sensors Article
A Low Noise CMOS Readout Based on a Polymer-Coated SAW Array for Miniature Electronic Nose Cheng-Chun Wu, Szu-Chieh Liu, Shih-Wen Chiu and Kea-Tiong Tang * Department of Electrical Engineering, National Tsing Hua University/No. 101, Sec. 2, Kuang-Fu Road, Hsinchu 30013, Taiwan; [email protected] (C.-C.W.); [email protected] (S.-C.L.); [email protected] (S.-W.C.) * Correspondence: [email protected]; Tel.: +886-3-516-2178 Academic Editor: Ki-Hyun Kim Received: 30 July 2016; Accepted: 19 October 2016; Published: 25 October 2016
Abstract: An electronic nose (E-Nose) is one of the applications for surface acoustic wave (SAW) sensors. In this paper, we present a low-noise complementary metal–oxide–semiconductor (CMOS) readout application-specific integrated circuit (ASIC) based on an SAW sensor array for achieving a miniature E-Nose. The center frequency of the SAW sensors was measured to be approximately 114 MHz. Because of interference between the sensors, we designed a low-noise CMOS frequency readout circuit to enable the SAW sensor to obtain frequency variation. The proposed circuit was fabricated in Taiwan Semiconductor Manufacturing Company (TSMC) 0.18 µm 1P6M CMOS process technology. The total chip size was nearly 1203 × 1203 µm2 . The chip was operated at a supply voltage of 1 V for a digital circuit and 1.8 V for an analog circuit. The least measurable difference between frequencies was 4 Hz. The detection limit of the system, when estimated using methanol and ethanol, was 0.1 ppm. Their linearity was in the range of 0.1 to 26,000 ppm. The power consumption levels of the analog and digital circuits were 1.742 mW and 761 µW, respectively. Keywords: low noise; low power; miniature electronic nose; surface acoustic wave sensor array
1. Introduction Electronic nose (E-Nose) systems, which are biomimetic olfactory systems for gas sensing [1], have been used in many applications such as food product quality control [2], indoor air quality monitoring, environmental monitoring [3,4], automotive industry production, and clinical diagnosis [5,6]. These applications involve a steady demand for a portable E-Nose system. One of the reasons for the large size of an E-Nose system is the complexity of the readout circuit [7,8]. Therefore, implementing a readout circuit with an application-specific integrated circuit (ASIC) reduces the size and power consumption of the system and facilitates moving toward a portable E-Nose system. Surface acoustic wave (SAW)-based gas sensors have the advantages of high sensitivity and fast response [9,10]. Readout circuits for SAW sensors have been developed for years; however, the corresponding power consumption is high and the data resolution is not sufficient for detecting gases at low concentration [11,12]. In our previous work, we implemented a noncontinuous-type readout circuit with integrated circuits (ICs) [10]. We successfully reduced the power consumption to 1.48 mW and improved the resolution to 10 Hz. However, all the sensors in the array did not operate continuously, and a time interval thus existed between the two data points in the measurement results of the same sensor; consequently, data integrity was incomplete. To improve the data integrity, we propose a continuous-type readout circuit realized by integrating the readout circuit and a cross-coupled pair into an ASIC. By downmixing the operating sensor frequency with a reference sensor frequency, we could further reduce power consumption. Moreover,
Sensors 2016, 16, 1777; doi:10.3390/s16111777
www.mdpi.com/journal/sensors
Moreover, using a deep N-well technique and modifying the switch timing of the multiplexer enabled us to reduce the noise level of the readout circuit. In general, an E-Nose system comprises three parts, namely sensor arrays, readout circuits for dataSensors acquisition, and data analysis and processing components. The present work focused on 2016, 16, 1777 2 of 13 implementing a sensor array and its readout circuit. Several sensor types are used for E-Nose systems [13]. Because the primary function of a sensor is to sense an external stimulus, such as detection of a using a deep N-well technique and modifying the switch timing of the multiplexer enabled us to low reduce concentration of gas, the sensor sensitivity is an essential consideration. In general, frequencythe noise level of the readout circuit. type sensors have the advantage of high sensitivity in thearrays, current work, circuits we applied In general, an E-Nose system comprises three [14]; parts,therefore, namely sensor readout an SAW sensor to our E-Nose interference from The certain environmental is for data acquisition, and datasystem. analysisHowever, and processing components. present work focusedfactors on implementing a sensor array and its readout circuit. Several sensor types are used for E-Nose more obvious [15]. To this end, a reference sensor was included to generate a frequency that is least systems [13]. Because the primary function of a sensor is to sense an external stimulus, such as affected by environmental factors. detection of a low concentration of gas, the sensor sensitivity is an essential consideration. In general, The SAW resonator structure was composed of a lithium niobate (LiNbO3) substrate and aurum frequency-type sensors have the advantage of high sensitivity [14]; therefore, in the current work, (Au)weinterdigital transducers Various polymer materials werefrom then coated onto the sensor applied an SAW sensor to(IDTs). our E-Nose system. However, interference certain environmental surface to is form sensor array [16]. Studies havesensor reported that coating polymers on sensors factors morethe obvious [15]. To this end, a reference was included to generate a frequency that can is leastthe affected by environmental factors.[13,17], and that the resulting sensor array can facilitate improve sensitivity of the sensors Thetypes SAW resonator structure was composed of a lithium niobate (LiNbO3 ) substrate and aurum identifying of gases [18]. (Au) interdigital transducers (IDTs). Various polymer materials were then coated onto the sensor In complementary metal oxide semiconductor (CMOS) process technology, substrate noise is an surface to form the sensor array [16]. Studies have reported that coating polymers on sensors essential consideration in IC design [19]. In [13,17], a sensorand array, frequencies of all sensors are can improve the sensitivity of the sensors that the the oscillator resulting sensor array can facilitate extremely close such that [18]. injecting current through the ASIC substrate significantly affects the identifying types of gases In complementary metaloutput oxide semiconductor (CMOS) process technology, noise is an performance of the oscillator signal. As the number of SAW sensorssubstrate increases, applying a essential consideration in IC design [19]. In a sensor array, the oscillator frequencies of all sensors higher substrate injection current among oscillators raises the output noise level of the SAW sensors. are extremely close such that injecting current through the ASIC substrate significantly affects the Therefore, we used the deep N-well technique to reduce the interference among different sensors. performance of the oscillator output signal. As the number of SAW sensors increases, applying a For data collection, wecurrent implemented a continuous-type interface circuit. measuring higher substrate injection among oscillators raises the output noise level ofCurrently, the SAW sensors. instruments areused typically and not portable; therefore, the different interface circuit was Therefore, we the deepexpensive N-well technique to reduce the interference among sensors. For data collection, we implemented a continuous-type interface circuit. Currently, measuring implemented using an ASIC to achieve reduced size, power consumption, and cost. The interface instruments are typically expensive portable; therefore, circuit was implemented circuit can convert an analog signaland to not a digital signal, andthe it interface comprises a mixer, low-pass filter, using an ASIC to achieve reduced size, power consumption, and cost. The interface circuit can convert multiplexer, counter, and parallel-to-serial circuit. To compress the interference between channels, an analog signal to a digital signal, and it comprises a mixer, low-pass filter, multiplexer, counter, we designed a low-noise analog multiplexer to switch the signalchannels, from each sensor. and parallel-to-serial circuit. To compress the interference between weSAW designed a low-noise analog multiplexer to switch the signal from each SAW sensor.
2. Proposed SAW Array and Its Interface ASIC 2. Proposed SAW Array and Its Interface ASIC
Figure 1 outlines a block diagram of the proposed E-Nose structure, which consists of two major Figure 1 outlines a block diagram of the proposed E-Nose structure, which consists of two major parts: the SAW sensor oscillator array and its readout ASIC. The SAW sensor oscillator is composed parts: the SAW sensor oscillator array and its readout ASIC. The SAW sensor oscillator is composed of of a apassive sensorand anda across-coupled cross-coupled pair. passiveSAW SAW sensor pair.
Figure 1. Surface acousticwave wave(SAW) (SAW) array array and application-specific integrated circuit Figure 1. Surface acoustic andits itsinterface interface application-specific integrated circuit (ASIC) block diagram. (ASIC) block diagram.
Sensors 2016, 16, 1777
3 of 13
Sensors2016, 2016,16, 16,1777 1777 Sensors
13 33ofof13
The SAW resonator structure is composed of a LiNbO3 substrate and Au IDTs. The center The SAWofresonator is determined composed of a LiNbO 3 substrate Au[10]. IDTs. center frequency the SAW structure sensor was to be approximately 114and MHz TheThe surface of the The SAW resonator structure is composed aapproximately LiNbO and Au IDTs. The center frequency of the SAW sensor determined to of be 114the MHz [10]. The surface of 14 the× 7.4 3 substrate piezoelectric substrate waswas coated with a thin film of polymer, and size of SAW sensor is frequency SAW sensor was with determined to be approximately 114 MHz [10]. sensor The surface of piezoelectric substrate was coated a thin film of polymer, and the size of SAW is 14 × 7.4 mm2. of the 2 the substrate was coated with aasthin film when of polymer, and theissize of SAW is mmpiezoelectric . The sensing mechanism is outlined follows: a gas sample adsorbed on sensor the substrate 2. 14 ×surface, 7.4 Themm sensing mechanism is outlined as follows: when a gas sample is adsorbed on the substrate the change of mass would cause a frequency shift. The sensor array comprises five sensors. The the sensing mechanism is outlined asa follows: when a gas sample is adsorbed on the substrate surface, of mass would cause frequency shift. The sensorpoly-N-vinylpyrrolidone, array comprises five sensors. Four of change them were coated with different polymers, namely poly-4surface, change ofcoated mass and would cause a polymers, frequencyanhydride; shift. The arraysensor comprises five Fourvinylphenol, of the them were with different namely poly-N-vinylpyrrolidone, polystyrene, polystyrene-co-maleic thesensor remaining waspoly-4not coated sensors. Four of them were coated with different polymers, namely poly-N-vinylpyrrolidone, vinylphenol, and polystyrene-co-maleic with any polystyrene, polymer, thus serving as a reference. anhydride; the remaining sensor was not coated poly-4-vinylphenol, polystyrene, and polystyrene-co-maleic anhydride; the remaining sensor was not with any polymer, thus serving as a reference. coated anyCross-Coupled polymer, thusPair serving as a reference. 2.1.with CMOS 2.1. CMOS Cross-Coupled Pair The sensor of thePair SAW resonator can be represented by an equivalent circuit model, as depicted 2.1. CMOS Cross-Coupled sensor resonator can be represented an equivalent as depicted inThe Figure 2. of Inthe theSAW model, the inductance, L1, and by capacitance, C1, circuit valuesmodel, determine the center The sensor of the SAW resonator can be represented by an equivalent circuit model, as depicted in Figure 2. Inofthe model, the inductance, L1,the and capacitance, C1, values determine the center R1, frequency the SAW resonator. Moreover, parasitic capacitance, C0, and parasitic resistance, in Figure 2.ofconsidered. In model, the inductance, L1,parasitic and capacitance, C1,C0, values determine the center frequency thethe SAW resonator. Moreover, the capacitance, and parasitic resistance, R1, must be frequency of the SAW resonator. Moreover, the parasitic capacitance, C0, and parasitic resistance, R1, must be considered. must be considered.
L1 L1 C1 C0 C1 C0 R1 R1 Figure 2. Equivalent circuit model of the SAW sensor. Figure Figure2.2. Equivalent Equivalentcircuit circuitmodel modelof ofthe theSAW SAWsensor. sensor.
R1 is closely related to the energy loss of the resonator. To reduce this loss, a CMOS crossR1 to to thethe energy loss loss of the resonator. To reduce this the loss, a CMOS cross-coupled R1isisclosely closely related the resonator. To with reduce this loss, aof CMOS cross- the coupled pairrelated was attached toenergy introduce aofnegative resistance objective eliminating pair was attached toattached introduce aintroduce negative The resistance with of eliminating the influence of in coupled pair was a negative resistance with the objective ofpair eliminating the influence of the parasitictoresistance. structure ofthe theobjective CMOS cross-coupled is illustrated the parasitic Theresistance. structure of thestructure CMOS cross-coupled is illustrated in is Figure 3. influence the parasitic The of the CMOS pair cross-coupled pair illustrated in Figureof 3.resistance. Figure 3.
Figure 3. Complementary semiconductor (CMOS) cross-coupled Figure 3. Complementary metalmetal oxideoxide semiconductor (CMOS) cross-coupled pair. pair. Figure 3. Complementary metal oxide semiconductor (CMOS) cross-coupled pair.
In the equivalent small-signal model, the input resistance, Rin, between the drain and the source In the equivalent small-signal model, the input resistance, Rin , between the drain and the source of small-signal model, the inputcan resistance, Rin, between drain and(1), the where source all ofIn the equivalent metal oxide semiconductor (MOS) be represented asthe Equation the metal oxide semiconductor (MOS) can be represented as Equation (1), where all transconductance of the metal oxide values semiconductor (MOS) be represented as Equation (1), where all transconductance are assumed to be gcan m: values are assumed to be gm : transconductance values are assumed to be gm: −2 Rin = Rin 2 (1) (1) gm2 g m Rin (1) gm
Sensors 2016, 16, 1777
4 of 13
The total equivalent resistance, RC , can then be calculated from Equation (2): Rc = ( Rin p //Rinn )= (
−2 −2 −2 )//( )= gmp gmn gmp + gmn
(2)
According to Equation (2), gmp and gmn are the transconductance values of a p-channel metal oxide semiconductor and an n-channel metal oxide semiconductor (NMOS), respectively. When a gas passes through the SAW sensor, the frequency of the oscillator output shifts. To improve gas detection sensitivity, the noise from the interface circuit should be as low as possible; thus, noise quantification is highly crucial. The noise can be equivalent to another frequency deviation (also called jitter) in the time domain. In the frequency domain, the noise-induced frequency deviation can be transferred to phase noise. Therefore, the phase noise is the most essential design parameter for an oscillator. In the present design, determining the relationship between the phase noise and frequency deviation is vital. Equation (3) shows the phase noise–jitter relationship [20]:
JRMS
v u f 2 u L( f ) 1 u w t2 · 10 10 d f = 2π f 0
(3)
f1
where JRMS represents the RMS jitter value (fs ), f 0 represents the center frequency (Hz), L(f ) represents the phase noise magnitude (dBc/Hz), and the integral range is from f 1 to f 2 in the frequency domain. According to the equation, the relationship between the jitter and phase noise is square. Equation (4) shows the period of the oscillator with jitter, where T0 represents the oscillator period with noise jitter, f 0 represents the oscillator center frequency, and ∆f represents the frequency deviation of the oscillator when jitter occurs; notably, the period is not a static value in a period range. According to Equation (5), the RMS jitter is the period deviation ∆T, defined as the difference between the maximum and minimum periods, as indicated in Equation (6). Finally, Equation (7) shows the relationship between the frequency deviation and the RMS jitter. 1 1 < T0 < f0 + ∆ f f0 − ∆ f
(4)
JRMS = ∆ T
(5)
JRMS =
1 1 − f0 − ∆ f f0 + ∆ f
JRMS =
2∆ f f0 2 − ∆ f 2
(6) (7)
Combining Equations (3) and (7) yields Equation (8), and simplifying Equation (8) yields Equation (9), where the frequency deviation and the phase noise also have a square relationship. On the basis of Equation (9), we can estimate the frequency deviation and calculate the phase noise value in this SAW sensor oscillator to suppress the noise. v u f 2 u L( f ) 2∆ f 1 u w t2 · = 10 10 d f 2 2 2π f f0 − ∆ f 0 f
(8)
1
Let M = 2 ·
wf2
10
L( f ) 10
df
f1
∆f =
−2 √ M
r
±
2
( √−2 ) + 4 · f 0 2 M
2
(9)
Sensors 2016, 16, 1777
5 of 13
2
Sensors 2016, 16, 1777
Δf
M
(
2 2 ) 4 f02 M 2
(9) 5 of 13
In In thisthis study, when thethe sensor was in in steady-state operation [21], thethe frequency variation was study, when sensor was steady-state operation [21], frequency variation was approximately 20 Hz. Therefore, we set Δf as 20 Hz. Moreover, when f 0 = 115 MHz, the phase noise approximately 20 Hz. Therefore, we set ∆f as 20 Hz. Moreover, when f 0 = 115 MHz, the phase noise was less than −177 dBc/Hz at 1atMHz. TheThe overall specifications of the SAW sensor oscillator areare was less than −177 dBc/Hz 1 MHz. overall specifications of the SAW sensor oscillator presented in Table 1. presented in Table 1. Table 1. Specifications of SAW sensor oscillator. Table 1. Specifications of SAW sensor oscillator.
Parameter Parameter Center Frequency (MHz) CenterVoltage Frequency Supply (V)(MHz) Supply Voltage (V) Power (mW) Power (mW) PhasePhase NoiseNoise (1 MHz) (dbc/Hz) (1 MHz) (dbc/Hz) Frequency Jitter 5 (fs) Frequency Jitter 5 (fs)
Specifications 110–120 110–120 1.8 1.8
A Low Noise CMOS Readout Based on a Polymer-Coated SAW Array for Miniature Electronic Nose Cheng-Chun Wu, Szu-Chieh Liu, Shih-Wen Chiu and Kea-Tiong Tang * Department of Electrical Engineering, National Tsing Hua University/No. 101, Sec. 2, Kuang-Fu Road, Hsinchu 30013, Taiwan; [email protected] (C.-C.W.); [email protected] (S.-C.L.); [email protected] (S.-W.C.) * Correspondence: [email protected]; Tel.: +886-3-516-2178 Academic Editor: Ki-Hyun Kim Received: 30 July 2016; Accepted: 19 October 2016; Published: 25 October 2016
Abstract: An electronic nose (E-Nose) is one of the applications for surface acoustic wave (SAW) sensors. In this paper, we present a low-noise complementary metal–oxide–semiconductor (CMOS) readout application-specific integrated circuit (ASIC) based on an SAW sensor array for achieving a miniature E-Nose. The center frequency of the SAW sensors was measured to be approximately 114 MHz. Because of interference between the sensors, we designed a low-noise CMOS frequency readout circuit to enable the SAW sensor to obtain frequency variation. The proposed circuit was fabricated in Taiwan Semiconductor Manufacturing Company (TSMC) 0.18 µm 1P6M CMOS process technology. The total chip size was nearly 1203 × 1203 µm2 . The chip was operated at a supply voltage of 1 V for a digital circuit and 1.8 V for an analog circuit. The least measurable difference between frequencies was 4 Hz. The detection limit of the system, when estimated using methanol and ethanol, was 0.1 ppm. Their linearity was in the range of 0.1 to 26,000 ppm. The power consumption levels of the analog and digital circuits were 1.742 mW and 761 µW, respectively. Keywords: low noise; low power; miniature electronic nose; surface acoustic wave sensor array
1. Introduction Electronic nose (E-Nose) systems, which are biomimetic olfactory systems for gas sensing [1], have been used in many applications such as food product quality control [2], indoor air quality monitoring, environmental monitoring [3,4], automotive industry production, and clinical diagnosis [5,6]. These applications involve a steady demand for a portable E-Nose system. One of the reasons for the large size of an E-Nose system is the complexity of the readout circuit [7,8]. Therefore, implementing a readout circuit with an application-specific integrated circuit (ASIC) reduces the size and power consumption of the system and facilitates moving toward a portable E-Nose system. Surface acoustic wave (SAW)-based gas sensors have the advantages of high sensitivity and fast response [9,10]. Readout circuits for SAW sensors have been developed for years; however, the corresponding power consumption is high and the data resolution is not sufficient for detecting gases at low concentration [11,12]. In our previous work, we implemented a noncontinuous-type readout circuit with integrated circuits (ICs) [10]. We successfully reduced the power consumption to 1.48 mW and improved the resolution to 10 Hz. However, all the sensors in the array did not operate continuously, and a time interval thus existed between the two data points in the measurement results of the same sensor; consequently, data integrity was incomplete. To improve the data integrity, we propose a continuous-type readout circuit realized by integrating the readout circuit and a cross-coupled pair into an ASIC. By downmixing the operating sensor frequency with a reference sensor frequency, we could further reduce power consumption. Moreover,
Sensors 2016, 16, 1777; doi:10.3390/s16111777
www.mdpi.com/journal/sensors
Moreover, using a deep N-well technique and modifying the switch timing of the multiplexer enabled us to reduce the noise level of the readout circuit. In general, an E-Nose system comprises three parts, namely sensor arrays, readout circuits for dataSensors acquisition, and data analysis and processing components. The present work focused on 2016, 16, 1777 2 of 13 implementing a sensor array and its readout circuit. Several sensor types are used for E-Nose systems [13]. Because the primary function of a sensor is to sense an external stimulus, such as detection of a using a deep N-well technique and modifying the switch timing of the multiplexer enabled us to low reduce concentration of gas, the sensor sensitivity is an essential consideration. In general, frequencythe noise level of the readout circuit. type sensors have the advantage of high sensitivity in thearrays, current work, circuits we applied In general, an E-Nose system comprises three [14]; parts,therefore, namely sensor readout an SAW sensor to our E-Nose interference from The certain environmental is for data acquisition, and datasystem. analysisHowever, and processing components. present work focusedfactors on implementing a sensor array and its readout circuit. Several sensor types are used for E-Nose more obvious [15]. To this end, a reference sensor was included to generate a frequency that is least systems [13]. Because the primary function of a sensor is to sense an external stimulus, such as affected by environmental factors. detection of a low concentration of gas, the sensor sensitivity is an essential consideration. In general, The SAW resonator structure was composed of a lithium niobate (LiNbO3) substrate and aurum frequency-type sensors have the advantage of high sensitivity [14]; therefore, in the current work, (Au)weinterdigital transducers Various polymer materials werefrom then coated onto the sensor applied an SAW sensor to(IDTs). our E-Nose system. However, interference certain environmental surface to is form sensor array [16]. Studies havesensor reported that coating polymers on sensors factors morethe obvious [15]. To this end, a reference was included to generate a frequency that can is leastthe affected by environmental factors.[13,17], and that the resulting sensor array can facilitate improve sensitivity of the sensors Thetypes SAW resonator structure was composed of a lithium niobate (LiNbO3 ) substrate and aurum identifying of gases [18]. (Au) interdigital transducers (IDTs). Various polymer materials were then coated onto the sensor In complementary metal oxide semiconductor (CMOS) process technology, substrate noise is an surface to form the sensor array [16]. Studies have reported that coating polymers on sensors essential consideration in IC design [19]. In [13,17], a sensorand array, frequencies of all sensors are can improve the sensitivity of the sensors that the the oscillator resulting sensor array can facilitate extremely close such that [18]. injecting current through the ASIC substrate significantly affects the identifying types of gases In complementary metaloutput oxide semiconductor (CMOS) process technology, noise is an performance of the oscillator signal. As the number of SAW sensorssubstrate increases, applying a essential consideration in IC design [19]. In a sensor array, the oscillator frequencies of all sensors higher substrate injection current among oscillators raises the output noise level of the SAW sensors. are extremely close such that injecting current through the ASIC substrate significantly affects the Therefore, we used the deep N-well technique to reduce the interference among different sensors. performance of the oscillator output signal. As the number of SAW sensors increases, applying a For data collection, wecurrent implemented a continuous-type interface circuit. measuring higher substrate injection among oscillators raises the output noise level ofCurrently, the SAW sensors. instruments areused typically and not portable; therefore, the different interface circuit was Therefore, we the deepexpensive N-well technique to reduce the interference among sensors. For data collection, we implemented a continuous-type interface circuit. Currently, measuring implemented using an ASIC to achieve reduced size, power consumption, and cost. The interface instruments are typically expensive portable; therefore, circuit was implemented circuit can convert an analog signaland to not a digital signal, andthe it interface comprises a mixer, low-pass filter, using an ASIC to achieve reduced size, power consumption, and cost. The interface circuit can convert multiplexer, counter, and parallel-to-serial circuit. To compress the interference between channels, an analog signal to a digital signal, and it comprises a mixer, low-pass filter, multiplexer, counter, we designed a low-noise analog multiplexer to switch the signalchannels, from each sensor. and parallel-to-serial circuit. To compress the interference between weSAW designed a low-noise analog multiplexer to switch the signal from each SAW sensor.
2. Proposed SAW Array and Its Interface ASIC 2. Proposed SAW Array and Its Interface ASIC
Figure 1 outlines a block diagram of the proposed E-Nose structure, which consists of two major Figure 1 outlines a block diagram of the proposed E-Nose structure, which consists of two major parts: the SAW sensor oscillator array and its readout ASIC. The SAW sensor oscillator is composed parts: the SAW sensor oscillator array and its readout ASIC. The SAW sensor oscillator is composed of of a apassive sensorand anda across-coupled cross-coupled pair. passiveSAW SAW sensor pair.
Figure 1. Surface acousticwave wave(SAW) (SAW) array array and application-specific integrated circuit Figure 1. Surface acoustic andits itsinterface interface application-specific integrated circuit (ASIC) block diagram. (ASIC) block diagram.
Sensors 2016, 16, 1777
3 of 13
Sensors2016, 2016,16, 16,1777 1777 Sensors
13 33ofof13
The SAW resonator structure is composed of a LiNbO3 substrate and Au IDTs. The center The SAWofresonator is determined composed of a LiNbO 3 substrate Au[10]. IDTs. center frequency the SAW structure sensor was to be approximately 114and MHz TheThe surface of the The SAW resonator structure is composed aapproximately LiNbO and Au IDTs. The center frequency of the SAW sensor determined to of be 114the MHz [10]. The surface of 14 the× 7.4 3 substrate piezoelectric substrate waswas coated with a thin film of polymer, and size of SAW sensor is frequency SAW sensor was with determined to be approximately 114 MHz [10]. sensor The surface of piezoelectric substrate was coated a thin film of polymer, and the size of SAW is 14 × 7.4 mm2. of the 2 the substrate was coated with aasthin film when of polymer, and theissize of SAW is mmpiezoelectric . The sensing mechanism is outlined follows: a gas sample adsorbed on sensor the substrate 2. 14 ×surface, 7.4 Themm sensing mechanism is outlined as follows: when a gas sample is adsorbed on the substrate the change of mass would cause a frequency shift. The sensor array comprises five sensors. The the sensing mechanism is outlined asa follows: when a gas sample is adsorbed on the substrate surface, of mass would cause frequency shift. The sensorpoly-N-vinylpyrrolidone, array comprises five sensors. Four of change them were coated with different polymers, namely poly-4surface, change ofcoated mass and would cause a polymers, frequencyanhydride; shift. The arraysensor comprises five Fourvinylphenol, of the them were with different namely poly-N-vinylpyrrolidone, polystyrene, polystyrene-co-maleic thesensor remaining waspoly-4not coated sensors. Four of them were coated with different polymers, namely poly-N-vinylpyrrolidone, vinylphenol, and polystyrene-co-maleic with any polystyrene, polymer, thus serving as a reference. anhydride; the remaining sensor was not coated poly-4-vinylphenol, polystyrene, and polystyrene-co-maleic anhydride; the remaining sensor was not with any polymer, thus serving as a reference. coated anyCross-Coupled polymer, thusPair serving as a reference. 2.1.with CMOS 2.1. CMOS Cross-Coupled Pair The sensor of thePair SAW resonator can be represented by an equivalent circuit model, as depicted 2.1. CMOS Cross-Coupled sensor resonator can be represented an equivalent as depicted inThe Figure 2. of Inthe theSAW model, the inductance, L1, and by capacitance, C1, circuit valuesmodel, determine the center The sensor of the SAW resonator can be represented by an equivalent circuit model, as depicted in Figure 2. Inofthe model, the inductance, L1,the and capacitance, C1, values determine the center R1, frequency the SAW resonator. Moreover, parasitic capacitance, C0, and parasitic resistance, in Figure 2.ofconsidered. In model, the inductance, L1,parasitic and capacitance, C1,C0, values determine the center frequency thethe SAW resonator. Moreover, the capacitance, and parasitic resistance, R1, must be frequency of the SAW resonator. Moreover, the parasitic capacitance, C0, and parasitic resistance, R1, must be considered. must be considered.
L1 L1 C1 C0 C1 C0 R1 R1 Figure 2. Equivalent circuit model of the SAW sensor. Figure Figure2.2. Equivalent Equivalentcircuit circuitmodel modelof ofthe theSAW SAWsensor. sensor.
R1 is closely related to the energy loss of the resonator. To reduce this loss, a CMOS crossR1 to to thethe energy loss loss of the resonator. To reduce this the loss, a CMOS cross-coupled R1isisclosely closely related the resonator. To with reduce this loss, aof CMOS cross- the coupled pairrelated was attached toenergy introduce aofnegative resistance objective eliminating pair was attached toattached introduce aintroduce negative The resistance with of eliminating the influence of in coupled pair was a negative resistance with the objective ofpair eliminating the influence of the parasitictoresistance. structure ofthe theobjective CMOS cross-coupled is illustrated the parasitic Theresistance. structure of thestructure CMOS cross-coupled is illustrated in is Figure 3. influence the parasitic The of the CMOS pair cross-coupled pair illustrated in Figureof 3.resistance. Figure 3.
Figure 3. Complementary semiconductor (CMOS) cross-coupled Figure 3. Complementary metalmetal oxideoxide semiconductor (CMOS) cross-coupled pair. pair. Figure 3. Complementary metal oxide semiconductor (CMOS) cross-coupled pair.
In the equivalent small-signal model, the input resistance, Rin, between the drain and the source In the equivalent small-signal model, the input resistance, Rin , between the drain and the source of small-signal model, the inputcan resistance, Rin, between drain and(1), the where source all ofIn the equivalent metal oxide semiconductor (MOS) be represented asthe Equation the metal oxide semiconductor (MOS) can be represented as Equation (1), where all transconductance of the metal oxide values semiconductor (MOS) be represented as Equation (1), where all transconductance are assumed to be gcan m: values are assumed to be gm : transconductance values are assumed to be gm: −2 Rin = Rin 2 (1) (1) gm2 g m Rin (1) gm
Sensors 2016, 16, 1777
4 of 13
The total equivalent resistance, RC , can then be calculated from Equation (2): Rc = ( Rin p //Rinn )= (
−2 −2 −2 )//( )= gmp gmn gmp + gmn
(2)
According to Equation (2), gmp and gmn are the transconductance values of a p-channel metal oxide semiconductor and an n-channel metal oxide semiconductor (NMOS), respectively. When a gas passes through the SAW sensor, the frequency of the oscillator output shifts. To improve gas detection sensitivity, the noise from the interface circuit should be as low as possible; thus, noise quantification is highly crucial. The noise can be equivalent to another frequency deviation (also called jitter) in the time domain. In the frequency domain, the noise-induced frequency deviation can be transferred to phase noise. Therefore, the phase noise is the most essential design parameter for an oscillator. In the present design, determining the relationship between the phase noise and frequency deviation is vital. Equation (3) shows the phase noise–jitter relationship [20]:
JRMS
v u f 2 u L( f ) 1 u w t2 · 10 10 d f = 2π f 0
(3)
f1
where JRMS represents the RMS jitter value (fs ), f 0 represents the center frequency (Hz), L(f ) represents the phase noise magnitude (dBc/Hz), and the integral range is from f 1 to f 2 in the frequency domain. According to the equation, the relationship between the jitter and phase noise is square. Equation (4) shows the period of the oscillator with jitter, where T0 represents the oscillator period with noise jitter, f 0 represents the oscillator center frequency, and ∆f represents the frequency deviation of the oscillator when jitter occurs; notably, the period is not a static value in a period range. According to Equation (5), the RMS jitter is the period deviation ∆T, defined as the difference between the maximum and minimum periods, as indicated in Equation (6). Finally, Equation (7) shows the relationship between the frequency deviation and the RMS jitter. 1 1 < T0 < f0 + ∆ f f0 − ∆ f
(4)
JRMS = ∆ T
(5)
JRMS =
1 1 − f0 − ∆ f f0 + ∆ f
JRMS =
2∆ f f0 2 − ∆ f 2
(6) (7)
Combining Equations (3) and (7) yields Equation (8), and simplifying Equation (8) yields Equation (9), where the frequency deviation and the phase noise also have a square relationship. On the basis of Equation (9), we can estimate the frequency deviation and calculate the phase noise value in this SAW sensor oscillator to suppress the noise. v u f 2 u L( f ) 2∆ f 1 u w t2 · = 10 10 d f 2 2 2π f f0 − ∆ f 0 f
(8)
1
Let M = 2 ·
wf2
10
L( f ) 10
df
f1
∆f =
−2 √ M
r
±
2
( √−2 ) + 4 · f 0 2 M
2
(9)
Sensors 2016, 16, 1777
5 of 13
2
Sensors 2016, 16, 1777
Δf
M
(
2 2 ) 4 f02 M 2
(9) 5 of 13
In In thisthis study, when thethe sensor was in in steady-state operation [21], thethe frequency variation was study, when sensor was steady-state operation [21], frequency variation was approximately 20 Hz. Therefore, we set Δf as 20 Hz. Moreover, when f 0 = 115 MHz, the phase noise approximately 20 Hz. Therefore, we set ∆f as 20 Hz. Moreover, when f 0 = 115 MHz, the phase noise was less than −177 dBc/Hz at 1atMHz. TheThe overall specifications of the SAW sensor oscillator areare was less than −177 dBc/Hz 1 MHz. overall specifications of the SAW sensor oscillator presented in Table 1. presented in Table 1. Table 1. Specifications of SAW sensor oscillator. Table 1. Specifications of SAW sensor oscillator.
Parameter Parameter Center Frequency (MHz) CenterVoltage Frequency Supply (V)(MHz) Supply Voltage (V) Power (mW) Power (mW) PhasePhase NoiseNoise (1 MHz) (dbc/Hz) (1 MHz) (dbc/Hz) Frequency Jitter 5 (fs) Frequency Jitter 5 (fs)
Specifications 110–120 110–120 1.8 1.8
