PROOF COPY [JESP-18-4491] 0061909JES

Journal of The Electrochemical Society, 166 (9) B1-B10 (2019) B1

1 JES FOCUS ISSUE ON 4D MATERIALS AND SYSTEMS

2 ABE-Stat, a Fully Open-Source and Versatile Wireless Potentiostat
3 Project Including Electrochemical Impedance Spectroscopy
4 Daniel M. Jenkins, 1,∗,z Bog Eum Lee,2 Soojin Jun,2 José Reyes-De-Corcuera,3
5 and Eric S. McLamore4
6
1 University of Hawaii Mānoa, Molecular Biosciences and Bioengineering, Honolulu, Hawaii, USA 96822
7
2 University of Hawaii Mānoa, Human Nutrition, Food, and Animal Sciences, Honolulu, Hawaii 96822, USA
8
3 University of Georgia, Food Science and Technology, Athens, Georgia 30602, USA
9
4 University of Florida, Agricultural and Biological Engineering, Gainesville, Florida 32611, USA
10

11 Exciting discoveries in material science and molecular interactions are resulting in many promising electrochemical biosensor
12 technologies. Compact, high-quality instrumentation is critical to adaptation of these new technologies especially for distributed
13 applications in the agriculture and food industries. To this end, we have developed ABE-Stat, a fully open-source, battery-powered
14 potentiostat project including a wireless Android interface. ABE-Stat is capable of conducting routine electrochemical analyses
15 including cyclic voltammetry (CV), differential pulse voltammetry (DPV), high impedance potentiometric measurements, and can
16 connect directly to the internet through WiFi or indirectly through the Android interface. Importantly it is the first fully open-
17 source potentiostat capable of evaluating electrochemical impedance spectroscopy (EIS) across a wide frequency spectrum (0.1 Hz to
18 100 kHz) with user selectable amplitude and bias. Current noise was observed to be over an order of magnitude larger than the nominal
19 resolution of the embedded 24-bit analog to digital converter (ADC), but were largely consistent with the actual ADC specifications.
20 In this manuscript we share detailed documentation for ABE-Stat including hardware design and source code, and evaluation of the
21 performance of all avaiable analyses. We also suggest design improvements that could improve the noise performance of ABE-Stat
22 and consistency of EIS measurements across the spectrum.
23 © The Author(s) 2019. Published by ECS. This is an open access article distributed under the terms of the Creative Commons
24 Attribution 4.0 License (CC BY, http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse of the work in any
25 medium, provided the original work is properly cited. [DOI: 10.1149/2.0061909jes]
26

27 Manuscript submitted January 2, 2019; revised manuscript received February 25, 2019. Published 00 0, 2019. This paper is part of
28 the JES Focus Issue on 4D Materials and Systems.

29 Rapid advances in the understanding of molecular interactions,1–4 technologies for food, agriculture and the environment. Our design is 63

30 development of novel materials,5–10 and stabilization of biologically largely inspired by the component selection and layout of DStat by 64

31 active compounds11–13 is increasing the scope for practical electro- Dryden and Wheeler,45 but like some other projects38,46–48 uses a wire- 65

32 analytical tools in the field.14,15 These advances coupled with sim- less/smart phone interface to improve portability of the system. This 66

33 ple electrochemical techniques are especially compelling for diag- inherently WiFi-connected platform facilitates sharing and aggrega- 67

34 nostic applications related to food,4,6,7,16–21 agriculture,16,22–28 and the tion of results and inclusion of geospatial and temporal metadata. It 68
35 environment22,28–31 that are highly sensitive to cost. In such applica- also potentiates the facile incorporation of new machine learning tools 69

36 tions effective management of resources requires molecular and bi- for classification of results and automated decision support,48,49 and 70

37 ological characterization across ecological scales and throughout a facilitates the integration of field data into the Internet of Things that 71

38 complex supply chain. is expected to have a tremendous impact in the food supply chain. Our 72

39 In our experience, unavailability of high quality, affordable ana- design is also the first such open-source project that has incorporated 73

40 lytical hardware/software can be a considerable barrier to commer- hardware and software features to enable EIS, with arbitrary DC bias 74

41 cialization of new diagnostic technologies.32–34 Our objective here and programmable AC amplitude. A parametric comparison of the 75

42 is to make available a new platform to facilitate the commercializa- ABE-Stat to other notable custom potentiostat networks described in 76

43 tion of electrochemistry-based diagnostics for food safety and qual- the recent literature is presented in Table I. 77

44 ity, agriculture and the environment, or for teaching and scientific in-
45 vestigations in under-resourced settings. An important sub-objective
46 is to extend capabilities of previously documented open-source po- Experimental 78
47 tentiostat projects by incorporating a wireless interface and enabling
48 electrochemical impedance spectroscopy (EIS). Numerous reports of Hardware design/approach.—The hardware design of the ABE- 79

49 customized devices for electrochemical analyses have been published Stat is largely inspired by Dryden and Wheeler’s DStat,45 with a 80

50 recently.35,36 Some of these approaches have leveraged powerful soft- versatile potentiostat network configurable through a series of high 81

51 ware tools available in proprietary platforms such as LabVIEW37 or impedance electronically addressable analog switches, significant at- 82

52 MATLAB,38,39 though for greater accessibility many devices are built tention to isolating noise of digital sections of the device from sensitive 83

53 on open-source hardware/software platforms such as Arduino,40,41 analog signals, and use of high resolution sensors (24 bit Sigma-Delta 84

54 single-chip potentiostats42 or commercially available systems-on-a- Analog to Digital Converter (ADC); ADS1220, Texas Instruments, 85

55 chip.43 Landmarks in the development of fully documented open- Dallas, TX, USA) and sources (16 bit Digital to Analog Converter 86

56 source potentiostat projects include the “CheapStat” from Rowe (DAC); AD5061, Analog Devices, Norwood MA, USA) communi- 87

57 et al.,44 fully operational versions of which are directly available at an cating to the microcontroller through the Serial Peripheral Interface 88

58 affordable price, and the “DStat” from Dryden and Wheeler45 which (SPI). The primary innovation of the analog network of ABE-Stat rela- 89

59 established a benchmark of performance for voltammetric applications tive to DStat is the incorporation of a network analyzer chip (AD5933, 90

60 rivalling proprietary commercial platforms. In the same spirit, we doc- Analog Devices) and reconfiguration of the control amplifier network 91

61 ument here a fully open-source potentiostat project “ABE-Stat”, which to enable EIS across the spectrum from 0.1 Hz to 100 kHz, with ar- 92

62 we hope can facilitate rapid commercialization of portable diagnostic bitrary bias and any of 4 pre-programmed AC amplitudes (Figure 1). 93

This limitation is constrained by the four preset amplitudes available 94

from the output of the AD5933 network analyzer used to evaluate EIS 95

∗ Electrochemical Society Member. at high frequencies, but the user can easily customize the AC am- 96
z
E-mail: [email protected] plitudes available for specialized applications by changing a resistor 97
 B2
Table I. Feature comparison of ABE-Stat to other recent notable custom potentiostats documented in peer-reviewed literature.

Journal of The Electrochemical Society, 166 (9) B1-B10 (2019)

PC/ Fully Fully
High-impedance Electrochemical 2 and 3 Wireless/ Windows/ documented documented Integration
Tool (year Voltammetric open-source Impedance electrode battery Linux/ Wi-Fi Language open-source open-source of machine
documented) applications potentiometry Spectroscopy configurations Android iOS powered USB Bluetooth enabled support hardware software learning
√ √ √ √ √ √ √ √ √ √ √
ABE-Stat (this ✗ ✗ ✗
work)
√ √ √ √ √ √
DStat (2015)45 ✗ ✗ ✗ ✗ ✗ ✗ ✗ ✗
√ √ √ √ √
CheapStat ✗ ✗ ✗ ✗ ✗ ✗ ✗ ✗ ✗
(2011)44
√ √ √ √ √ √ √§ √
UWED (2018)59 ✗ ✗ ✗ ✗ ✗ ✗
√ √ √
Pruna et al. ✗ ✗ ✗ ✗ ✗ ✗ ✗ ✗ ✗ ✗ ✗
(2018)39
√ √ √ √† √‡ √‡ √
uMED (2014)38 ✗ ✗∗ ✗∗ ✗ ✗ ✗ ✗
√ √ √ √ √ √ √§ √ √ √
“TongueMetrix” ✗ ✗ ✗ ✗
(2016)48,¶

∗ Digital information is communicated through frequency-shift keying over the audio jack of a generic phone and shuttled to a MATLAB program through Voice Over Internet Protocol, which returns processed
results back to phone by SMS.
† Device is wireless and battery powered, but requires cabled connection to audio jack of phone.
‡ Device connects to remote PC through generic phone and Voice Over Internet Protocol (VOIP).
¶ Builds on “CheapStat” as potentiostat hardware.
§ Wi-Fi enabled by virtue of interface to Android devices.
 Journal of The Electrochemical Society, 166 (9) B1-B10 (2019) B3

Figure 1. Abridged schematic of ABE-Stat analog networks. Analog switches (ADG715, ADG779, and ADG704) are controlled by ESP8266 (not shown). This
includes options to configure the electrode network (2 or 3 electrode, or open circuit) and composition of the excitation signal (from DAC and/or digital synthesizer
output of network analyzer AD5933) using single pole switches (ADG715). Another set of analog switches control the routing of working current to electrometer-
grade transimpedance amplifier (LMP7721) and 24-bit ADC (ADS1220), or alternatively to the inputs of the network analyzer IC (AD5933). The transimpedance
gain is controlled by one of four resistors selected by the switch ADG704.

98 value (i.e. resistor at top left of Figure 1) into the summing network to tary materials (Files S1) which also include a text file with links to the 133

99 the control amplifier. For system control we used the ESP8266 WiFi material on the internet (GitHub Inc., San Francisco, CA, USA). 134

100 microcontroller system (ESP12S Module, Espressif Systems, Shang-

101 hai, China) with embedded WiFi radio, 80 MHz 32-bit microprocessor Software design/approach.—While the ABE-Stat hardware is 135
102 and 4 MB of flash/programmable memory. For simple interfacing to a inherently WiFi-enabled through the ESP8266 microcontroller and 136
103 custom Android app described below, the hardware device also incor- could easily be automated to continuously upload data to a server, 137
104 porates a classic serial Bluetooth module (RN42, Microchip, Chan- for general purpose uses we have developed an Android app to inter- 138
105 dler AZ, USA). Due to the limited hardware IO of the ESP8266, a face with the device over Bluetooth (Figure 2). The app (ABE-Stat) 139
106 series of three 8-bit shift registers (74HC595, Fairchild Semiconduc- is freely available through Google Play, and the source code is in- 140
107 tor, Sunnyvale, CA, USA) are used to generate control signals for cluded as supplemental material (Android code as Files S2; Arduino 141
108 various chip selects and to route various network signals including to firmware operating on the hardware as Files S3) including text files 142
109 set the transimpedance amplifier gain and routing of the working cur- with links to most recent versions on GitHub. A quick guide to con- 143
110 rent through either a custom electrometer grade amplifier (LMP7721, necting and using ABE-Stat, including instructions for updating the 144
111 Texas Instruments) to the system ADC or to the amplifier terminals firmware and photographs of the populated PCB, is shared as sup- 145
112 of the network analyzer. Configuration of the electrode network and plemental material File S4. In general all commands originate from 146
113 the composition of the control signal is achieved by setting the states the Android side and conform to simple coded instructions. The hard- 147
114 of 8 analog switches (ADG715 octal analog switch, Analog Devices) ware then responds accordingly and returns coded data resulting from 148
115 through Inter-Integrated Circuit (i2c) communication. The i2c port is the requested analyses. Currently the interface defaults to a “multi- 149
116 also used to control and read data from the network analyzer AD5933. meter”-like interface when the device first connects, displaying the 150
117 Electrodes are connected to the device through spring loaded terminals open source potential on the electrode network, as well as device and 151
118 (COMBICON PTSM Series connectors, Phoenix Contact, Blomberg, thermocouple temperatures. Thermocouple temperature is estimated 152
119 Germany) which are also used to optionally connect a thermocou- assuming a type K thermocouple using software compensation rel- 153
120 ple for temperature measurement. A spare two outputs from the shift ative to the temperature value recorded on the ADC chip, and ITS- 154
121 registers are available and could potentially be used for feedback con- 90 standard polynomial coefficients.50 The user can then select other 155
122 trol of temperature for example. To enable truly portable operation, measurements including network current for a user defined potential 156
123 ABE-Stat can be powered directly with a single-cell lithium battery, (and electrode configuration), or resistance or capacitance for user 157
124 and the hardware includes a charging circuit for the battery that is defined electrode configuration. For more detailed analyses and/or 158
125 powered through a micro-USB jack. The battery connector is com- to create an electronic record of data the user can use the options 159
126 patible with many generic lithium polymer batteries resold through menu to create and edit methods based on several common techniques 160
127 a variety of electronics suppliers (i.e. we use a 2000 mA-hr battery, including Cyclic Voltammetry (CV), Differential Pulse Voltammetry 161
128 product ID 2011, from Adafruit, New York City, NY, USA). The full (DPV), Electrochemical Impedance Spectroscopy (EIS), and a “Static 162
129 EAGLE (Autodesk Inc., San Rafael, CA, USA) schematic and layout Logging” function suitable for implementing chronoamperometry or 163
130 files for the 4-layer printed circuit board (PCB) measuring 2.9 in × monitoring of potentiometric sensors, which records cell potential, 164
131 3.5 in (∼74 mm × 89 mm), along with gerber fabrication files, bill of current, and temperatures under static conditions for configurable 165
132 materials, and centroid files for assembly are available as supplemen- network (open circuit, or 2 or 3 electrodes with user defined bias). 166
 B4 Journal of The Electrochemical Society, 166 (9) B1-B10 (2019)

experience is only accurate to about 1%, is calibrated by connecting 205

the network to an oscilloscope and having the user enter the actual 206

frequency observed when the synthesizer is programmed to construct 207

a nominally 2047.85 Hz signal. Systematic step-by-step instructions 208

for calibration are shared as Supplemental Material File S6. 209

Evaluation of ABE-Stat current draw/power requirements.—To 210

estimate current draw from the battery (and correspondingly the esti- 211

mated battery life), power was supplied from a fully charged single-cell 212

lithium battery (4.2 V) and current was measured through a handheld 213

multimeter (Fluke-115, Fluke Corporation, Everett, WA, USA), for a 214

variety of operating conditions with the device firmware as shared in 215

the supplemental materials (Files S2). In the implementation described 216

here, average current requirements for ABE-Stat vary primarily de- 217

pending on the required rate of data transmission over the Bluetooth 218

radio, and whether or not the AD5933 network analyzer chip is pow- 219

ered (i.e. for EIS measurements above 60 Hz). 220

Evaluation of ABE-Stat noise.—To evaluate the noise on the 221

ABE-Stat, its firmware was customized to record cell potential and 222

current under static conditions, interfacing to a separate general pur- 223

pose data acquisition and control application ABE-VIEW51 described 224

elsewhere. In these experiments, the cell was configured for two elec- 225

trodes, and a nominally 1 M (5%) leaded resistor was used as the 226

load, applying a 0 V control signal. Current noise was estimated as the 227

root mean squared (RMS) error values for 1000 observations recorded 228

Figure 2. ABE-Stat (red enclosure at right) interfaced wirelessly to Android at each of the four transimpedance gains available on the device, for 229
app on smartphone (Google Pixel 1 pictured). the ADC configured in the default “normal” 20 Hz data rate used for 230

static logging and “multi-meter”-like functions, as well as in “turbo” 231

mode at 2 kHz data rate used for dynamic voltammetry functions (CV 232
167 The number of cycles or duration of each method can be controlled, and DPV) and low frequency EIS. Static voltage noise in the control 233
168 and results of analysis are displayed graphically in real time. Results circuit was estimated as the RMS error values of the corresponding 234
169 are stored as comma delimited (.csv) files which can be can be opened measurements recorded after each current measurement, and dynamic 235
170 for viewing later within the app, or shared by e-mail directly from voltage noise was estimated as the observed standard error in the least 236
171 the app. Currently the app includes separate string files to operate in squared error linear regression of observed cell potential against time 237
172 any of 6 languages (English, simplified Chinese, Spanish, Japanese, for the forward sweep of a CV scan at 0.1 V/s of a nominal 1 M (5%) 238
173 Korean, and Portuguese). resistor in two-electrode configuration. Thermocouple noise was esti- 239

mated as the RMS error value of 1000 voltage measurements at 20 Hz 240

174 System calibration.—Calibration of each ABE-Stat requires the data rate for a tin-plated solid copper wire shorted across the thermo- 241

175 use of a separate Android app (“ABE-Stat Cal”, available as Supple- couple inputs. 242

176 mental Material File S5). Calibration of the ABE-Stat system is pred-
177 icated on the presumed accuracy of the precision 2.048 V reference Voltammetric measurements.—For demonstrating the perfor- 243

178 on the ADS1220 ADC. System calibration first requires that the DAC mance of voltammetric functions of the ABE-Stat, single CV and DPV 244

179 be calibrated over its nominal dynamic range (−1.65 V to 1.65 V), scans were recorded for a three electrode cell in an electrolyte solution 245

180 which is achieved by communicating nominal system DAC outputs containing equimolar concentrations of K4 Fe(CN)6 and K3 Fe(CN)6 246

181 to the Android along with the corresponding observed network po- totaling 10 mM, and 0.1 M KCl.52 A saturated Ag/AgCl electrode (cat- 247

182 tential outputs at discrete intervals over the range. The calibration app alog # A57194, VWR International, LLC, Brisbane, CA) was used as 248

183 then returns least squared error linear regression coefficients which are the reference, platinum electrode (product # CHI115, CH Instruments, 249

184 stored into system EEPROM of the ABE-Stat hardware device. The Inc., Austin, TX) was used as the counter electrode, and gold wires 250

185 transimpedance gains are then calibrated by applying a (calibrated) po- (50 μm gold-plated tungsten wire, ESPI Metals, Ashland, OR) with 251

186 tential to a two-terminal network with a known resistor value (which and without dip coating in polyethyleneimine20 were used for work- 252

187 the user pre-calibrates and sends the value of from within the app), ing electrodes. Cells were equilibrated for 10 seconds at the starting 253

188 and measuring resulting voltage outputs from the transimpedance am- potentials (-0.2 V vs saturated Ag/AgCl), then scanned to +0.6 V and 254

189 plifier on the ADC. Calibration of the network analyzer is the most back to the starting potential at 0.1 V/s (CV) or with 2 mV steps/25 mV 255

190 tedious operation, as it requires calibration of digitized admittance and pulse amplitude at 25 Hz (DPV). All scans were compared with scans 256

191 system phase across frequencies for a total of 48 different settings (4 recorded under the same conditions/parameters with a reference lab- 257

192 different possible peak AC amplitude values of 10 mV, 20 mV, 50 mV, oratory potentiostat (μ-Autolab type III potentiostatic frequency re- 258

193 and 100 mV; 3 different transimpedance gains used in conjunction sponse analyzer (FRA) equipped with NOVA software version1.6, 259

194 with the network analyzer; internal x5 programmable gain amplifier Metrohm Autolab USA Inc., Riverview, FL). As the reference instru- 260

195 enabled or disabled, and; use of internal 16.776 MHz system clock ment reports DPV potentials from prior to the pulse and ABE-Stat 261

196 or external 250 kHz clock to set the frequency range for analysis). reports potentials after the pulse, potential data from ABE-Stat DPV 262

197 These calibrations are conducted using “real” resistive loads in a two- scans were shifted down by 25 mV for comparison. 263

198 electrode network, first having the system estimate the load resistance,
199 and then calibrating each setting expected to result in an unsaturated Potentiometric measurements.—To demonstrate high impedance 264

200 output with reasonably high signal to noise ratio (i.e. peak voltage potentiometric measurements with ABE-Stat measurements were 265

201 output above 100 mV). This is repeated for as many resistor values made using a combination pH electrode (Fisherbrand Accumet Cat- 266

202 required to achieve calibration of all available settings- typically nom- alog number 013-620-285, Thermo Fisher Scientific, Waltham, MA, 267

203 inal values of 100 k, 33 k, 10 k, 3.3 k, 1 k, 330 , and 100 USA), recording the average of 100 observed potentials with ABE- 268

204 . Finally the internal clock of the network analyzer, which in our Stat in each of three pH standard buffers (4, 7, 10), and comparing 269
 Journal of The Electrochemical Society, 166 (9) B1-B10 (2019) B5

Table II. ABE-Stat average current draw from single cell lithium
battery (4.2 V) for different conditions/analytical applications.

Condition Current Battery life∗

Idle† 33 mA 60 hr
Static Recordings‡ 55 mA 36 hr
Voltammetry¶ 65 mA 30 hr
EIS above 60 Hz 110 mA 18 hr

∗ Nominal value for continuous operation with 2000 mA-hr battery.

† Bluetooth not connected; note that ABE-Stat turns itself off if it is
“idle” for 60 seconds.
‡ Includes default multi-meter like measurements.
¶ CV, DPV, and EIS up to 60 Hz.

270 values to the “equilibrium” potential observed on a benchtop labora-

271 tory meter (Accumet Fisher AR 60). The temperature displayed by
272 the laboratory meter was recorded to evaluate the Nernstian behavior
Figure 3. Observed current noise on ABE-Stat measuring at “normal” (20 Hz)
273 of the system. data rate (black symbols) and “turbo” (2 kHz) data rate (red symbols).

274 Electrochemical impedance spectroscopy.—EIS measurements

275 were recorded using the same electrochemical cell and conditions applications like CV, DPV and low frequency (below 60 Hz) EIS, 318
276 described above for voltammetric measurements. For each working with RMS noise approximately equivalent to the quantization step 319
277 electrode (coated and uncoated gold wires) scans were recorded for of a 16-bit system with equivalent range. These levels of noise were 320
278 each combination of 0 and 200 mV bias vs. saturated Ag/AgCl, and greater than those reported by Dryden and Wheeler for the DStat,45 321
279 peak AC amplitudes of 10 mV and 100 mV. Prior to each scan, cells which were approximately equivalent to 23 bits of precision at lower 322
280 were equilibrated at the given bias potential for 10 seconds. For com- gains at a 60 Hz data rate. Much of the discrepancy can be attributed 323
281 parison, scans were also completed for the same signal settings using to limitations in the respective ADCs used in the projects even though 324
282 the same reference potentiostat as used for voltammetric scans, set to they are both nominally 24-bit Sigma-Delta systems. For example 325
283 estimate impedances from a single AC cycle at each frequency. the ADS1220 used for ABE-Stat has an Effective Number of Bits 326
284 To better understand some of the EIS performance limitations of (ENOB) of 20.01 when using a data rate of 20 Hz or 16.79 at 2 kHz,53 327
285 ABE-Stat, scans were also recorded for simple linear two-terminal net- compared to the ADS1255 (Texas Instruments) used for DStat with 328
286 works (in 2 electrode configuration), with 0 V bias. Loads evaluated reported ENOB of 23.8 at 60 Hz, 21.2 at 2 kHz, and 19.8 at 30 kHz.54 329
287 included a nominal 33 k (5%) leaded resistor, as well as other 33 k RMS noise in the voltage control circuit of ABE-Stat was about 330
288 resistors soldered in parallel to nominal 100 pF (20%) ceramic, 10 nF 50 μV under static conditions for all observed data rates, approxi- 331
289 (20%) ceramic, and 10 nF (20%) polyester film capacitors. Ideal EIS mately equivalent to the quantization step for the 3.3 V 16-bit DAC. 332
290 behavior was predicted for these loads by modeling each load as a re- This is consistent with an approximate current noise floor of 50 pA 333
291 sistor with value taken as the average impedance observed at the lowest which is near the value of about 100 pA observed at the highest gain 334
292 10 frequencies of the scan, and, where applicable, parallel capacitance (106 ) under the tested conditions (Figure 3). Under dynamic condi- 335
293 estimated as the equivalent capacitance for the impedance magnitude tions (i.e. CV) the RMS error of the control potential was observed to 336
294 observed at the highest factor of 10 frequency where impedance mag- be about 300 μV, potentially contributing additional noise to dynamic 337
295 nitudes were observed to be rolling off linearly on a logarithmic scale. measurements such as CV and DPV. 338
Noise at the thermocouple inputs was about 5 μV, equivalent to 339

296 Results roughly 0.1°C for a type-K thermocouple. This value is substantially 340

lower than the RMS noise observed on the control amplifier, and close 341
297 Evaluation of ABE-Stat current draw/power requirements.— to the manufacturers specifications (3.9 μV)53 under the settings used 342
298 The current from a single-cell lithium battery required to power ABE- for thermocouple measurement. 343
299 Stat varies depending on the analytical function being implemented
300 (Table II). In the firmware shared here it should be noted that the
301 WiFi radio of the ESP8266 is explicitly disabled, and that alternative Voltammetric measurements.—Cyclic voltammograms with 344

302 firmware implementations using WiFi capability could change the cur- ABE-Stat corresponded closely to measurements with a reference 345

303 rent draw significantly. Here, variations in current are largely governed instrument (Figure 4), with significantly larger redox current peaks 346

304 by the data transmission rate required for the given analysis, and im- for bare gold wire electrodes compared to electrodes coated in 347

305 portantly whether or not the analysis requires the network analyzer polyethyleneimine. Similarly, DPV scans resulted in expected charac- 348

306 AD5933 to be powered (i.e. only EIS at frequencies above 60 Hz). teristic shapes of redox current curves, with substantial correction for 349

307 Even for analytical applications with the highest power requirements the background currents from dielectric charging at the boundary lay- 350

308 (∼110 mA for EIS above 60 Hz) the nominally 2000 mA-hr battery we ers and larger current peaks on bare wires (Figure 5). All scans were 351

309 use should last for up to 18 hours of continuous operation on a single highly symmetric reflecting the highly reversible nature the selected 352

310 charge, well in excess of the typical battery life for smart-phones and redox reaction. 353

311 tablets with the screen illluminated.

Potentiometric measurements.—Potentiometric measurements 354

312 Evaluation of ABE-Stat noise.—Current noise observations on with the ABE-Stat corresponded closely to those of the reference 355

313 the ABE-Stat were approximately 21 times the value of the nominal instrument (Figure 6), with root mean squared error of 3.2 mV be- 356

314 quantization step of the 24-bit ADC even at the lowest gains and data tween corresponding readings, or standard error of 1.5 mV for the 357

315 rates (20 Hz) used for routine analysis (Figure 3)- equivalent to a loss linear calibration using the reference instrument as a standard. The 358

316 of more than 4 bits from the nominal system resolution. As expected potentials reported by ABE-Stat for different nominal pH standards re- 359

317 noise was higher at the higher data rates (2 kHz) used for dynamic sulted in a sensitivity of −57.9 mV/pH unit at 22°C, representing about 360
 B6 Journal of The Electrochemical Society, 166 (9) B1-B10 (2019)

Figure 6. Open circuit potentials observed from combination pH electrode

Figure 4. Cyclic voltammograms for bare gold wire (solid symbols) and with ABE-Stat, compared to reference instrument (Accumet Fisher AR60).
polyethyleneimine-coated gold wire (open symbols), for ABE-Stat (red sym- Error bars represent the standard deviations of 100 observations with ABE-
bols) compared to reference instrument (black symbols). All measurements Stat (∼0.5 mV). EABE-Stat = 0.532 + 0.984 Eref ; R2 = 0.99996.
made in 10 mM potassium ferricyanide/ferrocyanide, 0.1 M KCl solution, with
saturated Ag/AgCl reference.
as more of the control signal was pushed far from the redox potential 378

where incremental increases in current are lower. 379

361 1% deviation from the predicted Nernstian response of −58.5 mV/pH Removing the DC bias in EIS measurements (Figure 8) resulted 380
362 unit at that temperature. in significantly increased impedance for all conditions as redox cur- 381

rents were diminished, and correspondingly a more capacitive char- 382

363 Electrochemical impedance spectroscopy.—EIS scans of ferri- acteristic (larger phase lag of current) at low frequencies. Unbiased 383
364 cyanide solution with ABE-Stat resulted in data reasonably close to measurements also resulted in significantly more noise in the ABE- 384
365 those measured with the reference instrument, especially for small Stat impedance data, especially at low frequencies where network 385
366 AC amplitude signals (10 mV peak) with bias (200 mV) close to the impedance was highest (near or above 100 k in magnitude) and the 386
367 redox potential of ferricyanide (Figures 7a and 7c). At higher AC am- AC perturbation signal was small (10 mV; Figures 8a and 8c). 387
368 plitudes (100 mV) more significant deviations from the reference in- EIS scans of two-terminal linear passive networks (Figure 9) 388
369 strument could be observed (Figures 7b and 7d) which may partially mostly behaved as predicted, with a purely resistive characteristic at 389
370 be attributable to the non-linear nature of redox processes and cor- low frequencies transitioning to capacitive characteristics at higher 390
371 responding limitations in the approaches for estimating impedances frequencies for networks including a capacitor in parallel with the 391
372 with ABE-Stat (see discussion). As expected gold wires coated with load resistor. Nevertheless large discontinuities can be observed in 392
373 polyethyleneimine exhibited higher impedance than bare gold wires the ABE-Stat data, notably at several characteristic frequencies (2 Hz, 393
374 across the spectrum, and correspondingly capacitive characteristics 60 Hz, and 2 kHz) corresponding to different approaches that ABE- 394
375 and peak differences in phase of cell current and voltage occurred at Stat uses to estimate impedance. 395
376 lower frequencies. At lower frequencies especially the higher AC am-
377 plitude condition resulted in higher observed impedance magnitudes
Discussion 396

In the spirit of other fully open-source potentiostat projects like 397

CheapStat from Rowe et al.44 and DStat from Dryden and Wheeler45 398

we have shared an open-source hardware design with software tools 399

to make high quality electro-analysis more accessible to the commu- 400

nity. While ABE-Stat falls short of the exceptionally low-noise per- 401

formance of DStat, its performance is better in this respect than many 402

other published potentiostat projects.44,55,56 Importantly ABE-Stat can 403

be operated wirelessly through Bluetooth or WiFi, and is configured 404

to enable EIS measurements with user defined bias. The total cost of 405

components/bill of materials for ABE-Stat is about US$105, similar 406

to the component costs for DStat. Circuit board fabrication and assem- 407

bly can vary based on geographical factors and especially the scale of 408

production, but we have obtained quotes for fabrication and assem- 409

bly (excluding component costs) from a US-based manufacturer of 410

US$110 each for a quantity of 25, and US$47.50 each for a quantity 411

of 100. 412

While we believe that our existing hardware design is adequate for 413

many routine analytical applications, and could readily be adapted for 414

low-cost biological and chemical sensing in the field, there are sev- 415

eral important limitations that the user should be aware of and that 416

Figure 5. Differential pulse voltammograms from ABE-Stat (red symbols) could potentially be addressed with improvements in device design. 417

compared to reference instrument (black symbols), with bare gold wire (filled For precision voltammetric measurements of small signals the most 418

symbols) and polyethyleneimine-coated gold wire (open symbols). Only the obvious improvement would be to use the ADC from DStat (ADS1255 419

forward scans are shown for the reference instrument. family, Texas Instruments), which under similar conditions has 420
 Journal of The Electrochemical Society, 166 (9) B1-B10 (2019) B7

Figure 7. EIS magnitude and phase plots for 200 mV bias with 10 mV peak AC (left) and 100 mV peak AC (right). Red symbols are from ABE-Stat, black
symbols are from reference instrument, filled symbols are for bare gold wire, and open symbols are from “coated” gold wire. All measurements made in 10 mM
potassium ferricyanide/ferrocyanide and 100 mM KCl solution, with saturated Ag/AgCl reference. Vertical dashed lines represent major discontinuities in how
ABE-Stat determines impedance.

421 effectively 4 bits (16x) better precision and importantly has a max- to low energy Bluetooth itself might result in considerably diminished 444

422 imum data rate of 30 kHz compared to 2 kHz for the ADS1220 used noise. 445

423 in ABE-Stat. For measurement of small signals the stability of the con- One of the most compelling features of ABE-Stat is the ability to 446

424 trol amplifier also needs to be improved. Like Dryden and Wheeler, conduct EIS measurements. While other published projects such as 447

425 we made significant efforts to isolate noisy digital components of Pruna et al.39 and Zhang et al.57 nominally have the same capability, 448

426 the device and especially in the case of ABE-Stat the wireless ra- they are implemented on a connected PC operating with expensive pro- 449

427 dios from the analog circuitry. This included generous use of ferrite prietary software (MATLAB),39 or have relatively restricted frequency 450

428 beads and other filtering between the respective power supplies and range and do not allow user control over the applied DC bias.57 Even so 451

429 grounds. For the implementation shared in this work the WiFi radio of there are significant hardware and software improvements that could 452

430 ESP8266 was also disabled. The fact that noise on our thermocouple be made for EIS measurements with ABE-Stat including innovations 453
431 input was close to the rated value for our ADC suggests that isola- gleaned from the prior art. While the network analyzer chip AD5933 454

432 tion of the digital and analog power supplies was largely effective, but is powerful and is also used in the project of Zhang et al.,57 it has 455

433 the relatively high noise observed on the control amplifier of ABE- several limitations that can make non-linear analyses such as EIS dif- 456

434 Stat suggests that significant improvement could be made stabilizing ficult. On the AD5933 the frequency of AC signals (from a built-in 457

435 the potentials on this part of the design for both static and dynamic digital synthesizer) and the resulting impedance analyses are scaled 458

436 measurements. to the system clock, with synthesizer frequency governed by a user- 459

437 Several system improvements might be achieved by changing from defined phase accumulator value clocked in at a rate proportional to 460

438 an ESP8266 based microcontroller to the newer ESP32 (Espressif the system clock, and impedance is estimated from the coefficients of 461

439 Microsystems) based microcontroller, which has considerably more a 1024 point Discrete Fourier Transform (DFT) that is also sampled 462

440 general purpose input/output (GPIO), as well as integral Bluetooth at a rate proportional to the clock. Accurate measurements using the 463

441 (both classic and low energy standard) in addition to WiFi. All of internal 16.776 MHz clock therefore cannot be made for frequencies 464

442 these features could result in a considerably condensed footprint of the below about 1 kHz58 (corresponding to the 1024 measurements for the 465

443 digital circuits which may improve noise performance, and switching DFT sampled over a single cycle of the synthesizer frequency), and 466
 B8 Journal of The Electrochemical Society, 166 (9) B1-B10 (2019)

Figure 8. EIS magnitude and phase plots for 0 bias with 10 mV peak AC (left) and 100 mV peak AC (right). Red symbols are from ABE-Stat, black symbols
are from reference instrument, filled symbols are for bare gold wire, and open symbols are from “coated” gold wire. All measurements made in 10 mM potassium
ferricyanide/ferrocyanide and 100 mM KCl solution with saturated Ag/AgCl reference. Vertical dashed lines represent major discontinuities in how ABE-Stat
determines impedance.

467 similarly at the low end of frequencies major errors occur if frequen- controlling the AD5933 timing from a programmable clock, which 490

468 cies are evaluated that do not correspond to integer numbers of cycles in principle could be used to consistently control the sampling period 491

469 over the sampled period. In our case we force ABE-Stat to only eval- with respect to the applied frequency at any arbitrary frequency. This 492

470 uate integer harmonics of the minimum frequency below 10 kHz, and approach could potentially also address a limitation of the AD5933 493

471 as results were not consistent at the minimum frequency value we only for analyzing EIS, in that starting a new “frequency sweep” (measure- 494

472 used the internal clock for measurements starting at about 2 kHz. Be- ment at a new frequency if the frequency needs to be incremented on a 495

473 low this frequency we use an external 250 kHz clock with AD5933, logarithmic scale) requires a reset command which introduces a large 496

474 nominally extending the analysis range down to about 15 Hz. However, transient on the applied DC component of the signal.58 Some of this 497

475 due to AC coupling of the synthesizer signal into the control amplifier transient can be observed through the AC coupling to the control am- 498

476 (critical frequency 7 Hz) we only use this clock to evaluate frequencies plifier of ABE-Stat even when the synthesizer is disconnected through 499
477 down to about 60 Hz. Evaluation of impedances at lower frequencies it’s corresponding analog switch, resulting in undesired perturbations 500

478 is achieved by disconnecting the network analyzer and building the of electrochemical equilibrium that could affect measurements. 501

479 entire desired signal in software from the DAC. Impedances at the It should be noted that ABE-Stat measurements from the 502

480 desired frequency are then estimated from the corresponding DFT co- AD5933 using the 250 kHz clock (i.e. in the range from 60 Hz to 2 kHz; 503

481 efficients of the current signal from a 24-point sample. Below 2 Hz Figures 7, 8 and 9) are unreliable for non-linear non-resistive loads. 504

482 the samples are taken over a single cycle of the given frequency, but This is unfortunate because these correspond to very critical frequen- 505

483 since reading the ADC is the slowest part of the operation even at the cies (i.e. circular portions of Nyquist plot) in many electrochemical 506

484 highest data rate of ADS1220, above 2 Hz samples are distributed over systems. We are currently not able to explain this phenomenon, though 507

485 8 cycles (corresponding to 3 samples per cycle). These variations in in a previous iteration of the ABE-Stat design the analog power supply 508

486 the implementation of EIS measurements can result in large discon- to the network analyzer AD5933 was from a precision voltage refer- 509

487 tinuities at these critical frequencies (Figures 7, 8, and 9), especially ence that was overly taxed from the current demand of the AD5933. 510

488 when analyzing non-linear systems where the current signal is not This resulted in a relatively large AC signal on the voltage reference 511

489 a perfect sinusoid. Zhang et al.57 overcome many of these issues by and distortions in the control signal with a very large harmonic near 512
 Journal of The Electrochemical Society, 166 (9) B1-B10 (2019) B9

Figure 9. ABE-Stat EIS magnitude and phase plots for passive (2 terminal linear) RC networks for 10 mV peak (left) and 100 mV peak (right) AC signals. Closed
black, open black, closed red, and open red symbols respectively represent 33 k resistor alone, and in parallel with nominally 100 pF ceramic, 10 nF ceramic,
and 10 nF polyester film capacitors. Fitted data are represented by lines of the same color, solid for closed symbols and dashed for open symbols.

513 750 Hz.51 It is essential therefore that designers keep this constraint Conclusions 537

514 under consideration and power the AD5933 with a voltage source
We have designed a compact and versatile potentiostat device 538
515 with current capacity easily capable of supplying the rated require-
(ABE-Stat) and accompanying software to conduct many routine elec- 539
516 ment (10 mA for 3.3 V supply).58 For extending the frequency range
trochemical analyses, and made the designs and source code freely 540
517 and improving performance for the software-intensive EIS approach at
available for direct use or adaptation. The design was inspired largely 541
518 low frequencies of ABE-Stat, the higher speed ADC used by Dryden
by that of the DStat from Dryden and Wheeler,45 though with respect 542
519 and Wheeler45 (ADS1255 Series) could be useful.
to noise for voltammetric applications our design fell short of the DStat 543
520 While EIS can be conducted by ABE-Stat with arbitrary DC bias,
which we regard as a benchmark for performance of an open-source 544
521 one important limitation is that the analog reference for the AD5933
device. Even so the ABE-Stat represents some compelling advances 545
522 network analyzer chip (generated internal to the chip, and approxi-
in open-source potentiostat projects, in that it is also fully wireless en- 546
523 mately half of the analog supply voltage58 ) is not directly connected
abled (WiFi and Bluetooth), includes an Android interface to facilitate 547
524 to the reference for the rest of the analog circuitry (generated by a
field use, and in addition to basic voltammetric (CV and DPV) and 548
525 “rail” splitter chip from analog voltage supply and ground; TLE2426
potentiometric (open source voltage) measurements it incorporates 549
526 Texas Instruments). This results in discrepancies in bias voltage from
capability of conducting EIS measurements across a wide frequency 550
527 the calibrated values when impedance measurements are made with
spectrum (0.1 Hz to 100 kHz) and large range of impedance values 551
528 AD5933. We’ve observed these discrepancies to be about 20 to 30 mV,
(less than 100  to in excess of 100 k) with any of four available 552
529 which can be significant where precise biases relative to redox poten-
peak AC amplitudes (100 mV, 50 mV, 20 mV, or 10 mV). In contrast 553
530 tial is important. In future iterations this could be corrected by making
to other recent projects that incorporate EIS capability39,57 our imple- 554
531 a single voltage reference for the entire analog network using a voltage
mentation does not have additional proprietary software dependencies, 555
532 follower from the Vin pin of AD5933, as this pin is “virtually” con-
and allows user control of the DC bias. We have fully characterized all 556
533 nected to the AD5933 reference through an internal amplifier. Alterna-
of the capabilities of the device, and made rational recommendations 557
534 tively the AD5933 reference voltage could be routed through a voltage
for improving design with respect to current noise, and consistency of 558
535 follower to the system ADC to allow it to be measured/calibrated with
EIS measurements. 559
536 respect to the rest of the analog circuit.
 B10 Journal of The Electrochemical Society, 166 (9) B1-B10 (2019)

560 Acknowledgments 24. K. Siuzdak et al., Sensors Actuators B Chem., 280, 263 (2019). 597
25. J. Kwon, E.-M. Cho, P. Nandhakumar, S. I. Yang, and H. Yang, Anal. Chem., 90, 598
561 This work was a collegial initiative of the United States Department 13491 (2018). 599
562 of Agriculture (USDA) Multi-State Nanotechnology and Biosensors 26. E. M. Mateo et al., Food Chem., 267, 91 (2018). 600

563 Program (NC1194; project HAW05038-R), with some support from 27. T. H. V. Kumar and A. K. Sundramoorthy, J. Electrochem. Soc., 165, B3006 (2018). 601
28. G. Chen, S. Xiao, A. Lorke, J. Liu, and P. Zhang, J. Electrochem. Soc., 165, B787 602
564 the USDA Hatch Program (project HAW-5027H). (2018). 603
29. A. A. Ensafi, F. Akbarian, E. Heydari-Soureshjani, and B. Rezaei, Biosens. Bioelec- 604
tron., 122, 25 (2018). 605
ORCID 30. P. Kumar, P. Devi, R. Jain, A. Saini, and R. Noetzel, Mater. Lett., 236, 587 (2019). 606
31. N. R. Devi, M. Sasidharan, and A. K. Sundramoorthy, J. Electrochem. Soc., 165, 607
565 Daniel M. Jenkins https://orcid.org/0000-0003-1616-0845 B3046 (2018). 608
32. A. Larrea-Sarmiento et al., Sci. Rep., 8, 14298 (2018). 609
33. D. M. Jenkins, R. Kubota, J. Dong, Y. Li, and D. Higashiguchi, Biosens. Bioelectron., 610

566 References 30, 255 (2011). 611

34. M. L. Keremane et al., Crop Prot., 68, 41 (2015). 612

567 1. J. Slagboom et al., Toxicon, 148, 213 (2018). 35. P. L. Urban, Analyst, 140 (2015). 613

568 2. S. Amaya-González, N. de-los-Santos-Álvarez, A. Miranda-Ordieres, and 36. M. D. M. Dryden, R. Fobel, C. Fobel, and A. R. Wheeler, Anal. Chem., 89, 4330 614

569 M. Lobo-Castañón, Sensors, 13, 16292 (2013). (2017). 615

570 3. L. Zhao et al., Toxins (Basel)., 10, 427 (2018). 37. Y. C. Li et al., J. Chem. Educ., 95, 1658 (2018). 616

571 4. Z. Khoshbin, A. Verdian, M. R. Housaindokht, M. Izadyar, and Z. Rouhbakhsh, 38. A. Nemiroski et al., Proc. Natl. Acad. Sci., 111 (2014). 617

572 Biosens. Bioelectron., 122, 263 (2018). 39. R. Pruna et al., Biosens. Bioelectron., 100 (2018). 618

573 5. Q. Zhang et al., J. Electroanal. Chem., 823, 40 (2018). 40. G. N. Meloni, J. Chem. Educ., 93 (2016). 619

574 6. D. Vanegas et al., Biosensors, 8, 42 (2018). 41. K. Kellner, T. Posnicek, J. Ettenauer, K. Zuser, and M. Brandl, in Procedia Engineer- 620

575 7. K. D. Hills, D. A. Oliveira, N. D. Cavallaro, C. L. Gomes, and E. S. McLamore, ing„ vol. 120 (2015). 621

576 Analyst, 143, 1650 (2018). 42. A. F. D. Cruz, N. Norena, A. Kaushik, and S. Bhansali, Biosens. Bioelectron., 62 622

577 8. M. Taguchi et al., Analyst, 141, 3367 (2016). (2014). 623

578 9. S. M. Abdelbasir et al., ACS Sustain. Chem. Eng., 6, 12176 (2018). 43. P. Lopin, K. V. Lopin, and M. Rolandi, Editor. PLoS One, 13, e0201353 (2018). 624

579 10. S. L. Burrs et al., Biosens. Bioelectron., 85, 479 (2016). 44. A. A. Rowe et al., M. Wanunu and Editor, PLoS One, 6, e23783 (2011). 625

580 11. A. Halalipour, M. R. Duff, E. E. Howell, and J. I. Reyes-De-Corcuera, Biotechnol. 45. M. D. M. Dryden, A. R. Wheeler, and D. T. Eddington, Editor. PLoS One, 10, 626

581 Bioeng., 114, 516 (2017). e0140349 (2015). 627

582 12. A. Halalipour, M. R. Duff, E. E. Howell, and J. I. Reyes-De-Corcuera, Enzyme Mi- 46. P. B. Lillehoj, M.-C. Huang, N. Truong, and C.-M. Ho, Lab Chip, 13, 2950 (2013). 628

583 crob. Technol., 103, 18 (2017). 47. M. D. Steinberg, P. Kassal, I. Kereković, and I. M. Steinberg, Talanta, 143, 178 (2015). 629

584 13. J. I. Reyes-De-Corcuera, H. E. Olstad, and R. García-Torres, Annu. Rev. Food Sci. 48. G. F. Giordano et al., Electrochim. Acta, 219 (2016). 630

585 Technol., 9, 293 (2018). 49. Y. Rong et al., Analyst, 143, 2066 (2018). 631

586 14. S. Campuzano, P. Yáñez-Sedeño, and J. Pingarrón, Sensors, 17, 866 (2017). 50. Omega Engineering, https://www.omega.com/techref/pdf/z198-201.pdf. 632

587 15. S. Campuzano, P. Yáñez-Sedeño, and J. Pingarrón, Sensors, 17, 2533 (2017). 51. D. Jenkins and R. Kurasaki, Sensors, 18, 2647 (2018). 633

588 16. Z. Li, Z. Ye, Y. Fu, Y. Xiong, and Y. Li, Anal. Methods, 8, 548 (2016). 52. L. Lu, G. Chee, K. Yamada, and S. Jun, Biosens. Bioelectron., 42, 492 (2013). 634

589 17. M. Xu, R. Wang, and Y. Li, Talanta, 162, 511 (2017). 53. Texas Instruments, (2016) http://www.ti.com/lit/ds/symlink/ads1220.pdf. 635

590 18. L. Cao, Q. Zhang, H. Dai, Y. Fu, and Y. Li, Electroanalysis, 30, 517 (2018). 54. Texas Instruments, (2013) http://www.ti.com/lit/ds/symlink/ads1256.pdf. 636

591 19. T. Wen, R. Wang, A. Sotero, and Y. Li, Sensors, 17, 1973 (2017). 55. Š. Kubínová and J. Šlégr, J. Chem. Educ., 92 (2015). 637
592 20. I. Lee and S. Jun, J. Food Sci., 81, N1530 (2016). 56. H. Jin et al., J. Chem. Educ., 95, 326 (2018). 638
593 21. K. Yamada, W. Choi, I. Lee, B.-K. Cho, and S. Jun, Biosens. Bioelectron., 77, 137 57. D. Zhang et al., Sensors Actuators B Chem., 222, 994 (2016). 639
594 (2016). 58. Analog_Devices, (2017) http://www.analog.com/media/en/ 640
595 22. N. T. Garland et al., ACS Appl. Mater. Interfaces, 10, 39124 (2018). technical-documentation/data-sheets/AD5933.pdf. 641
596 23. E. Mikuła et al., BMC Vet. Res., 14, 328 (2018). 59. A. Ainla et al., Anal. Chem., 90 (2018). 642