Review

1 DOI: 10.1002/elan.201800474
2
3
4
Point-of-Care Smartphone-based Electrochemical
5
6
Biosensing
7
8 Alexander C. Sun[a] and Drew A. Hall*[a]
9
10
11 Abstract: Point-of-care (PoC) biosensors offer promising cal biosensors for PoC diagnostics. These peripherals
12 solutions to today’s adverse and costly healthcare issues utilize the various connectivity options (for example
13 by moving diagnostic tools closer to the patient. The proprietary ports, audio headphone-jack, or wireless
14 ubiquity of smartphones has brought about an emergence radio) to offload functionality to the smartphone. The
15 of PoC devices, which leverage the smartphone’s capabil- smartphone-based implementations of various electro-
16 ities, enabling the creation of low-cost and portable chemical techniques, such as amperometry, potentiometry,
17 biosensors. Electrochemical biosensors are well suited for and impedance spectroscopy are explored. Major chal-
18 PoC testing since the transducers can be miniaturized and lenges include reducing power, area, and cost of measure-
19 inexpensively fabricated. This review paper discusses ment circuitry, while maintaining adequate performance
20 recent developments in smartphone-based electrochemi- for PoC diagnostic applications.
21
Keywords: Smartphone · electrochemical biosensor · point-of-care
22
23
24
1 Introduction
25
26 Based on the most recent data from 2015, chronic
27 illnesses, such as heart disease, stroke, and diabetes, are
28 the leading cause of death and disability in the United
29 States, with over 50 % of the population having at least
30 one chronic illness [1–3]. Furthermore, 87 % of total
31 annual healthcare expenditures ($2.7 trillion in total) are
32 for chronic diseases with 35 % of this spending for a
33 disproportionate 8.7 % of the population [1]. The high
34 death rate and financial cost exist even though in many
35 cases these diseases are preventable or manageable with
36 proper equipment. On the other hand, in lower-income Fig. 1. Proposed concept for modern healthcare system aug-
37 countries, communicable disease, such as respiratory mented with PoC biosensors.
38 infections, human immunodeficiency virus (HIV), Ma-
39 laria, and tuberculosis (TB) are a much larger concern
40 and among the top causes of death [4]. Outbreaks of acute conditions. More comprehensive testing equipment also
41 infectious disease are also a major concern as they not can be carried by the physician or placed in the doctor’s
42 only can create a large death toll but also spread to other office for faster results, which can lead to earlier treatment
43 parts of the world if unchecked, such as the Ebola virus and less reliance on testing at centralized laboratories.
44 outbreak in 2014 that infected individuals in 10 different Finally, the same technology can also be used for rapid
45 countries, caused 11,310 verified deaths, and cost $3.6 and accurate ad hoc testing in remote settings. However,
46 billion dollars to control [5, 6]. Much like the issues in for these PoC platforms to be portable enough to be
47 developed countries, these diseases are often also prevent- useful, they need to be sensitive, compact, and low-power,
48 able or treatable, but, in this case, the lack of adequate all qualities in line with the World Health Organization’s
49 healthcare infrastructure makes such solutions infeasible. ASSURED criteria to evaluate PoC devices [9, 10].
50 A promising solution to these problems in both Fortunately, as of early 2018, 68 % of the world’s
51 developing and developed countries is to use point-of-care population use mobile phones [11]. In the United States,
52 (PoC) biosensors, as shown in Figure 1, which augments the number of people with a mobile phone has grown to
53 the current healthcare system by allowing for diagnostic 95 % with 77 % specifically using a smartphone [12]. This
54 tools to be brought closer to the patient providing more
55 rapid and frequent feedback loops [7, 8]. With PoC devices [a] A. C. Sun, D. A. Hall
56 located near a patient such as with at-home diagnostics or Electrical and Computer Engineering, University of California
57 wearables, the patient can be more involved in their own in San Diego, La Jolla, CA
58 health to help prevent serious issues and manage chronic E-mail: [email protected]

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 2
Review
1 growth can also be seen in other regions of the world such electrodes while simultaneously measuring the resulting
2 as Africa, Asia Pacific, and the Middle East that account current or voltage waveform at the same electrodes. For
3 for 80 % of all new mobile phone subscriptions [13]. This measurements that apply voltage and record current for
4 overwhelming data clearly shows that mobile phones and, example, any change in the cell that causes accumulation
5 increasingly so, smartphones have become a ubiquitous of charged molecules or promotes reduction and oxidation
6 part of everyday life around the world. Having what is reactions will produce a change in the measured current,
7 essentially a powerful, portable, and network connected since the electrical characteristics of the cell have been
8 device available to a majority of the world presents a altered. Electrochemical biosensors are essentially elec-
9 tremendous opportunity to leverage this technology to trochemical cells with an added “bioelement,” a partic-
10 develop PoC biosensors that hold promise for portable ularly chosen molecule (e. g., DNA, antibody, protein,
11 and convenient diagnostic testing [14–16]. peptide, etc.) whose specificity is used to isolate and
12 Hence, there has recently has been a large influx of detect the analyte (target molecule). The electrode or
13 biosensing peripheral modules that attach to smartphones transducer converts this detection of the analyte into an
14 to take advantage of their widespread availability, com- electrical signal that can be measured [46]. For example,
15 puting power, network connectivity, battery, and camera imagine that one of the electrodes, known as the working
16 to offload as much of the biosensor as possible onto the electrode (WE), has single-stranded DNA (ssDNA), the
17 phone itself [17–20]. The goal by doing so is to create a “bioelement”, immobilized on its surface meant to target
18 more streamlined device that takes up less area and is less its complimentary strand. When the WE is exposed to a
19 expensive compared to an equivalent stand-alone bio- solution full of a variety of other different ssDNA, ideally
20 sensor when combined with an already available smart- only those strands complimentary to the bound ones will
21 phone. It is worth noting here that, while much of the combine and stick to the surface. DNA is negatively
22 smartphone-based biosensor ecosystem consists of optical charged, so the WE has accumulated a significant amount
23 based sensing that uses the camera for microscopy [21– of charge. By using the potentiostat to apply a stimulus
24 30], colorimetric [31–38], and spectroscopic [39–42] assays, between the WE and the other electrode called the
25 the following sections will only focus on electrochemical reference electrode (RE), a difference in current will be
26 sensing. These optical techniques are typically limited by measured between the response before and after the
27 the resolution and focus of the smartphone camera as well ssDNA was bound.
28 as ambient lighting conditions [43]. On the other hand, A simplified version of the potentiostat circuit is
29 electrochemical measurement has the advantage of being shown in Figure 2 with a third electrode, counter electrode
30 mostly independent of the smartphone’s capabilities while (CE) used together in feedback with the RE to eliminate
31 still achieving a comparable or better formfactor than current following into the RE to reduce voltage errors
32 optical peripherals. [45]. The electrochemical cell itself can be modelled in
33 The electrochemical cell formed from electrodes terms of a network of passive electrical components
34 submerged in an electrolytic solution is the basis of an (Figure 2) that is known as Randles equivalent circuit
35 electrochemical sensor [44, 45]. Typically, a circuit called a [47]. The model consists of various impedances to
36 potentiostat is used to probe the cell by applying a represent the different electrochemical phenomena that
37 stimulus voltage or current waveform between two occur within the cell. As shown in Figure 2, a variety of
38
39
Alexander C. Sun received his B.S. degree Electrical and Computer Engineering, Uni-
40 (2012) in Electrical Engineering and Com- versity of California at San Diego, where he is
41 puter Science from the University of Califor- currently an Associate Professor. His research
42 nia, Berkeley, his M.S. degree (2014) and his interests include bioelectronics, biosensors,
43 Ph.D. degree (2018) in Electrical and Com- analog circuit design, medical electronics, and
44 puter Engineering from the University of sensor interfaces. Dr. Hall has been an
California at San Diego. His research focus Associate Editor of the IEEE TRANSAC-
45
is on electrochemical biosensors, electro- TIONS ON BIOMEDICAL INTEGRATED
46 chemical measurement techniques, and com- CIRCUITS since 2015 and has been a mem-
47 pact, low power circuit design for biomedical ber of the CICC Technical Program Commit-
48 and point-of-care devices. tee since 2017. Dr. Hall received the First
49 Place in the Inaugural International IEEE
Drew A. Hall received the B.S. degree in
50 Change the World Competition and First
computer engineering with honors from the
51 Place in the BME-IDEA invention competi-
University of Nevada, Las Vegas, NV, USA,
tion, both in 2009. He received the Analog
52 in 2005, and the M.S. and Ph.D. degrees in
Devices Outstanding Designer Award in
53 electrical engineering from Stanford Univer-
2011, an Undergraduate Teaching Award in
54 sity, Stanford, CA, USA, in 2008 and 2012,
2014, the Hellman Fellowship Award in 2014,
55 respectively. From 2011 to 2013, he was a
and an NSF CAREER Award in 2015. He is
Research Scientist in the Integrated Biosen-
56 also a Tau Beta Pi Fellow.
sors Laboratory at Intel Corporation. Since
57 2013, he has been with the Department of
58

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 3
Review
1 of charging or replacing an extra battery, which ideally
2 would improve patient adherence to testing. The lack of a
3 battery also can reduce the bulkiness and size of the
4 peripheral device. The amount of available power from
5 these interfaces can also be relatively high with USB rated
6 to provide a maximum of 2.5 W. This amount of power is
7 more than enough when designing electrochemical meas-
8 urement circuitry, which at most have been shown to
9 consume 400 mW [74]. However, interfacing with these
10 Fig. 2. Diagram of a simplified biosensor containing a potentiostat ports requires the addition of a USB interface chip (such
and the equivalent Randles circuit model (left). Illustration of
11 as the CH372 or FT232R), which is essentially a
various sensing modalities possible due to changes in the electro-
12
chemical cell (right).
component that handles the lower layer USB protocol
13 between the phone and onboard logic [71, 74]. Further-
14 more, compatibility between smartphone devices is lim-
15 ited when using only the proprietary interfaces. A device
16 different “bioelements” can affect changes in the cell such produced specifically for a certain smartphone type would
17 as charge displacement or accumulation, impeding or need to be redesigned in order to be compatible with a
18 encouraging the rate of reduction/oxidation reactions, or different make or model. For example, a different inter-
19 shifting the bulk ion concentration. It is worth noting that face would be required when switching between old and
20 the need for only the measurement circuitry and an new models of the iPhone, from iPhones to Android
21 electrochemically compatible electrode, which can be phones, from USB-C to USB-B devices, and even between
22 made from such materials as carbon, gold, silver, plati- various feature phones.
23 num, iridium tin oxide, or modified variants and can be
24 fabricated down to a miniature scale (1 mm2–0.045 mm2),
2.1.1.1 USB Communication and Power
25 makes electrochemical biosensors ideal for PoC applica-
26 tions [16, 48–55]. The sensors themselves have been In the literature, the USB 2.0 interface has been used
27 demonstrated to be highly scalable and can be fully much more often than other wired interfaces, i. e. the
28 integrated together with circuitry on the same chip lightning port on Apple devices. Several reasons for this
29 [51, 52, 56–62]. While it is a nonstandard process to skew exist such as Apple requiring licensing of both
30 fabricate electrodes onto an integrated circuit or plate the software and hardware in order to build a peripheral
31 top metal with a compatible material, it still typically device [76] and the usage of Android phones outnumber-
32 requires fewer steps than comparable optical-based chips ing that of iPhones [77]. Hence, this section will focus
33 [63–66]. Furthermore, electrodes have also been shown to solely on the USB interface, which uses a master-slave
34 be adaptable to a variety of different point-of-care settings protocol where the host or master controls all the two-
35 such as on flexible substrates or directly on the skin itself way communication between itself and multiple devices.
36 [67–69]. For USB to be used between a smartphone and a
37 peripheral device, the mobile must have the USB On-
38 The-Go (OTG) protocol built into it and enabled by the
2 Integration with Smartphone
39 firmware. OTG allows for the smartphone to switch
40 A number of different ways to integrate a peripheral between being master or slave depending on what is
41 module to a phone have been developed all with distinct connected at its USB port [78]. While USB is used across
42 trade-offs in terms of available power, data rate, and a variety of different devices, whether OTG is available
43 compatibility with different makes and models. For entirely depends on the make and model of the phone,
44 electrochemical biosensors specifically, the main function- thereby somewhat limiting its cross combability.
45 alities that can be offloaded to the smartphone are power, On the peripheral device itself, the physical USB
46 two-way data transmission, stimulus generation, and signal connector has 4 pins, two for ground and 5 V as well as
47 quantization using the variety of electrical interfaces two for differential data transfer, as is standard with slave
48 available on modern cellphones. devices. The smartphone senses via this hardware con-
49 nection that the peripheral is a slave and automatically
50 provides power to the device. The USB communication
51 2.1 Wired Peripherals protocol is strictly digital requiring that peripheral bio-
52 sensors have onboard digital-to-analog converters (DAC)
2.1.1 Proprietary Interfaces
53 and analog-to-digital converters (ADC) in order to
54 As shown in Figure 3, proprietary ports such as USB or generate the correct voltage stimulus on the device and
55 Apple’s lightning port have been widely used both to quantize the current response before transferring the
56 provide power to the periphral and to send bidirectional results to the phone. Also, the device must implement the
57 data [70–75]. Typically, no additional power source other high level framework, packet structure, and handshake
58 than the phone is required in this case removing the hassle protocol discussed in many technical white papers and

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 4
Review
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28 Fig. 3. Four different peripheral electrochemical biosensors that use the USB interface for data transfer and/or power. (A) A 3.5 3
29 5.7 cm2 sized device powered by a lithium polmer ion battery that transfers chronoampermetry data gathered by a Palm Instruments
30 potentiostat to a Nexus One [70]. The platform is paired with a microfluidic electrode chip to measure a Malaria biomarker. (B) A USB
powered module that uses a custom potentiostat design and screen-printed carbon electrodes [73]. (C) A 8.4 3 5.7 cm2 device that is also
31
USB powered and has has multiple channels [74]. (D) A similar biosensor module that uses a USB on-the-go chip to facilitate
32 communication [75].
33
34
35
36 specification documents [79], by programming the micro- remaining output channel. When used as the interface for
37 controller to handle queries from the host and relying on power and communication for a peripheral device, the
38 a separate USB interface chip paired with a crystal to headphone jack’s main advantages are that it is both the
39 handle the lower level data transfer. At this high level, the only truly universal I/O port that exists on all smartphones
40 protocol allows for a variety of transfer types depending and has remained entirely unchanged throughout many
41 on the frequency, quantity of data, and amount of error generations of smart devices (expect for the iPhone X,
42 correction desired. For electrochemistry, bulk transfers which requires an adapter). Hence, peripherals that use
43 can be used for sending raw ADC values, while interrupt the headphone jack are essentially hardware compatible
44 transfers can be triggered when the sensor has crossed a with all devices that have a three terminal audio port
45 predefined threshold. Overall, the USB port is a widely including laptops, tablets, mp3 players, and older mobile
46 used and familiar interface that offers both well-defined phones. Similar to devices that use the proprietary ports,
47 data transfer and plenty of power. However, these devices taking power from the audio jack means that no battery
48 have limited cross combability when compared to other other than phone’s is required. However, since this port is
49 interfaces to be discussed. AC coupled (bandpass from 20 Hz to 22 kHz) to prevent
50 damaging headphones, no DC signals can pass through
51 requiring both rectification of the audio output for power
2.1.2 Audio Headphone Port
52 and modulation/demodulation of control and data signals.
53 As demonstrated in Figure 4, the 3.5-mm audio port
54 intended for headphones can be used as an interface
2.1.2.1 Power Harvesting
55 between a peripheral device and mobile phone. This port
56 allows for both bidirectional data transfer via the micro- As is demonstrated in prior art, an output sinusoid from
57 phone terminal and one of the output channels (left or the audio channel can be rectified with a MOSFET H-
58 right ear) and power transfer from the phone using the Bridge and Schottky diode to obtain a 1.2–4 V signal that

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 5
Review
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25 Fig. 4. Two peripheral biosensors that transmit assay result data via the audio headphone port. (A) The first device quantizes the
26 measurements on-board and then transmits the signal digitally to the mobile, which then transmits the data over an audio phone call
27 [87]. (B) The second can harvest power from the headphone port and is able to conduct two-way communication between the
28 peripheral and the phone [86]. By frequency modulating the analog output signal from the potentiostat, the data can be quantized on
the phone side via the microphone port.
29
30
31
2.1.2.2 Data Transfer
32 can be further regulated and filtered [43, 80–85]. However,
33 since various makes and models of smartphones use Signal transmission via the audio port can be accom-
34 different drivers, the output characteristics of the head- plished in a number of different ways as shown in
35 phone jacks can vary. Figure 5(A) shows the measured Figure 6. Frequency shift keying (FSK) at ~ 17 bps
36 values of both the output resistance and available power handled by an on-board microcontroller can facilitate
37 for a wide variety of smartphones. The resistances and digital transmission between the smartphone and periph-
38 maximum power available range from 1–20 W and 3– eral device as is demonstrated by Nemiroski et al. and Sun
39 80 mW, respectively. The variation in output resistances et al. [81, 87]. Wang et al. [88] reported a phase-locked
40 can negatively affect the matching between the phone and loop (PLL) (TI CB4046B) with active filters to demodu-
41 the power harvesting circuitry leading to further reduced late a frequency signal from the smartphone to a voltage
42 power. To account for this, a tunable matching network, ramp stimulus for the potentiostat to run cyclic voltamme-
43 such as the one shown in Figure 5(B), can be added to the try. Another PLL is used to modulate the output of the
44 peripheral device in order to change the input impedance potentiostat back to a frequency signal to transmit it to
45 of the harvester and improve overall power transfer the phone to be quantized. This design does not require
46 efficiency. This technique has been demonstrated to an onboard DAC, ADC, or microcontroller, instead off-
47 improve efficiency from 52 % up to 85.4 % [83, 86]. Even loading most of this functionality to the smartphone.
48 with this improvement, however, the power consumption However, the ability to set the test parameters including
49 of biosensors that use the audio port must still be low scan rate and voltage range is limited because the absolute
50 enough in order to remain compatible with most smart- voltage and slope of the ramp are dependent on each
51 phones. As a result, all the peripherals in this space other. In Sun et al. [86], bidirectional data was instead
52 consume the least power (2.5–6.9 mW) when compared to handled by a low power microcontroller sending and
53 other smartphone-based biosensors, aside from those that receiving UART packets in order to set the test parame-
54 use NFC [80, 82, 86]. ters for cyclic voltammetry. The ramp signal was then
55 generated on-board using a PWM generator and integra-
56 tor with feedback to a course ADC to ensure accuracy.
57 The output of the potentiostat was then converted into a
58 frequency by a 555-timer-based voltage-controlled oscil-

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 6
Review
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Fig. 6. Modulated (A) analog signals using a voltage-controlled
33 oscillator and (B) digital signals using frequency shift keying
34 transfered from a peripheral devce to a smartphone [86].
35
36
Fig. 5. (A) A survey of the headphone port output resistance and
37 power available for a set of popular smartphones. (B) Simplified In general, audio port-based peripherals offer almost
38 schematic of typical power harvesting circuitry with transformer, universal cross compatibility at the cost of lower available
39 full-wave rectifier, and linear regulator. Also shown is a diagram power and more complicated analog circuitry compared
40 of the automatically tunable matching network to improve power to USB implementations. Unfortunately, while ports on
41 transfer efficiency [83]. the smartphone are currently widely available, we believe
42 that in the future smartphones could move completely
43 towards wireless interfaces making any wired peripherals
44 lator (VCO) modulated with marker tones to ensure that obsolete in these newer models. Hence, wireless interfaces
45 the measured data lines up with the stimulus voltage. This are a promising future technology for these dongles.
46 frequency signal was then sent to the phone to be
47 quantized. This design is able to independently set the
2.2 Wireless Peripherals
48 voltage range and scan rate, while still consuming low
49 enough power to be powered off of the audio jack. Jiang One of the main benefits of wireless schemes, i. e. Blue-
50 et al., similar to Wang et al., completely does away with tooth and NFC, is that the electronics can be separate
51 using a microprocessor, instead relying on frequency and from the actual electrodes themselves, an especially
52 voltage modulation of the audio output channel along important advantage in wearable-based biosensing sys-
53 with frequency and amplitude detection circuits to control tems where the sensor is intended to be placed directly on
54 the impedance biosensor [82]. This design allows the or near the body enabling long-term monitoring. While
55 peripheral to achieve a low 2.5 mW maximum power the topic of designing the wearable component in these
56 consumption and power itself directly from the audio platforms will not be expanded upon here because the
57 port. breadth of discussion on wearables is extensive and
58 outside the scope of this paper, one can find several

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 7
Review
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28 Fig. 7. Various battery powered peripherals that use Bluetooth for communication. (A) “UWED“ platform uses the RFDUINO to
29 manage the Bluetooth communication and consumes 20 mW with a size of 8 3 4 cm2 to run amperometric techniques [104]. Both (B) an
30 5.338.4 cm2 module that specically runs differential pulse voltammetry [98] and (C) a device for cyclic voltammetry of size 11 3 7 cm2
use the HC-06 shield for handling the wireless protocol [97].
31
32
33
34 detailed and high-quality review papers about the subject general-purpose benchtop potentiostat in many cases. The
35 [89–94]. Hence, these wireless systems can take advantage isolated wireless potentiostat from Pinnacle, Inc. is
36 of the many advances that have been made in the effort to another battery powered (9 V) Bluetooth potentiostat
37 improve the wearability and convenience of electrochem- with a smaller voltage range (4 V) and current range
38 ical sensors. (80 mA) than that of the EmStat, but with two channels.
39 However, all these wireless dongle devices require a
40 battery that must be either replaced or recharged in
2.2.1 Bluetooth
41 addition to the smartphone’s increasing the weight and
42 Recently more and more peripherals have begun using size of the device. Along with an extra battery and
43 Bluetooth to wirelessly transmit data to and from the Bluetooth radio module with antenna (such as HC-06,
44 smartphone, as seen in the examples in Figure 7 [17, 95– HM-10, TI CC2541, Bluegiga WT-12), these types of
45 105]. The obvious benefit is that these types of peripherals peripherals must also generate all the voltage stimulus
46 are compatible with all types of smartphones regardless of signals as well as quantize the potentiostat output on-
47 the make or model. Also, since some companies have board. In this way, platforms that use Bluetooth are
48 been shifting towards replacing most wired interfaces with similar to standalone biosensors and only leverage the
49 wireless ones, using Bluetooth connection appears to be smartphone as a readout tool.
50 more future proof than the headphone port. Companies
51 creating commercial potentiostats have also embraced
2.2.2 NFC
52 Bluetooth technology as a universal method to connect to
53 the smartphone for portable use. The EmStat Blue Alternatively, NFC-based peripherals, such as those
54 (PalmSens) is a battery powered and Bluetooth potentio- shown in Figure 8, also create a wireless data link between
55 stat capable of running a variety of electrochemical the sensor and smartphone, albeit at a much shorter
56 techniques. The 8 V voltage range and large current distance, while also transmitting a small amount of power.
57 measurement range (1 nA–100 mA, 1 pA resolution) The absence of a battery as well as the simplified
58 make it a good replacement for the traditional bulky measurement circuitry required to meet power constraints

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 8
Review
1 output current after amplification and buffering is applied
2 to an electrochromic device that visually changes color
3 when the signal surpasses a certain threshold. Hence, this
4 architecture only provides binary detection of an analyte
5 and does not transfer any data back to the smartphone.
6 The ability to receive power wirelessly allows for the
7 dongle that performs the sensing to be generally much
8 smaller when compared to Bluetooth modules. However,
9 the reduced available power and inability to run when not
10 near a reader limit the use of NFC to specific low-power
11 and low data rate applications where size is the most
12 important factor.
13
14
2.3 Internal Dedicated Hardware
15
16 While most smartphone-based biosensors are meant to be
17 external peripheral devices in order to take advantage of
18 the already saturated mobile phone market, there have
19 been a couple of examples of modules meant to be
20 integrated into the internal hardware of the phone as
21 shown in Figure 9. The desired advantage in this case is
22 that no device additional to the smartphone needs to be
23 managed by the user, which should improve convenience
24 and portability. However, due to the ever decreasing size
25 and available power of modern smartphones, any devices
26
27
28
29
30
31
32
33
34
35
36
37
38
39 Fig. 8. Two examples of NFC tag biosensors. (A) Sensor is in
40 parallel with antenna and detection of anlyte changes its
41 impedance thereby modulating backscatter when measured with
42 an NFC reader [107]. (B) Tag that uses the power from NFC to
run cyclic voltammetry with on-board signal generation. Readout
43
is through an electrochromic device [110].
44
45
46
47 in this scheme enables much smaller device sizes. Often
48 these types of platforms are created by modifying custom
49 flexible RFID tags with a sensing film in series or shunt
50 with the antenna and simply relying on an impedance
51 change caused by detection of the analyte to modulate the
52 backscatter [106–109]. Alternatively, Jung et al. reported
53 an NFC-based biosensor that implements cyclic voltam-
54 metry using circuits made from fabricated single walled
55 carbon nanotube network based thin film transistors [110].
56 This design rectifies the signal from the antenna to power Fig. 9. Photographs of PoC biosensors (A) integrated into the
57 a five-stage ring oscillator that generates the triangular hardware of a smartphone [112] and (B) as a swappable module
58 voltage waveform applied to the modified electrodes. The for a modular phone [111].

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 9
Review
1 integrated internally, especially noncrucial components, circuit similar to the standard transimpedance amplifier
2 must be implemented in a much smaller formfactor and (TIA) or current conveyor based designs [56, 113–129], is
3 consume an inconsequential amount of power relative to used to run electrochemical techniques that require
4 that of other smartphone functions. For example, in terms potential control and current measurement such as cyclic
5 of power, smartphones typically have a lithium-ion battery voltammetry, chronoamperometry, and pulse voltammetry
6 with capacity of approximately 1,500 mAh with idling and each with its own drawbacks and advantages. The
7 a phone call consuming 108 mW and 540 mW, respectively technique chosen typically depends on the application as
8 [111]. well as the type of assay. The measurement circuitry must
9 Guo et al. demonstrates an electrochemical measure- be able to apply a stimulus potential signal between the
10 ment module integrated into a smartphone through an WE and RE that can be any combination of triangle,
11 internal interface for power and communication [112]. square, step, pulse, or sinusoidal waveforms. The applied
12 Similar in design to previously discussed external electro- voltage range is traditionally limited to below 1 V with a
13 chemical dongles but with a small enough formfactor to fit resolution of at least 5 mV. The resulting generated
14 within the limited space inside the phone, this module current waveform is then measured, which corresponds to
15 contains a discrete ADC, DAC, voltage reference, and the amount of the biomarker detected by the sensor. The
16 microcontroller, which communicates to the smartphone signal portion of the waveform is highly dependent on the
17 via I2C. However, this configuration requires access to and assay, concentration range of the analyte, and size of the
18 understanding of the inner workings of the phone, both transducer. For the PoC screen-printed electrodes or
19 hardware and firmware, that often are proprietary and not microelectrodes such signal currents can range anywhere
20 openly available to third-party developers. While Guo has from femtoamperes to microamperes on top of back-
21 this knowledge for this specific smartphone, this imple- ground currents so both sensitivity and dynamic range are
22 mentation is more a proof of concept to demonstrate crucial parameters. Furthermore, the bandwidth require-
23 integrating electrochemical modules into the smartphone ments typically range from sub-Hz levels to 100 kHz, so
24 for possible future development. This module also con- contending with 1/f noise for some techniques is crucial,
25 nects to disposable electrochemical test strips via a slot while for others extending the bandwidth is necessary,
26 near the edge of the phone and runs an amperometric thus integrating more noise. Therefore, in the following
27 measurement technique for the detection of uric acid. sections, various potentiostat implementations designed
28 Sun et al. reports an electrochemical sensing module for a specific measurement mode are discussed. Table 1
29 as well but intended for use in a modular phone summarizes and compares these smartphone-based elec-
30 specifically for the Google Ara Project [111]. Modular trochemical biosensors sorted by interface and the electro-
31 smartphones allow and encourage the end user to custom- chemical techniques they can run.
32 ize the functionality of their mobile device by swapping in
33 various types of hardware modules. This 3.9 3 1.65 cm2
3.1 Amperometry
34 biosensor module with a max power consumption of
35 100 mW contains a reconfigurable, multi-technique poten- For potential controlled current measurement techniques
36 tiostat that repurposes components among three different such as chronoamperometry (step input), cyclic voltam-
37 measurement modes, amperometric, potentiometric, and metry (triangular waveform input), and pulse voltamme-
38 impedance spectroscopic. These multiple electrochemical try (pulse train input), most devices either use a
39 techniques allow for varying assays to be run making the commercially available AFE i. e., TI LMP91000 [130], or a
40 inclusion of a single biosensor module worth the space custom potentiostat circuit. The advantage of LMP91000
41 and power consumed. However, while both Guo’s and is that the entire AFE is completely integrated into a
42 Sun’s approaches eliminate the need for a separate single chip housed in a 4 3 4 mm2 package that consumes
43 external hardware dongle, they suffer from a major ~ 40 mW. The detectable current range for this chip is 5–
44 limitation when compared to the devices that use other 750 mA, which is acceptable for applications such as blood
45 interfaces. Since these designs are intended for a partic- glucose measurement where the analyte concentration is
46 ular smartphone, they are not compatible across different generally high [105], but not sensitive enough for many
47 makes and models unlike other wired or wireless inter- other assays. Instead, when power can be traded for lower
48 faces discussed. Also, the smartphone manufacturers and noise and smaller input bias current, custom potentiostat
49 designers need to allow and be a part of the incorporation circuits with resistive feedback transimpedance amplifiers
50 of the electrochemical module for these schemes to be (R-TIA) are designed to obtain a higher current reso-
51 practical. lution. The open source “CheapStat” potentiostat has a
52 current resolution of 1 nA with tunable gain from 33–
53 165 kW [96, 113]. Fan et al. demonstrates a custom Blue-
3 PoC Implementations of E-Chem Techniques
54 tooth potentiostat that also achieves a current resolution
55 While the interface used by a smartphone-based biosensor of 1 nA and limit-of-detection (LOD) for neuron-specific
56 sets the available resources for the system, the potentio- enolase of 22 pM using differential pulse voltammetry
57 stat design determines the measurement techniques that (DPV) [97]. Sun et al. obtains a ~ 300 pA current
58 can be run and the achievable sensitivity. A potentiostat resolution with a LOD for secretory leukocyte protease

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 10
Review
1 Table 1. Comparative summary of smartphone-based electrochemical biosensors.
2
Interface Technique Resolution Power Power Dimensions Analyte LOD Ref.
3
Source [mW] [cm2] [nM]
4
5 USB CA – lithium – 3.5 3 5.7 (20) PfHRP2 0.267 [70]
6 Pot. – USB – – a-amylase 89 [71]
7 CV – USB – – creatine kinase, 35.3, [72]
transaminase 7.14 ’’
8
CA – USB – – clenbuterol 0.274 [73]
9 CA – USB – – b-hydroxybutyrate 1,000 [75]
10 CA 122 pA USB 50 8.4 3 9.3 (78) hydrogen peroxide ~ [74]
11 Audio Jack CA 500 pA lithium 40.7 5.6 3 11 (59) glucose 500,000 [87]
12 SWASV ’’ ’’ ’’ ’’ heavy metal 61 ’’
13 Pot. ’’’ ’’ ’’ ’’ sodium – ’’
14 CA ’’ ’’ ’’ ’’ PfHRP2 1 ’’
CV – 9V – – nitrate 3 [88]
15
CV 300 pA audio 6.9 3.8 3 6.4 (24) SLPI 1.41 [86]
16 Pot. 3.6 mV audio 6.2 3.8 3 5.8 (22) pH – [81]
17 EIS – audio 2.5 8.1 3 6.6 (53) NeutrAvidin – [82]
18 Bluetooth CV 1,000 pA 9V – – honey – [96]
19 CV – battery – 5.3 3 8.4 (45) glucose 26,000 [98]
20 CV – battery – – white blood cell – [102]
CA 400 pA lithium – 5.9 3 3.3 (19) Ru(III)Cl – [99]
21
CA 6,400 pA lithium 20 8 3 4 (32) ferro/ferri – [104]
22
CA – lithium – 1.8 3 1.9 (3.4) Glucose – [105]
23 DPV 1,000 pA battery – 11 3 7 (77) neuron-specific enolase 0.022 [97]
24 EIS – battery – – E. Coli – [17]
25 EIS – battery – – 2,4,6-trinitrotoluene 709 [95]
26 EIS – battery – – acetone 26,900 [100]
27 EIS – lithium – 10 3 8 (80) blood cell – [101]
EIS – battery – – ferro/ferri – [103]
28
NFC Imp. – wireless – 4.52 (20) VCO – [106]
29 Imp. – wireless – 0.252 (0.06) biogenic amines – [108]
30 Imp. – wireless – 4.1 3 2.4 (9.8) ethenol 22,000 [107]
31 ’’ ’’ ’’ ’’ ’’ KCl 9,000,000 ’’
32 CV – wireless – 9.5 3 5.5 (52) TMPD 20,000 [110]
33 Integrated CA – internal – – uric acid – [112]
34 CV 1,000 pA internal 111 3.9 3 1.7 (6.4) lactoferrin 0.0125 [111]
CA ’’ ’’ ’’ ’’ glucose – ’’
35
EIS ’’ ’’ ’’ ’’ NeutraAvidin 10 ’’
36 Pot. 61 uV ’’ ’’ ’’ pH – ’’
37
38 CV – cyclic voltammetry; Pot. – potentiometric; CA – chronoamperometry; SWASV – sqare-wave anodic stripping voltammetry; EIS –
electrochemical impedance spectroscopy; Imp. – impedance measurement; DPV – differential pulse voltammetry, VCO – volatile
39
organic compounds; TMPD – tetramethyl-p-phenylenediamin
40
41
42 inhibitor (SLPI) of 1.4 nM using CV in an audio-jack buffer (600 fA), that has been used in audio-jack bio-
43 powered peripheral. Jung et al. and Pechlivanidis et al. sensors [81].
44 also both report current resolutions for their Bluetooth
45 based biosensors of 400 pA and 122 pA, respectively
3.3 Electrochemical Impedance Spectroscopy
46 [74, 99].
47 Electrochemical impedance spectroscopy (EIS) is typi-
48 cally measured by applying a small sinusoidal voltage
3.2 Potentiometry
49 stimulus between electrodes and measuring the magnitude
50 Potentiometric measurement circuitry typically only re- and phase of the resulting current signal at multiple
51 quires an amplifier with a large input impedance to frequencies in order to calculate an impedance spectrum.
52 measure the voltage from an ion selective electrode Due to the typical frequency range 1–100 kHz and the low
53 known for its high resistance (10 MW–1 GW). In Zhang 5-mV peak amplitude of the stimulus signal, EIS tends to
54 et al., a USB powered non-inverting amplifier with an be the most power consumptive measurement technique
55 input bias current of ~ 20 nA is used and can achieve a since it needs to measure both magnitude and phase
56 LOD for a-amylase of 89 nM [71]. There also exists a accurately from a small current signal at all frequencies
57 commercial potentiometric measurement chip, i. e. within the spectrum. Hence, most of the PoC oriented EIS
58 LMP91200 [131], which is essentially an ultra-low leakage sensors require power sources such as a 9 V lithium ion

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 11
Review
1 battery. As is true with the other two measurement modes, 3.4 Multi-technique
2 there exists a popular commercial IC solution, AD5933
3 [132], which is able to fit a complete digital signal Combining multiple techniques into a single platform
4 generator (DDS) core, DAC, ADC, and discrete Fourier allows users the versatility to run a variety of different
5 transform (DFT) hardware into a single small chip as is assays all just by changing the type of electrode. Nemir-
6 shown in Figure 10(A) [17, 95, 100]. However, the main oski et al. successfully demonstrates a peripheral device
7 issues with this IC are that the smallest potential stimulus that runs multiple amperometric and potentiometric
8 it can apply is ~ 200 mV peak-to-peak and the bias point techniques all with the same handheld-sized device paired
9 cannot be set independently of the amplitude. In most with a mobile-phone using the headphone jack for data
10 cases, EIS measurements need to be fitted to a linear transfer at a rate of 17 bps [87]. The device contains a
11 impedance model. With a high stimulus amplitude, it can custom designed potentiostat with a current resolution of
12 no longer be assumed that the data matches this linear 0.5 nA along with an on-board microcontroller, 16-bit
13 model. Furthermore, in assays where there are redox ADC, and 16-bit DAC all powered by a 3.7 V lithium
14 reactions, setting the bias point accurately is crucial to the polymer battery with a 210 mAh capacity. The module
15 measurement. Custom EIS circuits have been created to consumes at most 40.7 mW, and the board’s dimensions
16 obtain more accurate results that use on-board ADCs, are 5.6 3 10.6 cm2. It runs chronoamperometry for detec-
17 DACs, and microcontrollers relying on separate batteries tion of glucose, cyclic voltammetry for measuring P.
18 to provide enough power [101, 103]. However, Jiang et al., falciparum (PfHRP2), square-wave anodic stripping vol-
19 as illustrated in Figure 10(B), shifts the generation of the tammetry for heavy metal detection, and potentiometry
20 stimulus signal and quantization to the smartphone for sodium measurements in urine. The device also has a
21 through the headphone jack, negating the need for a vibration motor to increase metal deposition during
22 microcontroller and mixed-signal circuits allowing for a stripping voltammetry thereby amplifying the electro-
23 very low power (2.5 mW) EIS peripheral, which is mostly chemical signal. Sun et al., as discussed in a previous
24 analog, albeit with a frequency range limited to the audio section, also has a custom potentiostat with a 0.5 nA
25 band [82]. current sensitivity that takes up only a fraction (10 %) of
26 the area and consumes slightly more power (49.5 mW in
27 amperometric and 111 mW in impedance spectroscopic
28 modes) [111]. This device manages to implement cyclic
29 voltammetry for measuring lactoferrin for urinary tract
30 infections, chronoamperometry for blood glucose, poten-
31 tiometric for pH of sweat, and impedance spectroscopy
32 for a label-free NeutrAvidin assay.
33
34
4 Conclusion and Outlook for Future
35
36 The current ecosystem of smartphone-based PoC electro-
37 chemical biosensors was explored and the different types
38 of interfaces were compared to demonstrate the trade-offs
39 of each. Overall, compatibility with smartphones, amount
40 of available power, whether an external power source is
41 required, and the sensitivity and capabilities of the
42 potentiostats were examined. It is worth noting here that
43 it is often difficult to get FDA approval for smartphone-
44 based devices since smartphones themselves are not
45 permitted for medical use. Many commercial devices have
46 gotten around this by not using the phone in the critical
47 path and only having it display information. However,
48 more and more platforms are being approved by the FDA
49 that utilize different components of the smart device
50 giving us hope that biosensors based off the technology
51 discussed in this paper can one day become actual medical
52 devices. Some examples of devices that have received
53 approval are the Dip.io (Healthy.io), which is the
Fig. 10. PoC EIS implementations. (A) Bluetooth-based device
54 uses an impedance converter that has on-board signal generation smartphone-based urinalysis kit that uses the third-party
55 and Fourier transform capability [17]. (B) Audio-jack peripheral smartphone camera to measure the color change on a
56 uses the output of the audio channel as the stimulus signal and dipstick and the KardiaBand (AliveCor), a personal EKG
57 relies on the microphone on the smartphone to quantize the that is the first medical device accessory for the Apple
58 output [82]. Watch.

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 12
Review
1 In terms of interfaces, there seems to be more and fully will be accomplished using this current generation of
2 more wireless Bluetooth-based peripheral devices being smartphone-based biosensors. It is our hope that these
3 developed and published compared to wired versions that further improvements will promote the use of specialized,
4 use proprietary ports or the audio-jack. This trend is due portable, and practical medical devices well positioned to
5 to many factors including the increased popularity of be the first line of defense in the future of healthcare.
6 other Bluetooth connected peripherals such as wireless
7 headphones and credit card readers and the availability of
Acknowledgements
8 low-power commercial off the shelf Bluetooth shields that
9 greatly simplify making a biosensor wireless. However, This work was supported in part by Google’s Advanced
10 many Bluetooth implementations do not improve upon Technologies & Projects (ATAP), the US Agency for
11 the formfactor or convenience of their wired counterparts International Development (grant #AID-OAA-A-13-
12 due to the inclusion of an external battery and the need 00002), and the National Institutes of Health, Grant
13 for on-board signal generation. UL1TR000100. The content is solely the responsibility of
14 On the other hand, NFC-based sensors appear to be the authors and does not necessarily represent the official
15 one of the more promising interfaces for future develop- views of the NIH.
16 ment of smartphone-based biosensors that deserve further
17 attention. It allows for power transfer from the phone to
18 greatly reduce the size of the device while maintaining the References
19 convenience of wireless. However, the available power is
20 miniscule (< 1 mW), so innovative designs that either use [1] B. W. Ward, J. S. Schiller, R. A. Goodman, Prev. Chronic.
Dis. 2014, 11.
21 passive sensing techniques or low power integrated circuit [2] S. Murphy, J. Xu, K. Kochanek, S. Curtin, E. Arias, Deaths:
22 potentiostats are the clear next steps in this space. Final Data for 2015, National Center For Health Statistics
23 Eventually, the sensing circuit could be fully mounted on (U.S.). Division Of Vital Statistics, 2017.
24 the electrode itself without needing any additional exter- [3] Health, United States, 2016 With Chartbook on Long-Term
25 nal hardware. Trends in Health, National Center For Health Statistics
26 Furthermore, while each type of peripheral biosensor (U.S.). Centers For Disease Control And Prevention, 2017.
[4] Global Health Estimates 2016: Deaths by Cause, Age, Sex,
27 has its own advantages and optimal use cases, carrying
by Country and by Region, 2000–2016, World Health
28 around multiple dongles for different PoC assays in Organization, Geneva, 2018.
29 burdensome. Hence a device that combines all the [5] D. Gatherer, J. Gen. Virol. 2014, 95, 1619–1624.
30 techniques and benefits together into a single biosensing [6] H. A. Shiwani, R. B. Pharithi, B. Khan, C. B.-A. Egom, P.
31 platform integrated into a smartphone would greatly Kruzliak, V. Maher, E. E.-A. Egom, Asian Pac. J. Trop.
32 improve the practicality of PoC biosensors. However, Med. 2017, 10, 6–10.
33 since the modular smartphone has yet to be realized, the [7] E. Petryayeva, W. Russ Algar, RSC Adv. 2015, 5, 22256–
22282.
34 only possibility for full integration is directly interfacing
[8] L. Bueno, W. R. de Araujo, T. R. L. C. Paixão, in Med.
35 with the internal hardware of standard smartphones. For Biosens. Point Care POC Appl. (Ed.: R.J. Narayan), Wood-
36 this to be possible, a platform that can run multiple head Publishing, 2017, pp. 183–201.
37 techniques must be significantly miniaturized. Hence, the [9] G. Wu, M. H. Zaman, Bull. W. H. O. 2012, 90, 914–920.
38 next steps for a multi-technique smartphone-based bio- [10] P. K. Drain, E. P. Hyle, F. Noubary, K. A. Freedberg, D.
39 sensor module would be to shrink down the analog-front- Wilson, W. R. Bishai, W. Rodriguez, I. V. Bassett, Lancet
40 end and mixed signal circuitry, typically constructed from Infect. Dis. 2014, 14, 239–249.
[11] “Digital in 2018: World’s internet users pass the 4 billion
41 discrete components, into an integrated circuit version,
mark,” can be found under https://wearesocial.com/blog/
42 which would be much smaller and more suitable for 2018/01/global-digital-report-2018, 2018.
43 installing inside a regular non-modular smartphone. Fur- [12] Pew Res. Cent. Internet Sci. Tech 2018.
44 thermore, there is already an extensive amount of prior [13] K. Barboutox, “Key figures – Ericsson Mobility Report,”
45 art on designing integrated potentiostats that contain the can be found under https://www.ericsson.com/en/mobility-
46 analog front-end, quantizers, and digital control on a report/reports/november-2017/key-figures, 2017.
47 single silicon chip. These devices can outperform their [14] M.-P. Gagnon, P. Ngangue, J. Payne-Gagnon, M. Desmartis,
J. Am. Med. Inform. Assoc. 2016, 23, 212–220.
48 discrete counterparts with pA level sensitivities and low
[15] C. Jenkins, N.-S. Burkett, B. Ovbiagele, M. Mueller, S.
49 power consumption in the mW range [133–139]. Hence, on Patel, B. Brunner-Jackson, R. Saulson, F. Treiber, mHealth
50 the electronic technology side, there is a relatively low 2016, 2.
51 barrier of entry to be able to incorporate one of these [16] S. Nayak, N. R. Blumenfeld, T. Laksanasopin, S. K. Sia,
52 high-performance IC potentiostat designs into a smart Anal. Chem. 2017, 89, 102–123.
53 device such as a smartphone, tablet, or smartwatch. [17] J. Jiang, X. Wang, R. Chao, Y. Ren, C. Hu, Z. Xu, G. L.
54 However, to justify the high cost required to fabricate Liu, Sens. Actuators B Chem. 2014, 193, 653–659.
[18] D. Quesada-González, A. Merkoçi, Biosens. Bioelectron.
55 application specific integrated circuits and make these 2017, 92, 549–562.
56 platforms financially viable, there needs to be a high [19] A. Roda, E. Michelini, M. Zangheri, M. Di Fusco, D.
57 amount of consumer interest in them as well as enough Calabria, P. Simoni, TrAC Trends Anal. Chem. 2016, 79,
58 validated and proven biosensing applications that hope- 317–325.

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 13
Review
1 [20] J. Hu, X. Cui, Y. Gong, X. Xu, B. Gao, T. Wen, T. J. Lu, F. [49] K. Ino, Y. Kanno, T. Nishijo, H. Komaki, Y. Yamada, S.
2 Xu, Biotechnol. Adv. 2016, 34, 305–320. Yoshida, Y. Takahashi, H. Shiku, T. Matsue, Anal. Chem.
3 [21] D. N. Breslauer, R. N. Maamari, N. A. Switz, W. A. Lam, 2014, 86, 4016–4023.
D. A. Fletcher, PLoS One 2009, 4, e6320. [50] C. Ma, N. M. Contento, P. W. Bohn, J. Am. Chem. Soc.
4
[22] L. Guan, J. Tian, R. Cao, M. Li, Z. Cai, W. Shen, Anal. 2014, 136, 7225–7228.
5 [51] B. Nasri, T. Wu, A. Alharbi, M. Gupta, R. RanjitKumar, S.
Chem. 2014, 86, 11362–11367.
6 [23] M. K. Kanakasabapathy, H. J. Pandya, M. S. Draz, M. K. Sebastian, Y. Wang, R. Kiani, D. Shahrjerdi, in IEEE Int.
7 Chug, M. Sadasivam, S. Kumar, B. Etemad, V. Yogesh, M. Solid-State Circuits Conf., 2017, pp. 268–269.
8 Safavieh, W. Asghar, Lab Chip 2017, 17, 2910–2919. [52] D. A. Hall, J. S. Daniels, B. Geuskens, N. Tayebi, G. M.
9 [24] S. A. Lee, C. Yang, Lab Chip 2014, 14, 3056–3063. Credo, D. J. Liu, H. Li, K. Wu, X. Su, M. Varma, in IEEE
10 [25] X. Meng, H. Huang, K. Yan, X. Tian, W. Yu, H. Cui, Y. Int. Solid-State Circuits Conf., 2016, pp. 288–289.
Kong, L. Xue, C. Liu, S. Wang, Lab Chip 2016, 17, 104–109. [53] A. Hassibi, T. H. Lee, in IEEE Int. Pap. Solid-State Circuits
11
[26] I. Navruz, A. F. Coskun, J. Wong, S. Mohammad, D. Tseng, Conf., 2005, pp. 564–617 1.
12 [54] B. N. Kim, A. D. Herbst, S. J. Kim, B. A. Minch, M. Lindau,
R. Nagi, S. Phillips, A. Ozcan, Lab Chip 2013, 13, 4015–
13 Biosens. Bioelectron. 2013, 41, 736–744.
4023.
14 [55] S. K. Pandey, P. Singh, J. Singh, S. Sachan, S. Srivastava,
[27] D. Tseng, O. Mudanyali, C. Oztoprak, S. O. Isikman, I.
15 Sencan, O. Yaglidere, A. Ozcan, Lab Chip 2010, 10, 1787– S. K. Singh, Electroanalysis 2016, 28, 2472–2488.
16 1792. [56] A. Manickam, A. Chevalier, M. McDermott, A. D. Elling-
17 [28] H. Zhu, O. Yaglidere, T.-W. Su, D. Tseng, A. Ozcan, Lab ton, A. Hassibi, IEEE Trans. Biomed. Circuits Syst. 2010, 4,
Chip 2011, 11, 315–322. 379–390.
18
[57] M. J. Milgrew, M. O. Riehle, D. R. S. Cumming, Sens.
19 [29] Q. Wei, H. Qi, W. Luo, D. Tseng, S. J. Ki, Z. Wan, Z.
Actuators B Chem. 2005, 111–112, 347–353.
20 Göröcs, L. A. Bentolila, T.-T. Wu, R. Sun, ACS Nano 2013,
[58] J. Rothe, O. Frey, A. Stettler, Y. Chen, A. Hierlemann,
21 7, 9147–9155.
Anal. Chem. 2014, 86, 6425–6432.
[30] Z. J. Smith, K. Chu, A. R. Espenson, M. Rahimzadeh, A.
22 [59] B. Lim, S. Takahashi, M. Futagawa, F. Dasai, M. Ishida, K.
Gryshuk, M. Molinaro, D. M. Dwyre, S. Lane, D. Matthews,
23 Sawada, in 2014 IEEE Biomed. Circuits Syst. Conf., 2014,
S. Wachsmann-Hogiu, PLoS One 2011, 6, e17150.
24 pp. 113–116.
[31] S. Cho, A. Islas-Robles, A. M. Nicolini, T. J. Monks, J.-Y.
25 [60] K. Niitsu, S. Ota, K. Gamo, H. Kondo, M. Hori, K.
Yoon, Biosens. Bioelectron. 2016, 86, 697–705.
Nakazato, IEEE Trans. Biomed. Circuits Syst. 2015, 9, 607–
26 [32] F. Giavazzi, M. Salina, E. Ceccarello, A. Ilacqua, F. Damin,
619.
27 L. Sola, M. Chiari, B. Chini, R. Cerbino, T. Bellini, Biosens. [61] B. C. Cheah, A. I. Macdonald, C. Martin, A. J. Streklas, G.
28 Bioelectron. 2014, 58, 395–402. Campbell, M. A. Al-Rawhani, B. Nemeth, J. P. Grant, M. P.
29 [33] S. C. Kim, U. M. Jalal, S. B. Im, S. Ko, J. S. Shim, Sens. Barrett, D. R. S. Cumming, IEEE Trans. Biomed. Circuits
Actuators B Chem. 2017, 239, 52–59. Syst. 2016, 10, 721–730.
30
[34] E. Lebiga, R. E. Fernandez, A. Beskok, Analyst 2015, 140, [62] A. C. Sun, E. Alvarez-Fontecilla, A. G. Venkatesh, E.
31
5006–5011. Aronoff-Spencer, D. A. Hall, IEEE J. Solid-State Circuits
32 [35] V. Oncescu, D. O’Dell, D. Erickson, Lab Chip 2013, 13, 2018, 1–11.
33 3232–3238. [63] A. Hassibi, R. Singh, A. Manickam, R. Sinha, B. Kuimelis,
34 [36] A. Roda, M. Guardigli, D. Calabria, M. M. Calabretta, L. S. Bolouki, P. Naraghi-Arani, K. Johnson, M. McDermott,
35 Cevenini, E. Michelini, Analyst 2014, 139, 6494–6501. N. Wood, in IEEE Int. Solid-State Circuits Conf., 2017, pp.
36 [37] L. Shen, J. A. Hagen, I. Papautsky, Lab Chip 2012, 12, 68–69.
37 4240–4243. [64] L. Hong, H. Li, H. Yang, K. Sengupta, IEEE J. Solid-State
38 [38] A. K. Yetisen, N. Jiang, A. Tamayol, G. U. Ruiz-Esparza, Circuits 2017, 52, 2388–2406.
Y. S. Zhang, S. Medina-Pando, A. Gupta, J. S. Wolffsohn, [65] S. Koppa, Y. Joo, in 2013 IEEE Sens., 2013, pp. 1–4.
39
H. Butt, A. Khademhosseini, Lab Chip 2017, 17, 1137– [66] B. Jang, P. Cao, A. Chevalier, A. Ellington, A. Hassibi, in
40 1148. IEEE Int. Solid-State Circuits Conf., 2009, pp. 436-437.
41 [39] D. Gallegos, K. D. Long, H. Yu, P. P. Clark, Y. Lin, S. [67] J. Kim, I. Jeerapan, S. Imani, T. N. Cho, A. Bandodkar, S.
42 George, P. Nath, B. T. Cunningham, Lab Chip 2013, 13, Cinti, P. P. Mercier, J. Wang, ACS Sens. 2016, 1, 1011–1019.
43 2124–2132. [68] W. Gao, S. Emaminejad, H. Y. Y. Nyein, S. Challa, K. Chen,
44 [40] H. Guner, E. Ozgur, G. Kokturk, M. Celik, E. Esen, A. E. A. Peck, H. M. Fahad, H. Ota, H. Shiraki, D. Kiriya, Nature
45 Topal, S. Ayas, Y. Uludag, C. Elbuken, A. Dana, Sens. 2016, 529, 509–514.
46 Actuators B Chem. 2017, 239, 571–577. [69] A. Modali, S. R. K. Vanjari, D. Dendukuri, Electroanalysis
[41] X. Wang, T.-W. Chang, G. Lin, M. R. Gartia, G. L. Liu, 2016, 28, 1276–1282.
47
Anal. Chem. 2017, 89, 611–615. [70] P. B. Lillehoj, M.-C. Huang, N. Truong, C.-M. Ho, Lab Chip
48 2013, 13, 2950–2955.
[42] Y. Wang, X. Liu, P. Chen, N. T. Tran, J. Zhang, W. S. Chia,
49 S. Boujday, B. Liedberg, Analyst 2016, 141, 3233–3238. [71] L. Zhang, W. Yang, Y. Yang, H. Liu, Z. Gu, Analyst 2015,
50 [43] S. Choi, Biotechnol. Adv. 2016, 34, 321–330. 140, 7399–7406.
51 [44] Allen J. Bard, Larry R. Faulkner, Electrochemical Methods [72] W. Deng, Y. Dou, P. Song, H. Xu, A. Aldalbahi, N. Chen,
52 Fundamentals and Applications, Wiley, 2001. N. N. El-Sayed, J. Gao, J. Lu, S. Song, J. Electroanal. Chem.
53 [45] J. Wang, Analytical Electrochemistry, John Wiley & Sons, 2016, 777, 117–122.
Inc., 2006. [73] Y. Dou, Z. Jiang, W. Deng, J. Su, S. Chen, H. Song, A.
54
[46] N. J. Ronkainen, H. B. Halsall, W. R. Heineman, Chem. Aldalbahi, X. Zuo, S. Song, J. Shi, J. Electroanal. Chem.
55 2016, 781, 339–344.
Soc. Rev. 2010, 39, 1747–1763.
56 [74] N. G. Pechlivanidis, K. I. Papadimitriou, D. Evans, N.
[47] J. E. B. Randles, Discuss. Faraday Soc. 1947, 1, 11–19.
57 Vasilakis, T. Prodromakis, in IEEE Int. Symp. Circuits Syst.,
[48] M. Hu, I. Fritsch, Anal. Chem. 2015, 87, 2029–2032.
58 2017, pp. 1–4.

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 14
Review
1 [75] J. Guo, Anal. Chem. 2017, 89, 8609–8613. [104] A. Ainla, M. P. S. Mousavi, M.-N. Tsaloglou, J. Redston,
2 [76] “MFi Program - Apple Developer,” can be found under J. G. Bell, M. T. Fernández-Abedul, G. M. Whitesides, Anal.
3 https://developer.apple.com/programs/mfi/, 2018. Chem. 2018, 90, 6240–6246.
[77] “Mobile Operating System Market Share Worldwide,” can [105] A. J. Bandodkar, S. Imani, R. Nuñez-Flores, R. Kumar, C.
4
be found under http://gs.statcounter.com/os-market-share/ Wang, A. M. V. Mohan, J. Wang, P. P. Mercier, Biosens.
5
mobile/worldwide, 2018. Bioelectron. 2018, 101, 181–187.
6 [78] USB Implementers Forum, Inc., “USB On-The-Go and [106] J. M. Azzarelli, K. A. Mirica, J. B. Ravnsbæk, T. M. Swager,
7 Embedded Host,” can be found under http://www.usb.org/ Proc. Natl. Acad. Sci. 2014, 111, 18162–18166.
8 developers/onthego/, 2018. [107] G. Xu, Q. Zhang, Y. Lu, L. Liu, D. Ji, S. Li, Q. Liu, Sens.
9 [79] USB Implementers Forum, Inc., “USB.org – USB 2.0 Actuators B Chem. 2017, 246, 748–755.
10 Documents,” can be found under http://www.usb.org/devel- [108] L. K. Fiddes, J. Chang, N. Yan, Sens. Actuators B Chem.
11 opers/docs/usb20_docs/, 2017. 2014, 202, 1298–1304.
[80] A. Sun, T. Wambach, A. G. Venkatesh, D. A. Hall, in IEEE [109] L. K. Fiddes, N. Yan, Sens. Actuators B Chem. 2013, 186,
12
Biomed. Circuits Syst. Conf., 2014, pp. 312–315. 817–823.
13
[81] A. Sun, T. Phelps, C. Yao, A. G. Venkatesh, D. Conrad, [110] Y. Jung, H. Park, J.-A. Park, J. Noh, Y. Choi, M. Jung, K.
14 D. A. Hall, Sensors 2017, 17, 1245. Jung, M. Pyo, K. Chen, A. Javey, Sci. Rep. 2015, 5,
15 [82] H. Jiang, A. Sun, A. G. Venkatesh, D. A. Hall, IEEE Sens. srep08105.
16 J. 2017, 17, 589–597. [111] A. Sun, A. G. Venkatesh, D. A. Hall, IEEE Trans. Biomed.
17 [83] C. Yao, A. Sun, D. A. Hall, in IEEE Glob. Humanit. Circuits Syst. 2016, 10, 945–954.
18 Technol. Conf., 2015, pp. 219–225. [112] J. Guo, Anal. Chem. 2016, 88, 11986–11989.
19 [84] T. Phelps, H. Jiang, D. A. Hall, in IEEE Eng. Med. Biol. [113] A. A. Rowe, A. J. Bonham, R. J. White, M. P. Zimmer, R. J.
20 Soc. Conf, 2017, pp. 3297–3300. Yadgar, T. M. Hobza, J. W. Honea, I. Ben-Yaacov, K. W.
[85] Y. S. Kuo, T. Schmid, P. Dutta, ACM DEV ’10 Proceedings Plaxco, PLoS One 2011, 6, e23783.
21
of the First ACM Symposium on Computing for Develop- [114] M. Vergani, M. Carminati, G. Ferrari, E. Landini, C.
22 ment Article No. 24 London, United Kingdom – December Caviglia, A. Heiskanen, C. Comminges, K. Zor, D. Sabour-
23 17–18, 2010 doi > 10.1145/1926180.1926210. in, M. Dufva, IEEE Trans. Biomed. Circuits Syst. 2012, 6,
24 [86] A. C. Sun, C. Yao, A. G. Venkatesh, D. A. Hall, Sens. 498–507.
25 Actuators B Chem. 2016, 235, 126–135. [115] L. Li, X. Liu, W. A. Qureshi, A. J. Mason, IEEE Trans.
26 [87] A. Nemiroski, D. C. Christodouleas, J. W. Hennek, A. A. Biomed. Circuits Syst. 2011, 5, 439–448.
27 Kumar, E. J. Maxwell, M. T. Fernández-Abedul, G. M. [116] S. Hwang, S. Sonkusale, Ultra Low Impedance CMOS
28 Whitesides, Proc. Natl. Acad. Sci. 2014, 111, 11984–11989. Potentiostat for Environmental Sensing Applications, IEEE
[88] X. Wang, M. R. Gartia, J. Jiang, T.-W. Chang, J. Qian, Y. Sensors Journal, vol. 10, no. 4, pp. 820–821.
29
Liu, X. Liu, G. L. Liu, Sens. Actuators B Chem. 2015, 209, [117] I. Ramfos, N. Vassiliadis, S. Blionas, K. Efstathiou, A.
30 677–685. Fragoso, C. K. O’Sullivan, A. Birbas, Biosens. Bioelectron.
31 [89] J. R. Windmiller, J. Wang, Electroanalysis 2013, 25, 29–46. 2013, 47, 482–489.
32 [90] A. J. Bandodkar, W. Jia, J. Wang, Electroanalysis 2015, 27, [118] C. Ionescu, P. Svasta, C. Tamas, C. Bala, L. Rotariu, in
33 562–572. IEEE Des. Technol. Electron. Packag. Symp., 2010, pp. 215–
34 [91] J. Heikenfeld, Electroanalysis 2016, 28, 1242–1249. 218.
35 [92] D. Pankratov, E. González-Arribas, Z. Blum, S. Shleev, [119] A. F. D. Cruz, N. Norena, A. Kaushik, S. Bhansali, Biosens.
36 Electroanalysis 2016, 28, 1250–1266. Bioelectron. 2014, 62, 249–254.
[93] M. D. Steinberg, P. Kassal, I. M. Steinberg, Electroanalysis [120] C.-Y. Huang, H.-Y. Lin, Y.-C. Wang, W.-Y. Liao, T.-C.
37
2016, 28, 1149–1169. Chou, in IEEE Asia-Pac. Conf. Circuits Syst., 2004, pp. 633–
38
[94] A. J. Bandodkar, D. Molinnus, O. Mirza, T. Guinovart, J. R. 636 vol. 2.
39 Windmiller, G. Valdés-Ramı́rez, F. J. Andrade, M. J. Schön- [121] C.-Y. Huang, M.-H. Lee, Z.-H. Wu, H.-Y. Tseng, Y.-C.
40 ing, J. Wang, Biosens. Bioelectron. 2014, 54, 603–609. Huang, B.-D. Liu, H.-Y. Lin, in IEEE Circuits Syst. Int.
41 [95] D. Zhang, J. Jiang, J. Chen, Q. Zhang, Y. Lu, Y. Yao, S. Li, Conf. Test. Diagn., 2009, pp. 1–4.
42 G. Logan Liu, Q. Liu, Biosens. Bioelectron. 2015, 70, 81–88. [122] J. R. Blanco, F. J. Ferrero, J. C. Campo, J. C. Anton, J. M.
43 [96] G. F. Giordano, M. B. R. Vicentini, R. C. Murer, F. Augusto, Pingarron, A. J. Reviejo, J. Manso, in Proc. IEEE Instrum.
44 M. F. Ferrão, G. A. Helfer, A. B. da Costa, A. L. Gobbi, Meas. Technol. Conf., 2006, pp. 690–694.
L. W. Hantao, R. S. Lima, Electrochim. Acta 2016, 219, [123] M. M. Ahmadi, G. A. Jullien, IEEE Trans. Circuits Syst.
45
170–177. Regul. Pap. 2009, 56, 1339–1348.
46
[97] Y. Fan, J. Liu, Y. Wang, J. Luo, H. Xu, S. Xu, X. Cai, [124] B. Goldstein, D. Kim, J. Xu, T. K. Vanderlick, E. Culurciel-
47 Biosens. Bioelectron. 2017, 95, 60–66. lo, IEEE Trans. Biomed. Circuits Syst. 2012, 6, 111–119.
48 [98] D. Ji, L. Liu, S. Li, C. Chen, Y. Lu, J. Wu, Q. Liu, Biosens. [125] R. G. Kakerow, H. Kappert, E. Spiegel, Y. Manoli, in 8th
49 Bioelectron. 2017, 98, 449–456. Int. Conf. Solid-State Sens. Actuators, 1995, pp. 142–145.
50 [99] J. Jung, J. Lee, S. Shin, Y. T. Kim, Sensors 2017, 17, 2416. [126] T. Luo, H. Wang, H. Song, J. B. Christen, in 2014 IEEE
51 [100] L. Liu, D. Zhang, Q. Zhang, X. Chen, G. Xu, Y. Lu, Q. Liu, Biomed. Circuits Syst. Conf., 2014, pp. 336–339.
52 Biosens. Bioelectron. 2017, 93, 94–101. [127] S. M. Martin, F. H. Gebara, T. D. Strong, R. B. Brown,
53 [101] N. Talukder, A. Furniturewalla, T. Le, M. Chan, S. Hirday, IEEE Sens. J. 2009, 9, 135–142.
X. Cao, P. Xie, Z. Lin, A. Gholizadeh, S. Orbine, Biomed. [128] H. S. Narula, J. G. Harris, in Int. Symp. Circuits Syst., 2004,
54
Microdevices 2017, 19, 36. pp. I-457–60.
55 [102] X. Wang, G. Lin, G. Cui, X. Zhou, G. L. Liu, Biosens. [129] R. F. B.Turner, D. J. Harrison, H. P. A Baltes, CMOS poten-
56 Bioelectron. 2017, 90, 549–557. tiostat for amperometric chemical sensors IEEE J. Solid-
57 [103] L. Yang, T. Chen, in IEEE Biomed. Circuits Syst. Conf., State Circuits, (1987), pp. 473–478.
58 2017, pp. 1–4.

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 15
Review
1 [130] Texas Instruments, “Configurable AFE Potentiostat for [136] M. H. Nazari, H. Mazhab-Jafari, L. Leng, A. Guenther, R.
2 Low-Power Chemical Sensing Applications,” LMP91000 Genov, IEEE Trans. Biomed. Circuits Syst. 2013, 7, 338–
3 datasheet, December 2014. 348.
[131] Texas Instruments, “Integrated AFE for Low-Power pH [137] M. Punjiya, C. H. Moon, S. S. Nanolab, in IEEE Int. Symp.
4
Sensing Applications,” LMP91200 datasheet, February Circuits Syst., 2016, pp. 2883–2886.
5 2016. [138] M. M. Ahmadi, G. A. Jullien, IEEE Trans. Biomed. Circuits
6 [132] Analog Devices, “1 MSPS, 12-Bit Impedance Converter, Syst. 2009, 3, 169–180.
7 Network Analyzer,” AD5933 datasheet, April 2017. [139] M. Razzaghpour, S. Rodriguez, E. Alarcon, A. Rusu, in
8 [133] M. H. Nazari, M. Mujeeb-U-Rahman, A. Scherer, in IEEE IEEE Biomed. Circuits Syst. Conf., 2011, pp. 5–8.
9 Symp. VLSI Circuits, 2014, pp. 1–2.
10 [134] A. Agarwal, A. Gural, M. Monge, D. Adalian, S. Chen, A.
Scherer, A. Emami, in IEEE Symp. VLSI Circuits, 2017, pp.
11
C108–C109. Received: July 11, 2018
12 [135] L. Zuo, S. K. Islam, I. Mahbub, F. Quaiyum, IEEE Trans. Accepted: August 23, 2018
13 Circuits Syst. II Express Briefs 2015, 62, 204–208. Published online on November 26, 2018
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58

www.electroanalysis.wiley-vch.de 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Electroanalysis 2019, 31, 2 – 16 16