This article has been accepted for inclusion in a future issue of this journal.

Content is final as presented, with the exception of pagination.

IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS 1

A Multi-Technique Reconfigurable Electrochemical

Biosensor: Enabling Personal Health
Monitoring in Mobile Devices
Alexander Sun, Student Member, IEEE, A. G. Venkatesh, and Drew A. Hall, Member, IEEE

Abstract—This paper describes the design and characterization

of a reconfigurable, multi-technique electrochemical biosensor
designed for direct integration into smartphone and wearable
technologies to enable remote and accurate personal health mon-
itoring. By repurposing components from one mode to the next,
the biosensor’s potentiostat is able reconfigure itself into three
different measurements modes to perform amperometric, poten-
tiometric, and impedance spectroscopic tests all with minimal
redundant devices. A 3.9 × 1.65 cm2 PCB prototype of the
module was developed with discrete components and tested using
Google’s Project Ara modular smartphone. The amperometric
mode has a ±1 nA to ±200 μA measurement range. When used
to detect pH, the potentiometric mode achieves a resolution of
< 0.08 pH units. In impedance measurement mode, the device
Fig. 1. Illustration showing potential uses for the multi-technique biosensor
can measure 50 Ω–10 MΩ and has been shown to have < 6◦ of platform integrated into smartphones and wearable devices.
phase error. This prototype was used to perform several point-of-
care health tracking assays suitable for use with mobile devices:
1) Blood glucose tests were conducted and shown to cover the di-
agnostic range for Diabetic patients (∼200 mg/dL). 2) Lactoferrin, equipment today is confined to centralized laboratories and
a biomarker for urinary tract infections, was detected with a limit hospitals. Furthermore, this equipment is too expensive and
of detection of approximately 1 ng/mL. 3) pH tests of sweat were bulky for direct point-of-care (POC) use.
conducted to track dehydration during exercise. 4) EIS was used to
determine the concentration of NeutrAvidin via a label-free assay. Fortunately, recent advances in portable electronics and
sensor miniaturization have allowed for the development and
Index Terms—Electrochemical biosensor, glucose, Google Ara, proliferation of mobile health (mHealth) technologies that can
lactoferrin, mHealth, pH, point-of-care, potentiostat, smartphone,
sweat, wearables. continuously monitor patients at the POC, away from tra-
ditional hospital settings. Many mobile devices have fitness
oriented sensors built-in, such as accelerometers for tracking
I. I NTRODUCTION physical activity, electrocardiograms (ECG) to record the elec-
trical signals of the heart, and photoplethsymogram (PPG) to
C HRONIC illnesses, such as heart disease, stroke, cancer,
and diabetes, are not only the leading cause of death and
disability in the US, but also the most commonly diagnosed
determine heart rate as well as the blood oxygenation level.
Unfortunately, these sensors offer limited medically actionable
and expensive health issues to treat [1]. One of the many data, especially for those with chronic diseases. Biomolecular
reasons for this phenomena is the heavy reliance on periodic sensors, on the other hand, that measure the constituents of
hospital checkups as the sole mechanism to determine one’s biological samples (e.g., blood, urine, saliva, etc.) provide a
well-being. While remote and at-home testing is a promising much more complete and medically relevant picture of the
solution to help alleviate this burden on the healthcare system user’s health. Such sensors could be used for at-home diag-
and potentially improve one’s health, most medical diagnostic nosis of infection, monitoring of treatment progression [2]–[4],
hydration and fatigue tracking during exercise [5], and testing
food and water safety [2], [6]–[9].
While several add-on biosensing modules for mobile phones
Manuscript received February 8, 2016; revised April 24, 2016; accepted
May 28, 2016. This work was supported in part by Google’s Advanced have been developed that leverage intrinsic hardware such as
Technologies & Projects (ATAP). The views and opinions expressed in this the camera, Bluetooth, USB, and audio port [2]–[18], these
article are those of the authors and do not necessarily reflect that of the funding devices are still external to the phone making them more
agency. This paper was recommended by Associate Editor D. Demarchi.
The authors are with the Department of Electrical and Computer Engineer- burdensome to manage and transport than a fully integrated
ing, University of California, San Diego, La Jolla, CA 92093 USA (e-mail: solution, dissuading frequent use. By integrating biosensors di-
[email protected]; [email protected]; [email protected]). rectly into a smartphone or smartwatch (Fig. 1) and leveraging
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org. the scalability, cost-effectiveness, and accuracy of electrochem-
Digital Object Identifier 10.1109/TBCAS.2016.2586504 ical biosensing, which led to the success of glucose meter,
1932-4545 © 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

2 IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS

one can develop much more accessible and seamless mHealth

applications that promote adherence to frequent or continuous
testing. Furthermore, in addition to being a boon for those who
live with chronic illnesses, biosensors integrated into everyday
items also enable other individuals who are either at risk for
disease, trying to improve health and fitness, or curious about
their well-being to routinely monitor themselves. Fig. 2. Randles equivalent circuit model for a three electrode system.
To this end, we describe the design of an electrochemical
biosensor module for direct integration into a smartphone or
wearable through the use of a reconfigurable bipotentiostat for characterizing the electrode-solution interface, and contains
capable of both supporting an extended range of techniques four main components: double-layer capacitance (Cdl ), charge
and, at the same time, conforming to the challenging size transfer resistance (Rct ), Warburg impedance (ZWarg ), and
and power consumption constraints set by continually shrink- solution resistance (Rs ). Cdl is a combination of the capaci-
ing portable devices. While enabling a wide variety of tests tance of the electrode itself and the capacitance generated by
would typically consume additional area and power, this design layers of ions and charged molecules forming at the surface
alleviates the problem by repurposing the same components of the electrode due to electrostatic forces. Cdl is not a strict
in different measurement modes, ultimately reducing the re- capacitance, and is typically modelled as a constant phase
dundancy. The entire platform (Fig. 1) consists of the sensing element with an impedance of Zdl = 1/((jω)m Cdl ), where m
module that houses the reconfigurable potentiostat that is meant is the phase parameter. Cdl ranges from 0.1–1 μF/mm2 and
to be built into a mobile device, an external sensor (disposable is highly dependent on the salt concentration in solution as
test strips, screen-printed electrodes, ion selective electrodes, well as the voltage of the electrode [19], [20]. Rct captures
etc.), and the mobile device itself. Since the external sensor the transfer of electrons between the solution and electrode
component is in contact with the biological sample and is from reduction and oxidation reactions of molecules close to
meant to be disposable, it is not permanently integrated into the surface. This resistance is ∼10–100 kΩ or approximately
the smartphone like the rest of the module. However, when infinite in cases without the presence of redox molecules (non-
compared to non-integrated biosensors, which have this same faradic measurements) and varies with the concentration and
external sensor constraint, smartphone integration ultimately type of molecule as well as the materials and voltage bias of
eliminates having to carry around an extra hardware component the electrode. ZWarg models the diffusion of redox molecules
thereby increasing accessibility. The mobile platform used in to and from the surface. Similar to Cdl , it also is a constant
this work is Google’s Project Ara modular smartphone, which phase element component, but always with a 45◦ phase shift.
allows the user to swap out different components and customize Finally, Rs models the ions drifting in bulk solution and is set
the phone’s hardware. This platform is ideal for biosensor by the solution conductivity and applied voltage. Depending on
integration because of its open and high-speed interface as the measurement technique, different components of this model
well as its modularity that enables the smartphone to have become important to the design of the potentiostat.
biosensing, amongst many other, capabilities. Amperometry is the standard method to perform most sen-
sitive labelled assays, which use enzymatic tags that transduce
and amplify a detection event into a measurable electrochemical
II. E LECTROCHEMICAL S ENSING BACKGROUND
signal. The circuitry for amperometric techniques [21]–[22]
As with other state-of-the-art POC electrochemical biosen- applies a voltage waveform between the WE and RE using the
sors, the most crucial component is the potentiostat, or the ana- CE to reduce voltage error while measuring the corresponding
log front-end, that interfaces with and controls the electrodes in generated current signal at the WE, which is proportional to the
contact with the sample. A typical electrochemical cell consists concentration of the biomarker. For example, cyclic voltamme-
of a working electrode (WE), where the biochemical reaction try (CV) and linear sweep voltammetry (LSV) both use slow
occurs, and a reference electrode (RE), usually working in (10–100 mV/s) ramps (< 1 V sweep range) to stimulate the
tandem with a counter electrode (CE) to set the potential of the electrochemical cell, while step-techniques such as chronoam-
cell. While there are numerous types of techniques which the perometry (CA) and square wave voltammetry (SWV) instead
potentiostat can conduct, each with varying sets of parameters, use pulsed voltages (a single step for CA and 10–100 Hz for
requirements, and advantages, all these methods essentially SWV). In the majority of amperometry, the objective is to
measure different aspects of the same phenomenon: the move- measure the current due to a particular redox reaction rather
ment and displacement of charge at the interface between an than from the faster charging and discharging of Cdl , referred to
electrode and an electrolytic solution, also known as an electro- as background current. Even in pulsed techniques, the sections
chemical cell. Equivalent circuit models of this electrochemical of the current measurement that contain the signal occur after
cell can be used to better understand the sensing mechanisms of the signal has settled. Hence, amperometry necessitates precise
various electrochemical techniques, thereby guiding the design voltage control and high measurement sensitivity for slow, large
and implementation as well as setting the requirements of the signal currents.
circuits tailored for each distinct test type. While ions cannot be easily measured with traditional la-
Randles equivalent circuit [19], shown in Fig. 2 for a three belled assays or DC current measurements, their inherent
electrode system, is the most widely used electrical model charge and size allows them to be detected via potentiometric
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

SUN et al.: A MULTI-TECHNIQUE RECONFIGURABLE ELECTROCHEMICAL BIOSENSOR 3

Fig. 3. Schematic of reconfigurable potentiostat where VCM , VIN , VWE1 , and Fig. 4. Simplified schematic of the potentiostat in amperometric mode.
VWE2 are DAC outputs and VADC1−4 are ADC inputs.

one mode to the next while maintaining performance across dif-

tests. Ion-selective electrodes (ISEs) separate specific ions with ferent techniques. Hence, the potentiostat is designed to support
a semi-permeable membrane between two electrodes, thereby three distinct techniques: 1) amperometric, 2) potentiometric,
creating a potential difference (∼0.1–100 mV) proportional to and 3) impedance spectroscopy.
the amount of that ion concentration in the solution. However, To further increase the flexibility and compatibility of the
due to the nature of these sensors, their impedance is very high, platform with POC type of tests, the potentiostat includes
roughly on the order of 100 MΩ, necessitating high resolution dual WEs each with its own resistive feedback transimpedance
sampling of the electrode voltage with a high input impedance. amplifier (TIA), which is based on circuit topology commonly
For label-free electrochemical assays, electrochemical im- used in potentiostats [30]–[32]. Using networks of switches that
pedance spectroscopy (EIS) is most often used since it measures can switch between a range of different resistors and capacitors,
changes in impedance on the surface of an electrode due to each TIA has independently adjustable gain (10 kΩ, 100 kΩ,
displacement of charge (ΔCdl ) or impeding of redox reactions and 1 MΩ) and bandwidth (1 Hz–100 kHz), expanding its
(ΔRct ). Although there are many different circuit topologies dynamic range and allowing it to measure different types of
that can implement EIS [24]–[26], generally they all apply small biomarkers that have varying sensitivity requirements. This
amplitude (<10 mV) voltage sinusoids of varying frequencies dual WE functionality also enables two tests of the same tech-
(0.1 Hz–100 kHz) between a two electrode cell and record the nique to be run simultaneously on the same sample, allowing
resulting current. For each frequency, the magnitude and phase one to either be a control to compensate for factors such as
change is calculated and used to find the complex impedance temperature variation or background signals, or an additional
ultimately forming an impedance spectrum that can be fitted to sensor for another biomarker. In order to take advantage of this
the Randles circuit. Unlike in traditional amperometry where parallel testing, an assay must either generate no free-roaming
aligning the timing of the input and output waveforms is often redox molecules that can diffuse between electrodes and cause
not necessary, EIS circuitry must not only have high enough interference (e.g., label free assays) or use an electrode design
bandwidth to measure these small signal AC currents but also that physically isolates or spreads out the sensing surfaces using
have the ability to accurately track phase change between the wells or additional sample collection channels. Alternatively,
applied voltage and measured current. Furthermore, any fre- the two electrodes can be used together for redox cycling with
quency dependent phase shift introduced by the measurement an interdigitated electrode in order to chemically amplify the
circuitry must be calibrated out. signal for higher sensitivity, particularly when dealing with
micro- and nano-scale sensors [33]. The common-mode voltage
III. D ESIGN OF R ECONFIGURABLE M ODULE is adjustable to accommodate and optimize the various current
and voltage ranges, which can be skewed either to the positive
In this work, the potentiostat discussed is based on a well- or negative side depending on the expected response. The dif-
studied and commonly used topology in electrochemistry [27], ferent configurations and respective performance are discussed
[28], and is an expanded and improved version of our previous in the following sections.
work [29]. However, to enable a large set of possible mHealth
applications, the potentiostat must be able to run multiple types
A. Amperometric
of techniques discussed above, which require different sensing
modes and additional circuitry. Each of these various types The potentiostat configuration for this mode is shown in
of tests would typically require a different and separate set Fig. 4. A voltage signal is applied to the three electrode sensor
of circuitry. However, space and power are highly constrained between the RE and the WE, with the CE supplying the current
resources on a mobile device and commodities must be shared to set the solution potential. This voltage waveform, which
with the device’s other components. Therefore, in order to varies depending on the technique chosen, generates a current
reduce the area and, more importantly when moving to an signal in the solution that is measured at the WE, in this
integrated circuit implementation, the power, a single recon- case, with a resistive feedback TIA. To expand the possible
figurable design (Fig. 3), rather than three different sets of applications of the device, this potentiostat has two working
potentiostat circuits, is used that repurposes components from electrodes with each channel having TIAs with independently
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

4 IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS

Fig. 5. Simplified schematic of the potentiostat in potentiometric mode. Fig. 6. Simplified schematic of the potentiostat in EIS mode.

configurable gain and bandwidth (adding either 1, 10, and an additional sensor on the other WE. Small signal (20 mV)
100 nF capacitor in parallel with the feedback resistance). The voltage sinusoids, with varying frequency from 1 Hz–10 kHz,
variable gain allows the device to adjust for the different base- are applied between the two terminals and the WE measures
line currents and varying physiological concentration ranges the resulting current. The gain and bandwidth of the WE TIA
of different biomarkers, assays, and sensor areas. Also, since is adjusted depending on the impedance and frequency being
different amperometric techniques excite the electrochemical measured, changing if the signal is too small or if the channel
cell with different input voltage waveforms, the bandwidth of becomes saturated. In the two other modes, the open switches
the generated current signal can vary. and unused electrodes are always low impedance nodes set to
Since the sensitivity of these measurements depends on how known voltages in order to avoid instability and interference.
accurately current can be measured, the most important design However, in the EIS mode, the RE input is left floating in the
considerations for this mode are the input-referred noise of circuitry in order to avoid the leakage current from adding a
the TIA and the current leakage at the WE node. Hence, all switch at this node, which is crucial for accurate amperometric
the switches where chosen to have low leakage (< 20 pA) and potentiometric measurements. However, the RE can be tied
and the opamps (U1 and U2, Analog Devices AD8552) were directly to the CE through a short on the electrode without
selected to balance the power, input bias current (160 pA), affecting the impedance measurements as it can be incorporated
and noise. The requirements at the other electrodes are less into the calibration.
constrained. The input bias current of the RE circuitry must Making the approximation that the system is linear, due
be minimized in order to reduce the IR error of the applied to the stimulus being small, the complex impedance ZCell is
voltage. By using a very low input bias opamp (U3, Analog computed as
Devices AD8691) chosen specifically for the potentiometric
mode (described later), this design achieves an RE leakage of VIN (jω)
ZCell (jω) = H(jω) (1)
200 fA, which, with a typical solution resistance of 100 Ω, VOUT (jω)
contributes a negligible 100 nV error. Furthermore, since the
CE, which is controlled by U4 (Texas Instruments OPA2333), where H(jω) is the transfer function that converts the current
only needs to be able to supply the necessary current to the cell to voltage, VIN is the voltage sinusoid applied to the ZCell ,
for this mode, the parameters for the control circuitry are set by and VOUT is the voltage read by the ADC. H(jω) is not
the EIS mode. only dependent on the feedback network of the TIA, which
changes depending on the cell impedance, but also other factors
B. Potentiometric such as parasitics in the switching networks and phase shift
in the signal path. Hence, to account for varying H(jω) and
In the potentiometric mode (Fig. 5), the voltage generated compensate the channel accordingly, known test impedances
between two electrodes in a solution is measured. Typically, an (ZKnown ) measured prior to use, one for each WE, are switched
ISE requires measurement circuitry with an input bias current in between the two electrodes, given the same stimulus, and
of less than 1 pA to ensure that measurement error is less than measured at each frequency before the actual sensor is tested
1%. Without adding a new set of components, the input buffer
used for RE in the amperometric mode is switched into the VIN (jω)
H(jω) = ZKnown (jω)
signal path for use as a high impedance input with a working VMeasured (jω)
electrode operating as the other terminal. By adjusting the 1
ZKnown (jω) = Rs + ||Rct . (2)
bandwidth switches to provide a short, the WE circuitry in this jωC
case operates as a buffer and allows the voltage from the sensor
Using known impedance measurements, the transfer function
to either be sampled single-ended or pseudo-differentially to
of the channel can be determined for each frequency [Eq. (2)]
reject common-mode signals.
and used to calibrate the impedance measurements in software
on the host device [34]. Furthermore, to ensure that the input
C. Impedance Spectroscopy
signal is correctly aligned with the output, the ADC simulta-
In the EIS mode (Fig. 6), a two-electrode sensor is attached neously measures the CE voltage, thereby reducing phase error
between the CE and a single WE, with the option of attaching introduced by the control circuitry.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

SUN et al.: A MULTI-TECHNIQUE RECONFIGURABLE ELECTROCHEMICAL BIOSENSOR 5

application [Fig. 7(c)]. Furthermore, this current prototype

is small enough to be considered compatible with wearable
devices as well. The module communicates with the Spiral 1
Ara platform via the I2 C serial communication pins of the
microcontroller. For testing purposes, we used several off-
the-shelf sensors that each have varied connectors. Hence, an
interposer board to accommodate all the electrodes was also
constructed and attaches to the top or bottom of the module.
Since the sensing areas are smaller than the module itself, actual
developed and complete mobile devices can have the sensor or
the sensor holder, in the case of disposable test strips, mounted
directly into the shell of the device without altering the form
factor.

V. E LECTRICAL M EASUREMENT DATA

Each of the three modes were characterized and tested to
verify their functionality. For the amperometric mode, since
the sensitivity of these measurements depends on how accu-
rately current can be measured, the most important design
Fig. 7. (a) Block diagram of the entire module. (b) Photograph of PCB next to considerations for this mode are the input-referred noise, which
a US penny for scale. (c) Screenshot of smartphone application interface. was measured with 100 kΩ gain and a 1 kHz bandwidth to
be 216 pARMS , and the leakage current at the input of the
TIAs or WE. Since low leakage switches for selecting the
IV. I NTEGRATION W ITH M OBILE T ECHNOLOGY
gain and bandwidth are used and the number of connections
Shown in Fig. 7(a), aside from the potentiostat, the module to the inverting node are minimized, the overall input leakage
also contains a power regulation network, a DAC (Analog (∼180 pA) is dominated by the input bias current of the opamp.
Devices AD5685R), an ADC (Analog Devices AD7682), and Hence, we can measure bidirectional currents ranging from
a microcontroller (Microchip dsPIC33EP256MC204). This pe- ∼500 pA to 200 μA, which is ideal for most POC applications.
riphery circuitry can be easily tailored to the specifications of For the potentiometric mode, the input bias current of the
the wearable or mobile device. The design takes its power from measurement circuitry is ∼200 fA, setting the approximate
the host device with an input voltage anywhere from 2.5 to input impedance at 5 TΩ. The input referred voltage noise
5.5 V and with a light-load efficient buck-boost DC-DC con- is 1.06 μVRMS (10 Hz bandwidth), and the voltage offset is
verter (Texas Instruments LM366SD) in series with two LDOs ∼400 μV. For EIS, when testing a known impedance of
(Texas Instruments TPS79101) regulates it to both a 4 V and 100 kΩ in parallel with 1 μF from 1 Hz–10 kHz, the module
a 3.3 V thereby isolating the analog and digital supplies. The was measured to have a 5% magnitude and a 6◦ phase error
DAC (14-bits) and ADC (16-bits) both have 4 channels, and, as compared with a benchtop EIS tool. This configuration
via SPI, their maximum update and sample rates are ∼200 kHz. and calibration scheme described previously can measure an
The microcontroller controls the potentiostat during testing impedance range of 50 Ω–10 MΩ.
by updating and sampling from the proper DAC and ADC To demonstrate the reproducibility and stability of all modes,
channels respectively. The microcontroller also communicates a series of repeated measurements (N = 100) on known inputs
with the host device via serial communication (either SPI, was performed using both the smartphone integrated platform
I2 C, or RS-232 depending on the mobile device interface) and and a benchtop potentiostat (CH Instruments 750E) also re-
configures the potentiostat with the proper settings. In order to ferred to in this paper as CHI. For the amperometric mode, a
integrate this module into a mobile device that is not a modular signal current of 10 μA was generated by applying a voltage
smartphone, an internal I/O port needs to be accessible. While signal across a model of an electrochemical cell made from
this communication between the module and processor would circuit components resembling a simplified version of Randles
usually be through a proprietary communication protocol, it is equivalent circuit model (Rct = 200 kΩ, Cdl = 2 μF, and
reasonable to expect that for mobile devices that use different Rs = 1 kΩ). For the potentiometric mode, an input voltage
types of sensors (such as accelerometers and pulse oximeters) of 0.5 V was applied directly from the sourcemeter across
the translation hardware is already available that implements two electrodes. Finally, for EIS, the same circuit model was
the required communication interface between sensors and the measured by both instruments to find the value of the charge
high-speed processor bus. Hence, adding this module would be transfer resistance. The results, shown in Fig. 8, show the mean
as simple as integrating any other sensor. and standard deviation of the measurements normalized to the
A 3.9 × 1.65 cm2 4-layer PCB [Fig. 7(b)] with discrete CHI data. While the variance in the data from the module is
ICs was fabricated to fit into the Google Project Ara smart- larger than that of the benchtop potentiostat (1.41 nA versus
phone as a 2 × 1 sized module and work with an Android 88.6 pA, 93.7 μV versus 16.3 μV, and 0.630 Ω versus 0.186 Ω),
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

6 IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS

Fig. 8. (a) Plots of the amperometric, potentiometric, and (b) EIS mode
repeated measurements for both the CHI and module potentiostat for N = 100
normalized to the CHI average.
Fig. 9. (a) Chronoamperometry curves for glucose measured by the sensing
module. (b) Calibration curves for both the biosensor and CHI with the positive
each is still within acceptable bounds for that particular tech- and negative diagnosis ranges annotated.
nique and matches well with the CHI measurements.
These modes also consume varying amounts of power, due VI. T ESTING POC A PPLICATIONS
to the different ADC sampling and data transfer rates required
by each mode. Also, since the module can disable the potentio- While the device itself can perform many types of elec-
stat, ADC, and DAC, as well as make the microcontroller sleep, trochemical tests, the biomarkers detected in the following
essentially shutting itself off when not measuring (< 100 μW), assays were chosen due to their POC applications. All these
the average runtimes of each technique also determine the over- experiments, while some taking more effort and materials than
all energy used by each mode. The entire potentiostat’s peak others, do not require lab equipment to pre-process the samples
power consumption including the switches and multiplexors and are possible to measure at the POC.
is 9.6 mW. To conserve space, many of the parts used in the
potentiostat contain more than one device in a single package
A. Amperometric Testing
making it difficult to power gate individual unused compo-
nents, so the power consumption of the potentiostat remains 1) Glucose: For Glucose experiments, PBS was spiked with
approximately constant across different modes. The digital and various concentrations of Dextrose from Marcon (4912-12) to
mixed signal circuitry including the microcontroller, ADC, and create the test solutions. Commercial glucose test strips (True
DAC consume a maximum of 49.5 mW in amperometric mode Test Blood Glucose Strips) based on Glucose dehydrogenase-
with a runtime of 10–200 s and 46.2 mW in potentiometric PQQ (GDH) were applied with the various test solutions
mode for tests that last approximately 10 s. In EIS mode, this (27-450 mg/dL) and measured with chronoamperometry (0.5 V
power consumption is 111 mW for an average of 130 s. To step for 10 seconds) with both a benchtop instrument (CHI
put these numbers into context, the lithium ion battery found 750E) and the biosensor module. Since commercial glucose
in most of today’s smartphones has a capacity of approximately strips are optimized for small droplets of blood (a few micro-
1500 mAh. Average idle time is ∼50 hours (@ 108 mW), while liters), 1 μL of each of the test solutions were used in these mea-
talk time is ∼10 hours (@ 540 mW). Hence, at the very worst, surements. The results (Fig. 9) show that the measured currents
this module would about match the power consumption of the (taken after 10 seconds) for each concentration measured by
phone while idling, and consume 80% less than a phone call. both instruments follow the same trend. The calibration curve
Therefore, making a couple of several minute-long measure- demonstrates that the assay is in the correct region to be able to
ments per day should not add noticeably to the battery drain of diagnose or monitor diabetes (positive > 200 mg/dL according
the mobile device. to the American Diabetes Association).
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

SUN et al.: A MULTI-TECHNIQUE RECONFIGURABLE ELECTROCHEMICAL BIOSENSOR 7

Fig. 11. Plot of pH levels of sweat from subject during exercise and the pH test
strip result for each sample.

Fig. 10. (a) CV curves for LTF measured by the sensing module. (b) Calibra-
tion curves for the LTF assay with the positive and negative diagnosis ranges
annotated.

2) Lactoferrin: Lactoferrin (LTF) is a common biomarker

for infection found in various concentrations in bodily fluid
such as sweat [35], saliva [36], urine [37], tears [38], and
stool [39]. In this case, the detection of LTF in urine is used
to diagnose urinary tract infection. Unlike the detection of
glucose which is enzymatic, the detection method used here is
a sandwich assay similar to an ELISA.
Gold DropSens electrodes were functionalized for detection
of LTF. Anti-human LTF (Abcam #ab10110) was mixed with
Traut’s reagent (Pierce 26101), dropped on the gold work-
ing electrodes, and incubated overnight at −4 ◦ C. 2% BSA
(Thermo Scientific 37525) was applied for 1 hour at room
Fig. 12. (a) Nyquist plot of each serial dilution of NeutraAvidin. (b) Concen-
temperature to block the surface. Afterwards, various concen- tration curve after fitting data to the Randles circuit to find charge transfer
trations of LTF (Abcam #a78526) in 20 μL droplets were resistance with baseline drawn below.
added to each electrode before adding the secondary antibody
(Abcam #ab25811) and then the NeutrAvidin conjugated al- of this assay is approximately ∼1 ng/mL. Hence, LTF can be
kaline phosphatase (Thermo Scientific #31002). Each binding detected by this device in the diagnostically relevant range.
step lasted an hour and included washing in between. Finally, Furthermore, the average LTF concentration in various bod-
before running cyclic voltammetry on each electrode, the sub- ily fluids in healthy patients is annotated on the same plot,
strate, p-AminoPhenyl Phosphate (Santa Cruz Biotechnology demonstrating that this device could also be used to measure
sc-281392) was added and allowed to react for 10 minutes. The physiological LTF concentrations in these other fluids.
sweep range and scan rate were −0.2 V to 0.3 V and 25 mV/s,
respectively.
B. pH Measurements in Sweat
The concentration of LTF in the urine of a patient with and
without an UTI is 3300 ± 646.3 ng/mL and 60.3 ± 14.9 ng/mL, pH levels in sweat secreted from the skin have been shown
respectively [37]. As shown in Fig. 10, the limit of detection to correlate with hydration levels in the body [5]. The higher
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

8 IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS

TABLE I
C OMPARISON W ITH S TATE OF THE A RT FOR A LL E LECTROCHEMICAL M ODES

the pH the more dehydrated someone is. Hence, by monitoring with biotin, a pair commonly used as a preliminary model for
sweat during exercise, hydration can be tracked allowing the label-free detection assays.
user to act accordingly to optimize his or her workout and avoid Prior to the start of the assay, the electrode, 100 nm of
dangerous over exertion. gold sputtered onto a glass substrate, was cleaned with 1 mM
In order to first test the potentiostatic mode’s accuracy when KOH/H2 O2 and functionalized with a 100 μM thiolated-biotin
interfaced with a high impedance sensor, standard pH buffers (Sigma-Aldrich #746622) reagent solution. After performing
from Thermo Scientific (910104, 910107, and 910110) were a washing and blocking, the electrode was ready for use.
used as well as separately prepared phosphate buffers adjusted 20 μL droplets of different concentrations of NeutrAvidin
to specific values ranging from pH 4–10. All measurements (Thermo Scientific #31000) in a 1 mM ferro/ferri-cyanide
were taken with an Oakton pH Probe (EW-35811-74). These (K4 [Fe(CN)6 ])/K3 [Fe(CN)6 ]) PBS buffer were added to the
buffers were measured with the biosensor module in poten- electrode, allowed to bind for 10 minutes, and then measured
tiometric mode and verified with a table top pH meter (Orion using EIS (1 Hz–10 kHz) with a Ag wire pseudo RE. These
Star A211). The maximum deviation was found to be 1.2% or data were then fitted against the standard Randles circuit [19]
0.08 pH between the two measurement methods. to determine the change in charge transfer resistance, relevant
Next, 75 μL of sweat was collected at 10-minute intervals in faradaic impedance measurements. The Nyquist plot of the
from a volunteer running at a steady pace for an hour. After- results as well as the concentration curve, shown in Fig. 12,
wards, the sweat was tested with the module using a small pH clearly demonstrate that this module can be used as a label-
electrode (VersaFlex VNIS/LD). Each sample was also tested free biosensor. While NeutrAvidin itself is not a particularly
using standard pH test strips (pHydrion Vivid 67). As shown useful biomarker, due to the mechanism of the biotin-avidin
in Fig. 11, the pH level increases steadily as more sweat is lost bonding, the results of this model assay demonstrate that this
during the exercise as expected when compared to published device can be generalized and used in most label-free affinity
data [5]. The test strips line up with the pH levels measured by assays already developed [46]–[48].
the device and serve to confirm this trend as well.
D. Comparison With Literature
C. Label-Free Assay
These POC applications experiments demonstrate both the
The sensing of certain biomarkers, such as ions (H+, Na+, performance and the extensive functionality of the reconfig-
etc.) and some metabolites (glucose, lactose, etc.) especially urable module. To closely examine the performance, Table I
those with large physiological concentrations, can be easily shows a comparison with state-of-the-art portable biosensors
designed for portable POC use without an abundance of steps that have been previously published. For each mode, our mod-
or reagents. However, assays for more complex molecules (pep- ule approximately matches the performance of other platforms
tides, proteins, DNA, etc.) that require much higher sensitivity in terms of dynamic range, sensitivity, and error, while at the
to detect can be more cumbersome and time consuming for same time being able to reconfigure itself into these three dif-
a user to conduct. For infrequent diagnostic tests, such as the ferent sensing modes. Hence, whereas other devices only have
labelled and highly sensitive UTI test discussed previously, the one or two of these measurement capabilities, this device is able
additional washing and reagent steps in the assay are manage- to package all these multiple techniques with approximately
able in the case of at-home testing. However, for more remote equivalent performance into a single small form factor module.
applications that require equally high sensitivity and increased
portability, label-free techniques, such as EIS, are a promising
VII. C ONCLUSION
solution as they do not use enzymatic labels to indirectly mea-
sure the biomarker, but rather physical and chemical changes, We have built and demonstrated a reconfigurable, multi-
allowing for faster results with fewer assay steps [44], [45]. To technique biosensor platform specially designed for integration
demonstrate our module’s label-free capabilities, we conducted directly into mobile devices for diagnosing and monitoring the
an assay for the detection of NeutrAvidin using biotin immobi- health of a user at the POC. By reusing components in different
lized on the surface of a gold electrode. NeutrAvidin is version measurement modes, we can minimize the size and power of
of avidin, a protein that forms a specific and high-affinity bond the design while at the same time keeping performance and
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

SUN et al.: A MULTI-TECHNIQUE RECONFIGURABLE ELECTROCHEMICAL BIOSENSOR 9

expanding the functionality of the module for use in most POC [16] E. H. Doeven, G. J. Barbante, A. J. Harsant, P. S. Donnelly, T. U. Connell,
applications. By adding this dedicated hardware directly into C. F. Hogan, and P. S. Francis, “Mobile phone-based electrochemilu-
minescence sensing exploiting the ‘USB On-The-Go’ protocol,” Sens.
every day carry electronics, we hope to promote the use of spe- Actuators B, Chem., vol. 216, pp. 608–613, Sep. 2015.
cialized, portable, and practical medical devices well positioned [17] A. Sun, T. Wambach, A. G. Venkatesh, and D. A. Hall, “A low-
to be the first line of defense in the future of healthcare. cost smartphone-based electrochemical biosensor for point-of-care
diagnostics,” in Proc. IEEE Biomedical Circuits Systems Conf., 2014,
pp. 312–315.
[18] A. C. Sun, C. Yao, V. A. G., and D. A. Hall, “An efficient power harvesting
mobile phone-based electrochemical biosensor for point-of-care health
ACKNOWLEDGMENT monitoring,” Sensors and Actuators B: Chemical, vol. 235, pp. 126–135,
Nov. 2016.
The authors acknowledge T. Phelps for his help program- [19] A. J. Bard and L. R. Faulkner, Electrochemical Methods Fundamentals
ming the Android smartphone application. and Applications, 2nd ed. Hoboken, NJ, USA:Wiley, 2001.
[20] H.-J. Butt, K. Graf, and M. Kappl, Physics and Chemistry of Interfaces,
2nd ed. Weinheim, Germany: Wiley-VCH, 2006.
[21] M. H. Nazari, H. Mazhab-Jafari, L. Leng, A. Guenther, and R. Genov,
R EFERENCES “CMOS neurotransmitter microarray: 96-channel integrated potentiostat
with on-die microsensors,” IEEE Trans. Biomed. Circuits Syst., vol. 7,
[1] B. W. Ward, J. S. Schiller, and R. A. Goodman, “Multiple chronic con- no. 3, pp. 338–348, Jun. 2013.
ditions among US adults: A 2012 update,” Prev. Chronic. Dis., vol. 11, [22] M. M. Ahmadi and G. A. Jullien, “Current-mirror-based potentiostats
Apr. 2014. for three-electrode amperometric electrochemical sensors,” IEEE Trans.
[2] A. Nemiroski, D. C. Christodouleas, J. W. Hennek, A. A. Kumar, Circuits Syst. I, Reg. Papers, vol. 56, no. 7, pp. 1339–1348, Jul. 2009.
E. J. Maxwell, M. T. Fernández-Abedul, and G. M. Whitesides, “Univer- [23] M. Stanacevic, K. Murari, A. Rege, G. Cauwenberghs, and N. V. Thakor,
sal mobile electrochemical detector designed for use in resource-limited “VLSI potentiostat array with oversampling gain modulation for wide-
applications,” Proc. Nat. Acad. Sci., vol. 111, no. 33, pp. 11984–11989, range neurotransmitter sensing,” IEEE Trans. Biomed. Circuits Syst.,
Aug. 2014. vol. 1, no. 1, pp. 63–72, Mar. 2007.
[3] P. B. Lillehoj, M.-C. Huang, N. Truong, and C.-M. Ho, “Rapid elec- [24] A. Manickam, A. Chevalier, M. McDermott, A. D. Ellington, and
trochemical detection on a mobile phone,” Lab. Chip, vol. 13, no. 15, A. Hassibi, “A CMOS Electrochemical Impedance Spectroscopy (EIS)
pp. 2950–2955, Jul. 2013. Biosensor Array,” IEEE Trans. Biomed. Circuits Syst., vol. 4, no. 6,
[4] B. Berg, B. Cortazar, D. Tseng, H. Ozkan, S. Feng, Q. Wei, R. Y.-L. Chan, pp. 379–390, Dec. 2010.
J. Burbano, Q. Farooqui, M. Lewinski, D. Di Carlo, O. B. Garner, and [25] A. Carullo, F. Ferraris, M. Parvis, A. Vallan, E. Angelini, and P. Spinelli,
A. Ozcan, “Cellphone-based hand-held microplate reader for point- “Low-cost electrochemical impedance spectroscopy system for corrosion
of-care testing of enzyme-linked immunosorbent assays,” ACS Nano, monitoring of metallic antiquities and works of art,” IEEE Trans. Instrum.
vol. 9, no. 8, pp. 7857–7866, Aug. 2015. Meas., vol. 49, no. 2, pp. 371–375, Apr. 2000.
[5] V. Oncescu, D. O’Dell, and D. Erickson, “Smartphone based health acces- [26] H. Jafari, L. Soleymani, and R. Genov, “16-Channel CMOS impedance
sory for colorimetric detection of biomarkers in sweat and saliva,” Lab. spectroscopy DNA analyzer with dual-slope multiplying ADCs,” IEEE
Chip, vol. 13, no. 16, pp. 3232–3238, Jul. 2013. Trans. Biomed. Circuits Syst., vol. 6, no. 5, pp. 468–478, Oct. 2012.
[6] X. Wang, M. R. Gartia, J. Jiang, T.-W. Chang, J. Qian, Y. Liu, X. Liu, and [27] M. Vergani, M. Carminati, G. Ferrari, E. Landini, C. Caviglia,
G. L. Liu, “Audio jack based miniaturized mobile phone electrochemical A. Heiskanen, C. Comminges, K. Zor, D. Sabourin, M. Dufva, M. Dimaki,
sensing platform,” Sens. Actuators B, Chem., vol. 209, pp. 677–685, R. Raiteri, U. Wollenberger, J. Emneus, and M. Sampietro, “Multichannel
Mar. 2015. bipotentiostat integrated with a microfluidic platform for electrochemical
[7] C. Ionescu, P. Svasta, C. Tamas, C. Bala, and L. Rotariu, “Portable mea- real-time monitoring of cell cultures,” IEEE Trans. Biomed. Circuits Syst.,
suring and display unit for electrochemical sensors,” in Proc. IEEE 16th vol. 6, no. 5, pp. 498–507, Oct. 2012.
Int. Symp. Design Technol. Electron. Packaging, 2010, pp. 215–218. [28] L. Li, X. Liu, W. A. Qureshi, and A. J. Mason, “CMOS amperometric
[8] S. K. J. Ludwig, C. Tokarski, S. N. Lang, L. A. van Ginkel, H. Zhu, instrumentation and packaging for biosensor array applications,” IEEE
A. Ozcan, and M. W. F. Nielen, “Calling biomarkers in milk using a Trans. Biomed. Circuits Syst., vol. 5, no. 5, pp. 439–448, Oct. 2011.
protein microarray on your smartphone,” PLoS ONE, vol. 10, no. 8, [29] A. Sun, T. Wambach, A. G. Venkatesh, and D. A. Hall, “A multitechnique
Aug. 2015, Art. no. e0134360. reconfigurable electrochemical biosensor for integration into mobile tech-
[9] S. K. J. Ludwig, H. Zhu, S. Phillips, A. Shiledar, S. Feng, D. Tseng, nologies,” in Proc. IEEE Biomedical Circuits Syst. Conf., 2015, pp. 1–4.
L. A. van Ginkel, M. W. F. Nielen, and A. Ozcan, “Cellphone-based [30] A. A. Rowe, A. J. Bonham, R. J. White, M. P. Zimmer, R. J. Yadgar,
detection platform for rbST biomarker analysis in milk extracts using a T. M. Hobza, J. W. Honea, I. Ben-Yaacov, and K. W. Plaxco,
microsphere fluorescence immunoassay,” Anal. Bioanal. Chem., vol. 406, “CheapStat: An open-source, ‘Do-It-Yourself’ potentiostat for analytical
no. 27, pp. 6857–6866, Jun. 2014. and educational applications,” PLoS ONE, vol. 6, no. 9, Sep. 2011,
[10] C. M. McGeough and S. O’Driscoll, “Camera phone-based quantitative Art. no. e23783.
analysis of c-reactive protein ELISA,” IEEE Trans. Biomed. Circuits Syst., [31] S. Hwang and S. Sonkusale, “CMOS VLSI potentiostat for portable
vol. 7, no. 5, pp. 655–659, Oct. 2013. environmental sensing applications,” IEEE Sens. J., vol. 10, no. 4,
[11] L. Cevenini, M. M. Calabretta, G. Tarantino, E. Michelini, and A. Roda, pp. 820–821, 2010.
“Smartphone-interfaced 3D printed toxicity biosensor integrating bio- [32] I. Ramfos, N. Vassiliadis, S. Blionas, K. Efstathiou, A. Fragoso,
luminescent ‘sentinel cells,” Sens. Actuators B, Chem., vol. 225, C. K. O’Sullivan, and A. Birbas, “A compact hybrid-multiplexed po-
pp. 249–257, Mar. 2016. tentiostat for real-time electrochemical biosensing applications,” Biosens.
[12] K. Su, Q. Zou, J. Zhou, L. Zou, H. Li, T. Wang, N. Hu, and P. Wang, Bioelectron., vol. 47, pp. 482–489, Sep. 2013.
“High-sensitive and high-efficient biochemical analysis method using a [33] J. Das, K. Jo, J. W. Lee, and H. Yang, “Electrochemical immunosen-
bionic electronic eye in combination with a smartphone-based colorimet- sor using p-aminophenol redox cycling by hydrazine combined with a
ric reader system,” Sens. Actuators B, Chem., vol. 216, pp. 134–140, low background current,” Anal. Chem., vol. 79, no. 7, pp. 2790–2796,
Sep. 2015. Apr. 2007.
[13] W. Xu, S. Lu, Y. Chen, T. Zhao, Y. Jiang, Y. Wang, and X. Chen, [34] J. S. Daniels, E. P. Anderson, T. H. Lee, and N. Pourmand, “Simultaneous
“Simultaneous color sensing of O2 and pH using a smartphone,” Sens. measurement of nonlinearity and electrochemical impedance for protein
Actuators B, Chem., vol. 220, pp. 326–330, Dec. 2015. sensing using two-tone excitation,” in Proc. 30th Annu. Int. Conf. IEEE
[14] M. Zangheri, L. Cevenini, L. Anfossi, C. Baggiani, P. Simoni, F. Di Nardo, Engineering Medicine Biology Soc., 2008, pp. 5753–5756.
and A. Roda, “A simple and compact smartphone accessory for quan- [35] J.-H. Park, G.-T. Park, I. H. Cho, S.-M. Sim, J.-M. Yang, and D.-Y. Lee,
titative chemiluminescence-based lateral flow immunoassay for salivary “An antimicrobial protein, lactoferrin exists in the sweat: proteomic analy-
cortisol detection,” Biosens. Bioelectron., vol. 64, pp. 63–68, Feb. 2015. sis of sweat,” Exp. Dermatol., vol. 20, no. 4, pp. 369–371, 2011.
[15] D. Zhang, Y. Lu, Q. Zhang, L. Liu, S. Li, Y. Yao, J. Jiang, G. L. Liu, and [36] F. Mizuhashi, K. Koide, S. Toya, M. Takahashi, R. Mizuhashi, and
Q. Liu, “Protein detecting with smartphone-controlled electrochemical H. Shimomura, “Levels of the antimicrobial proteins lactoferrin and chro-
impedance spectroscopy for point-of-care applications,” Sens. Actuators mogranin in the saliva of individuals with oral dryness,” J. Prosthet. Dent.,
B, Chem., vol. 222, pp. 994–1002, Jan. 2016. vol. 113, no. 1, pp. 35–38, 2015.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

10 IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS

[37] S. Arao, S. Matsuura, M. Nonomura, K. Miki, K. Kabasawa, and A. G. Venkatesh received the B.Sc. degree in bio-
H. Nakanishi, “Measurement of urinary lactoferrin as a marker of urinary chemistry from the University of Madras, Tamil
tract infection,” J. Clin. Microbiol., vol. 37, no. 3, pp. 553–557, Mar. 1999. Nadu, India, the M.Sc. degree in biotechnology
[38] A. Kijlstra, S. H. Jeurissen, and K. M. Koning, “Lactoferrin levels in from Bharathidasan University, Tamil Nadu, India,
normal human tears.,” Br. J. Ophthalmol., vol. 67, no. 3, pp. 199–202, the M.Tech. degree in bioelectronics from Tezpur
Mar. 1983. University, Assam, India, and the Ph.D. degree
[39] M. Joishy, I. Davies, M. Ahmed, J. Wassel, K. Davies, A. Sayers, and in physics from Bielefeld University, Bielefeld,
H. Jenkins, “Fecal Calprotectin and Lactoferrin as noninvasive markers Germany.
of pediatric inflammatory bowel disease,” J. Pediatr. Gastroenterol. Nutr., During his doctoral research, he developed a
vol. 48, no. 1, pp. 48–54, Jan. 2009. novel platform to monitor DNA-protein interactions
[40] A. F. D. Cruz, N. Norena, A. Kaushik, and S. Bhansali, “A low-cost minia- in real-time and, as a Postdoctoral Researcher at
turized potentiostat for point-of-care diagnosis,” Biosens. Bioelectron., the University of Freiburg, Freiburg im Breisgau, Germany, he developed
vol. 62, pp. 249–254, Dec. 2014. low-cost, smartphone-based devices for biomedical applications. Currently,
[41] M. D. Steinberg, P. Kassal, I. Kereković, and I. M. Steinberg, “A wire- he is involved in developing smartphone-based electrochemical assays for
less potentiostat for mobile chemical sensing and biosensing,” Talanta, clinical applications in the point-of-care domain at the University of California,
vol. 143, pp. 178–183, Oct. 2015. San Diego, La Jolla, CA, USA. As an Interdisciplinary Researcher, he is
[42] E. Angelini, S. Corbellini, M. Parvis, F. Ferraris, and S. Grassini, interested in research that involves the integration of physics, chemistry, and
“An Arduino-based EIS with a logarithmic amplifier for corrosion biology.
monitoring,” in Proc. IEEE Int. Instrum. Meas. Technol. Conf., 2014,
pp. 905–910.
[43] J. Punter-Villagrasa, B. del Moral-Zamora, J. Colomer-Farrarons, Drew A. Hall (S’07–M’12) received the B.S. degree
P. Miribel-Catala, J. Cid, I. Rodriguez-Villarreal, and B. Prieto-Simon, (honors) in computer engineering from the Univer-
“A portable point-of-use EIS device for in-vivo biomédical applications,” sity of Nevada, Las Vegas, Las Vegas, NV, USA, in
in Proc. Conf. Design Circuits Integr. Circuits, 2014, pp. 1–6. 2005, and the M.S. and Ph.D. degrees in electrical
[44] J. S. Daniels and N. Pourmand, “Label-Free Impedance Biosensors: engineering from Stanford University, Stanford, CA,
Opportunities and Challenges,” Electroanalysis, vol. 19, no. 12, USA, in 2008 and 2012, respectively.
pp. 1239–1257, 2007. From 2011 to 2013, he was a Research Scientist
[45] E. Katz and I. Willner, “Probing biomolecular interactions at conduc- at the Intel Corporation in the Integrated Biosensors
tive and semiconductive surfaces by impedance spectroscopy: Routes Laboratory. Since 2013, he has been with the Uni-
to impedimetric immunosensors, DNA-sensors, and enzyme biosensors,” versity of California, San Diego, La Jolla, CA, USA,
Electroanalysis, vol. 15, no. 11, pp. 913–947, 2003. as an Assistant Professor in the Department of Elec-
[46] M. Xu, X. Luo, and J. J. Davis, “The label free picomolar detection of trical and Computer Engineering. His research interests include bioelectronics,
insulin in blood serum,” Biosens. Bioelectron., vol. 39, no. 1, pp. 21–25, biosensors, analog circuit design, medical electronics, and sensor interfaces.
Jan. 2013. Dr. Hall was the corecipient of First Place in the inaugural International
[47] R. Ohno, H. Ohnuki, H. Wang, T. Yokoyama, H. Endo, D. Tsuya, and IEEE Change the World Competition and First Place in the BME-IDEA inven-
M. Izumi, “Electrochemical impedance spectroscopy biosensor with inter- tion competition, both in 2009. He received the Analog Devices Outstanding
digitated electrode for detection of human immunoglobulin A,” Biosens. Designer Award in 2011, an undergraduate teaching award in 2014, the Hellman
Bioelectron., vol. 40, no. 1, pp. 422–426, Feb. 2013. Fellowship Award in 2014, and an NSF CAREER Award in 2015. He is a Tau
[48] T. Bryan, X. Luo, P. R. Bueno, and J. J. Davis, “An optimised electro- Beta Pi Fellow.
chemical biosensor for the label-free detection of C-reactive protein in
blood,” Biosens. Bioelectron., vol. 39, no. 1, pp. 94–98, Jan. 2013.

Alexander Sun (S’13) received the B.S. degree in

electrical engineering and computer science from the
University of California, Berkeley, Berkeley, CA,
USA, and the M.S. degree in electrical and com-
puter engineering from the University of California,
San Diego (UCSD), La Jolla, CA, USA, in 2012 and
2014, respectively.
Currently, he is working toward the Ph.D. degree
at UCSD. His research focuses on electrochemi-
cal biosensors, electrochemical measurement tech-
niques, and compact, low power circuit design for
biomedical, and point-of-care devices.