Preparation of MWCNT-polymer gas sensor-array chip

Figure 3-3 shows a stainless steel testing stand with a Si (001) chip, consisting of eight gas sensors, one humidity sensor, two temperature sensors, one heating device and the corresponding 12 independent sensing electrical leads in a 2×6 arrangement to facilitate measurement of the resistance response of each sensor. The humidity sensors were reserved for future use. The chip is 34 mm× 20 mm in size. Each circular membrane sensor was limited to 2 mm in diameter to minimize heat loss from the silicon substrate (Chang et al. 2004; Chang and Yuan 2009)

.

Fig. 3-3 Gas testing stand with a sensing chip

The eight polymer types were selected according to their linear salvation energy relationship (LSER), and physical absorption bonding properties. First, the LSER theory is applying to enhance the device’s recognition capability for different odors

(Grate et al. 1998; Grate et al. 1995; Liron et al. 1997)

. The LSER model has been widely used to explain equilibrium partition coefficients in gas–liquid chromatography (GLC), thermogravimetry (Artzi-Gerlitz et al. 2009)

, and quartz crystal microbalance (QCM) and 1

4

5 8

2 3

6 7

surface acoustic wave (Murugaraj et al. 2010)

sensing experiments (Star et al. 2006)

. In this model, each interaction term on the right of Eq. (3-1) is expressed as product of two solvation parameters (one for solute and one for solvent) that quantify complementary characteristics of vapor and polymer molecules for a particular interaction process.

The LSER is described by the expression:

Log K

＝c＋rR＋sπ

＋aΣα

＋bΣβ

＋lLog L

---

where c, r, s, a, b and l characterize solubility properties of the solvent (polymer), and R、、、、π^H、、Σα、、 ^H、、Σβ、、 ^H and Log L¹⁶ are the complementary solute (vapor) parameters for specific solvation processes. These are usually referred to as salvation parameters.

The r and R measure abilities to interact with dispersion forces; s and π^H measure abilities to solvate with dipole interactions; a measures hydrogen bond basicity of the polymer and Σα^H measures hydrogen bond acidity of the vapor; b measures hydrogen bond acidity of the polymer and Σβ^H measures hydrogen bond basicity of the vapor;

l and Log L¹⁶ measure combined effects of dispersion interaction and cavity formation in hexadecane; and c is a regression constant representing residual effects not covered by other terms. LSER equations correlate the log of the partition coefficient of a vapor in a polymer with the vapor salvation parameters using a series of LSER coefficients related to the polymer’s solubility properties.

For example, applying LSER equation to estimate the solvation coefficient between ethanol gas and various polymer membranes, and considering parameters such as dispersive interactions, dipole interactions, hydrogen bond acidity, and hydrogen bond basicity, the solvation coefficients would be estimated as PMVEMA <

PEA < HPMC < PVBC. A smaller solvation coefficient corresponds to a larger

interaction between ethanol gas and polymer, resulting in a larger resistance variation.

Therefore, solvation coefficient provides a good estimation. However, carbon nanotubes may change the physical characteristics of the film (glass transition temperature, rigidity, and density), and possibly the chemical characteristics.

Therefore, solvation energies may not be sufficient for complete description of the sensor–analyte interaction. Hierlemann et al. 2001; Ryan et al. 1998)

When using carbon nanotube-polymer composites as gas sensors, to use the sensors repetitively with a short recovery time, the interaction between the gas and polymer membrane is usually reversible physical absorption bonding. There are five kinds of physical absorption bonding: (1) hydrogen-bond acidic (HBA), (2) hydrogen-bond basic (HBB), (3) dipolar and hydrogen-bond basic (D-HBB), (4) moderately dipolar and weakly H-bond basic or acidic (MD-HB), and (5) weakly dipolar with weak or no hydrogen-bond properties (WD) (Hierlemann et al. 2001; Ryan et al. 1998)

. Consequently, eight polymer materials were selected for this work. These were styrene/allyl alcohol copolymer (SAA) (WD), polyvinylpyrrolidone (PVP) (HBB), poly(vinylidene chloride-co-acrylonitrile) (P(VDC-AN)), poly(methyl vinyl ether-alt-maleic acid) (PMVEMA), poly(alpha-methylstyrene)(PMS) (WD), hydroxypropyl methyl cellulose (HPMC) (D-HBB), poly(ethylene adipate) (PEA) (HBA), and poly(vinyl benzyl chloride) (PVBC) (D-HBB). The selected polymer materials used in this work are listed in Table 3-1.

Table 3- 2 Sensing polymer materials selected in this work Sensor

number* Polymer / acronym Supplier

S1 Styrene Allyl Alcohol copolymer/SAA Arcos

S2 Polyvinylpyrrolidone/ PVP Arcos

S3 Poly(vinylidene chloride-co-acrylonitrile)/

P(VDC-AN) Arcos

S4 Poly(methyl vinyl ether-alt-maleic acid)/

PMVEMA Arcos

S5 hydroxypropyl methyl cellulose / HPMC Arcos

S6 Poly(alpha-methylstyrene)/ PMS Arcos

S7 Poly(ethylene adipate)/ PEA Arcos

S8 Poly(vinyl benzyl chloride)/ PVBC Arcos

*Sensor numbers of S1, S2 … correspond to the sensor position numbers of 1, 2 … in Fig. 3-2, respectively.

Fig. 3-4 Two different solution drop casting processes to fabricate the gas sensors

Two different solution drop casting processes (Processes 1 and 2) were used to fabricate eight gas sensors on the sensing chip, as shown in Figs. 3-4a and 3-4b, respectively, to illustrate the processing steps.

Process 1: PVP polymer+MWCNT composite film sensor

In this process, the MEK (methyl ethyl ketone) solution was prepared by mixing 1 wt% PVP and 1 wt% MWCNTs. The solution was magnetically and ultrasonically stirred to maximize uniform dispersion. The dispersed composite precursor solution was then drop cast onto the desired spot of an interdigitated microelectrode (IME) on chip using an HPLC syringe, and followed by air drying and monitoring the sensor resistance. The processes were repeated several times to control the sensor resistance within the range of 1 kΩ to 10 kΩ. The remaining solvent in the sensor was finally removed to form composite membrane by cooling in a vacuum oven for 24 hours.

Heat up to 40℃℃ and cool in ℃℃ vacuum for 24 hrs.

Step1: solution drop casting mixed (MWCNT+Polymer) solution

Step 1: solution drop casting MWCNT solution

Step 2: solution drop casting polymer solution

Process 2: polymer/MWCNT stacked film sensor

The stacked sensor was fabricated a two-step drop casting method. A MWCNTs-modified electrode layer was prepared by drop-casting a MEK solution with 1 wt% MWCNTs onto the desired spot of an interdigitated microelectrode (IME) on chip using an HPLC syringe. The MEK solvent was finally evaporated in air at room temperature for two hr to yield the MWCNTs film. One of the eight different polymer MEK solutions was then repeatly deposited by drop casting onto the MWCNTs layer, and the sensor resistance after each casting step was monitored to control its value within the range of 1 kΩ to 10 kΩ. The drop cast sensor was followed by a final drying step in vacuum for 24 h to remove the remaining solvent to form a gas stacked sensor.