DataPDF Available

Supporting Info for: Early detection and monitoring of chronic wounds using low-cost, omniphobic paper-based smart bandages

Supporting Information
Early detection and monitoring of chronic wounds using low-cost, omniphobic
paper-based smart bandages
Aniket Pal
, Debkalpa Goswami
, Hugo E. Cuellar
, Beatriz Castro
, Shihuan Kuang
and Ramses V. Martinez
School of Industrial Engineering, Purdue University, 315 N. Grant Street, West Lafayette, IN 47907,
Department of Animal Sciences, Purdue University, 270 S. Russell St, West Lafayette, IN 47907, USA
Weldon School of Biomedical Engineering, Purdue University, 206 S. Martin Jischke Drive, West
Lafayette, IN 47907, USA
* Corresponding author.
E-mail address:
Keywords: smart bandages, chronic wound monitoring, pressure ulcers, wearable sensor, pH sensor, uric
acid sensor.
1. Stencil printing of flexible conductive electrodes on omniphobic paper
We used AutoCAD 2016 (Autodesk Inc.) to design the electrodes embedded in the omniphobic
paper-based smart bandages (OPSBs). We fabricated the stencil masks used to print the electrodes by
cutting the designs into transparency sheets (VWO100C-BE, Apollo Inc.) using a commercial CO
cutter (MT3050D, Morntech Inc.). After cutting, the transparency sheets served as the stencil masks used
to print the electrodes on omniphobic Whatman #1 paper with conductive inks (AGCL-675 and C-200,
Applied Ink Solutions). Fig. S1 shows the dimensions of the different electrodes embedded in the OPSBs.
To pattern the electrodes used to measure uric acid (UA), the working and counter electrodes (WE and
CE) were first printed using carbon ink, followed by the reference electrode (RE) and contact pads using
Ag/AgCl ink. We overlapped the electrodes and contact pads to ensure good electrical conductivity.
Fig. S1. Schematics and pictures of the stencil-printed electrodes embedded in the OPSBs. (a) Carbon
and Ag/AgCl electrodes used for the electrochemical quantification of UA at the wound site.
(b) Electrodes used to monitor pH levels at the wound site. (c) Electrode array used for the early detection
and monitoring of pressure ulcers on an in-vivo mouse model.
2. Reproducibility of stencil printed electrodes
We measured the resistance of 25 sets of electrodes used to measure UA (Fig. S1a). The measured
resistances for each of the electrodes followed a normal distribution with low variability (coefficients of
variation for WE, RE and CE are 16%, 11%, and 12% respectively, Fig. S2). The WE and CE, printed
with carbon ink, have a higher resistance (mean of 120 Ω and 225 Ω, respectively). The RE, which was
printed with Ag/AgCl ink, has a much lower resistance (mean of 1.1 Ω).
Fig. S2. Reproducibility of stencil printed electrodes. The distribution of the measured resistances of
(a) working electrode, (b) reference electrode, and (c) counter electrode used to measure UA with
3. Gas Permeability of omniphobic paper-based sensors
We fabricated our sensors on Whatman #1 paper, which was rendered omniphobic by spraying a 2%
solution of R
in IPA. This solution makes the paper omniphobic by chemically altering the surface
of cellulose fibers without clogging the pores of the paper. The porosity of the omniphobic paper ensures
that the OPSBs remains breathable and do not impede the oxygenation of the wound. We demonstrate
the breathability of omniphobic paper by covering a vial of ammonium hydroxide (NH
OH, 35.50 g/mol,
Consolidated Chemical and Solvents LLC.) with a piece of the omniphobic paper and depositing a drop
of pH indicator (pH = 5, pH indicator A20, Water Company Inc.) on it. Fig. S3 and Movie S1 shows that
the color of the of the indicator changes due to the NH
gas flow of through the omniphobic paper.
Fig. S3. Snapshots from Movie S1 demonstrates the gas permeability of omniphobic paper. (a) We
placed a drop of a liquid pH indicator on top of omniphobic paper, and separated it from an open
scintillation vial containing NH
OH by a plastic film (gas barrier). (b) When the plastic film is removed,
in the gas phase crosses the porous structure of the omniphobic paper and reacts with the pH
indicator at time t = 0 s. (c, d) The color of the pH indicator gradually changes from yellow (pH = 5) to
green (pH = 8) in 20 s as the NH
gas permeates through the omniphobic paper.
4. Design of the housing for the wearable potentiostat
The components of the wearable potentiostat were enclosed in a foam-based housing to keep it light in
weight (~8 g, Fig. S4). A “smiley face” sticker was placed on top to increase patient compliance and
Fig. S4. (a) Foam-based housing for the wearable potentiostat. (b) Top sticker used to improve patient
compliance and adherence.
5. Diffusion controlled electrochemical reactions governed by the Cottrell equation
During chronoamperometry measurements, we applied a constant voltage (300 mV) and measured the
resulting current due cause by redox reactions between the electrochemical species. Diffusion controlled
reactions follow the Cottrell Equation (Eq. S1) (Hamedi et al., 2016), where the current i is proportional
to the initial concentration C
of the analyte:
 
Here n is the number of electrons, F is the Faraday constant (96,485 C/mol), A is the area of the electrode,
t is time, and D is the diffusion coefficient. Using the Cottrell equation, we can calibrate the concentration
of the analyte by measuring the current of the diffusion controlled electrochemical reaction. Integrating
the measured current with respect to time, we obtain the net charge exchanged (Q) during the redox
reaction, which also is proportional to the initial concentration. (Eq. S2).
  
;  
We used Q to calculate the concentration of UA using chronocoulometry.
6. Oxidation of uric acid in OPSBs
We used uricase in the OPSBs, so that it oxidises UA and produces allantoin (Eq. S3). Potassium
ferricyanide is reduced to potassium ferrocyanide to complete the redox couple (Pal et al., 2017).
  
//010  
 !"
$ (S3)
7. pH sensing mechanism of OPSBs
We fabricated omniphobic paper-based pH sensors using a pH sensitive polymer (polyaniline, PANi).
We mixed 150 mg of PANi and 250 mg of silver microflakes in 5 mL of IPA and pipetted 10 µ L of the
solution between the electrodes to fabricate the pH sensors. Polyaniline changes between two forms—
the blue colored emeraldine base (PANi-EB), and the green colored emeraldine salt (PANi-ES).
Fig. S5 shows how, in the presence of protonic acids, PANi-EB transforms to PANi-ES. The ES phase
of PANi is more conductive (~10 S/m) than the EB phase (~10
S/m) (Sukitpaneenit et al., 2007). This
difference in conductivity allows us to calibrate the impedance of the sensor to the pH of the wound
exudate model.
Fig. S5. Chemical structures of PANi-EB and PANi-ES and their reversible transformation by protonic
8. Wound exudate model used to measure pH using OPSBs
Table S1 shows the volumes in which we mixed 0.2 M disodium phosphate and 0.1 M citric acid to obtain
20 mL Mcllavine buffers of different pH values (Pearse, 1968). We used solutions with 5.5−8.5 pH
values—covering the clinical range of pH values in open wounds—to calibrate the pH sensors embedded
in the OPSBs.
Table S1. Mcllavine buffer compositions used to generate the 20-mL-solutions with pH values ranging
5.5 to 8.5.
pH 0.2M disodium phosphate
0.1M citric acid
5.5 11.5 8.5
6.0 12.6 7.4
6.5 14.5 5.5
7.0 16.5 3.5
7.5 18.0 2.0
8.0 19.5 0.5
8.5 20.0 0.0
9. Scanning Electron Microscopy
We used silver microflakes to fabricate a Ag/PANi-ES composite for the pH sensors. We took scanning
electron microscope (SEM) images of the microflakes (Fig. S6a, b) and the Ag/PANi-ES composite
(Fig. S6c, d) using an FEI Nova NanoSEM. The composites were prepared on Whatman #1 paper and
sputter coated with a ~10-nm-thick layer of Pt using a 20 mA filament current for 60 s. The unprocessed
silver microflakes were imaged without requiring any metallization (Fig. S6a). We used a 5.0 kV electron
accelerating potential for all images. A 4.8 mm working distance with an Everhart-Thornley Detector
(ETD) was used to image all the Ag/PANi composite samples. For the silver microflakes, we used a
5.7 mm working distance with a Through the Lens Detector (TLD).
Fig. S6. Scanning electron microscopy images of Ag/PANi layers. (a) Ag microflakes used to increase
the conductivity of PANi-EB. (b) High resolution SEM image of a representative microflake. (c) Image
of Ag/PANi-ES layer printed on Whatman #1 paper. (d) High resolution SEM image of the Ag/PANi-ES
10. In-vivo detection of pressure ulcers
Laboratory mice (C57B6J, 8 to 15 weeks old, male) were used in our experiments (Fig. S7a). Mice were
of mixed backgrounds and housed in the animal facility with free access to water and standard rodent
chow food. The mice were anaesthetized using ketamine-xylazine cocktail (0.1 g/kg of body weight). We
shaved the back of the mice and then used Nair depilatory cream to remove all the hair (Fig. S7b) and
cleaned the area with Dawn detergent. Fig. S7c shows how the shaved skin was gently tented up and
placed between two disc magnets (10 mm diameter, 2 mm thickness). The magnets did not interfere with
the normal activity of the mice once the anesthesia had worn off. The magnets were kept in place for 1 h
or 3 h to create different degrees of tissue damage. Each mouse received only one pressure induced
wound. The mice were always anesthetized prior to performing impedance spectroscopy. All procedures
involving mice were performed in accordance with Purdue University’s Animal Care and Use
Fig. S7. Mice preparation for the in-vivo early detection of pressure ulcers. (a) Mouse model used in the
experiments. (b) The hair covering the back of the mouse is shaved using a trimmer and depilatory cream.
(c) The shaved skin is pressed between two disc magnets to create an in-vivo pressure ulcers model.
11. Impedance spectroscopy using an AD5933 IC
The wearable potentiostat uses an integrated circuit chip, AD5933 (Analog Devices Inc.) to perform
impedance spectroscopy. The AD5933 chip outputs an AC excitation signal, which passes through the
tissue under the OPSB and is received by the chip. The impedance is calculated by taking the voltage to
current ratio. The excitation signal is created at a certain frequency by a 27-bit phase accumulator direct
digital synthesizer (DDS) core in the AD5933 chip. During the impedance spectroscopy, the frequency
was swept from 10 to 100 kHz with an AC voltage of 0.97 V with a DC bias of 0.76 V. The response
signal from the tissue passes through a gain amplifier and a low pass filter to increase the signal-to-noise
ratio. The signal is then sampled by the on-board analog-to-digital converter (ADC). Then a discrete
Fourier transform (DFT) is processed by an on-board digital signal processing (DSP) engine. Finally, the
DSP engine returns the real and imaginary components of the signal.
12. Fabrication costs
We estimate the cost of each OPSB to be less than $0.10 without considering labor costs or capital
expenses (Table S2, S3).
Table S2. Itemized cost of materials to fabricate an OPSB for monitoring open wounds.
Cost of OPSB for open wounds
UA sensor pH sensor
Ag/AgCl ink
C ink
Whatman #1 paper
Omniphobic solution
Commercial bandage
Total $0.07462
Table S3. Itemized cost of materials to fabricate an OPSB for monitoring pressure ulcers.
Cost of OPSB for pressure ulcers
Ag/AgCl ink
Whatman #1 paper
Omnphobic solution
Conductive hydrogel
Commercial bandage
Total $0.08910
The wearable potentiostat to work with the OPSB costs ~$18, the individual component costs are shown
in Table S4.
Table S4. Cost of components to fabricate the wearable potentiostat.
Cost of components of the wearable potentiostat
Electrochemical sensor (LMP91000)
3.6V coin battery (LIR2032)
Arduino Nano microcontroller
Impedance analyser (AD5933
Total $18.84
13. Disposability of the OPSBs
OPSBs can be easily disposed by incineration, generating minimal amounts of solid byproducts (Fig. S8).
Fig. S8. Demonstration of the disposal of an OPSB by incineration. All scale bars are 2 cm.
14. Arduino code to operate the wearable potentiostat
14.1. Arduino code to perform chronoamperometry using the LMP91000 chip
#include <Wire.h>
#include "LMP91000.h"
int NumberOfScans = 1;
int Amp350Time = 10; //in minutes
int Amp120Time = 4; //in minutes
int Amp35Time = 6; //in minutes
int InitialDelayTime = 10; //in milliseconds
int ScanNumber = 0;
LMP91000 lmp91000;
int NoOf350Samples = Amp350Time * 600; // 10 readings per second
int NoOf120Samples = Amp120Time * 600; // 10 readings per second
int NoOf35Samples = Amp35Time * 600; // 10 readings per second
void setup(void) {
Serial.println("LMP91000 Test");
void loop(void) {
LMP91000_TIA_GAIN_350K | LMP91000_RLOAD_10OHM,
if (millis() > InitialDelayTime) {
if (ScanNumber < NumberOfScans) {
LMP91000_TIA_GAIN_350K | LMP91000_RLOAD_10OHM,
Serial.print("Config Result: ");
for (int i = 0; i < NoOf350Samples; i++) {
LMP91000_TIA_GAIN_120K | LMP91000_RLOAD_10OHM,
Serial.print("Config Result: ");
for (int i = 0; i < NoOf120Samples; i++) {
LMP91000_TIA_GAIN_35K | LMP91000_RLOAD_10OHM,
Serial.print("Config Result: ");
for (int i = 0; i < NoOf35Samples; i++) {
ScanNumber = ScanNumber + 1;
Serial.print("Scan ");
Serial.print(" Complete");
14.2. Arduino code to perform impedance spectroscopy using the AD5933 chip
#include "Wire.h"
#define button 2
#define SLAVE_ADDR 0x0D
#define ADDR_PTR 0xB0
#define START_FREQ_R1 0x82
#define START_FREQ_R2 0x83
#define START_FREQ_R3 0x84
#define FREG_INCRE_R1 0x85
#define FREG_INCRE_R2 0x86
#define FREG_INCRE_R3 0x87
#define NUM_INCRE_R1 0x88
#define NUM_INCRE_R2 0x89
#define NUM_SCYCLES_R1 0x8A
#define NUM_SCYCLES_R2 0x8B
#define RE_DATA_R1 0x94
#define RE_DATA_R2 0x95
#define IMG_DATA_R1 0x96
#define IMG_DATA_R2 0x97
#define TEMP_R1 0x92
#define TEMP_R2 0x93
#define CTRL_REG 0x80
#define CTRL_REG2 0x81
#define STATUS_REG 0x8F
const float MCLK = 16.776 * pow(10, 6); // AD5933 Internal Clock Speed
16.776 MHz
const float start_freq = 60 * pow(10, 3); // Set start freq, < 100Khz
const float incre_freq = 1 * pow(10, 3); // Set freq increment
const int incre_num = 40; // Set number of increments; < 511
char state;
void setup() {
pinMode(button, INPUT);
writeData(CTRL_REG, 0x0); //clear ctrl-reg
writeData(CTRL_REG2, 0x10); //reset ctrl register
void loop() {
if (Serial.available() > 0) {
state =;
switch (state) {
case 'A': //Program Registers
case 'B': //Measure Temperature
case 'C':
void programReg() {
writeData(CTRL_REG, 0x01); // Set Range 1, PGA gain 1
writeData(NUM_SCYCLES_R1, 0x07); // Set settling cycles
writeData(NUM_SCYCLES_R2, 0xFF);
// Start frequency of 1kHz
writeData(START_FREQ_R1, getFrequency(start_freq, 1));
writeData(START_FREQ_R2, getFrequency(start_freq, 2));
writeData(START_FREQ_R3, getFrequency(start_freq, 3));
// Increment by 1 kH
writeData(FREG_INCRE_R1, getFrequency(incre_freq, 1));
writeData(FREG_INCRE_R2, getFrequency(incre_freq, 2));
writeData(FREG_INCRE_R3, getFrequency(incre_freq, 3));
// Points in frequency sweep (100), max 511
writeData(NUM_INCRE_R1, (incre_num & 0x001F00) >> 0x08 );
writeData(NUM_INCRE_R2, (incre_num & 0x0000FF));
void runSweep() {
short re;
short img;
double freq;
double mag;
double phase;
double gain;
double Impedance;
double GF;
double FFW;
double wt;
double BF;
double totmag = 0;
double count = 0;
double avgmag;
double totimp = 0;
double avgimp;
int i = 0;
double Resistance;
double Reactance;
// 1. Standby '10110000' Mask D8-10 of avoid tampering with gains
writeData(CTRL_REG, (readData(CTRL_REG) & 0x07) | 0xB0);
// 2. Initialize sweep
writeData(CTRL_REG, (readData(CTRL_REG) & 0x07) | 0x10);
// 3. Start sweep
writeData(CTRL_REG, (readData(CTRL_REG) & 0x07) | 0x20);
while ((readData(STATUS_REG) & 0x07) < 4 ) {
delay(100); // delay between measurements
int flag = readData(STATUS_REG) & 2;
if (flag == 2) {
byte R1 = readData(RE_DATA_R1);
byte R2 = readData(RE_DATA_R2);
re = (R1 << 8) | R2;
R1 = readData(IMG_DATA_R1);
R2 = readData(IMG_DATA_R2);
img = (R1 << 8) | R2;
freq = start_freq + i * incre_freq;
mag = sqrt(pow(double(re), 2) + pow(double(img), 2));
totmag = totmag + mag;
//Gain factor calibration
GF = 103.496 * (pow(10, -11)) + 811.151 * (pow(10, -18)) * (freq -
start_freq); //100kohm 40-100kHz
Impedance = 1 / (GF * mag);
totimp = totimp + Impedance;
count = count + 1;
Resistance = 1 / (GF * re);
Reactance = 1 / (GF * img);
phase = atan(double(img) / double(re));
phase = (180.0 / 3.1415926) * phase; //convert phase angle to
//Phase Calibration
System phase = 118;
phase = phase - 118;
Serial.print("Frequency: ");
Serial.print(freq / 1000);
Serial.print(" Magnitude: ");
Serial.print(" Impedance: ");
Serial.print(Impedance / 1000);
//Increment frequency
if ((readData(STATUS_REG) & 0x07) < 4 ) {
writeData(CTRL_REG, (readData(CTRL_REG) & 0x07) | 0x30);
avgmag = totmag / count;
avgimp = totimp / count;
Serial.print(" Avg Mag: ");
Serial.print(" Avg Impedance: ");
Serial.print(avgimp / 1000);
writeData(CTRL_REG, (readData(CTRL_REG) & 0x07) | 0xA0); //Power
down; writeData(CTRL_REG,0xA0);
void writeData(int addr, int data) {
int readData(int addr) {
int data;
Wire.requestFrom(SLAVE_ADDR, 1);
if (Wire.available() >= 1) {
data =;
else {
data = -1;
return data;
boolean measureTemperature() {
writeData(CTRL_REG, 0x90);
int flag = readData(STATUS_REG) & 1; //Check status reg for temp
measurement available
if (flag == 1) { // Temperature is available
int temperatureData = readData(TEMP_R1) << 8;
temperatureData |= readData(TEMP_R2);
temperatureData &= 0x3FFF; // remove first two bits
if (temperatureData & 0x2000 == 1) { // negative temperature
temperatureData -= 0x4000;
double val = double(temperatureData) / 32;
temperatureData /= 32;
Serial.print("Temperature: ");
writeData(CTRL_REG, 0xA0); // Power Down '10100000'
return true;
} else {
return false;
byte getFrequency(float freq, int n) {
long val = long((freq / (MCLK / 4)) * pow(2, 27));
byte code;
switch (n) {
case 1:
code = (val & 0xFF0000) >> 0x10;
case 2:
code = (val & 0x00FF00) >> 0x08;
case 3:
code = (val & 0x0000FF);
code = 0;
return code;
Hamedi, M.M., Ainla, A., Güder, F., Christodouleas, D.C., Fernández-Abedul, M.T., Whitesides, G.M.,
2016. Integrating Electronics and Microfluidics on Paper. Adv. Mater. 28, 5054–5063.
Pal, A., Cuellar, H.E., Kuang, R., Caurin, H.F.N., Goswami, D., Martinez, R. V., 2017. Self-Powered,
Paper-Based Electrochemical Devices for Sensitive Point-of-Care Testing. Adv. Mater. Technol.
Pearse, A.G., 1968. Histochemistry, theoretical and applied.
Sukitpaneenit, P., Thanpitcha, T., Sirivat, A., Weder, C., Rujiravanit, R., 2007. Electrical conductivity
and mechanical properties of polyaniline/natural rubber composite fibers. J. Appl. Polym. Sci.
106, 4038–4046.

File (1)

ResearchGate has not been able to resolve any citations for this publication.
Full-text available
This work describes the fabrication of self-powered, paper-based electrochemical devices (SPEDs) designed for sensitive diagnostics in low-resource settings and at the point of care. SPEDs are inexpensive, lightweight, mechanically flexible, easy to use, and disposable by burning. The top layer of the SPED is fabricated using cellulose paper with patterned hydrophobic domains that delineate hydrophilic, wicking-based microfluidic channels for accurate colorimetric assays, and self-pipetting test zones for electrochemical detection. The bottom layer of the SPED is a triboelectric generator (TEG) fabricated on hydrophobic paper and capable of harvesting electric energy from the user's interaction with the SPED. An inexpensive and rechargeable handheld potentiostat is fabricated to interface with the SPED, enabling the accurate quantitative electrochemical detection of glucose, uric acid, and l-lactate. The battery powering the potentiostat can be recharged by the user, using the sequential discharge of a capacitor previously charged with the TEG built into the SPED. A machine-vision diagnostic application is created to automatically identify and quantify each of the colorimetric tests from a digital image of the SPED, taken under a wide range of ambient light conditions, in order to provide fast diagnostic results to the user as well as to facilitate remote expert consultation.
Paper microfluidics and printed electronics have developed independently, and are incompatible in many aspects. This work demonstrates monolithic integration of microfluidics and electronics on paper. This integration makes it possible to print two- and 3D fluidic, electrofluidic, and electrical components on paper, and to fabricate devices using them.
Electrically conducting elastomer fibers based on natural rubber (NR) and up to 10% w/w polyaniline (PANI) in its emeraldine base (EB) form were fabricated by a wet spinning process. The resulting fibers at various PANI contents were doped by immersion in aqueous HCl solution, which converted the PANI to the electrically conductive emeraldine salt (ES) form. The morphology of the composite fibers was studied by scanning electron microscopy (SEM). PANI particles were inhomogeneously distributed in the NR matrix. The electrical conductivity of the fibers increased with the increasing PANI-ES content and leveled off at a value of around 10−3 S/cm at PANI-ES concentration of 5% w/w. The fibers retained most of their elasticity upon doping, while the tenacity was somewhat reduced. Gratifyingly, the electrical conductivity of the new elastomer fibers was preserved upon elongational deformation, even if strains as large as 600% were applied. © 2007 Wiley Periodicals, Inc. J Appl Polym Sci, 2007