independent PtNP synthesis, 1.0 mL of 0.01 M H₂PtCl₆ was added to 9.0 mL of distilled water containing 1% polyvinylpyrrolidone. The mixture was cooled in an ice bath, and freshly prepared 0.01 M NaBH₄ was added dropwise under stirring at 500 rpm. Reduction was continued for 30 min. The product was collected by centrifugation at 10,000 rpm for 10 min, washed four times with distilled water, and dried at 60 °C.
For the ternary composite, G5-PAMAM (100 mg) was dissolved in 5.0 mL methanol containing 1% polyvinylpyrrolidone and cooled in an ice bath. H₂PtCl₆ solution (1.0 mL) was added and stirred for 10 min before dropwise addition of cold NaBH₄. Separately, f-MWCNTs (50 mg) were ultrasonically dispersed in 10 mL deionized water. The nanotube dispersion was gradually added to the G5-PAMAM/PtNP dispersion and stirred for 6 h. The product was centrifuged at 10,000 rpm for 10 min, washed three to four times with deionized water, and vacuum-dried at room temperature for at least 24 h.

Electrode modification and GOx immobilization
The working electrode was rinsed with deionized water and pH 7.4 buffer, immersed in 30% HNO₃ for 1 h, washed three times with deionized water and three times with buffer, and dried at room temperature. The ternary nanocomposite was dispersed in pH 7.4 buffer at 10 mg mL⁻¹. A 5 µL aliquot was drop-cast onto the working electrode, dried, and gently rinsed. This sequence was repeated twice to form three nanocomposite layers.
GOx was prepared at 10 mg mL⁻¹ in pH 7.4 buffer. A 5 µL aliquot was applied to the nanocomposite-modified electrode and held at 8 °C for 24 h; deposition was repeated twice to form three GOx layers. After rinsing, 5 µL of 5% chitosan was applied and dried at 8 °C for 24 h. Finally, 5 µL of 25% glutaraldehyde was applied for 30 min, followed by rinsing with pH 7.4 buffer.

Structural and morphological characterization
X-ray diffraction patterns were collected using a Shimadzu XRD-6000 diffractometer with Cu Kα radiation (λ = 1.5406 Å) at 40 kV and 30 mA over 5-80° (2θ) with a 0.05° step. FTIR spectra were recorded using a Shimadzu IRAffinity-1 spectrometer over 400-4000 cm⁻¹. FESEM images were obtained using a TESCAN MIRA3 LMU microscope at 20 kV. TEM was used to examine Pt-containing nanodomains after ultrasonication and deposition on carbon-coated copper grids. Particle diameters were measured using FIJI/ImageJ.

Electrochemical measurements and data analysis
Electrochemical measurements were performed with a computer-controlled potentiostat using the integrated three-electrode SPCE. Ferricyanide characterization used K₃[Fe(CN)₆] in 0.2 M KCl supporting electrolyte at pH 7.4. A 150 µL aliquot covered the three-electrode area. Layer-number optimization compared one to four sequential nanocomposite coatings. Scan-rate measurements were recorded at 10, 25, 50, 100, and 200 mV s⁻¹.
H₂O₂ CV responses were recorded in 0.1 M PBS (pH 7.4) at 100 mV s⁻¹ using 5, 10, 25, 50, 100, 250, and 500 µM H₂O₂. H₂O₂ amperometry was performed at -100 mV versus integrated Ag/AgCl using 25, 50, 100, 250, and 500 µM. Glucose CV measurements used 10, 20, 40, 60, 80, 100, and 200 µM glucose in pH 7.4 buffer, 150 µL per measurement, a scan rate of 100 mV s⁻¹, and a potential sweep from -0.40 to +0.70 V and back. The magnitude of the selected reverse-scan cathodic peak was used as the analytical signal.
Linear regression used the ranges that remained linear after inspection of the full concentration series. The regression residual standard deviation was calculated as σ = [Σ(yi - ŷi)²/(n - 2)]¹ᐟ². LOD and LOQ were estimated as 3.3σ/S and 10σ/S, respectively, where S is the slope, following the general ICH Q2(R2) framework [14]. Repeatability was summarized as the sample standard deviation and relative standard deviation (RSD).

RESULTS AND DISCUSSION
Structural and morphological characterization
Pristine MWCNTs showed a major XRD reflection at 25.9° assigned to the graphitic (002) plane and a weaker reflection at 42.9° associated with the (100) plane. Independently prepared PtNPs exhibited reflections at 39.7°, 46.2°, and 67.6°, consistent with the (111), (200), and (220) planes of face-centered-cubic Pt. In the ternary material, the MWCNT-related reflection occurred at 25.7° and Pt-related reflections at 39.6°, 46.1°, and 67.5°. These results support the coexistence of the graphitic and metallic-Pt crystalline components and are consistent with reported crystalline Pt nanomaterials used in electrochemical sensing [8,15,16]. No crystalline Pt oxide peak was detected. G5-PAMAM is predominantly amorphous and therefore was not expected to produce a sharp diffraction peak.
FTIR provided complementary evidence of surface functionalization and retention of PAMAM-related groups [10,13].
Acid-functionalized MWCNTs showed a broad O-H feature near 3390 cm⁻¹ and a carbonyl band near 1729 cm⁻¹. PAMAM-containing materials retained N-H stretching at approximately 3280-3400 cm⁻¹, Amide I absorption near 1630-1650 cm⁻¹, and C-N features in the 1110-1200 cm⁻¹ region. Changes in band position and intensity are consistent with interfacial association, but they do not independently establish a specific Pt-N or Pt-O covalent bond.
FESEM showed preservation of an interconnected tubular MWCNT network together with a rougher nanogranular morphology after composite formation. TEM gave more direct evidence of Pt-containing nanoscale domains distributed within the f-MWCNT framework. Independently prepared PtNPs had a mean apparent diameter of 58.32 ± 18.67 nm (n = 51), whereas Pt-containing domains in the ternary composite had a mean apparent diameter of 41.45 ± 19.66 nm, median 38.52 nm, and range 15.83-108.30 nm (n = 21). The approximately 28.9% lower mean apparent diameter is consistent with interfacial stabilization by PAMAM and f-MWCNT functional groups [10,11,16], although localized aggregation remained visible.

Optimization of nanocomposite layer number
The three-layer coating produced the largest peak currents among the tested conditions: Ipa = 377.02 µA and |Ipc| = 327.43 µA. The corresponding values were 221.53/154.99 µA for one layer, 254.47/178.45 µA for two layers, and 349.01/293.20 µA for four layers. Relative to one layer, three layers increased Ipa by 70.2% and |Ipc| by 111.3%. A fourth layer reduced Ipa by 7.4% and |Ipc| by 10.5% compared with the three-layer coating. The decline is consistent with increased transport resistance in a thicker film, although the experiment does not independently identify the mechanism. Three layers were therefore used in subsequent measurements. 

Scan-rate behavior
Both anodic and cathodic peak-current magnitudes increased as the scan rate rose from 10 to 200 mV s⁻¹. Linear regression against the square root of scan rate gave Ipa (µA) = 78.05v¹ᐟ² + 0.32 (R² = 0.9946) and |Ipc| (µA) = 80.45v¹ᐟ² - 87.96 (R² = 0.9993), where v is expressed in mV s⁻¹. These relationships support a major diffusion-related contribution. Peak-to-peak separation increased with scan rate, however, indicating non-ideal reversibility. The data were therefore not used to calculate an electroactive area with the classical reversible Randles-Sevcik equation.

H₂O₂ transduction
H₂O₂ was examined before enzyme immobilization because it is the electroactive product of GOx-catalyzed glucose oxidation. The selected CV peak-current magnitude increased from 5 to 500 µM. The principal linear interval was 5-250 µM and followed |Ip| (µA) = 1.2263C (µM) + 118.905 with R² = 0.9983. The residual-regression estimates were LOD = 14.16 µM and LOQ = 42.91 µM. The 500 µM point deviated from the fitted trend, consistent with the onset of a non-linear response region. Because the 5 µM calibration point lies below the regression-derived LOD, it supports the fitted trend but should not be interpreted as a separately validated quantitative detection limit.
At -100 mV versus integrated Ag/AgCl, the steady-state current shifted systematically from -41.27 µA at 25 µM to -29.60 µA at 500 µM H₂O₂. The regression was I (µA) = 0.0240C - 41.364 with R² = 0.9875. The monotonic concentration-dependent response establishes the ability of the Pt-containing interface to transduce H₂O₂ under the tested conditions. This behavior is consistent with reports describing PtNP electrocatalysis and Pt/carbon interfaces for peroxide detection [8,9,15,16].

Enzymatic glucose sensing
After GOx immobilization, reverse-scan cathodic peak-current magnitudes were 12.80, 27.28, 47.30, 84.20, 113.79, 131.71, and 147.83 µA at 10, 20, 40, 60, 80, 100, and 200 µM glucose, respectively. The 10-100 µM interval was linear, while the small additional response at 200 µM indicated the onset of saturation or another transport/transduction limitation. Regression over 10-100 µM gave |Ip| (µA) = 1.3715C (µM) - 1.3457 with R² = 0.9916 and a non-area-normalized slope of 1.3715 µA µM⁻¹. Using σ = 4.9222 µA, the estimated LOD and LOQ were 11.84 and 35.89 µM, respectively. Blank-based limits from independent replicates would provide stronger validation of low-concentration performance.
The response is consistent with GOx-mediated conversion of glucose to D-glucono-δ-lactone and H₂O₂, followed by electrochemical transduction at the PtNP-containing interface. PAMAM supports nanoparticle organization and enzyme immobilization, f-MWCNTs provide conductive pathways, and PtNPs provide peroxide electrocatalysis. Recent GOx biosensors using Pt nanostructures, carbon nanotubes, screen-printed electrodes, and chitosan-based enzyme immobilization support this functional interpretation [4,6,7,17-19]. 
Earlier Pt/CNT and dendrimer-containing platforms established the usefulness of combining nanoscale platinum with conductive nanotube networks [20-23]. while other hybrid nanostructured GOx electrodes demonstrate the broader importance of high-area enzyme-supporting interfaces [23-25]. Direct numerical ranking remains inappropriate because technique, electrode area, operating potential, matrix, and the calculation of sensitivity and detection limit differ among studies.

Same-electrode repeatability and analytical scope
Nine valid consecutive CV measurements were analyzed on the same glucose biosensor. A tenth stored trace was excluded because it was an exact point-by-point duplicate of the fifth trace and therefore was not an independent observation. The nine cathodic peak-current magnitudes had a mean of 147.04 µA, sample SD of 7.32 µA, and RSD of 4.98%. This experiment demonstrates intra-electrode repeatability; it does not establish reproducibility across independently fabricated electrodes.
The present dataset supports proof-of-concept H₂O₂ transduction and enzymatic glucose sensing. It does not yet provide complete analytical validation for practical or clinical use. Inter-electrode reproducibility, quantitative selectivity at documented interferent levels, storage stability, pH optimization, and recovery in a representative sample matrix remain to be established. Claims are therefore limited to the measured buffer-system performance.

CONCLUSION
A G5-PAMAM/PtNPs/f-MWCNT ternary nanocomposite was integrated with a screen-printed carbon electrode and coupled to GOx for H₂O₂-mediated glucose sensing. Structural and morphological characterization supported coexistence of a graphitic nanotube framework, crystalline FCC Pt, and a chemically functional PAMAM-containing interface. Three nanocomposite layers produced the largest ferricyanide peak currents among the tested coatings. The modified electrode showed linear H₂O₂ CV behavior over 5-250 µM with R² = 0.9983. After GOx immobilization, the glucose response was linear over 10-100 µM with R² = 0.9916, and nine valid same-electrode measurements gave an RSD of 4.98%. These results establish the feasibility of the ternary material as a peroxide-transducing interface for enzymatic glucose detection. 

CONFLICT OF INTEREST
The authors declare that there is no conflict of interests regarding the publication of this manuscript.