Integrated Ink Printing Paper Based Self‐Powered Electrochemical Multimodal Biosensing (IFP−Multi) with ChatGPT–Bioelectronic Interface for Personalized Healthcare Management

Abstract Personalized healthcare management is an emerging field that requires the development of environment‐friendly, integrated, and electrochemical multimodal devices. In this study, the concept of integrated paper‐based biosensors (IFP−Multi) for personalized healthcare management is introduced. By leveraging ink printing technology and a ChatGPT–bioelectronic interface, these biosensors offer ultrahigh areal‐specific capacitance (74633 mF cm−2), excellent mechanical properties, and multifunctional sensing and humidity power generation capabilities. More importantly, the IFP−Multi devices have the potential to simulate deaf‐mute vocalization and can be integrated into wearable sensors to detect muscle contractions and bending motions. Moreover, they also enable monitoring of physiological signals from various body parts, such as the throat, nape, elbow, wrist, and knee, and successfully record sharp and repeatable signals generated by muscle contractions. In addition, the IFP−Multi devices demonstrate self‐powered handwriting sensing and moisture power generation for sweat‐sensing applications. As a proof‐of‐concept, a GPT 3.5 model‐based fine‐tuning and prediction pipeline that utilizes recorded physiological signals through IFP−Multi is showcased, enabling artificial intelligence with multimodal sensing capabilities for personalized healthcare management. This work presents a promising and ecofriendly approach to developing paper‐based electrochemical multimodal devices, paving the way for a new era of healthcare advancements.

Figure S2a shows the CV curves of ZnCl2/PVA-IFP at the scan rates of 1-100 mV s -1 .It can be seen that there is an approximately ideal rectangular shape, indicating that the ZnCl2/PVA-IFP has good double-layer capacitance characteristics, mainly due to the existence of carbonaceous materials.Furthermore, as can be seen from Figure S2d, two pairs of redox peaks corresponding to PANI appeared after the introduction of PANI on the surface of ZnCl2/PVA-IFP.However, in the CV comparison diagram as shown in Figure S2g, it is obvious that the absolute area of the CV curve of ZnCl2/PVA-IFP@PANI is larger than that of ZnCl2/PVA-IFP at the same scan rate of 50 mV s -1 .This means that ZnCl2/PVA-IFP@PANI has a relatively high specific capacitance.Moreover, according to the specific capacitance calculation formula, the areal specific capacitance of ZnCl2/PVA-IFP is 8.632 F cm -2 , while that of ZnCl2/PVA-IFP@PANI is 74.633F cm -2 , which is due to the good synergistic effect between PANI and ZnCl2/PVA-IFP.Furthermore, by calculation, the area specific energy densities of the two were 24 mwh cm -2 and 34.08 mwh cm -2 , respectively.These excellent electrochemical properties can be compared with some excellent work in the past, as displayed in Table S1.In addition, Figure S2b, e show the GCD curves of ZnCl2/PVA-IFP and ZnCl2/PVA-IFP@PANI materials at current densities of 3, 5 and 7 mA cm -2 , respectively.Obviously, the GCD curves in Figure S2e under different current densities are approximately linear and have good symmetry, which is attributed to the typical pseudo-capacitance contribution of PANI.Furthermore, the GCD curves of both are compared at a current density of 5 mA cm -2 as shown in Figure S2h, where ZnCl2/PVA-IFP@PANI has a longer discharge time at the same current density, indicating that the ZnCl2/PVA-IFP@PANI has a larger specific capacitance.This is consistent with the result of CV curve.In addition, as can be seen from the EIS impedance diagram, the arc diameter of ZnCl2/PVA-IFP@PANI is larger than that of ZnCl2/PVA-IFP, which is due to the enhanced contribution of pseudocapacitance brought by PANI, thus increasing the polarization and internal impedance of the material.Moreover, by testing the resistance of different parts of ZnCl2/PVA-IFP material, it was found that the resistance was basically maintained between 30 and 50 ohms, showing good stability (See the Figure S3).In addition,

Figure S7
The PCA analysis shows that the first six principal components can explain 98.04% of the variance in the dataset, while the first three components alone account for 95.7% of the total variance.

Figure
Figure S3a, b display the change of areal specific capacitance and capacitance

Figure
Figure S2 The electrochemical properties of integrated ZnCl2/PVA-IFP and ZnCl2/PVA-IFP@PANI supercapacitors were tested and compared.(a-c) and (d-f) are the CV curves (a and d), GCD curves (b and e) and EIS measurements (c and f) of ZnCl2/PVA-IFP and ZnCl2/PVA-IFP@PANI at different scan rates and current densities, respectively.(g-h) is the comparison of CV curve, GCD curve and EIS measurements at 50 mV s -1 and 5 mA cm -2 .

Figure
Figure S3 (a) The change of areal specific capacitance of ZnCl2/PVA-IFP and ZnCl2/PVA-IFP@PANI hybrid at different scan rates.(b) The capacitance retention of ZnCl2/PVA-IFP and ZnCl2/PVA-IFP@PANI hybrid at a scan rate of 50 mV s -1 after experiencing 5000 cycles.

Figure
Figure S4 (a) A tensile test was conducted to evaluate the adaptability of the ZnCl2/PVA-FP composites to external forces.(b) Based on the excellent tensile properties of the flexible IFP -Multi material, a simple electrical test was conducted by bending the IFP -Multi from 0 to 180°, which was stable and sensitive to continuous changes in action.

Figure
Figure S5 Current (Voltage): (a) Description of moisture generate current (voltage) process of ZnCl2/PVA-IFP, (b) The relationship of ZnCl2/PVA-IFP between moisture component and generating capacity.

Figure S6
Figure S6 Signal segmentation by findpeaks function of matlab.

Figure
Figure S8 29 'Hi' signals and 17 simulated 'Hi' signals were introduced, along with 20 sets of random noise.

Figure S9
Figure S9ChatGPT conversation interface for model fine tuning and predicting.

Table S1
Comparison maps of areal specific capacitance and specific energy density with others' work.