Article

pubs.acs.org/acssensors

Noninvasive Alcohol Monitoring Using a Wearable Tattoo-Based

Iontophoretic-Biosensing System
Jayoung Kim,†,∥ Itthipon Jeerapan,†,∥ Somayeh Imani,‡,∥ Thomas N. Cho,† Amay Bandodkar,†
Stefano Cinti,† Patrick P. Mercier,*,‡ and Joseph Wang*,†
†
Department of Nanoengineering and ‡Department of Electrical & Computer Engineering, University of California, San Diego, La
Jolla, California 92093, United States
*
S Supporting Information
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

ABSTRACT: In this paper we demonstrate a wearable tattoo-

based alcohol biosensing system for noninvasive alcohol
monitoring in induced sweat. The skin-worn alcohol
Downloaded via UNIV OF OKLAHOMA on August 21, 2024 at 18:30:51 (UTC).

monitoring platform integrates an iontophoretic-biosensing

temporary tattoo system along with flexible wireless
electronics. The wearable prototype enables the transdermal
delivery of the pilocarpine drug to induce sweat via
iontophoresis and amperometric detection of ethanol in the
generated sweat using the alcohol-oxidase enzyme and the
Prussian Blue electrode transducer. The new skin-compliant
biosensor displays a highly selective and sensitive response to ethanol. On-body results with human subjects show distinct
differences in the current response before and after alcohol consumption, reflecting the increase of ethanol levels. The skin-worn
alcohol sensor is coupled with a flexible electronics board, which controls the iontophoresis/amperometry operation and
transmits data wirelessly in real time via Bluetooth communication. The new wireless epidermal iontophoretic-biosensing system
offers considerable promise for noninvasive monitoring of alcohol consumption in practical settings and can be readily expanded
toward the monitoring of additional analytes.
KEYWORDS: sweat alcohol, tattoo sensor, wearable, wireless electronics, iontophoresis

U nsafe levels of alcohol consumption can lead to vehicle

crashes, violence, and the degenerated health of heavy
drinkers.1,2 Such alcohol-related incidents and health concerns
alcohol concentration (BrAC). BrAC instruments calculate
BAC by following Henry’s law, but may compromise accuracy
since the results can be easily affected by humidity, temper-
continue to rise rapidly across the globe, leading to ature, and individual traits.7,8 BrAC can also generate false
considerable socioeconomic costs. Accordingly, there are alarms due to alcohol vapor associated with consumer products
tremendous needs for an accurate easy-to-use alcohol (e.g., mouthwash and breath freshener), and environmental
measuring device for use by law enforcement personnel, factors (e.g., paint fume, varnish, and chemicals).9 BAC can also
service/hospitality industry, or individual drinkers to provide a be estimated by measuring transdermal alcohol concentration
convenient means to monitor alcohol consumption. Although (TAC), because a person’s perspiration can contain traces of
different methods have been used for determining alcohol alcohol after alcohol consumption.10 Two wearable transdermal
consumption, including direct measurements of urine, blood, alcohol sensor devices (SCRAM, Giner WrisTAS) have been
saliva, breath, or sweat, or verbal/interview measures, Blood developed for detecting local ethanol vapor concentration in
Alcohol Concentration (BAC) is the most commonly used insensible perspiration over the skin, which is not secreted by
indicator of alcohol intoxication.3−6 Unfortunately, blood sweat glands.11,12 However, these devices suffer from long time
samples cannot currently be obtained without penetrating delays (0.5−2 h) compared to BAC estimated by breath-
skin, typically via a lancet pricking a finger or earlobe, which can alyzers.11 Therefore, the development of a new prototype to
be painful and inconvenient, and demands user compliance. monitor BAC noninvasively in real time is highly desired.
Such blood sampling methods limit the general use of alcohol Recent work13 has combined iontophoresis and biosensing
monitoring devices to the most extreme cases (e.g., law processes for measuring alcohol level in induced sweat (sensible
enforcement), rather than as a general-purpose tool that can sweat) based on the established correlation between the two
monitor a user’s alcohol levels and warn or prevent what could fluids during alcohol consumption.14,15 Such an operation
amount to a catastrophe. Accordingly, there are considerable showed that the maximum ethanol level in both fluids, blood
demands for developing an alternative approach for measuring
BAC indirectly in a noninvasive and real-time manner. Received: May 29, 2016
Currently, breathalyzers are the most commonly used devices Accepted: July 12, 2016
to indirectly estimate BAC, and operate by measuring breath Published: July 12, 2016

© 2016 American Chemical Society 1011 DOI: 10.1021/acssensors.6b00356

ACS Sens. 2016, 1, 1011−1019
ACS Sensors Article

Figure 1. Tattoo-based transdermal alcohol sensor. (A) Schematic diagram of an iontophoretic-sensing tattoo device, containing the iontophoretic
electrodes (IEs; anode and cathode) and the three sensing electrodes (working, reference, and counter electrodes: WE, RE, and CE, respectively).
(B) Photograph of an alcohol iontophoretic-sensing tattoo device with integrated flexible electronics applied to a human subject. (C) Schematic
diagram of a wireless operation of the iontophoretic-sensing tattoo device for transdermal alcohol sensing. In the diagrams of the tattoo-base device,
blue and red highlights show the active zones during iontophoresis and amperometric detection, respectively. (D) Schematic diagram of constituents
in the iontophoretic system (left) and of the reagent layer and processes involved in the amperometric sensing of ethanol on the working electrode
(right).

and collected sweat induced by pilocarpine iontophoresis, screen-printed electrode to monitor chloride ion in sweat for
reached at nearly the same time, compared to the time delays of early determination of cystic fibrosis.29
common TAC sweat sensors using insensible sweat (Giner In the following sections we describe an integrated tattoo-
WrisTAS). However, the electrochemical sensing system was based wearable system for effective noninvasive ethanol
combined with a bulky nonwearable iontophoretic system, and monitoring based on coupling sweat-inducing iontophoresis
required replacement of electrodes between the iontophoresis and amperometric enzymatic biosensing (Figure 1A), along
and amperometric detection steps. with a flexible supporting electronic readout module featuring
Here we demonstrate a wearable temporary-tattoo biosens- wireless telemetry (Figure 1B). The new skin-worn low-cost
ing system capable of real-time noninvasive alcohol monitoring noninvasive alcohol monitoring device enables real-time alcohol
via integration of printed and flexible iontophoretic-sensing measurements in induced sweat and obviates the need for
electrodes with wireless electronics (Figure 1). Our study is lengthy and costly procedures. This represents the first example
built upon recently developed noninvasive wearable chemical of integrating a drug-loaded iontophoretic operation and
sensors for monitoring chemical markers in sweat toward health electrochemical amperometric biosensing of metabolite on a
or fitness applications.16−22 Our laboratory has developed wearable tattoo platform, in general, and for tattoo-based
body-compliant tattoo-based epidermal electrochemical sensors flexible epidermal alcohol sensor system, in particular.
for noninvasive monitoring of sweat electrolytes, metabolites, The new wearable epidermal alcohol sensor system uses
or trace metals.23−28 Recently, we reported a tattoo-based constant-current iontophoresis for inducing sweat by delivering
noninvasive glucose monitoring system that combines reverse- the drug pilocarpine across the skin, followed by amperometric
iontophoresis with an amperometric enzyme electrode.25 biosensing of the sweat ethanol (Figure 1C). The latter relies
Gonzalo-Ruiz et al.29 also demonstrated the coupling of a on an alcohol-oxidase (AOx) enzymatic electrode along with a
potentiometric ion selective electrode with iontophoresis on a printed Prussian Blue (PB) electrode transducer. Such
1012 DOI: 10.1021/acssensors.6b00356
ACS Sens. 2016, 1, 1011−1019
ACS Sensors Article

integration of the iontophoretic and amperometric operations solution) were mixed together in a ratio of 8:1:1 (v/v/v). Afterward, a
onto a single flexible tattoo platform obviates the need for 4 μL droplet of the mixed solution was casted on the electrode. After
replacing the electrodes between iontophoresis and detection air-drying, the electrode was covered with a 2 μL droplet of chitosan
steps as in prior work.13 This coupling has been accomplished solution (0.5 wt %). Then, the electrode was dried under ambient
conditions. The agarose hydrogel was prepared by heating a
by adding a pair of conductive Ag/AgCl iontophoretic continuously stirred agarose solution (2% w/v) with 0.1 M potassium
electrodes to the three-electrode amperometric sensing system phosphate buffer (pH 7.0) until its complete dissolution. The
using a specific design pattern essential for integrating the two dissolved agarose hydrogel was then casted on the working enzyme
systems (Figure 1A). All the electrodes were fabricated by a electrode. Subsequently, the prepared cryogels (2.0 × 1.5 cm2) were
screen-printing technique on the wearable temporary-tattoo soaked in 1% pilocarpine nitrate and 1% sodium nitrate (for 1 h) and
paper to offer cost-effective mass production along with then covered on the anode and cathode compartments, respectively.
convenient placement and removal from the skin. For The cryogels were made as following procedure by freezing and
realization of real-time alcohol monitoring with the wearable thawing sequences of PVA solution. First, a 5.0% w/v PVA solution
sensor, the tattoo alcohol sensor was integrated with flexible was prepared in deionized water by heating the solution to 120 °C and
then cooled down to room temperature. After the mixture was cooled
printed electronic circuitry (Figure 1B), which controls the
down, it was placed in the ice bath and the pH was adjusted to pH 1.0
entire iontophoretic-sensing operation wirelessly and transmits by adding 5 M hydrochloric acid. Subsequently, the cross-linker
the data to laptop or mobile devices via Bluetooth glutaraldehyde was added to give a final concentration 0.5% w/v. The
communication (Figure 1C). Both the skin-worn sensor and final mixture was then stirred for 1 min and poured onto the glass
electronic board are flexible and compatible with the non- mold to set in −20 °C freezer overnight.
planarity of the epidermis and offer resistance to mechanical Evaluation of Sensor Performance in Buffer Medium. The
stress from the wearer’s movement. electrochemical performance of the alcohol tattoo biosensor was
The performance of the developed sensor was evaluated first evaluated first in vitro in a 0.1 M phosphate buffer (pH 7.0) solution
in a buffer medium for its sensitivity, selectivity, and potential (PBS). The chronoamperometric response was measured, after 1 min
cross talk between the iontophoretic and sensing operations. immersion in the test solution, by stepping the potential to −0.2 V (vs
Ag/AgCl) for 60 s. Calibration plots were obtained using 3 mM
Subsequently, the on-body iontophoresis operation was increments of the ethanol concentration, up to 36 mM in the buffer
optimized on the human skin and the epidermal sensor was solution. The selectivity was examined by measuring the response to
tested with human subjects under ingestion of alcoholic 10 mM ethanol in the presence of relevant electroactive species: 0.2
beverages. The on-body results (including corresponding mM glucose, 10 mM lactate, 84 μM creatine, 10 μM ascorbic acid, and
control experiments) indicate that the new wearable epidermal 60 μM uric acid.30 The influence of pilocarpine upon the activity of the
alcohol sensor system holds considerable promise for reliable PB transducer was examined using cyclic voltammetry (CV) of the
decentralized noninvasive alcohol monitoring in diverse bare PB electrode; the CV was recorded in PBS solution containing
practical settings toward saving lives and supporting the 1% pilocarpine solution (dissolved in PBS). To address the pilocarpine
criminal justice system. interference, the working enzyme electrode was modified by drop-

■
casting 2% agarose gel containing PBS. The recovered PB activity with
agarose gel modification was confirmed by CV in 1% pilocarpine
EXPERIMENTAL SECTION solution (dissolved in DI water) in comparison with a bare PB
Chemicals and Instruments. Alcohol oxidase (AOx, from Pichia electrode (without the agarose gel modification).
pastoris, 10−40 units/mg protein), chitosan, bovine serum albumin The resilience of the tattoo biosensor to mechanical strain was
(BSA), potassium phosphate monobasic (K 2PO4), potassium examined by measuring the ethanol response after repeated 90°
phosphate dibasic (K2HPO4), ethanol, acetic acid, pilocarpine nitrate, bending. The tattoo sensors were transferred to transparent plastic
sodium nitrate, L(+)-ascorbic acid, uric acid, agarose type IV, and substrate to mimic the flexible properties of the skin. The
poly(vinyl alcohol) (PVA, average molecular weight = 70 000− amperometric response was recorded every 20 times of repeated
100 000) were purchased from Sigma-Aldrich (St. Louis, MO). All bending up to 120 times.
reagents were used without further purification. A μAutolab type III On-Body Evaluation of a Tattoo Alcohol Biosensor. The
PGSTAT302N (Metrohm), controlled by Autolab NOVA software v epidermal evaluation on human subjects was conducted in strict
1.11.2, was applied for the electrochemical characterization in buffer compliance following a protocol approved by the institutional review
medium and in on-body evaluation. board (IRB) at the University of California, San Diego. A total of nine
Fabrication, Chemical Modification, and Transfer Process of healthy volunteers were recruited for on-body evaluation of the
a Tattoo Sensor. Patterns for printing the sensor were designed in developed sensor before and after consumption of alcoholic beverages.
AutoCAD (Autodesk, San Rafael, CA) and transferred to stainless In all experiments, the tattoo biosensors were placed on the subjects’
steel plates (12 × 12 in.2) etched to fabricate stencils (Metal Etch arms, and ethanol sweat measurement was performed using the
Services, San Marcos, CA). Temporary transfer tattoo paper kits were iontophoresis-amperometry operation to obtain the current response
obtained from HPS Papilio (Rhome, TX). The silver/silver chloride at BAC 0.000% (“before drinking”). During the iontophoresis process,
(Ag/AgCl) ink (E2141 Ercon Inc., Wareham, MA) and PB conductive a constant current of 0.6 mA (0.2 mA cm−2) was applied for 5 min
carbon (C2070424P2, Gwent Group, UK) were screen-printed on the through a cryogel layer between the two iontophoresis (i.e., anode and
substrate by using an MPM-SPM semiautomatic screen-printer cathode) electrodes to deliver pilocarpine and induce sweat. The
(Speedline Technologies, Franklin, MA). As illustrated in Figure 1A, applied current density and duration were optimized and selected
the tattoo sensor design consists of iontophoresis and pseudo using on-body test as a trade-off between efficient sweat generation
reference electrodes patterned from Ag/AgCl ink, and the working and the subject’s compliant. For example, high currents that result in
and counter electrodes patterned using the PB conductive carbon ink. efficient pilocarpine drug delivery, also lead to skin burning/irritation.
The diameter of the printed working electrode was 3 mm. A The applied time (5 min) was also selected based on the on-body
transparent insulator was screen-printed over the surface of the results; 5 min is minimum duration to induce sweat and detect the
electrode pattern to confine the electrodes and contact areas. The Ag/ alcohol sweat current signal. This was evaluated by multiple subjects (n
AgCl ink was cured at 90 °C for 10 min, while the PB conductive = 9), since skin permeability varies among individuals. It was followed
carbon ink was cured at 80 °C for 10 min in a convection oven. by a 5 min rest period during which the sweat was generated.
In order to prepare the alcohol biosensor, the printed working Subsequently, the amperometric response of ethanol in sweat (BAC
electrode was functionalized with an enzymatic layer. The AOx 0.000%) was recorded at an applied potential of −0.2 V (vs Ag/AgCl)
enzyme, BSA stabilizer, and chitosan solution (0.5 wt % in acetic acid for 150 s with a benchtop potentiostat. The subject then consumed an

1013 DOI: 10.1021/acssensors.6b00356

ACS Sens. 2016, 1, 1011−1019
ACS Sensors Article

alcoholic beverage (12 oz. of beer or 5 oz. of table wine) for 5 min and graphical interface. The integration of the wearable tattoo sensor with
waited for 10 min to allow alcohol diffusion in blood. Upon wireless electronics and its operation are shown in the Supporting
consumption, the alcohol passes through the stomach and gastro- Information.

■
intestinal (GI) tract into the bloodstream. Afterward, it diffuses to the
surrounding body tissues, including the skin. The iontophoresis/
detection cycle was then repeated to measure the corresponding
RESULTS AND DISCUSSION
ethanol sweat concentration. Rationale for an Iontophoretic-Biosensing System of
Along with the above on-body iontophoresis-sensing experiments a Tattoo-Based Alcohol Sensor. A novel aspect of the new
involving alcohol consumption, three different control experiments wearable alcohol biosensor is the integration of pilocarpine-
(without drinking, without enzyme modification, and without
ionotphoretic and amperometric detection systems and their
iontophoresis) were performed. The sensor response toward sweat
ethanol was confirmed first in comparison with measurements without coupling onto a single flexible tattoo-based device. Combining
drinking alcoholic beverages. The selectivity of the sensor toward both systems and operations into a single skin-worn platform
ethanol was evaluated also by comparison with the response of the requires a specific electrode design. The alcohol tattoo sensor
enzyme-free electrodes. The effect of iontophoresis was verified by system thus consisted of the iontophoretic electrodes (anode
control experiments without the iontophoresis step. These three and cathode) and the three amperometric sensing electrodes
different control experiments were carried out under otherwise (WE, RE, and CE) located in the anode compartment (Figure
identical conditions, ensuring that the observed response is solely 1A). The pair of iontophoretic electrodes is responsible for
due to increased alcohol level following drinking. Each set of
inducing sweat by delivering the pilocarpine drug from the
measurement cycles was accompanied by simultaneous BAC measure-
ments using a commercial FDA-approved breathalyzer (Alcovisor anode compartment. These sweat-generating electrodes are
Mars Breathalyzer, Hong Kong) to validate the sensor performance. thus positioned in the middle of sensing electrodes. Such a
Several additional on-body tests involved estimating the correlation specific electrode layout is essential for an optimal integration
between BAC and the current response of the tattoo sensor with of iontophoretic and detection systems and its skin-worn
repeated consumption of wine. These involved the same iontopho- operation. The screen-printed three-electrode detection system
resis-amperometry experimental procedure using two sets of repeated consisted of the PB-based working electrode, the pseudo Ag/
drinking and measurement. The current response was measured AgCl reference electrode, and the PB conductive carbon
initially before drinking (when BAC is 0.000%) using the counter electrode. Adding the third counter electrode to our
iontophoresis-sensing operation. Then, the current response of such
operation was recorded following the first and second drinking. earlier two-electrode tattoo biosensor design25 offers improved
Flexible Wireless Instrumentation Electronics. The flexible potentiostatic control. In addition, in order to achieve the fully
tattoo-based iontophoretic alcohol monitoring patch must connect to integrated system, with simultaneous iontophoresis, enzymatic
instrumentation electronics to control the iontophoretic and reaction, and amperometric detection, the position of the
amperometric electrodes and readout data. While initial testing and iontophoresis electrode and its distance from sensing electrodes
characterization of the biosensor system have been carried out using are crucial factors considering the small volume of generated
benchtop equipment (CH Instruments), such a bulky electrochemical sweat. The iontophoresis protocol was optimized in preliminary
analyzer limits the system-level attractiveness of having a thin experiments for effective sweat generation avoiding skin
temporary-tattoo design. Accordingly, the tattoo sensor was connected
directly to a flexible printed circuit board (PCB) containing irritation. Cryogel layers covered all the iontophoretic electro-
commercial off-the-shelf (COTS) integrated circuits for instrumenta- des for ensuring proper contact with the skin, which was soaked
tion, control, and telemetry in a thin, wearable form-factor. Specifically, with 1% pilocarpine nitrate and sodium nitrate on the anode
the PCB employs a Texas Instrument (TI) CC2541 Bluetooth Low and cathode, respectively, to deliver these species across the
Energy (BLE) System-on-Chip for communication and processing. skin by applying a mild electrical current. Cryogel has been
Iontophoretic current injection is achieved via a TI LM334 current selected as the drug sorbent based on our preliminary
source, which applies a 0.6 mA current between the iontophoretic experiments (compared with filter paper and agarose gel),
cathode and anode electrodes. After 5 min of current injection, the because it offers a porous sponge structure, elasticity, and
LM334 was disconnected by control signals from the CC2541.
Following an additional 5 min, a TI LMP91000 chemical sensing
biocompatibility. These properties were obtained from ice
analog front end was enabled, and applied a −0.2 V potential across porogen of PVA gel (i.e., ice crystals acting as highly porous
the amperometric electrodes for constant-potential amperometry molds under subzero temperature; they are melted away after
experiments. The potential was held for up to 60 s, or until the the gel has set) which allows high loading of pilocarpine and
resulting amperometric current, also read by the LMP91000, stabilized. good permeation of sweat across the gel.31 During the
The resulting current throughout the 60 s experiment was sampled iontophoresis operation, the cationic pilocarpine drug was
and digitized by an on-board analog-to-digital converter (ADC) on the delivered across the skin at the anode electrode (Figure 1D).32
CC2541, and was transmitted via a Johanson Technology 2.45 GHz As expected, sweat was generated at the pilocarpine-loaded
chip antenna (2450AT42A100) and impedance matched balun
(2450BM15A0002) in a 2-byte format to a Bluetooth 4.0-enabled
iontophoresis electrode a short time after delivering the drug,
receiver. The receiver decoded the data and presented the results in a causing no apparent skin irritation and burning.
Python-based graphical interface on a desktop or laptop. The flexible The amperometric biosensing of alcohol in the generated
PCB was powered by two 396/397 watch batteries (2 × 1.55 V, 33 sweat relied on an AOx-coated screen-printed PB transducer
mAh each) mounted in series and conditioned via a TPS61220 boost that offers specific low-potential electrocatalytic detection of the
converter and an LM4120 low-dropout voltage regulator. The PCB hydrogen peroxide product of the AOx enzymatic reaction.33
prototype, shown in Figure 1B, measured 2 cm × 5 cm. The AOx was immobilized on the working electrode by drop-
A stainless steel sheet was cut in 2 mm × 10 mm size and attached casting with BSA and chitosan, and then covered with agarose
on connection point of electrodes using conductive silver epoxy
adhesive for detachable magnetic connection with magnets soldered
gel containing PBS (K+) to provide consistent and stable
on flexible board. The sensor-board assembly is shown in Figure 1B; electrolyte level essential for electron-shuttling activity of PB
on-body evaluation proceeded following the same protocol described (Figure 1D) and the electrochemical detection. The resulting
in the previous section. The resulting current response was transmitted effect of the agarose gel in the iontophoretic-sensing system is
to a laptop via a Bluetooth 4.0 and was displayed using a custom-made discussed and confirmed experimentally in following sections.
1014 DOI: 10.1021/acssensors.6b00356
ACS Sens. 2016, 1, 1011−1019
ACS Sensors Article

Figure 2. (A) Amperometric response of the tattoo-based alcohol sensor to increasing ethanol concentrations from 0 mM (a) to 36 mM (m) in pH
7 buffer solution with 3 mM increments. The inset shows the corresponding calibration plot. Error bars represent the standard deviations (n = 3).
Potential step to −0.2 V (vs Ag/AgCl). (B) Interference study in the presence of 10 mM ethanol (a), followed by additions of 0.2 mM glucose (b),
10 mM lactate (c), 84 μM creatine (d), 10 μM ascorbic acid (e), and 60 μM uric acid (e); potential step to −0.2 V (vs Ag/AgCl); medium, pH 7
phosphate buffer solution. (C) Mechanical deformation tests of the alcohol tattoo sensor involving 90° bending. Relative amperometric current
response to 10 mM ethanol during a series of 120 deformations, recorded after 20 such repeated bending tests in pH 7 buffer solution. Photograph
of the 90° bending of the transferred tattoo electrode on a transparent plastic substrate. Potential step to −0.2 V (vs Ag/AgCl). (D) Cyclic
voltammograms of the tattoo-based alcohol sensor in pH 7 buffer solution (a) and 1% pilocarpine dissolved pH 7 buffer solution (b). (E) Cyclic
voltammograms of the tattoo-based bare Prussian Blue (PB) electrode (a) and at the PB electrode modified with 2% agarose gel (b) recorded in an
in 1% pilocarpine solution (dissolved in DI water).

Characterization of an Alcohol Tattoo Sensor in electrodes against mechanical deformations expected in

Buffer Medium. The electrochemical performance of the practical wearing situations.
alcohol biosensor was validated first in buffer medium over a The new printable electrode system represents the first
wide dynamic concentration range of 0−36 mM ethanol, attempt to combine iontophoresis and detection operations on
corresponding to the physiological level of ethanol in sweat. a single skin-worn tattoo platform. Therefore, potential cross
Figure 2A displays well-defined chronoamperometric response talk between the iontophoresis and detection electrodes should
to 12 successive increments of 3 mM ethanol in a potassium be critically assessed before integration of the two systems.
phosphate buffer solution (b-m). These additions result in well- Accordingly, we examined the effect of pilocarpine upon the PB
defined amperometric response over the entire ethanol range, activity (Figure 2D). Pilocarpine is a large positively charged
with convenient quantitation (stable current) observed ∼30 s molecule, and the electron shuttling activity of PB can be
after the potential step. The corresponding calibration plot, influenced by pilocarpine due to the vulnerability of PB against
shown in the inset of Figure 2A, is highly linear over the entire other positive ions in electrolytes.32 Pilocarpine can thus
range examined (slope, 0.362 ± 0.009 μA/mM; intercept, 1.810 decrease the electron shuttling ability of PB, as indicated from
± 0.170 μA; correlation coefficient, R2 = 0.993; n = 3). the CVs of Figure 2D. These data illustrate 39.7% diminished
Since human sweat contains various physiologically relevant redox peak currents of PB in phosphate buffer medium in the
interferences (e.g., glucose, uric acid, lactate, ascorbic acid, and presence of pilocarpine, along with slight shifts in the peak
creatine), it is important to examine the selectivity of the new potentials (a vs b).
biosensor toward the target ethanol. Figure 2B evaluates the In order to prevent interference of pilocarpine upon the
influence of several potential interferences upon the response electron shuttling activity of PB, the enzyme-coated PB working
to 10 mM ethanol. These data demonstrate that the various electrode was covered with a 2% agarose gel. The agarose gel
coexisting compounds have a negligible effect upon the ethanol layer containing buffer electrolytes provides sufficient potas-
response (a vs b-f). Such high selectivity reflects coupling of the sium ions to address and “buffer” the pilocarpine effect on the
specific enzymatic reaction with the extremely low detection PB electrode and fully restore its electron shuttling ability.
potential of the peroxide product at the PB transducer. Figure 2E illustrates the recovered CV redox peaks in 1%
Wearable applications of the tattoo alcohol biosensor require pilocarpine solution (dissolved in DI water) following
resilience against mechanical stress. The endurance of tattoo modification of the enzyme-modified PB working electrode
alcohol sensor against mechanical deformations was evaluated with a 2% agarose gel. While no PB peaks are observed at the
by measuring its performance under repeated 90° bending. The unmodified electrode immersed in the 1% pilocarpine solution
tattoo-based biosensor underwent 120 times of repeated (dissolved in DI water) due to the absence of electrolytes (a),
bending and their amperometric response was measured after the PB redox signals can be observed at the electrolyte-
every 20 times of bending. As illustrated in Figure 2C, the containing agarose modified electrode (b). It indicates that the
current response to 10 mM ethanol was retained after such agarose gel layer eliminates the pilocarpine interference by
repeated bending, indicating the robustness of the tattoo impeding diffusion of pilocarpine to working electrode and
1015 DOI: 10.1021/acssensors.6b00356
ACS Sens. 2016, 1, 1011−1019
ACS Sensors Article

Figure 3. (A) Schematic procedures of on-body study of alcohol sensing. Chronoamperograms obtained for noninvasive alcohol detection obtained
from human subjects wearing the alcohol tattoo sensors and operated by the benchtop potentiostat; (B) experiments with consumption of 12 oz of
beer measured from three different human subjects before (plot “a”) and after drinking alcohol beverage (plot “b”); (C) control experiments without
drinking (left: first measurement, plot “a”; second measurement, plot “b”), without enzyme immobilization (middle: before drinking, plot “a”; after
drinking, plot “b”), and without iontophoresis (IP) (right: before drinking, plot “a”; after drinking, plot “b”); (D) experiments to demonstrate
correlation between BAC level and current response from the tattoo biosensor measured. The left plot shows chronoamperograms obtained before
drinking (a, BAC: 0%), after drinking 5 oz of wine (b, BAC: 0.025%) and 10 oz of wine (c, BAC: 0.062%). The right plot shows the correlation
between current response obtained from the tattoo sensor and BAC level. The BAC level was obtained by breathalyzer. Potential step to −0.2 V (vs
Ag/AgCl).

providing additional electrolytes to facilitate the redox function alcohol sweat concentration can be estimated from BAC
of PB. through the following equation:13
On-Body Alcohol Monitoring on Human Subjects.
After evaluation of the tattoo sensor performance in vitro, we BAC (g L−1) = 0.71 × sweat ethanol concentration (g L−1)
tested the on-body operation of the skin-worn alcohol tattoo r = 0.912
sensor with human subjects. First, the detection ability of the
tattoo alcohol biosensor was confirmed under consumption of The BAC of the above subjects can thus be converted to 6.7
alcoholic beverage. Figure 3B displays on-body amperometric mM, 5.2 mM, and 7.9 mM alcohol sweat concentrations,
results obtained by three different subjects before and after respectively. Note also the different current signals for subjects
alcohol consumption, using the protocol described in the with the same BAC values that reflect their different skin
Experimental Section. A distinct increase of the current permeabilities and sweat compositions. None of the subjects
response is observed by comparing the signals from these reported perceptible discomfort during these on-body measure-
ments. Three control experiments (no drinking, no enzyme
three subjects. A second amperometric response (after
modification, and no iontophoresis) were performed to verify
drinking) was measured 20 min after alcohol consumption, the absence of potential false alarms from the on-body results
along with a parallel BAC measurement. Even though three obtained under alcohol consumption. None of the control
subjects consumed the same amount of alcohol, their BAC experiments showed a current difference (Figure 3C). For
valuesmeasured by breathalyzervary due to their different example, the control experiment without enzyme modification
alcohol metabolism rates and body weights. For example, three (Figure 3C, No enzyme modification) clearly demonstrates the
subjects in Figure 3B have 0.022%, 0.017%, and 0.026% BAC, high specificity of the tattoo sensor toward sweat alcohol.
respectively, after consuming the same amount of alcohol. The Similarly, the control experiment without drinking (Figure 3C,
1016 DOI: 10.1021/acssensors.6b00356
ACS Sens. 2016, 1, 1011−1019
ACS Sensors Article

Figure 4. (A) Photograph of an alcohol iontophoretic-sensing tattoo device with an integrated flexible electronics applied to a human subject and the
wireless communication with a laptop computer via Bluetooth. The inset shows an image of the flexible wireless electronics; (B) experiments with
consumption of 12 oz of beer measured from two different human subjects, before (plot “a”) and after drinking alcohol beverage (plot “b”); (C)
control experiments without drinking, first measurement (plot “a”), and second measurement (plot “b”). BAC was recorded by breathalyzer.
Potential step to −0.2 V (vs Ag/AgCl).

No drinking) indicates that the sensor response is caused by compact printed electronic circuitry has been developed for
the alcohol consumption. The data also indicates that the controlling the iontophoretic and detection processes along
iontophoresis process does not affect amperometric current with a real-time wireless data transmission via a BLE radio.
signal despite the loss of pilocarpine (and its electrolyte from Details of this board are given in the Experimental Section. The
the hydrogel) during the iontophoresis. The last control performance of the integrated wearable device was evaluated
experiments, performed without iontophoresis (Figure 3C, No following the same protocol as previous on-body experiments.
IP), show no current signal increment due to the absence of A BLE-enabled printed circuit board was employed for
induced sweat, although BAC was spiked up to 0.018% iontophoresis by applying 0.6 mA current and amperometric
following the alcohol ingestion. Overall, the results of the measurements at −0.2 V for 150 s. The current response was
experiments described in Figure 3B and C suggests that the sampled with a frequency of 1 Hz, and transmitted in real time
current output of the new biosensor system corresponds to via Bluetooth 4.0 to laptop/mobile devices and plotted on
alcohol in the iontophoresis-induced sweat and not to any screen with graphical interface developed using Python. Figure
other sources including interference from oxygen reduction 4B, C illustrates the application of the integrated flexible tattoo
toward PB. iontophoresis/sensor-electronic platform for on-body testing of
Additional on-body experiments were performed to evaluate alcohol consumption using two human subjects. Figure 4B
correlation between the BAC breathalyzer output and the displays the wirelessly transmitted current response and
current response of the new alcohol tattoo sensor following a demonstrates a clear difference in the signals before (a) and
serial consumption of wine. As illustrated in Figure 3D, the first after (b) consumption of the alcohol beverage. As illustrated in
current response recorded corresponded to a BAC 0.000% Figure 4C, no current response is observed in analogous on-
(before drinking); this was followed by two sets of drinking/ body control experiments without drinking. This performance
measurement cycles using the same electrode. Following the corresponds to that observed in the on-body data of Figure 3B,
first drink ingestion, a distinct current response was observed C involving the lab-scale potentiostatic analyzer. Overall, Figure
and the BAC turned out to be 0.025%; the BAC increased 4B, C demonstrates the practicality of the skin-worn integrated
further to 0.062% after the second drinking, and also led to a flexible alcohol sensor-electronic wireless system toward
larger amperometric current signal. The resulting plot (BAC v noninvasive BAC monitoring.
current response) is shown in Figure 3D. The current response
increases linearly with the BAC level (slope, 102 μA/BAC %;
correlation coefficient, R2 = 0.999), indicating good correlation
■ CONCLUSIONS
We have demonstrated the first example of a completely
between the sweat ethanol level and the BAC in the same wearable tattoo-based alcohol biosensor system, combining
subject. The results of Figure 3B and D indicate that future three important functions in a single wearable tattoo platform:
efforts will focus on personalizing the sensor to different sweat-inducing direct iontophoresis with amperometric enzyme
subjects. While the concept has been demonstrated in biosensing, toward noninvasive alcohol monitoring from
connection to the consumption of wine and beer, it can be human sweat, along with a thin flexible printed electronic
applied to the consumption of other alcoholic beverages circuitry for controlling the entire operation and a wireless real-
offering a broad noninvasive route to estimate BAC. time data collection. The tattoo-based biosensor system
Characterization of an Integrated Wireless Tattoo- provides reliable monitoring of alcohol consumption in real-
Based Alcohol Sensor. The practical use of the new wearable world settings, as confirmed using a variety of control
alcohol biosensor requires integration with a body-compliant experiments. Such development of a low-cost single-use
wireless circuit board. As illustrated in Figure 4A, a flexible and epidermal alcohol biosensor brings a highly useful tool for
1017 DOI: 10.1021/acssensors.6b00356
ACS Sens. 2016, 1, 1011−1019
ACS Sensors Article

monitoring alcohol on our highways, in crime scenes, or in the (6) Bates, M. E.; Martin, C. S. Immediate, quantitative estimation of
workplace. The new sensor system could thus be fitted in blood alcohol concentration from saliva. J. Stud. Alcohol 1997, 58,
vehicles to measure the drivers’ blood alcohol concentration, 531−538.
and integrated with the vehicles’ alcohol ignition interlocks to (7) Simpson, G. Accuracy and precision of breath-alcohol measure-
prevent driving while intoxicated. Such single-use printed ments for a random subject in the postabsorptive state. Clin. Chem.
1987, 33, 261−268.
tattoos could be used by bartenders or friends for identifying
(8) Dumett, M. A.; Rosen, I. G.; Sabat, J.; Shaman, A.; Tempelman,
patrons that become intoxicated and preventing drunk driving. L.; Wang, C.; Swift, R. M. Deconvolving an estimate of breath
The future systems would require additional efforts for measured blood alcohol concentration from biosensor collected
calibrating the devices and ensuring data security and privacy transdermal ethanol data. Appl. Math. Comput. 2008, 196, 724−743.
safeguards. The calibration will offer real-time BAC value, and (9) Worner, T. M.; Prabakaran, J. The accuracy of breath alcohol
will be obtained by “personalizing” the device (to account for analysis using the breathalyzer. Alcohol Alcohol. 1985, 20, 349−350.
variability in the skin permeability and sweat composition (10) Swift, R. Transdermal alcohol measurement for estimation of
among individuals). The attractive design and performance of blood alcohol concentration. Alcohol.: Clin. Exp. Res. 2000, 24, 422−
the new wearable iontophoresis-sensing platform also offers 423.
considerable promise for noninvasive monitoring of additional (11) Swift, R. M.; Martin, C. S.; Swette, L.; Laconti, A.; Kackley, N.
target analytes. Studies on a wearable, electronic, transdermal alcohol sensor. Alcohol.:

■
Clin. Exp. Res. 1992, 16, 721−725.
(12) Leffingwell, T. R.; Cooney, N. J.; Murphy, J. G.; Luczak, S.;
ASSOCIATED CONTENT Rosen, G.; Dougherty, D. M.; Barnett, N. P. Continuous objective
*
S Supporting Information monitoring of alcohol use: twenty-first century measurement using
The Supporting Information is available free of charge on the transdermal sensors. Alcohol.: Clin. Exp. Res. 2013, 37, 16−22.
ACS Publications website at DOI: 10.1021/acssen- (13) Gamella, M.; Campuzano, S.; Manso, J.; de Rivera, G. G.;
sors.6b00356. López-Colino, F.; Reviejo, A. J.; Pingarrón, J. M. A novel non-invasive
electrochemical biosensing device for in situ determination of the
Integration of the wearable alcohol tattoo sensor with alcohol content in blood by monitoring ethanol in sweat. Anal. Chim.
wireless electronics and its operation in real time via Acta 2014, 806, 1−7.
Bluetooth communication (AVI) (14) Buono, M. J. A novel non-invasive electrochemical biosensing

■
device for in situ determination of the alcohol content in blood by
monitoring ethanol in sweat. Exp. Physiol. 1999, 84, 401−404.
AUTHOR INFORMATION (15) Nyman, E.; Palmlöv, A. The elimination of ethyl alcohol in
Corresponding Authors sweat. Skand. Arch. Physiol. 1936, 74, 155−159.
*E-mail: [email protected]. (16) Matzeu, G.; Florea, L.; Diamond, D. Advances in wearable
*E-mail: [email protected]. chemical sensor design for monitoring biological fluids. Sens. Actuators,
B 2015, 211, 403−418.
Author Contributions (17) Gao, W.; Emaminejad, S.; Nyein, H. Y. Y.; Challa, S.; Chen, K.;
∥
J.K., I.J., and S.I. contributed equally to this work. Peck, A.; Fahad, H. M.; Ota, H.; Shiraki, H.; Kiriya, D.; Lien, D.-H.;
Notes Brooks, G. A.; Davis, R. W.; Javey, A. Fully integrated wearable sensor
The authors declare no competing financial interest. arrays for multiplexed in situ perspiration analysis. Nature 2016, 529,

■
509−514.
(18) Bandodkar, A. J.; Wang, J. Non-invasive wearable electro-
ACKNOWLEDGMENTS chemical sensors: a review. Trends Biotechnol. 2014, 32, 363−371.
This work was supported by the National Institute of (19) Steinburg, M. D.; Kassal, P.; Steinberg, I. M. System
Biomedical Imaging and Bioengineering of NIH Architectures in Wearable Electrochemical Sensors. Electroanalysis
(R21EB019698), from the Defense Threat Reduction Agency 2016, 28, 1149.
Joint Science and Technology Office for Chemical and (20) Glennon, T.; O’Quigley, C.; McCaul, M.; Matzeu, G.; Beirne, S.;
Wallace, G. G.; Stroiescu, F.; O’Mahoney, N.; White, P.; Diamond, D.
Biological Defense (grants nos. HDTRA1-16-1-0013), and
‘SWEATCH’: A Wearable Platform for Harvesting and Analysing
from the UCSD Center of Wearable Sensors. I.J. acknowledges Sweat Sodium Content. Electroanalysis 2016, 28, 1283.
support from the Thai Development and Promotion of Science (21) Windmiller, J. R.; Bandodkar, A. J.; Valdés-Ramírez, G.;
and Technology Talents (DPST). Parkhomovsky, S.; Martinez, A. G.; Wang, J. Electrochemical sensing

■ REFERENCES
(1) Litten, R. Z.; Allen, J. P. Measuring alcohol consumption:
based on printable temporary transfer tattoos. Chem. Commun. 2012,
48, 6794−6796.
(22) Imani, S.; Bandodkar, A. J.; Mohan, A. V.; Kumar, R.; Yu, S.;
psychosocial and biochemical methods; Springer Science & Business Wang, J.; Mercier, P. P. A wearable chemical-electrophysiological
Media: New York, 2012. hybrid biosensing system for real-time health and fitness monitoring.
(2) Zador, P. L.; Krawchuk, S. A.; Voas, R. B. Alcohol-related relative Nat. Commun. 2016, 7, 11650.
risk of driver fatalities and driver involvement in fatal crashes in (23) Jia, W.; Bandodkar, A. J.; Valdes- Ramírez, G.; Windmiller, J. R.;
relation to driver age and gender: an update using 1996 data. J. Stud. Yang, Z.; Ramírez, J.; Chan, G.; Wang, J. Electrochemical tattoo
Alcohol 2000, 61, 387−395. biosensors for real-time noninvasive lactate monitoring in human
(3) Swift, R. Direct measurement of alcohol and its metabolites. perspiration. Anal. Chem. 2013, 85, 6553−6560.
Addiction 2003, 98, 73−80. (24) Bandodkar, A. J.; Hung, V. W. S.; Jia, W.; Valdes-Ramírez, G.;
(4) Babor, T. F.; Stephens, R. S.; Marlatt, G. A. Verbal report Windmiller, J. R.; Martinez, A. G.; Ramírez, J.; Chan, G.; Kerman, K.;
methods in clinical research on alcoholism: response bias and its Wang, J. Tattoo-based potentiometric ion-selective sensors for
minimization. J. Stud. Alcohol 1987, 48, 410−424. epidermal pH monitoring. Analyst 2013, 138, 123−128.
(5) Hawthorne, J. S.; Wojcik, M. H. Transdermal alcohol (25) Bandodkar, A. J.; Jia, W.; Yardımcı, C.; Wang, X.; Ramirez, J.;
measurement: A review of the literature. J. - Can. Soc. Forensic Sci. Wang, J. Tattoo-based noninvasive glucose monitoring: a proof-of-
2006, 39, 65−71. concept study. Anal. Chem. 2015, 87, 394−398.

1018 DOI: 10.1021/acssensors.6b00356

ACS Sens. 2016, 1, 1011−1019
ACS Sensors Article

(26) Bandodkar, A. J.; Molinnus, D.; Mirza, O.; Guinovart, T.;

Windmiller, J. R.; Valdes-Ramírez, G.; Andrade, F. J.; Schoning, M. J.;
Wang, J. Epidermal tattoo potentiometric sodium sensors with wireless
signal transduction for continuous non-invasive sweat monitoring.
Biosens. Bioelectron. 2014, 54, 603−609.
(27) Guinovart, T.; Bandodkar, A. J.; Windmiller, J. R.; Andrade, F.
J.; Wang, J. A potentiometric tattoo sensor for monitoring ammonium
in sweat. Analyst 2013, 138, 7031−7038.
(28) Kim, J.; de Araujo, W. R.; Samek, I. A.; Bandodkar, A. J.; Jia, W.;
Brunetti, B.; Paixão, T. R.; Wang, J. Wearable temporary tattoo sensor
for real-time trace metal monitoring in human sweat. Electrochem.
Commun. 2015, 51, 41−45.
(29) Gonzalo-Ruiz, J.; Mas, R.; de Haro, C.; Cabruja, E.; Camero, R.;
Alonso-Lomillo, M. A.; Muñoz, F. J. Early determination of cystic
fibrosis by electrochemical chloride quantification in sweat. Biosens.
Bioelectron. 2009, 24, 1788−1791.
(30) Harvey, C. J.; LeBouf, R. F.; Stefaniak, A. B. Formulation and
stability of a novel artificial human sweat under conditions of storage
and use. Toxicol. In Vitro 2010, 24, 1790−1796.
(31) Kueseng, P.; Thammakhet, C.; Thavarungkul, P.; Kanatharana,
P. Multiwalled carbon nanotubes/cryogel composite, a new sorbent
for determination of trace polycyclic aromatic hydrocarbons. Micro-
chem. J. 2010, 96, 317−323.
(32) Gibson, L. E.; Cooke, R. E. A test for concentration of
electrolytes in sweat in cystic fibrosis of the pancreas utilizing
pilocarpine by iontophoresis. Pediatrics 1959, 23, 545−549.
(33) Ricci, F.; Palleschi, G. Sensor and biosensor preparation,
optimization and applications of Prussian Blue modified electrodes.
Biosens. Bioelectron. 2005, 21, 389−407.

1019 DOI: 10.1021/acssensors.6b00356

ACS Sens. 2016, 1, 1011−1019