IEEE SENSORS JOURNAL, VOL. 22, NO.

19, 1 OCTOBER 2022 18253

Toward a Gas Sensor Interface

Circuit—A Review
Mingyuan Ren , Honghai Xu, Changchun Dong, and Zhu Zhang

1 Abstract —With the development of science and technol-

2 ogy, gas sensors have gradually expanded to many vertical
3 fields. However, each application requires different technical
4 requirements for gas sensors, such as low cost, small size,
5 high sensitivity, high precision, and low power consumption.
6 To create gas sensors that meet the performance require-
7 ments, researchers worldwide are working on solutions in
8 fields such as sensitive materials, sensor arrays, and interface
9 circuits. However, there are bottlenecks in the research of
10 interface circuits, which limits the volume, power consump-
11 tion, and intelligent design development of gas sensors.
12 Therefore, this article reviews various types of gas sensors
13 and interface circuits implemented with different technolo-
14 gies, analyzes their research significance in various areas, and provides ideas for future gas sensor research directions.
15 Index Terms — CMOS integrated circuit, gas sensor, interface circuit.

16 I. I NTRODUCTION a single-chip method that is compatible with interface circuit 34

technology [11], [12], [13] and the use of the flip chip, wafer
W ITH gradual improvements in living conditions and
17 35

18 industrial safety, new requirements for functional gas stitching, and other technologies to achieve the same package. 36

19 sensors are continuously being put forward [1], [2], [3], includ- The multichip methods have been developed [14], [15], but 37

20 ing indoor air quality management, toxic gas monitoring, these technologies are not sufficiently mature, resulting in 38

21 diagnosis of environmental pollution, and breathing issues. high process manufacturing costs. In recent years, sensor 39

22 For example, the spread of the new coronavirus has sounded miniaturization technology has gradually improved and can 40

23 the alarm for some people. Even during the outbreak of the be applied to complex gas sensing systems [16], such as 41

24 new coronavirus, if a gas sensor could detect breathing or the mass spectrometry [17], infrared spectrometry [18], atomic 42

25 presence of the virus in the air, it could play a critical role absorption spectrometry [19], and gas chromatography [20]. 43

26 in preventing the epidemic. As a result, in various IoT envi- For different work scenarios in daily life, gas sensors have 44

27 ronments, the demand for gas monitoring is growing rapidly also adopted various technologies and have been gradually 45

28 [4], [5]. A traditional gas sensor has been used to establish an developed into catalytic combustion [21], infrared [22], elec- 46

29 independent sensing system on a PCB board [6], [7]. However, trochemical [23], metal oxide semiconductors (MOXs) [24], 47

30 considering its large area and high cost, it is still necessary thermal conductivity [25], and photoionization detector (PID) 48

31 to develop a portable, low-cost gas sensing system [8], [9], gas sensors [26]. 49

32 [10]. Researchers have attempted to miniaturize gas sensors In addition to the sensor, a complete gas detection system 50

33 and proposed a variety of solutions and technologies, including also needs to communicate with the computer [27], [28]. 51

This requires a signal conditioning circuit that is responsible 52

for converting sensor signals into electrical signals that the 53

Manuscript received 13 July 2022; revised 11 August 2022;
accepted 13 August 2022. Date of publication 29 August 2022; computer can recognize. The sensor interface circuit not only 54

date of current version 30 September 2022. This work was sup- plays a connecting role but also has the functions of signal 55
ported in part by the Jinhua Public Welfare Technology Appli- detection and signal preprocessing. Among the above sensing 56
cation Research Project under Grant 2022-4-064 and in part by
the Scientific Research Project of the Jinhua Advanced Research technologies, MOX can integrate small sensing systems and 57

Institute under Grant GYY202102. The associate editor coordinat- interface circuits on a single chip due to their good per- 58
ing the review of this article and approving it for publication was formance and CMOS compatibility [10]. This requires the 59
Dr. Ferran Reverter. (Corresponding author: Zhu Zhang.)
Mingyuan Ren, Changchun Dong, and Zhu Zhang are with the interface circuit to have a simple structure, small size, low 60

Jinhua Advanced Research Institute, Jinhua 321013, China (e-mail: power consumption, and other requirements. Since many solid 61
rmy2000@[Link]; hitdongcc@[Link]; [Link]@[Link]). gas-sensitive materials need to be operated at high temper- 62
Honghai Xu is with the School of Computer Science and Technology,
Harbin University of Science and Technology, Harbin 150080 China. atures (300 ◦ C–500 ◦ C), higher temperatures are required 63

Digital Object Identifier 10.1109/JSEN.2022.3199254 to activate the gas-sensitive mechanism. Users also want to 64

1558-1748 © 2022 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See [Link] for more information.
 18254 IEEE SENSORS JOURNAL, VOL. 22, NO. 19, 1 OCTOBER 2022

65 be able to detect a plurality of different types of gases at

66 the same time. In the meantime, countries stipulating index
67 requirements for portable gas sensing applications [29], [30]
68 and new requirements for their performance are put forward
69 every year. For example, according to the literature [31], the
70 dynamic range of MOX sensors must be at least four orders
71 of magnitude, the accuracy has to be less than 0.5%, and
72 the power consumption has to be less than 1 mW to meet Fig. 1. Interface circuit of the catalytic combustion gas sensor.
73 the dynamic characteristics (dynamic range, accuracy, and
74 baseline resistance) of these gas sensors. This undoubtedly
as well as the prerelease of respiratory diseases, toxic, harmful, 122
75 increases the difficulty of designing the interface circuit, so it
flammable, explosive gas detection, and gas sensors, which are 123
76 is necessary to study the high-performance sensor interface
a key part of gas detection devices. Gas detection devices can 124
77 circuit. The high-performance sensor interface circuit is the
roughly be divided into six types according to their working 125
78 core technology of the high-performance sensor and detec-
mechanism: catalytic combustion gas sensors, electrochemical 126
79 tion system. Its performance directly affects the measurement
gas sensors, thermal conductivity gas sensors, infrared gas sen- 127
80 accuracy and sensitivity of the whole system. For the interface
sors, PID photoionization gas sensors, and MOX gas sensors. 128
81 circuit, the common methods of resistance interface circuit
This section introduces the working principle of various gas 129
82 include oscillation and analog-digital converter (ADC) tech-
sensors and a simplified interface circuit topology diagram. 130
83 nology [34]. The former is affected by its parasitic capacitance
84 and needs to constantly modify the time constant, which will
85 increase the measurement error [15], [31]. In the working A. Catalytic Combustion Gas Sensor 131

86 process of the latter, the voltage division of the reference In catalytic combustion gas sensors, the working principle 132

87 resistor will change with the working time, which affects utilizes a gas-sensitive material (usually platinum coil) that 133

88 the accuracy of the sensor. Therefore, a high-precision ADC is energized, and as the temperature rises (normally main- 134

89 structure is required; however, this usually comes at the cost tained at 450 ◦ C–500 ◦ C), the combustible gas and detection 135

90 of design cost and complexity [35]. In addition to these element will undergo flameless combustion under the action 136

91 two technologies, microcontroller, current transmitter, phase- of a catalyst [37]. As the temperature of the gas-sensitive 137

92 locked amplifier, autobalance interface, and micro-electro- material increases, the sensitive resistance changes and gas 138

93 mechanical system (MEMS) technology are also used in concentration are measured by a back-end interface circuit. 139

94 the gas sensor interface circuit. The development of these This type of gas sensor is mainly used for the detection of 140

95 technologies and their respective advantages and disadvantages flammable gases [38] and has the advantages such as high 141

96 will be presented in the following. accuracy, fast response, and long service life, but it can only be 142

97 If the sensors need to respond in a small temperature used to measure flammable gases, resulting in low selectivity 143

98 range, so precise temperature control loop is indispensable. and risk of poisoning. 144

99 Traditional temperature controller combines high-efficiency The simplified interface circuit topology is shown in Fig. 1. 145

100 pulsewidth modulation (PWM) control with excellent tempera- This includes a detection resistor R D , a compensation resistor 146

101 ture stability [4]. The self-adaptive modulation method exploits RC , and two gas-sensitive resistors (R1 , R2 ). Following the 147

102 the changes in the sensor resistance to drive its operating principle of a Wheatstone bridge, when R D and the com- 148

103 temperature [36]. bustible gas in the environment undergo flameless combustion, 149

104 This article introduces the main improvement measures R X will change, break the balance of the bridge, generate 150

105 taken by researchers to deal with various challenges faced by a stable voltage signal at the output, and then process the 151

106 the interface circuit, studies the development of the interface data through a signal amplifier circuit, where the measurement 152

107 circuit, and provides relevant analysis. This article can be results can be obtained. The Wheatstone bridge circuit is used 153

108 summarized as follows. Section II briefly summarizes the in the interface circuit, which can accurately measure the small 154

109 working principle of six types of gas sensors and the functions change of the thermistor and convert the resistance change into 155

110 of corresponding interface circuits. Section III summarizes the voltage change of the sensor. This combination of bridge 156

111 improvements of sensitive circuits based on oscillating and and operational amplifier is widely used in the industry for 157

112 ADC technology to solve various problems over the past ten various sensors. 158

113 years. Section IV summarizes improvements in temperature

114 modulation technology. Section V summarizes the research B. Electrochemical Gas Sensor 159

115 and puts forward possible development trends of application Electrochemical gas sensors are essentially miniature fuel 160

116 specific integrated circuit (ASIC) chips for gas sensors in the cell components, in which the gas to be detected produces 161

117 future. a chemical reaction (oxidation/reduction) on the surface of 162

the working electrode (WE), generating a sensitive current 163

118 II. I NTERFACE C IRCUITS OF D IFFERENT G AS S ENSORS between the electrodes [39]. The magnitude of the current is 164

119 As introduced in Section I, early warning and protec- proportional to the gas concentration [40] and can be measured 165

120 tive measures have attracted attention for maintaining indus- by the back-end interface circuit. These types of sensors 166

121 trial production safety and reducing environmental pollution, have high sensitivity and good long-term stability but have 167
 REN et al.: TOWARD A GAS SENSOR INTERFACE CIRCUIT 18255

Fig. 2. Interface circuit of the electrochemical gas sensor.

Fig. 4. Interface circuit of the infrared gas sensor.

Fig. 3. Interface circuit of the thermal conductivity gas sensor.

Fig. 5. Interface circuit of the PID photoionization gas sensor.

168 disadvantages such as a short lifespan and a slow response
169 time.
170 Fig. 2 shows the simplified interface circuit topology of the concentration and type of the gas to be measured [42]. 205

171 an electrochemical gas sensor. As shown in Fig. 2, the elec- This type of sensor can exhibit good antitoxicity and high 206

172 trochemical gas sensor contains three pins, namely, the WE, sensitivity but has a complex structure and high cost. 207

173 the reference electrode (RE), and the counter electrode (CE). Fig. 4 shows a block diagram of the gas detection system for 208

174 Among these, the generated sensitive current flows from WE to this type of sensor. The infrared gas sensor interface circuit 209

175 CE, and RE only provides a reference voltage. This chemical in the picture contains a lamp drive circuit controlled by a 210

176 reaction is extremely weak, and the sensitivity of the generated low-frequency pulse signal, where the pyroelectric detection 211

177 current signal is very low (from tens of nA/ppm to hundreds layer is not sensitive to the gas (RP), and a pyroelectric 212

178 of nA/ppm). Therefore, this type of interface circuit requires detection layer is sensitive to the gas (AP) and it also 213

179 a transimpedance amplifier (TIA). The weak chemical signal includes a temperature sensor. In addition, there are two active 214

180 can be converted into an electrical signal in the sensor, and bandpass filter circuits and an ADC. The pulse signal for 215

181 the signal can be amplified with a certain intensity and low infrared radiation will be reflected inside the sensor, producing 216

182 noise. channels of different wavelengths. At this time, the AP layer 217

will absorb the gas to be measured, while the RP layer will 218

remain unchanged. The electrical signals converted by the two 219

183 C. Thermal Conductivity Gas Sensor
detection layers pass through the active bandpass, amplify and 220
184 The working principle of a thermal conductivity gas sensor reduce noise in the filter circuit, and, finally, are input into 221
185 is the same as a catalytic combustion gas sensor, but a thermal ADC for analog-to-digital conversion. Because both channels 222
186 conductivity gas sensor has a wider detection range, which can are sensitive to temperature, a temperature sensor is usually 223
187 detect both combustible and inert gases [37]. To achieve this added inside or outside of the infrared sensor to properly adjust 224
188 broad spectrum, as shown in Fig. 3, a servo circuit is added the background temperature. 225
189 to adjust the sensitive current flowing through the Wheatstone
190 bridge. When the conductivity is high, the thermistor is easy to
191 dissipate heat, and its resistance will also decrease. Therefore, E. Photoionization Detector Gas Sensor 226
192 the interface circuit can change the resistance through the
A PID photoionization gas sensor is composed of the main 227
193 signal adjustment and conversion circuit (the electrical signal
parts of an ultraviolet light source and an ion chamber [43]. 228
194 can be converted into a useful signal for display, recording,
First, the positive and negative electrodes in the ion chamber 229
195 and control), thereby converting the Wheatstone bridge into
form an electric field, and then, under irradiation from an 230
196 an unbalanced state.
ultraviolet lamp, the gas to be measured is ionized, generating 231

positive and negative ions; a current is formed between the 232

197 D. Infrared Gas Sensor electrodes, finally amplifying the output signal. 233

198 Infrared gas sensors use continuously varying frequency As shown in Fig. 5, the circuit portion of the PID pho- 234

199 infrared light to illuminate the gas to be measured and use toionization sensor is similar to the infrared gas sensor. The 235

200 different infrared absorption spectra of different gas mole- photoionization sensor only includes a signal detection circuit, 236

201 cules [41]. Thus, when the infrared wavelength matches the as well as an ultraviolet light power drive circuit, a fan control 237

202 absorption spectrum of the gas to be measured, the infrared circuit, and other controls. Because of the need for photoion- 238

203 energy will be absorbed by the sensor and processed by the ization, the current detected by the operational amplifier is 239

204 interface circuit in the form of electrical signals to determine extremely weak (less than 1 nA). For this reason, this type 240
 18256 IEEE SENSORS JOURNAL, VOL. 22, NO. 19, 1 OCTOBER 2022

(RV-ADC) to represent this type of technology. However, the 284

sensitive circuit based on oscillation uses the principle that 285

the charge and discharge of the RC circuit will produce a 286

time constant change, and the sensor resistance change will 287

be displayed in a periodic form and then combined with the 288

signal processing circuit to adjust it into a digital output signal. 289

This article also uses resistance to period, frequency signal into 290

a digital signal (RP-FDC) to represent this type of technology. 291

Fig. 6. Interface circuit of the MOX gas sensor. A. RV-ADC Technology 292

The sensitive circuit based on the RV-ADC method was 293

proposed by Grassi et al. [14], whose team at the University 294

241 of interface circuit requires a combination of weak signal of Pavia in Italy attempted to improve the dynamic range 295

242 detection technology. and measurement accuracy of the gas sensor. As shown in 296

Fig. 7, the circuit consisted of programmable gain operational 297

amplifiers (PGAs), one ADC, and two digital-analog convert- 298

243 F. Metal Oxide Semiconductor Gas Sensor
ers (DACs). As the working principle, the potential at one end 299
244 An MOX sensor is mainly composed of a heating element
of the sensitive resistor is fixed to convert the resistance change 300
245 and a gas sensor [44]. At a certain temperature, the gas sensor
into a current change, which is then converted into a voltage 301
246 can adsorb the gas, and this adsorption capacity changes with
signal through the PGA circuit, and it, finally, enters the ADC. 302
247 gas composition and concentration; thus, it can show different
The generated digital signal can be used to calibrate the circuit, 303
248 conductivity [45]. The interface circuit obtains the relevant
which can effectively weaken the offset signal generated by 304
249 information of the gas to be measured by detecting the changes
the baseline drift and operational amplifier work. However, 305
250 in resistance. The MOX gas sensor has the advantages such
it can also be used to control the PGA circuit to select the 306
251 as high sensitivity, a fast response time, a long life span,
appropriate circuit gain to increase the measurement range of 307
252 and low cost. However, it requires heating, so it has high
the sensitive resistance value. Experiments have verified that 308
253 power consumption and is easily affected by the environment,
the measurement accuracy can reach 0.1%, with a measure- 309
resulting in low gas selectivity.
ment resistance range within 100 –20 M. In 2012, Ha et al.
254
310
255 Fig. 6 shows a simplified topology diagram of the MOX
[46] at the Pohang University of Science and Technology used 311
256 sensor interface circuit, where R H is the heating resistance,
this method to design a low-power interface circuit based on 312
257 V H is the heating voltage, R S is the gas-sensitive resistance,
CDS technology. The CDS technology reduced the operational 313
258 R L is the load resistance, and VL is the voltage across the
amplifier to offset and improve the noise reduction capability 314
259 load resistance. This type of interface circuit can integrate
of the circuit. In addition, the 12-bit SAR ADC adopted the 315
260 the heating module with the sensor module. When the sensor
working mode of time-interleaved single-differential sampling, 316
261 starts to measure the gas, the heating resistor will control
which was conducive to improving work efficiency and reduc- 317
262 the temperature until the resistance of the sensitive resistor
ing power consumption. Finally, resistance changes of 23 k– 318
263 becomes stable. At this point, the sensor will convert chemical
4.6 M were realized, and the power consumption was only 319
signals into electrical signals to obtain VL . At this time,
11.3 μW under a power supply voltage of 0.5 V.
264
320
265 the operational amplifier will process the signal further to
In 2019, the Moscow State Bauman University of Tech- 321
266 obtain VO .
nology in Russia designed a low-noise, wide-bandwidth, and 322
267 The detection of toxic and harmful gases has become an
wide-dynamic-range-sensitive circuit based on RV-ADC tech- 323
268 important branch of environmental monitoring and chemical
nology [47]. This sensitive circuit optimized the traditional 324
269 analysis applications. The development of these types of
circuit by cascading two PGA circuits with different functions. 325
270 sensors is also changing with the application requirements.
The first-level circuit added a capacitor array to the traditional 326
271 However, this also poses challenges for signal processing
structure and used digital unit control to achieve variable gains 327
272 of the back-end interface circuit. In Section III, we use a
while also achieving adjustable bandwidth, avoiding signal 328
273 semiconductor gas sensor as an example to introduce its
aliasing, and reducing noise. The purpose of the second-level 329
274 typical development in an interface circuit.
circuit was to achieve high gain and high slew rates. Finally, 330

after testing, a resistance range of 50 k–1 G was real- 331

275 III. D EVELOPMENT OF S ENSITIVE C IRCUITS ized, and the input reference noise was only 2.14 pArms at 332

276 Gas sensor interface circuits can be divided into two cate- 1 kHz with an input reference noise of only 0.84 nArms at 333

277 gories: the sensitive part and the temperature modulation part. 1 MHz. Traditional circuits based on RV-ADC technology 334

278 The sensitive circuit is usually based on ADC technology and have gradually developed toward low noise and low power 335

279 oscillation technology. The sensitive circuits based on ADC consumption, and with a wide dynamic range. Of course, there 336

280 technology use ADC structures or other circuits to realize are also some teams, such as Cheng-Ta Chiang, from National 337

281 the analog signal conversion to the digital signal output by Taiwan University, which are researching performance, such 338

282 measuring changes in voltage across the sensor resistance. This as reducing the chip area and complexity [48], [49]. In 2016, 339

283 article uses resistance to voltage and analog into digital signals Chiang’s team designed a digital gas sensor to measure CO2 340
 REN et al.: TOWARD A GAS SENSOR INTERFACE CIRCUIT 18257

Fig. 7. Sensitive circuit based on RV-ADC technology.

TABLE I
P ERFORMANCE I NDICATORS OF S ENSITIVE C IRCUITS B ASED ON RV-ADC T ECHNOLOGY IN THE PAST 15 Y EARS

341 concentrations. The sensor used a PGA module to discretely is proposed in [52], where the circuit power consumption 357

342 output the voltage and then used a fourth-order continuous- was only 15.51 μW. Considering all related circuits, the 358

343 time (CT) sigma–delta modulation (SDM) circuit to quantify sensitive circuit based on RV-ADC technology has the most 359

344 and reduce the noise of the voltage signal and, finally, through significant advantages in terms of measurement accuracy. 360

345 latching a comparator obtained the digital output. Due to the Except for the research conducted by Dudina et al. [47] and 361

346 reduction in ADC modules, the chip area was only 2.92 mm2 . Shin et al. [53], the accuracy of other circuits is approximately 362

347 Many sensitive circuits have been made using similar methods. less than 1%, with some circuits even less than 0.023% [49], 363

348 Related circuits from the past 15 years are sorted from top [50], [54]. In terms of chip area, the chip area of [53], 364

349 to bottom in chronological order in Table I. The table shows [55], and [56] was less than 1 mm2 . Due to the popular- 365

350 that most of the circuits from the past ten years have been ity of gas sensors in portable applications, the chip area 366

351 based on 180-, 350-, and even 500-nm CMOS processes [50]. is the primary issue that restricts the development of this 367

352 In recent years, 55-nm CMOS processes [51] have started to technology. 368

353 appear, as well as sensitive circuits based on a 65-nm CMOS

354 process [52]. With process advancements, the sensitive circuit B. RP-FDC Technology 369

355 can be driven by a smaller power supply voltage, offering great Because not all gas sensors require high accuracy of 0.1%, 370

356 progress in terms of power consumption. A typical example researchers have proposed RP-FDC technology. This type of 371
 18258 IEEE SENSORS JOURNAL, VOL. 22, NO. 19, 1 OCTOBER 2022

Fig. 8. Sensitive circuit based on RP-FDC technology.

372 sensitive circuit has a simple structure, is easy to design, and resistor branch, as shown in Fig. 8. The switch always acted 410

373 does not introduce additional error signals; therefore, in appli- on the high-impedance node without introducing additional 411

374 cations that do not require high precision, this technology can error signals. Furthermore, capacitive sensing technology was 412

375 save design time and costs. Fig. 8 shows a circuit schematic of introduced into the integrating circuit, which made the sensi- 413

376 the sensitive circuit. The working principle is given as follows. tive circuit suitable for both resistance and capacitance readout 414

377 First, the potential at both ends of the sensitive resistor is circuits. Thus, a 130-nm CMOS process was adopted, where 415

378 fixed through the operational amplifier, and as long as the the dynamic range was 148 dB, the accuracy was 0.4%, and 416

379 sensitive resistance changes, a corresponding sensitive current the overall layout area was 130 000 μm2 . 417

380 will be generated. Then, it is proportionally copied to the In 2019, Ciciotti et al. [31] used 130-nm CMOS technology 418

381 integrating circuit through the current mirror structure, and this to design a low-power, small-area, and high-linearity ASIC 419

382 circuit converts the current signal into a triangle wave. Finally, circuit. This circuit could detect sensitive resistors in the range 420

383 the voltage signal is modulated by a comparator, producing of 100 –1 M, with a chip area of 0.13 mm2 , and the power 421

384 a square wave digital output signal. Based on this method, consumption was only 0.68 mW. The experiment also verified 422

385 in 2007, Grassi et al. [15] designed a sensitive circuit with a the ability of this circuit to detect CO at 5 ppm in the air. 423

386 wide dynamic range. Circuits related to this technology over the past 15 years are 424

387 When a circuit is not calibrated, the dynamic range can chronologically sorted and presented in Table II. As shown in 425

388 reach 141 dB, and the measurement accuracy can reach Table II, this technique offers significant advantages over the 426

389 0.4%. To create a gas sensor that always works at the RV-ADV approach in terms of both dynamic range and power 427

390 optimal temperature, in 2013, the team designed a sensi- consumption. Considering measurement range, this technol- 428

391 tive circuit based on this method, which could dynamically ogy can generally achieve changes of four to six decades. 429

392 adjust the temperature at 1.5 ◦ C increments in the range of Specifically, Ciciotti et al. [31] had the largest detection range, 430

393 25 ◦ C–500 ◦ C [65]. In addition, to enhance the matching reaching 7.5 decades. Moreover, the dynamic range of [65], 431

394 accuracy of the current mirror, two additional operational [66], and [68] exceeded 150 dB, which was suitable for 432

395 amplifiers were added to control the grid voltage of the current detecting the concentration changes of various gases. In terms 433

396 mirror branch tube, and these structures were used to fix the of power consumption, the power consumption of [31], [67], 434

397 voltage across the sensitive resistor, which could effectively [69], [70], [71], [72], and [73] was all less than 1 mW and 435

398 isolate the noise caused by the ground and the power supply. could be used in portable electronic devices. However, this 436

399 The dynamic range of the final measurement reached 162 dB, type of technology does not perform very well in terms of 437

400 and the accuracy was 0.85%. accuracy. The best example over 15 years is [71], which 438

401 In 2018, Hijazi et al. [66] designed a sensitive circuit reached 0.015%. Although the accuracy requirements for this 439

402 that met the requirements for advanced portable gas sensors type of technology are not high, the long-term operation of 440

403 based on the above methods. Experiments showed that the the circuit will continue to introduce new errors into the 441

404 sensitive circuit could measure resistance in the range of sensitive circuit. Therefore, to be better used for gas detection, 442

405 10 k–10 G, with an accuracy of less than 0.9% and the a breakthrough in accuracy is necessary. 443

406 dynamic range reaching 161 dB. To avoid the parasitic effect
407 caused by the selection of different sensitive resistors by C. Other Technology 444

408 the gas sensor with a multiplex structure, in the same year, In addition to RV-ADC and RP-FDC technologies, 445

409 Ciciotti et al. [67] optimized the circuit and added a reference some other technologies can also combine the above two 446
 REN et al.: TOWARD A GAS SENSOR INTERFACE CIRCUIT 18259

TABLE II
P ERFORMANCE I NDICATORS OF S ENSITIVE C IRCUITS B ASED ON RP-FDC T ECHNOLOGY IN THE PAST 15 Y EARS

TABLE III
P ERFORMANCE I NDICATORS OF S ENSITIVE C IRCUITS B ASED ON O THER T ECHNOLOGY IN THE PAST 15 Y EARS

447 technologies to improve the performance of gas sensors. error of less than 0.03% in the full span and did not require 475

448 As shown in Table III, they are the use of microcontrollers, LPF and ADC circuits, which had a simple structure and 476

449 current conveyors, lock-in amplifiers (LIAs), autobalancing low cost. This research also provided theoretical support for 477

450 interfaces, and MEMS technology. RP-FDC technology. 478

451 In 2009, Reverter and Casas [83] at the University of A useful block for high-frequency current mode applications 479

452 Girona, Spain, proposed an MOX sensor that connected a is the second-generation current conveyor (CCII) proposed by 480

453 differential resistor directly to a microcontroller. The front- Sedra and Smith [86]. CCII topology has been widely used in 481

454 end-sensitive circuit had a simple structure [83], consisting of portable systems due to its low noise, low bias, low voltage, 482

455 only a few RC circuits and two sensor resistors. The working and low power consumption. In 2006, Ferri et al. [87] pre- 483

456 principle was also simple. When the sensor resistance changed, sented an integrated improved CCII topology for the resistive 484

457 the discharge time of the RC circuit was measured, which sensor. In 2009, the team proposed a low-cost fully integrable 485

458 was then converted into an electrical signal. Finally, the ADC CM interface for gas sensors [88]. The read-out circuit is 486

459 circuit is used for signal transmission and adjustment. The able to evaluate the resistive behavior of the utilized sensor. 487

460 experiment showed that the nonlinear error of the full-scale In 2011, Mu et al. [50] at Michigan State, USA, proposed 488

461 circuit was 0.01%. However, the measurement time of the sen- a sensitive circuit based on current conveyor technology to 489

462 sitive circuit is as long as 100 ms, and the mismatch between eliminate issues related to different baseline values generated 490

463 the internal resistance of the microcontroller will also affect by different sensor operations [50]. As shown in Fig. 9, this 491

464 the system error seriously. In 2010, the team proposed an MOX circuit was composed of a source bias circuit, a subtractive 492

465 sensor that connected the differential capacitor directly to the gain circuit, and a DAC module. This circuit could realize 493

466 microcontroller [84]. Experiments verified that the nonlinearity the function of using an adaptive CR array to select different 494

467 error of the full-scale circuit reached 1%. Although the above current mirror branches according to the specific applications. 495

468 two sensitive circuits were not suitable for modern high- The subtractive gain circuit did not have feedback resistors 496

469 speed and high-precision sensing applications, they provided or other circuits. When a gain stage was used to eliminate 497

470 valuable scientific research experience for the related research circuit noise, it greatly reduced the chip area and circuit power 498

471 of MOX sensors based on resistance and capacitance. In 2020, consumption. Finally, experiments showed that the power 499

472 Reverter and Gasulla [85] evaluated a digital sensor-sensitive consumption was only 66 μW when the power supply voltage 500

473 circuit based on timer demodulation. Experiments showed that was 3.3 V. In 2020, Ahmad et al. [89] designed an ASIC 501

474 the sensitive circuit using timer demodulation had a nonlinear circuit for an automatic calibration resistance measurement 502
 18260 IEEE SENSORS JOURNAL, VOL. 22, NO. 19, 1 OCTOBER 2022

combines multiple disciplines, advanced technology can be 545

used to integrate the ADC. Also, sensors and sensitive circuits 546

can be designed on a chip, resulting in reduced chip area, 547

and saving power consumption loss between the modules. 548

For example, in 2018, Wang et al. [28] combined CMOS 549

technology and integrated an on-chip ADC module using 550

MEMS technology to design a high-sensitivity circuit that 551

could detect different types of gases with a concentration 552

of <1 ppm. The chip area was also less than 3 mm2 , 553

which improved the performance of the sensitive circuit in 554

the RV-ADC structure, the three-electrode structure heating 555

technology, and the programmable current source bias tech- 556

nology. The RP-FDC technology also improved the accuracy 557

Fig. 9. Programmable current conveyors’ technology.
of the chip due to the integration of the on-chip ADC 558

module. In addition, the advantages of high integration also 559

enabled the sensor to integrate more modules and become 560

503 system. The system adopted programmable current source bias intelligent and multifunctional. For example, in 2020, Prof. 561

504 and chopping modulation technology to realize the four-phase Yeongjin Lim and Prof. Professor Heungjoo Shin at the 562

505 automatic calibration function, which was used to compensate Ulsan University of Science and Technology, South Korea, 563

506 for the resistance baseline value with a high degree of accu- proposed a multifunctional, efficient, and multichannel gas 564

507 racy. When the measured resistance was 0.5 M, a relative sensing detection system that combined MEMS technology 565

508 error was less than ±1%. with deep learning optimization algorithms [64]. Deep learning 566

509 A LIA has been quite a versatile measurement instru- can provide advantages such as accuracy and power saving for 567

510 ment for extracting low-level signals buried in noise. multimode operation (pattern recognition and self-calibration). 568

511 The importance of LIA has been realized as a sensor- Through experiments that compared, the researchers showed 569

512 interfacing platform, and a number of portable implementa- that, compared to a system without a deep learning algorithm, 570

513 tions targeting sensing applications have been reported [90]. the efficiency of the system increased by 2.6 times. However, 571

514 De Marcellis et al. [6] presented a low-voltage low-power the shortcoming was a combination of hardware and software 572

515 phase-locked amplifier for low-frequency gas sensors. The technologies, which increased the design difficulty and cost; 573

516 amplifier adopts standard CMOS technology and has analog however, it did advance the development of intelligent gas 574

517 characteristics. This team proposed a novel fully analog LIA, sensors. 575

518 which is suitable for the accurate measurements of low quan-

519 tities of gas [91]. In 2016, this team presented an improved IV. D EVELOPMENT OF T EMPERATURE 576

520 LIA that was designed as an integrated circuit using 0.35-μm M ODULATION C IRCUITS 577

521 standard CMOS technology with low-current and low-power MOX gas sensors normally work at temperatures between 578

522 characteristics, and it could automatically calibrate the relative 200 ◦ C and 500 ◦ C. The sensing process of MOX gas 579

523 phase between the input signal and the reference signal at any sensors strongly depends on the substrate temperature. The 580

524 change during system startup and operation [92]. temperature modulation that exploits the dynamic properties 581

525 The autobalancing bridge interface circuit was realized by of the sensor changes its selectivity and also improves the 582

526 analog technology in the early stage. Mantenuto et al. [93] stability, the time response, and the power consumption [36]. 583

527 presented a simple analog interface for automatically and The conventional temperature control circuit is based on a 584

528 continuously estimating large range resistance changes of constant power heating scheme instead of directly controlling 585

529 sensors whose baseline is unknown. This team presented a the temperature [99]. However, since the heater temperature 586

530 suitable Wheatstone bridge circuit to adopt a new interface depends on external factors, such as humidity, ambient tem- 587

531 for grounding and floating wide-range resistance sensor esti- perature, or airflow, the circuit technology based on constant 588

532 mation [94]. In 2014, they proposed an improved solution heating power has been abandoned [100]. Therefore, in order 589

533 to an automatic fully analog uncalibrated Wheatstone bridge to continuously control the temperature using voltage or 590

534 [95]. Digital technology further improved the autobalancing current, researchers have adopted a variety of methods, such 591

535 interface. Jain and George [96] presented a self-balancing as PWM technology and self-adaptive technology [57], [61], 592

536 digital method of resistance sensor based on a half-bridge [36], [101]. 593

537 structure. Miyazaki et al. [97] proposed a Wheatstone bridge

538 with -modulated digitally controlled variable resistor. A. PWM Technology 594

539 Areekath et al. [98] proposed a novel capacitance-pulsewidth PWM is considered an efficient technique to apply power 595

540 self-balancing converter (CPC), which used sinusoidal excita- to load without dissipating much heat [102]. PWM is imple- 596

541 tion and works in a closed-loop structure. mented by series switching with microheater, which includes 597

542 In addition to the abovementioned sensor-sensitive circuit the ON – OFF, proportional (P), proportional–integral (PI), 598

543 technology, the emergence of MEMS technology has acceler- PI–derivative (PID), and Fuzzy modes [103], [104], [105], 599

544 ated the development of gas sensors. Because this technology [106], [107], [108], [109], [110]. 600
 REN et al.: TOWARD A GAS SENSOR INTERFACE CIRCUIT 18261

Fig. 10. Temperature modulation technology with PWM.

Fig. 11. Self-adaptive temperature modulation technology.

601 The ON – OFF control technique is a simple way to realize operation of the sensor with the reduction of one electrode, 645

602 closed-loop temperature control. However, its pulse current also ensured the symmetry of the circuit, and reduced the 646

603 and persistent jittering may cause instability and even damage introduced error. Experiments showed that the interface circuit 647

604 to the microheater. Malfatti et al. [103] proposed an ON / OFF could measure resistance from 100  to 20 M and consumed 648

605 control that reduced the power consumption of the circuits. 2.5 mW. In 2017, the team improved the three-electrode 649

606 The temperature control circuit power is 0.65 mW at 3.3 V. interface circuit [57] and successfully solved several problems, 650

607 It can reach an accuracy of 2.5 ◦ C in the temperature range such as lack of insulating layer, small resistance coverage, and 651

608 of 100 ◦ C–400 ◦ C. The temperature control loop can set the electrical asymmetry. By introducing ADC and DAC modules, 652

609 temperature of the microheater within the ambient temperature the sensor performance was significantly improved. Finally, 653

610 range of up to 300 ◦ C, and the maximum peak-to-peak experiments showed that the interface circuit could control the 654

611 deviation is less than 2.0 ◦ C [104]. temperature up to 250 ◦ C, in which the measurable resistance 655

612 An analog P temperature controller was proposed by range was 80 –2 M, the power consumption was only 656

613 Barrettino et al. [105]. The on-chip temperature controller can 1.8 mW, and the chip area was 0.64 mm2 . 657

614 adjust the temperature of the film to 350 ◦ C, and the resolution
615 is 0.5 ◦ C. However, the analog P controller is affected by
B. Self-Adaptive Technology 658
616 component aging, component process changes and temperature
617 drift. The adaptive temperature modulation is based on the fol- 659

618 Singh et al. [106] proposed the closed-loop switched PI lowing evidence: the sensitivity of the sensor resistance to the 660

619 temperature control scheme for platinum microheaters. The gas depends on the operating temperature, and conversely, the 661

620 results show that, compared with PI, the rapid initial rise of temperature sensitivity depends on the gas. 662

621 open-loop platinum heater can achieve a faster settling time. In 2012, Martinelli et al. [101] proposed a self-adapted tem- 663

622 In 2005, Conrad et al. [107] proposed a system that mainly perature modulation in MOX gas sensors. The main limitation 664

623 relies on PWM signals for temperature control. The PWM of the self-adapted temperature modulation must be that it is 665

624 signal generated by MCU makes this method very suitable for difficult to change the modulation parameters of the sensor 666

625 industrial applications. The system provides great flexibility in real-time measurement. Therefore, Di Giuseppe et al. [36] 667

626 by adjusting software components to adapt to a wide range improved this adaptive temperature modulation circuit using 668

627 of applications. In 2009, Lombardi et al. [108] proposed a digital platform in 2021. As shown in Fig. 11, the digital 669

628 a temperature control circuit with a digital set point. The platform actually replicated the behavior of the multivibrator, 670

629 temperature control circuit can adjust the temperature of the limited the use of analog components of the heater current 671

630 sensor within the range of 250 ◦ C, with an accuracy of better transmission part, and improved the flexibility of the modula- 672

631 than 1.5 ◦ C and a maximum peak-to-peak ripple of 1.0 ◦ C. tion interface. 673

632 In 2011, Bissi et al. [82] proposed a three-terminal ultralow

633 power MOX gas sensor electronic system based on micro- V. D ISCUSSION 674

634 control architecture, which eliminated the electrical isolation A. Sensitive Materials 675

635 between the heating module and the sensor device. In 2015, In addition to the interface circuit of the gas sensor, the 676

636 Jeong-Ho Park and other members of the Korean University sensitive materials of the gas sensor are also evolving, which 677

637 of Aeronautics and Astronautics designed a three-electrode develops from metal oxide (still the subject of many studies, 678

638 sensor using this method [61]. As shown in Fig. 10, the system especially nanometer structure) to the conductive polymers to 679

639 consisted of a heating module, a sensitive signal measurement carbon nanotubes and their composite materials, as well as the 680

640 module, a temperature control module, and a clock for system latest 2-D materials [111]. 681

641 operations. The control module used pulse signals with differ- 2-D materials, such as MoS2 and graphene, are consid- 682

642 ent clock phases to control the interface circuit so that it could ered attractive candidates for gas sensors due to their strong 683

643 self-adjust to ensure that the sensor always operated within the adsorption of gas molecules and their sensitivity to changes in 684

644 most appropriate temperature range. This ensured the normal the test environment [112], [113], [114], [115], [116]. When 685
 18262 IEEE SENSORS JOURNAL, VOL. 22, NO. 19, 1 OCTOBER 2022

686 there is a specific gas, the adsorption of gas molecules will technology based on the application type. Among these, RP- 741

687 change the carrier concentration of the thin film, resulting in FDC technology is suitable for applications with low power 742

688 the change of the resistivity and dielectric constant of the consumption, small chip area, and low precision. However, 743

689 device, thus realizing the detection of the specific gas. The for sensors, high precision is a rigid requirement. Therefore, 744

690 unique structural characteristics of 2-D materials make them a this technology must be combined with error elimination 745

691 candidate material for gas sensors. In order to improve the technology, or the technology must be improved to match 746

692 sensitivity of gas sensors, 2-D materials are processed by the current mirror in order to ensure a small surface area, 747

693 doping, modifying, and changing the number of layers. reduce offset errors, and improve precision. RV-ADC tech- 748

694 However, due to some difficulties in the preparation process, nology is suitable for applications that require high precision 749

695 2-D materials have the disadvantages of small scale, low and a large dynamic range. This technology requires high- 750

696 market share, and high cost. Especially for the detection of performance ADC, large areas, and complex structures. Thus, 751

697 strongly adsorbed gas molecules, the recovery characteristics studies should focus on reducing the area and complexity 752

698 of the sensor are poor. Therefore, it is necessary to find a without affecting the performance of the ADC. Capacitor- 753

699 method to quickly remove the residual gas on the surface of resistance hybrid SAR ADC can also be used as the core 754

700 the sensor material in order to restore the performance of the module. This structure can reasonably design the number of 755

701 sensor in a short time. Although 2-D transition metal chalco- capacitors and resistors and uses a large number of digital 756

702 genide materials have high gas sensitivity, their selectivity to MOS tubes, which have significant advantages in reducing 757

703 different gases should also be considered in order to accurately power consumption and size. In addition to RV-ADC and RP- 758

704 detect the concentration of specific gas in the mixture. FDC technologies, microcontrollers, current conveyors, LIAs, 759

autobalancing interfaces, and MEMS technologies are also 760

used to improve the performance of gas sensors. At the 761

705 B. Interface Circuits
end of this article, temperature modulation is introduced, 762
706 MOX gas sensors are the ideal choice for realizing gas which improves the performance of the sensor by using its 763
707 sensing systems due to their advantages such as low cost, dynamic characteristics, and it is also an indispensable part 764
708 long service life, high sensitivity, small size, simple operation, of the interface circuit of the gas sensor. PWM is a general 765
709 flexible production, and compatibility with standard CMOS temperature modulation technology. The power consumption 766
710 processes. The requirements of the MOX sensors’ interface of the gas sensor is reduced by controlling the duty cycle of 767
711 circuits include wide DR (over four decades), high accuracy, PWM. Self-adaptive temperature modulation techniques can 768
712 high precision (below 0.5%), and low power consumption adapt to different sensors or dynamically reduce the baseline 769
713 (below 1mW) [31]. drift by modifying the software parameters. 770
714 RV-ADC technology is an implemented solution for the
715 MOX sensor interface circuit. A typical sensitive circuit con-
716 sists of a voltage divider or a Wheatstone bridge, which R EFERENCES 771

717 is connected to the ADC through a buffer. Although the [1] M. I. A. Asri, M. N. Hasan, M. R. A. Fuaad, Y. Md. Yunos, and 772

718 implementation of this method is very simple, it requires M. S. M. Ali, “MEMS gas sensors: A review,” IEEE Sensors J., vol. 21, 773
no. 17, pp. 18381–18397, Sep. 2021. 774
719 a high-power ADC whose resolution is proportional to the [2] Y. Kang, H.-I. Jung, D.-H. Baek, K. Lee, S. Pyo, and J. Kim, 775
720 required DR. In addition, in order to maintain high sensitivity “Highly sensitive detection of VOC using impact ionization induced 776

721 and high accuracy, the gain of the programmable transresis- by photoelectron,” in Proc. IEEE 30th Int. Conf. Micro Electro Mech. 777
Syst. (MEMS), Las Vegas, NV, USA, Jan. 2017, pp. 183–186. 778
722 tance amplifier (PTA) is adjusted to realize the autoranging
[3] Y. Sadaoka and M. Mori, “Detection of VOC in air with a planar- 779
723 procedure. type potentiometric gas sensor based on YSZ with a PT electrode 780
724 A more effective method is based on the RP-FDC technolo- modified with TiO2 ,” Sens. Actuators B, Chem., vol. 248, pp. 878–885, 781

725 gies. Such circuits are based on a relaxed or ring oscillator Sep. 2017. 782
[4] J. W. Gardner, P. K. Guha, F. Udrea, and J. A. Covington, “CMOS 783
726 structure, whose output signal period (or frequency) carries interfacing for integrated gas sensors: A review,” IEEE Sensors J., 784
727 the indirect measurement value of the sensor. Since DR is vol. 10, no. 12, pp. 1833–1848, Dec. 2010. 785

728 unlimited by voltage, the system is not that complicated. [5] J. Wüsten and K. Potje-Kamloth, “Chalcogenides for thin film NO 786
sensors,” Sens. Actuators B, Chem., vol. 145, no. 1, pp. 216–224, 787
729 On the basis of simple topology, RP-FDC has a high dynamic Jan. 2010. 788
730 range and good linearity. [6] A. D’Amico et al., “Low-voltage low-power integrated analog lock- 789
in amplifier for gas sensor applications,” Sens. Actuators B, Chem., 790
vol. 144, no. 2, pp. 400–406, Feb. 2010. 791
731 VI. C ONCLUSION [7] R. Khakpour, M. N. Hamidon, and G. A. E. Vandenbosch, “Develop- 792
ment of an auto-calibrated interfacing circuit for thick film multi-gas 793
732 This article first introduced the working principles of six sensor,” Sens. Actuators A, Phys., vol. 204, pp. 48–57, Dec. 2013. 794
733 different types of gas sensors and the corresponding interface [8] E. Yoon and K. D. Wise, “An integrated mass flow sensor with on- 795
734 circuit functions, including metal semiconductor gas sensors, chip CMOS interface circuitry,” IEEE Trans. Electron Devices, vol. 39, 796

735 electrochemical gas sensors, catalytic combustion gas sensors, no. 6, pp. 1376–1386, Jun. 1992. 797
[9] C. Hagleitner, D. Lange, A. Hierlemann, O. Brand, and H. Baltes, 798
736 infrared gas sensors, thermal conductivity gas sensors, and PID “CMOS single-chip gas detection system comprising capacitive, calori- 799
737 gas sensors. Then, this article summarized sensitive circuits metric and mass-sensitive microsensors,” IEEE J. Solid-State Circuits, 800

738 based on semiconductor gas sensor interface circuits and vol. 37, no. 12, pp. 1867–1878, Dec. 2002. 801
[10] P. Malcovati, M. Grassi, and A. Baschirotto, “Towards high-dynamic 802
739 analyzed the research significance of various technologies, range CMOS integrated interface circuits for gas sensors,” Sens. 803
740 which could be divided into RP-FDC technology and RV-ADC Actuators B, Chem., vol. 179, pp. 301–312, Mar. 2013. 804
 REN et al.: TOWARD A GAS SENSOR INTERFACE CIRCUIT 18263

805 [11] M. Y. Afridi et al., “A monolithic CMOS microhotplate-based gas [34] S. W. Chiu and K. T. Tang, “Towards a chemiresistive sensor-integrated 880
806 sensor system,” IEEE Sensors J., vol. 2, no. 6, pp. 644–655, Dec. 2002. electronic nose: A review,” Sensors, vol. 13, no. 10, pp. 14214–14247, 881
807 [12] M. Graf, D. Barrettino, M. Zimmermann, A. Hierlemann, H. Baltes, Oct. 2013. 882
808 and S. Hahn, “CMOS monolithic metal–oxide sensor system com- [35] C. Wang, L. Yin, L. Zhang, D. Xiang, and R. Gao, “Metal oxide gas 883
809 prisinga microhotplate and associated circuitry,” IEEE Sensors J., sensors: Sensitivity and influencing factors,” Sensors, vol. 10, no. 3, 884
810 vol. 4, no. 1, pp. 9–16, Feb. 2004. pp. 2088–2106, Mar. 2010. 885

811 [13] B. Guo, A. Bermak, P. C. H. Chan, and G.-Z. Yan, “A monolithic [36] D. Di Giuseppe, A. Catini, E. Comini, D. Zappa, C. Di Natale, and 886
812 integrated 4 × 4 tin oxide gas sensor array with on-chip multiplexing E. Martinelli, “Optimizing MOX sensor array performances with a 887
813 and differential readout circuits,” Solid-State Electron., vol. 51, no. 1, reconfigurable self-adaptive temperature modulation interface,” Sens. 888
814 pp. 69–76, Jan. 2007. Actuators B, Chem., vol. 333, Art. no. 129509, Apr. 2021. 889

815 [14] M. Grassi, P. Malcovati, and A. Baschirotto, “A 160 dB equivalent [37] T. Hübert, L. Boon-Brett, V. Palmisano, and M. A. Bader, “Develop- 890

816 dynamic range auto-scaling interface for resistive gas sensors arrays,” ments in gas sensor technology for hydrogen safety,” Int. J. Hydrogen 891

817 IEEE J. Solid-State Circuits, vol. 42, no. 3, pp. 518–528, Mar. 2007. Energy, vol. 39, no. 35, pp. 20474–20483, Dec. 2014. 892

818 [15] M. Grassi, P. Malcovati, and A. Baschirotto, “A 141-dB dynamic [38] E. Lee et al., “Micromachined catalytic combustible hydrogen gas 893

819 range CMOS gas-sensor interface circuit without calibration with 16- sensor,” Sens. Actuators B, Chem., vol. 153, no. 2, pp. 392–397, 894

820 bit digital output word,” IEEE J. Solid-State Circuits, vol. 42, no. 7, Apr. 2011. 895

821 pp. 1543–1554, Jul. 2007. [39] H. Li, X. Mu, Y. Yang, and A. J. Mason, “Low power multimode 896

822 [16] J. Fraden, Handbook of Modern Sensors, 2nd ed. New York, NY, USA: electrochemical gas sensor array system for wearable health and 897

823 Springer, 2010, pp. 399–429. safety monitoring,” IEEE Sensors J., vol. 14, no. 10, pp. 3391–3399, 898

824 [17] O. Gould, T. Wieczorek, B. de Lacy Costello, R. Persad, and Oct. 2014. 899

825 N. Ratcliffe, “Assessment of a combined gas chromatography mass [40] X. Mu, Z. Wang, X. Zeng, and A. J. Mason, “A robust flexible 900

826 spectrometer sensor system for detecting biologically relevant volatile electrochemical gas sensor using room temperature ionic liquid,” IEEE 901

827 compounds,” J. Breath Res., vol. 12, no. 1, Dec. 2017, Art. no. 016009. Sensors J., vol. 13, no. 10, pp. 3976–3981, May 2013. 902

828 [18] R. S. El Shamy, D. Khalil, and M. A. Swillam, “Mid infrared optical [41] T.-V. Dinh, I.-Y. Choi, Y.-S. Son, and J.-C. Kim, “A review on 903

829 gas sensor using plasmonic Mach–Zehnder interferometer,” Sci. Rep., non-dispersive infrared gas sensors: Improvement of sensor detection 904

830 vol. 10, no. 1, p. 1293, Jan. 2020. limit and interference correction,” Sens. Actuators B, Chem., vol. 231, 905

831 [19] Y. Toya, T. Itagaki, and K. Wagatsuma, “Correlation between the gas pp. 529–538, Aug. 2016. 906

832 temperature and the atomization behavior of analyte elements in flame [42] D. Gibson and C. MacGregor, “A novel solid state non-dispersive 907

833 atomic absorption spectrometry estimated with a Continuum-light- infrared CO2 gas sensor compatible with wireless and portable deploy- 908

834 source spectrometer system,” Spectrochimica Acta B, At. Spectrosc., ment,” Sensors, vol. 13, no. 6, pp. 7079–7103, May 2013. 909

835 vol. 125, pp. 146–151, Nov. 2016. [43] L. Spinelle, M. Gerboles, G. Kok, S. Persijn, and T. Sauerwald, 910

836 [20] T. Itoh et al., “Development of an exhaled breath monitoring system “Review of portable and low-cost sensors for the ambient air mon- 911

837 with semiconductive gas sensors, a gas condenser unit, and gas itoring of benzene and other volatile organic compounds,” Sensors, 912

838 chromatograph columns,” Sensors, vol. 16, no. 11, p. 1891, 2016. vol. 17, no. 7, p. 1520, Jun. 2017. 913

839 [21] H. Hadano, A. Miyagi, T. Okuno, Y. Nagawa, and Y. Ishiguro, “Devel- [44] N. Samotaev et al., “Technology of SMD MOX gas sensors rapid 914

840 opment of a catalytic combustion type gas sensor with low power prototyping,” Mater. Sci. Forum, vol. 977, pp. 231–237, Feb. 2020. 915

841 consumption,” ECS Trans., vol. 75, no. 16, pp. 195–198, Aug. 2016. [45] J. G. Monroy, J. González-Jiménez, and J. L. Blanco, “Overcoming 916
the slow recovery of MOX gas sensors through a system modeling 917
842 [22] L. Kassa-Baghdouche and E. Cassan, “Mid-infrared gas sensor based
approach,” Sensors, vol. 12, no. 10, pp. 13664–13680, Oct. 2012. 918
843 on high-Q/V point-defect photonic crystal nanocavities,” Opt. Quantum
844 Electron., vol. 52, no. 5, pp. 1–13, May 2020. [46] H. Ha, Y. Suh, S. O. Lee, H. Park, and J. Sim, “A 0.5V, 11.3-μW, 1- 919
kS/s resistive sensor interface circuit with correlated double sampling,” 920
845 [23] A. Setiabudi, G. Yuliani, and Z. Pawitan, “An oxygen sensor using
in Proc. IEEE Custom Integr. Circuits Conf., San Jose, CA, USA, 921
846 zinc-air electrochemical cells: A simple device to demonstrate applied
Sep. 2012, pp. 1–4. 922
847 technology in electrochemistry teaching,” School Sci. Rev., vol. 97,
[47] A. Dudina, F. Seichepine, Y. Chen, A. Stettler, A. Hierlemann, 923
848 no. 361, pp. 30–34, 2016.
and U. Frey, “Monolithic CMOS sensor platform featuring an array 924
849 [24] D. Oletić, V. Jelicic, D. Antolović, and V. Bilas, “Energy-efficient of 9’216 carbon-nanotube-sensor elements and low-noise, wide- 925
850 atmospheric CO concentration sensing with on-demand operating MOX bandwidth and wide-dynamic-range readout circuitry,” Sens. Actuators 926
851 gas sensor,” in Proc. IEEE Sensors, Valencia, Spain, Nov. 2014, B, Chem., vol. 279, pp. 255–266, Jan. 2019. 927
852 pp. 795–798.
[48] C.-T. Chiang and C.-C. Hsieh, “Design of a CMOS digitized gas 928
853 [25] H.-Q. Zhang, B. Shen, and X.-L. Liu, “Fabrication and character- transducer with noise shaping for CO2 concentration monitoring appli- 929
854 izations of microelectromechanical systems (MEMS) thermal con- cations,” IEEE Sensors J., vol. 16, no. 4, pp. 975–982, Feb. 2016. 930
855 ductivity gas sensor,” J. Nanoelectron. Optoelectron., vol. 13, no. 7, [49] Z. Cai et al., “A ratiometric readout circuit for thermal-conductivity- 931
856 pp. 1076–1081, Jul. 2018. based resistive CO2 sensors,” IEEE J. Solid-State Circuits, vol. 51, 932
857 [26] S. O. Agbroko and J. Covington, “A novel, low-cost, portable PID no. 10, pp. 2463–2474, Oct. 2016. 933
858 sensor for the detection of volatile organic compounds,” Sens. Actuators [50] X. Mu, D. Rairigh, and A. J. Mason, “125ppm resolution and 120dB 934
859 B, Chem., vol. 275, pp. 10–15, Dec. 2018. dynamic range nanoparticle chemiresistor array readout circuit,” in 935
860 [27] J. Wang et al., “A fully integrated gas detection system with program- Proc. IEEE Int. Symp. Circuits Syst. (ISCAS), Rio de Janeiro, Brazil, 936
861 mable heating voltage and digital output rate for gas sensor array,” May 2011, pp. 2213–2216. 937
862 IEEE Sensors J., vol. 21, no. 5, pp. 6821–6829, Mar. 2021. [51] M. Konijnenburg et al., “22.1 A 769μW battery-powered single-chip 938
863 [28] J. Wang et al., “Gas detection microsystem with MEMS gas sensor and SoC with BLE for multi-modal vital sign health patches,” in Proc. 939
864 integrated circuit,” IEEE Sensors J., vol. 18, no. 16, pp. 6765–6773, IEEE Int. Solid-State Circuits Conf. (ISSCC), San Francisco, CA, USA, 940
865 Apr. 2018. Feb. 2019, pp. 360–362. 941
866 [29] T. Lin, X. Lv, Z. Hu, A. Xu, and C. Feng, “Semiconductor metal oxides [52] R. Agarwala, P. Wang, H. L. Bishop, A. Dissanayake, and B. H. Cal- 942
867 as chemoresistive sensors for detecting volatile organic compounds,” houn, “A 0.6 V 785-nW multimodal sensor interface IC for ozone 943
868 Sensors, vol. 19, no. 2, p. 0233, Jan. 2019. pollutant sensing and correlated cardiovascular disease monitoring,” 944
869 [30] D. Rüffer, F. Hoehne, and J. Bühler, “New digital metal-oxide (MOx) IEEE J. Solid-State Circuits, vol. 56, no. 4, pp. 1058–1070, Apr. 2021. 945
870 sensor platform,” Sensors, vol. 18, no. 4, p. 1052, Mar. 2018. [53] Y.-S. Shin, S. Lee, J.-K. Wee, and I. Song, “A small-area low-power 946
871 [31] F. Ciciotti et al., “A 450-μA 128-dB dynamic range A/D CMOS current readout circuit using two-stage conversion method for 64- 947
872 interface for MOX gas sensors,” IEEE Sensors J., vol. 19, no. 24, channel CNT sensor arrays,” IEEE Trans. Biomed. Circuits Syst., vol. 7, 948
873 pp. 12069–12078, Dec. 2019. no. 3, pp. 276–284, Jun. 2013. 949
874 [32] A. de Marcellis, G. Ferri, and P. Mantenu, “Uncalibrated operational [54] D. Lin, C. Tu, S. Kuo, and T. Lin, “An analog front-end circuit for 950
875 amplifier-based sensor interface for capacitive/resistive sensor applica- CO2 sensor readout in 0.18-μm CMOS process,” in Proc. Int. Symp. 951
876 tions,” IET Circuits, Devices Syst., vol. 9, no. 4, pp. 249–255, Jul. 2015. VLSI Design, Autom. Test, Taiwan, Apr. 2019, pp. 1–4. 952
877 [33] P. R. Nagarajan, B. George, and V. J. Kumar, “A linearizing digitizer [55] T. S. Cho, K. Lee, J. Kong, and A. Chandrakasan, “A 32 W 1.83-kS/s 953
878 for wheatstone bridge based signal conditioning of resistive sensors,” carbon nanotube chemical sensor system,” IEEE J. Solid-State Circuits, 954
879 IEEE Sensors J., vol. 17, no. 6, pp. 1696–1705, Mar. 15, 2017. vol. 44, no. 2, pp. 659–669, Feb. 2009. 955
 18264 IEEE SENSORS JOURNAL, VOL. 22, NO. 19, 1 OCTOBER 2022

956 [56] T. Tzeng et al., “A portable micro gas chromatography system for [77] A. Depari, A. Flammini, E. Sisinni, A. De Marcellis, G. Ferri, and 1031
957 volatile compounds detection with 15ppb of sensitivity,” IEEE J. Solid- P. Mantenuto, “Fast, versatile, and low-cost interface circuit for elec- 1032
958 State Circuit, vol. 51, no. 1, pp. 259–272, Jan. 2016. trochemical and resistive gas sensor,” IEEE Sensors J., vol. 14, no. 2, 1033
959 [57] J.-H. Park, K.-M. Park, T.-W. Kim, S. Shin, C.-O. Park, and H.-J. Yoo, pp. 315–323, Feb. 2014. 1034
960 “Three-electrode metal-oxide gas sensor system with CMOS interface [78] C. Li et al., “A high linearity detection circuit with constant detection 1035
961 IC,” IEEE Sensors J., vol. 17, no. 3, pp. 784–793, Feb. 2017. voltage for resistive gas sensor,” in Proc. IEEE 4th Adv. Inf. Technol., 1036
962 [58] C. Chiang, M. Huang, and M. Chung, “A CMOS CO2 concentration to Electron. Autom. Control Conf. (IAEAC), Chengdu, China, Dec. 2019, 1037
963 frequency converter with calibration circuits,” in Proc. IEEE Int. Freq. pp. 1051–1058. 1038
964 Control Symp. (FCS), Taiwan, May 2014, pp. 1–5. [79] R. Puyol, S. Petre, Y. Danlee, T. Walewyns, L. A. Francis, and 1039
965 [59] P. Livi et al., “Monolithic integration of a silicon nanowire FET array D. Flandre, “An ultra-low-power read-out circuit for interfacing novel 1040
966 on a CMOS chip for bio-chemical sensor applications,” Anal. Chem., gas sensors matrices,” IEEE Sensors J., vol. 22, no. 10, pp. 9521–9533, 1041
967 vol. 87, no. 19, pp. 9982–9990, Sep. 2015. May 2022. 1042

968 [60] M. Choi, J. Gu, D. Blaauw, and D. Sylvester, “Wide input range [80] P. M. Maya-Hernández, L. C. Álvarez-Simón, M. T. Sanz-Pascual, and 1043
969 1.7μW 1.2kS/s resistive sensor interface circuit with 1 cycle/sample B. Calvo, “An integrated low-power lock-in amplifier and its applica- 1044

970 logarithmic sub-ranging,” in Proc. IEEE Symp. VLSI Circuits (VLSI tion to gas detection,” Sensors Basel, vol. 14, no. 9, pp. 15880–15899, 1045

971 Circuits), Kyoto, Japan, Jun. 2015, pp. C330–C331. 2014. 1046

972 [61] J. Park, K. Park, T. Kim, C. Park, and H. Yoo, “Interface circuit for [81] M. Chen, S. Peng, N. Wang, L. Xu, F. Lin, and F. Wu, “A wide- 1047

973 three-electrode metal-oxide (MOX) gas sensor,” in Proc. IEEE Sensors, range and high-resolution detection circuit for MEMS gas sensor,” 1048
974 Busan, South Korea, Nov. 2015, pp. 1–4. IEEE Sensors J., vol. 19, no. 8, pp. 3130–3137, Apr. 2019. 1049

975 [62] J. Xu et al., “A 36 μW 1.1 mm2 reconfigurable analog front-end [82] L. Bissi, M. Cicioni, P. Placidi, S. Zampolli, I. Elmi, and A. Scorzoni, 1050

976 for cardiovascular and respiratory signals recording,” IEEE T. Biomed. “A programmable interface circuit for an ultralow power gas sensor,” 1051

977 Circuits Syst., vol. 12, no. 4, pp. 774–783, Aug. 2018. IEEE Trans. Instrum. Meas., vol. 60, no. 1, pp. 282–289, Jan. 2011. 1052

978 [63] Q. Lin et al., “A 119 dB dynamic range charge counting Light-to- [83] F. Reverter and Ò. Casas, “Interfacing differential resistive sensors 1053

979 Digital converter for wearable PPG/NIRS monitoring applications,” to microcontrollers: A direct approach,” IEEE Trans. Instrum. Meas., 1054

980 IEEE Trans. Biomed. Circuits Syst., vol. 14, no. 4, pp. 800–810, vol. 58, no. 10, pp. 3405–3410, Oct. 2009. 1055

981 Aug. 2020. [84] F. Reverter and O. Casas, “Interfacing differential capacitive sensors 1056

982 [64] K. Park et al., “An energy-efficient multimode multichannel gas-sensor to microcontrollers: A direct approach,” IEEE Trans. Instrum. Meas., 1057

983 system with learning-based optimization and self-calibration schemes,” vol. 59, no. 10, pp. 2763–2769, Oct. 2010. 1058

984 IEEE Trans. Ind. Electron., vol. 67, no. 3, pp. 2402–2410, Mar. 2020. [85] F. Reverter and M. Gasulla, “Demodulating AM square signals via 1059

985 [65] F. Conso, M. Grassi, C. De Berti, P. Malcovati, and A. Baschirotto, a digital timer for sensor applications,” IEEE Trans. Instrum. Meas., 1060

986 “A 0.85%-precision, 6-decade-range, I2 C—Programmable front-end vol. 69, no. 5, pp. 2593–2601, May 2020. 1061

987 ASIC for resistive gas-sensor arrays with 20-input analog MUX, digital [86] A. Sedra and K. C. Smith, “A second generation current conveyor 1062
988 output, and 1.5◦ C-accuracy dynamic temperature synthesis,” in Proc. and its applications,” IEEE Trans. Circuit Theory, vol. CT-17, no. 1, 1063
989 IEEE Sensors, Baltimore, MD, USA, Nov. 2013, pp. 1–4. pp. 132–134, Feb. 1970. 1064

990 [66] Z. Hijazi, M. Grassi, D. D. Caviglia, and M. Valle, “Time-based [87] G. Ferri, V. Stornelli, and M. Fragnoli, “An integrated improved CCII 1065
991 calibration-less read-out circuit for interfacing wide range MOX gas topology for resistive sensor application,” Anal. Integr. Circuits Signal 1066
992 sensors,” Integration, vol. 63, pp. 232–239, Sep. 2018. Process., vol. 48, no. 3, pp. 247–250, Sep. 2006. 1067

993 [67] F. Ciciotti, C. Buffa, R. Gaggl, and A. Baschirotto, “A program- [88] G. Ferri et al., “A CCII-based low-voltage low-power read-out circuit 1068
994 mable dynamic range and digital output rate oscillator-based readout for DC-excited resistive gas sensors,” IEEE Sensors J., vol. 9, no. 12, 1069
995 interface for MEMS resistive and capacitive sensors,” in Proc. Int. pp. 2035–2041, Dec. 2009. 1070
996 Conf. IC Design Technol. (ICICDT), Villach, BGL, Austria, 2018, [89] M. Ahmad et al., “An auto-calibrated resistive measurement system 1071
997 pp. 41–44. with low noise instrumentation ASIC,” IEEE J. Solid-State Circuits, 1072
998 [68] Z. Hijazi, M. Grassi, D. D. Caviglia, and M. Valle, “153 dB dynamic vol. 55, no. 11, pp. 3036–3050, Nov. 2020. 1073
999 range calibration-less gas sensor interface circuit with quasi-digital [90] K. Kishore and S. A. Akbar, “Evolution of lock-in amplifier as portable 1074
1000 output,” in Proc. New Gener. CAS (NGCAS), Genova, Italy, Sep. 2017, sensor interface platform: A review,” IEEE Sensors J., vol. 20, no. 18, 1075
1001 pp. 109–112. pp. 10345–10354, Sep. 2020. 1076
1002 [69] M. R. Valero, S. Celma, B. Calvo, and N. Medrano, “CMOS voltage-to- [91] A. De Marcellis, G. Ferri, A. D’Amico, C. Di Natale, and E. Martinelli, 1077
1003 frequency converter with temperature drift compensation,” IEEE Trans. “A fully-analog lock-in amplifier with automatic phase alignment for 1078
1004 Instrum. Meas., vol. 60, no. 9, pp. 3232–3234, Sep. 2011. accurate measurements of PPB gas concentrations,” IEEE Sensors J., 1079
1005 [70] J. H. L. Lu, M. Inerowicz, S. Joo, J.-K. Kwon, and B. Jung, “A low- vol. 12, no. 5, pp. 1377–1383, May 2012. 1080
1006 power, wide-dynamic-range semi-digital universal sensor readout cir- [92] A. De Marcellis, G. Ferri, and A. D’Amico, “One-decade frequency 1081
1007 cuit using pulsewidth modulation,” IEEE Sensors J., vol. 11, no. 5, range, in-phase auto-aligned 1.8 V 2 mW fully-analog CMOS inte- 1082
1008 pp. 1134–1144, May 2011. grated lock-in amplifier for small/noisy signal detection,” IEEE Sensors 1083
1009 [71] C. Azcona, B. Calvo, S. Celma, N. Medrano, and P. A. Martínez, J., vol. 16, no. 14, pp. 5690–5701, Jul. 2016. 1084
1010 “Ratiometric voltage-to-frequency converter for long-life autonomous [93] P. Mantenuto, A. De Marcellis, and G. Ferri, “Uncalibrated analog 1085
1011 portable equipment,” IEEE Sensors J., vol. 13, no. 6, pp. 2382–2390, bridge-based interface for wide-range resistive sensor estimation,” 1086
1012 Mar. 2013. IEEE Sensors J., vol. 12, no. 5, pp. 1413–1414, May 2012. 1087
1013 [72] K. C. Koay and P. K. Chan, “A low-power resistance-to-frequency [94] A. De Marcellis, G. Ferri, and P. Mantenuto, “Analog wheatstone 1088
1014 converter circuit with wide frequency range,” IEEE Trans. Instrum. bridge-based automatic interface for grounded and floating wide-range 1089
1015 Meas., vol. 64, no. 12, pp. 3173–3182, Dec. 2015. resistive sensors,” Sens. Actuators B, Chem., vol. 187, pp. 371–378, 1090
1016 [73] L. Álvarez-Simón, E. Gómez-Ramírez, and M. Sanz-Pascual, “Low- Oct. 2013. 1091
1017 power highly robust resistance-to-period converter,” Sensors, vol. 19, [95] P. Mantenuto, G. Ferri, and A. De Marcellis, “Uncalibrated auto- 1092
1018 no. 1, p. 8, Dec. 2018. matic bridge-based CMOS integrated interfaces for wide-range resis- 1093
1019 [74] A. Depari et al., “An uncalibrated wide-range single-supply integrable tive sensors portable applications,” Microelectron. J., vol. 45, no. 6, 1094
1020 front-end for resistance and capacitance estimation,” in Proc. Trans- pp. 589–596, Jun. 2014. 1095
1021 ducers Int. Solid-State Sensors, Actuat. Microsyst. Conf., Lyon, France, [96] V. Jain and B. George, “Self-balancing digitizer for resistive half- 1096
1022 Jun. 2007, pp. 2031–2034. bridge,” in Proc. IEEE Int. Instrum. Meas. Technol. Conf. (IMTC), 1097
1023 [75] F. Conso, M. Grassi, P. Malcovati, and A. Baschirotto, “A very high May 2018, pp. 1–5. 1098
1024 dynamic range interface circuit for resistive gas sensor matrix read- [97] T. Miyazaki, S. Nakagawa, and H. Ishikuro, “High-resolution auto- 1099
1025 out,” in Proc. IEEE Int. Symp. Circuits Syst. (ISCAS), Rio de Janeiro, balancing wheatstone-bridge with successive approximation of 1100
1026 Brazil, May 2011, pp. 2209–2212. -modulated digitally controlled variable resistor,” in Proc. 26th 1101
1027 [76] K.-T. Tang, S.-W. Chiu, M.-F. Chang, C.-C. Hsieh, and J.-M. Shyu, IEEE Int. Conf. Electron., Circuits Syst., Nov. 2019, pp. 474–477. 1102
1028 “A low-power electronic nose signal-processing chip for a portable [98] L. Areekath, B. George, and F. Reverter, “An auto-balancing 1103
1029 artificial olfaction system,” IEEE Trans. Biomed. Circuits Syst., vol. 5, capacitance-to-pulse-width converter for capacitive sensors,” IEEE Sen- 1104
1030 no. 4, pp. 380–390, Aug. 2011. sors J., vol. 21, no. 1, pp. 765–775, Jan. 2021. 1105
 REN et al.: TOWARD A GAS SENSOR INTERFACE CIRCUIT 18265

1106 [99] S. S. W. Chan and P. C. H. Chan, “A resistance-variation-tolerant Mingyuan Ren received the master’s and Ph.D. 1166
1107 constant-power heating circuit for integrated sensor applications,” IEEE degrees in microelectronics and solid-state elec- 1167
1108 J. Solid-State Circuits, vol. 34, no. 4, pp. 432–439, Apr. 1999. tronics from the Harbin Institute of Technology, 1168
1109 [100] S. A. Bota, A. Dieguez, J. L. Merino, R. Casanova, J. Samitier, and Harbin, China, in 2007 and 2016, respectively. 1169
1110 C. Cane, “A monolithic interface circuit for gas sensor arrays: Control He is currently an Associate Professor with 1170
1111 and measurement,” Analog Integr. Circuits Signal Process., vol. 40, the Jinhua Advanced Research Institute, Jinhua, 1171
1112 no. 2, pp. 175–184, 2004. China. His research interests include analog inte- 1172
1113 [101] E. Martinelli, D. Polese, A. Catini, A. D’Amico, and C. Di Natale, grated circuits and sensor interface circuits. 1173
1114 “Self-adapted temperature modulation in metal-oxide semiconductor
1115 gas sensors,” Sens. Actuators B, Chem., vol. 161, no. 1, pp. 534–541,
1116 Jan. 2012.
1117 [102] A. Bora, “A temperature modulation circuit for metal oxide semicon-
1118 ductor gas sensor,” Indian J. Sci. Technol., vol. 8, no. 1, pp. 1–7,
1119 Jan. 2015.
1120 [103] M. Malfatti and D. A. L. A. A. Stoppa Simoni Lorenzelli Adami
1121 Baschirotto, “A CMOS interface for a gas-sensor array with a 0.5%-
1122 linearity over 500k-to-1G range and +2.5◦ C temperature control
Honghai Xu was born in Liaoning, China. 1174
1123 accuracy,” in Proc. IEEE Int. Conf. Solid-State Circuits (ISSCC),
He received the bachelor’s degree in integrated 1175
1124 San Francisco, CA, USA, Feb. 2006, pp. 294–295, and 654.
circuit design and integrated systems and the 1176
1125 [104] D. K. Sharma, R. S. V. Dwara, B. A. Botre, and S. A. Akbar,
master’s degree in integrated circuit engineer- 1177
1126 “Temperature control and readout circuit interface for mox based
ing from the Harbin University of Science and 1178
1127 NH3 gas sensor,” Microsyst. Technol., vol. 23, no. 5, pp. 1575–1583,
Technology, Harbin, China, in 2018 and 2021, 1179
1128 May 2017.
respectively. 1180
1129 [105] D. Barrettino et al., “A smart single-chip micro-hotplate-based chemi-
He is mainly engaged in research in the field of 1181
1130 cal sensor system in CMOS-technology,” in Proc. IEEE Int. Sym. Circ.
gas sensor interface circuits. 1182
1131 Syst., May 2002, p. II.
1132 [106] S. Singh, D. K. Sharma, K. Kishore, B. A. Botre, and S. A. Akbar,
1133 “Modeling, simulation, and implementation of fast settling switched
1134 PI controller for MOX integrated PT microheater,” IEEE Sensors J.,
1135 vol. 18, no. 20, pp. 8549–8557, Oct. 2018.
1136 [107] T. Conrad, P. Hiry, and A. Schutze, “PuMaH—A temperature control
1137 and resistance read-out system for microstructured gas sensors based
1138 on PWM signals,” in Proc. IEEE SENSORS, Irvine, CA, USA, Oct.
1139 2005, pp. 377–384. Changchun Dong received the master’s and 1183
1140 [108] A. Lombardi et al., “A CMOS integrated interface circuit for metal- Ph.D. degrees in microelectronics and solid-state 1184
1141 oxide gas sensors,” Sens. Actuators B, Chem., vol. 142, no. 1, electronics from the Harbin Institute of Technol- 1185
1142 pp. 82–89, Oct. 2009. ogy, Harbin, China, in 2007 and 2012. 1186
1143 [109] U. Frey, M. Graf, S. Taschini, K. U. Kirstein, and A. Hierlemann, He is currently an Associate Professor with 1187
1144 “A digital CMOS architecture for a micro-hotplate array,” IEEE the Jinhua Advanced Research Institute, Jinhua, 1188
1145 J. Solid-State Circuits, vol. 42, no. 2, pp. 441–450, Feb. 2007. China. His research interests are analog inte- 1189
1146 [110] A. H. Falah, M. Rivai, and D. Purwanto, “Implementation of gas and grated circuits, fluxgate interface ASIC, and sen- 1190
1147 sound sensors on temperature control of coffee roaster using fuzzy sor interface circuits. 1191
1148 logic method,” in Proc. Int. Seminar Intell. Technol. Appl. (ISITIA),
1149 Surabaya, Indonesia, Aug. 2019, pp. 80–85.
1150 [111] M. V. Nikolic, V. Milovanovic, Z. Z. Vasiljevic, and Z. Stamenkovic,
1151 “Semiconductor gas sensors: Materials, technology, design, and appli-
1152 cation,” Sensors, vol. 20, no. 22, p. 6694, Nov. 2020.
1153 [112] S. Yang, C. Jiang, and S. H. Wei, “Gas sensing in 2D materials,” Appl.
1154 Phys. Rev., vol. 4, no. 2, Jun. 2017, Art. no. 021304.
1155 [113] M. Donarelli and L. Ottaviano, “2D materials for gas sensing appli- Zhu Zhang received the Ph.D. degree in com- 1192
1156 cations: A review on graphene oxide, MoS2 , WS2 and phosphorene,” munication engineering from the Harbin Institute 1193
1157 Sensors, vol. 18, no. 11, p. 3638, Oct. 2018. of Technology, Harbin, China, in 2013. 1194
1158 [114] S. Yuan and S. Zhang, “Recent progress on gas sensors based on She was a Visiting Scholar with the University 1195
1159 graphene-like 2D/2D nanocomposites,” J. Semicond., vol. 40, no. 11, of Oulu, Oulu, Finland, from 2014 to 2015. She is 1196
1160 Nov. 2019, Art. no. 111608. currently an Associate Professor with the Depart- 1197
1161 [115] J. Cao et al., “Recent development of gas sensing platforms based on ment of Artificial Intelligence, Jinhua Advanced 1198
1162 2D atomic crystals,” Research, vol. 2021, Apr. 2021, Art. no. 9863038. Research Institute, Jinhua, China. Her research 1199
1163 [116] S. Das et al., “Transistors based on two-dimensional materials interests include machine learning, deep learn- 1200
1164 for future integrated circuits,” Nature Electronics, vol. 4, no. 11, ing, and sensor interface circuits. 1201
1165 pp. 99–786 Nov. 2021.