Pt modified carbon nanotube networked layers for enhanced gas microsensors

Serra ENEA, Brindisi Technical Unit of Technologies for Materials, PO Box 51-Br4, I-72100 Brindisi, Italy a b s t r a c t a r t i c l e i n f o Available online 30 April 2011 Keywords: C

Pt-modi fied carbon nanotube networked layers for enhanced gas microsensors

M Penza ⁎ , R Rossi, M Alvisi, D Suriano, E Serra

ENEA, Brindisi Technical Unit of Technologies for Materials, PO Box 51-Br4, I-72100 Brindisi, Italy

a b s t r a c t

a r t i c l e i n f o

Available online 30 April 2011

Keywords:

Carbon nanotubes networks

Platinum-decoration

Gas sensors

Carbon nanotubes (CNTs) networkedfilms have been grown by chemical vapor deposition (CVD) technology onto miniaturized low-cost alumina substrates, coated by nanosized Co-catalyst for growing CNTs, to perform chemical detection of toxic gasses (NO2and NH3), greenhouse gasses (CO2and CH4) and domestic safety gasses (CO and C2H5OH) at an operating sensor temperature of 120 °C The morphology and structure of the CNTs networks have been characterized by scanning electron microscopy (SEM) A dense network of bundles of multiple tubes consisting of multi-walled carbon nanostructures appears with a maximum length of 1–5 μm and single-tube diameter varying in the range of 5–40 nm Surface modifications of the CNTs networks with sputtered Platinum (Pt) nanoclusters, at tuned loading of 8, 15 and 30 nm, provide higher sensitivity for significantly enhanced gas detection compared to un-decorated CNTs This could be caused by a spillover of the targeted gas molecules onto Pt-catalyst surface with a chemical gating into CNTs layers The measured electrical conductance

of the functionalized CNTs upon exposures of a given oxidizing and reducing gas is modulated by a charge transfer model with p-type semiconducting characteristics The effect of activated carbons as chemicalfilters to reduce the influence of the domestic interfering alcohols on CO gas detection has been studied Functionalized CNT gas sensors exhibited better performances compared to unmodified CNTs, making them highly promising candidates for functional applications of gas control and alarms

© 2011 Elsevier B.V All rights reserved

1 Introduction

Carbon nanotubes (CNTs)[1,2]are very promising one-dimensional

nanomaterials, which have attracted intense interest due to their unique

mix of special electrical, optical, mechanical and thermal properties[3]

and the large variety of potential electronic applications[4–7]including

gas sensing [8–12] A strong demand of cost-effective and high

performance gas sensors involves several technological sectors such as

air quality control, domestic safety, homeland security, environmental

monitoring, global warming and climate changes

Carbon nanotubes are interesting sensor materials due to wide

properties of the carbon chemistry providing broad opportunities for gas

sensing applications CNTs offer high gas adsorption capability, large

specific surface area, high sensitivity, various chemical interactions, lower

operating temperature, and low electrical consumption These properties

make CNTs very attractive as gas sensing nanomaterials [8–25]

The sensing mechanism of the CNTs gas sensors is essentially based on

the p-type semiconducting property[8,10,13,15,16,18,19,21,25]

Gener-ally, their electrical conductance is modified by the electron transfer

between CNTs and oxidizing or reducing gas molecules adsorbed on their

surface If p-type semiconducting CNTs adsorb oxidizing molecules, the

electrical resistance of the CNTs decreases with the increase of the

adsorbed gas molecules [15,16,19,25] In the contrast, the electrical resistance of the CNTs increases with the increasing content of adsorbed reducing gas molecules

Decoration of CNTs is of great interest to modify the properties of raw carbon nanotubes for enhanced gas detection and versatile functional applications[21,23–26] Attaching metal nanoparticles to nanotube sidewalls can be used as innovative hybrid material for gas sensing and catalysis

CNT-based sensors suffer of a lack of selectivity such as other gas sensor functional materials A strategy to overcome these limitations has been investigated by surface-modifications Kong et al [27] studied CNTs functionalized with Pd for molecular hydrogen sensors, at room temperature Kumar et al.[28]used a wet chemistry route to prepare Pt-decorated CNTs highly sensitive to hydrogen Star et al.[29] decorated CNTs by selective electroplating either with Pt, Pd, Au or Rh for detecting CO, NO2, CH4, H2S, NH3and H2 W Wongwiriyapan et al.[30] demonstrated CO detection at room temperature using Pt-decorated single-walled carbon nanotubes E Llobet[31]and co-workers studied the gas sensing of oxygen plasma-treated multiwalled carbon nanotubes decorated with metal nanoparticles (Rh, Pd, Au or Ni) for selective detection of benzene at trace levels with a low detection limit below

50 ppb Also, the authors demonstrated CNT layers decorated with sputtered Au and Pt clusters for enhancing response towards NO2and

NH3, at 150 °C[26] Recently, they decorated vertically-aligned CNTs with metal nanoclusters of Pt, Ag, and Ru for gas detection[32]and landfill gas monitoring applications[33]

⁎ Corresponding author Tel.: +39 0831 201422; fax: +39 0831 201423.

E-mail address: michele.penza@enea.it (M Penza).

0040-6090/$ – see front matter © 2011 Elsevier B.V All rights reserved.

Contents lists available atScienceDirect Thin Solid Films

j o u r n a l h o m e p a g e : w w w e l s ev i e r c o m / l o c a t e / ts f

In this work, CNTs networked layers have been prepared by CVD

onto alumina substrates, coated by nanosized Co-catalyst, and used

for a simple gas sensor device Their surface has been functionalized

with sputtered Pt-nanoclusters at tuned loading of 8, 15 and 30 nm

The aim of the paper is the investigation of the CNTs sensitivity,

recovery and selectivity towards toxic gasses (NO2, NH3), greenhouse

gasses (CO2, CH4) and domestic safety gasses (CO, C2H5OH), at a low

sensor temperature of 120 °C

2 Experimental

2.1 Sensor fabrication

Fig 1 shows the scheme of the fabricated chemiresistor The

multiwalled carbon nanotubes (MWCNTs) networked films were

grown by chemical vapor deposition (CVD) technology onto alumina

substrates previously coated with Cobalt (Co) sputtered nanosized

catalyst of nominal thickness of 7.5 nm The substrate size was 5 mm

width × 5 mm length × 0.6 mm thickness A thermal CVD apparatus

was used for growing carbon nanotubes The substrates were placed

in a quartz boat and then inserted into the center of a 1-inch diameter

quartz tube reactor housed in a furnace The tube was evacuated at a

base pressure of 5 × 10− 3Torr by a rotative pump Hence the

substrates were heated up to 550 °C in a H2flux of 100 sccm at a

working pressure of 100 Torr Then, acetylene (C2H2) was introduced

at aflow rate of 20 sccm added to H2at aflow rate of 80 sccm The

flow rate was controlled using two separate mass flow controllers The

CNT growth was performed with a total pressure of 100 Torr for

30 min After growth, the furnace was cooled to room temperature in

H2atmosphere Then, a pair of metal strips of Cr/Au (20 nm/300 nm)

was vacuum sputtered onto CNTfilms to serve as electrical contacts

for two-pole chemiresistor The Cr/Au electrode sizes were 1 mm

width × 5 mm length The gap between two electrodes was 3 mm The

CNT-based chemiresistors have been functionalized with a tuned

loading of Platinum (Pt) nanoclusters, prepared by RF magnetron

sputtering, with a nominal thickness of 8, 15 and 30 nm The Cr/Au

electrodes were protected by a shadow-mask during

Pt-functionali-zation of the CNT layers The Arflux to supply gas-discharge was fixed

at 4.6 sccm and the process Apressure was 2.2 × 10−2Torr The RF

power supply of Pt target wasfixed at 36 W No substrate heating was

applied during film deposition The electrical resistance, at room

temperature and upon ambient atmosphere, of the un-modified CNT,

Pt8nm loaded-CNT (CNT:Pt8 nm), Pt15 nm loaded-CNT (CNT:

Pt15 nm), and Pt30 nm loaded-CNT (CNT:Pt30 nm) sensors was

measured as 0.96, 0.60, 0.46 and 0.20 kΩ, respectively The presence

of a Pt loading at increasing thickness onto the surface of the CNTsfilm

decreases the electrical resistance in the chemiresistor proportionally

to the thickness of the deposited Pt clusters This effect is caused by an enhanced electrical transport produced by conductive channels of the partially-interconnected Pt metal clusters with an increased electrical conduction at higher Pt thickness The sheet resistance of the CNTs films was estimated at about 1 kΩ/square

The morphology of the CNTs was examined by means of a Field Emission (FE) Gun Scanning Electron Microscope (SEM) (Leo, model 1530), equipped with a high-resolution secondary electron detector (in-lens detector)

2.2 Chemical sensing measurements Sensors have been located in a thermally isolated and sealed test cell (500 ml volume) endowed with gas inlet and outlet The sensors have been placed to thermal contact with a heater-sink to calibrate the sensor operating temperature Cell case can host up to four chemiresistive sensors Dry air was used as reference gas, carrier gas and diluting gas to air-conditioning the sensors The gasflow rate was controlled by mass flow controllers (MFC) (MKS 1179A) regulated and driven by a controller-unit (MKS, GGK DMFC-5) The gas management system was equipped with a multiplexing/conversion system communicating with a desktop-host via standard RS-485/RS-232 serial bus The total flow rate was kept constant at 1500 ml min− 1 The gas sensing experiments have been performed by measuring the electrical conductance of CNTs thinfilms upon controlled concentrations in air

of nitrogen dioxide (NO2), ammonia (NH3), carbon dioxide (CO2), methane (CH4), carbon monoxide (CO) and ethanol (C2H5OH) in the range of 0.01–10 ppm, 10–1000 ppm, 10–1200 ppm, 90–5600 ppm,

100–1000 ppm, and 40–120 ppm, respectively Tested vapor of ethanol (C2H5OH) as domestic interfering analyte of CO was generated by a bubbling method mixing the saturated vapor stream with a dry air stream tofix a controlled gas concentration for exposure testing A chemical filter based on a metal cylinder (9 cm length and 3 cm diameter)filled with a mixed solid solution of activated carbon was used

to study the effects of cut of the domestic interfering ethanol on CO gas response All gas sensing measurements were performed at a sensor temperature of 120 °C The temperature was measured by a J-type thermocouple, allocated to thermal contact with heater-sink The output dc voltage from thermocouple was monitored by a multimeter (Agilent, 34401A) and a digital thermometer (Tersid, 705 Series) The gas sensing cycle consisted of a period (at least 60 min) in the stabilization of the sensor signals upon dry airflowing, of an exposure time of 5 or 10 min to various targeted gas concentrations,finally of a recovery time (at least 30 min) to restore the sensor signals upon dry air flowing to clean the test-cell and sensor-surface The dc electrical conductance of the CNT-sensors has been measured by the volt-amperometric technique with a dc voltage of 1 V applied in the two-pole format by means of a multimeter (Agilent, 34401A) Sensors conduc-tance was scanned by a switch system (Keithley, 7001) equipped by a low-current scanner card (Keithley, 7158) with a multiplexed read-out All data were acquired and stored in a PC-based workstation, interfaced with instrumentation by a USB-GPIB interface-card (Agilent 82357A), upon software compiled in Agilent-VEE The sensor response to a given gas concentration was defined as the percentage relative resistance change,ΔR/Ri(%), whereΔR is the change in resistance between the values of steady-state of the electrical resistance, Rfand Ri, of the sensor upon a target gas and in air, respectively

3 Results and discussion 3.1 SEM characterization of Pt-modified carbon nanotubes Fig 2shows typical scanning electron microscopy (SEM) images of the carbon-based nanostructures grown on alumina substrate coated with 7.5 nm-thick Co catalyst The CNTs appear to be networked bundles

Fig 1 Cross-section of the Pt-modified CNT-based chemiresistor.

of multiple nanostructures consisting of several nanotubes in the

multi-walled format Moreover, amorphous carbon and metal catalyst particles

are also present in the CNTs networkedfilms with many structural

defects Carbon nanostructures exhibit a maximum length of 1–5 μm and

single-tube diameter varying in the range of 5–40 nm

Surface modifications of the CNTs networks with sputtered

Platinum (Pt) nanoclusters, at tuned loading of 15 and 30 nm, have

been shown at low and high magnifications inFig 2 The Pt-metal

catalysts were deposited onto the surface of the CNT networks with

non-uniform distribution at dispersed islands Pt cluster diameter

ranges from 3 to 30 nm with increasing coverage of the CNTs surface

In the case of Pt 30 nm, a total cover of the surface has been found

This Pt-coverage of the CNTs-surface promotes enhanced gas

adsorption and thus better gas sensitivity

3.2 Sensor characterization: toxic gasses (NO2, NH3)

Fig 3(a) shows the typical time response in terms of electrical

resistance change for the four chemiresistors based on unmodified

CNTs (CNT), and functionalized with a tuned loading of Pt of 8 nm

(CNT:Pt8 nm), 15 nm (CNT:Pt15 nm) and 30 nm (CNT:Pt30 nm),

exposed to various 5-minute pulses of gas concentrations ranging

from 10 and 0.05 ppm NO2, at a sensor temperature of 120 °C This

temperature was selected as optimal trade-off between low

temper-ature (low power consumption) and maximum sensitivity of the

metal-modified CNT-sensors towards toxic gasses (i.e., NO2), as

reported by the authors in the literature[33] It is found that the

electrical resistance of the unmodified and Pt-functionalized

CNT-sensors decreases upon a single gas exposure of the NO2oxidizing gas

due to molecules adsorption In fact, the adsorption of

electron-withdrawing (NO) gas molecules into CNTs layer causes electron

charge transfer from the CNTs to gas molecules, with a consequent increase of hole density, and thus increased conductivity in the p-type semiconducting CNTs

The related calibration curves of the four CNTs-based chemiresis-tors towards NO2have been shown inFig 3(b) The sensor signal increases with the tested gas concentration and the highest gas sensitivity has been measured for the CNTs functionalized with a Pt loading of 8 nm This enhanced gas sensitivity compared to unmodi-fied CNTs can be attributed to a spillover effect of the NO2 gas molecules onto Pt catalyst A decreased gas response of the CNTs modified with Pt loadings of 15 and 30 nm has been recorded as well Probably, Pt cluster size affects the NO2gas sensitivity and optimal gas sensitivity is matched with a Pt loading of 8 nm

Additionally, the NH3gas sensing characteristics of the four CNT-based chemiresistors have been shown inFig 4 The sensor transients exhibit increasing electrical resistance upon ammonia exposure caused by the electron-donating features of the NH3 This trend in the response of the sensors is compatible to the p-type character in the unmodified and Pt-modified CNTs layers The dynamic range of the all sensors is quite large covering two orders of magnitude for ammonia gas concentration from 10 to 1000 ppm The calibration curves of the Pt-modified CNTs gas sensors exhibit a response proportional to NH3gas concentration, at the given sensor temper-ature of 120 °C The gas response of the Pt-modified CNT sensors results to be higher than unmodified CNT layers, with the highest NH3

sensitivity measured for a Pt loading of 15 nm A lower NH3 gas response for Pt-modified CNTs with a loading of 30 nm has been found as well In this case, the Pt cluster size produces optimal NH3

sensitivity for a Pt loading of 15 nm

These results demonstrate that the Pt-modified CNTs-based sensors are useful detectors for NO and NH gas sensing In particular, both gasses

Fig 2 FE-SEM images of a CNT film, grown by CVD onto Co-coated alumina substrate and functionalized with a loading of Pt (15 nm) at (a) low and (b) high magnifications; and with

a loading of Pt (30 nm) at (c) low and (d) high magnifications.

are toxic and dangerous for human health, even at low concentrations.

These NO2and NH3sensing microdevices are interesting for

environ-mental air monitoring and industrial process control, respectively

3.3 Sensor characterization: greenhouse gasses (CO2, CH4)

Actually, global warning and climate changes are real phenomena to

be monitored Some primary reasons are anthropic actions, e.g.,

industrial combustion processes, with consequent air-emission of

various greenhouse gasses such as carbon dioxide (CO2), methane

(CH4), nitrous oxide (N2O) Therefore, new cost-effective gas

micro-sensors are needed for environmental monitoring and air quality control

The time responses of the four CNTs-based chemiresistors are

shown inFigs 5(a) and6(a) for CO2 and CH4greenhouse gasses,

respectively For both reducing gasses, the electrical resistance

increases upon a given gas exposure of 5-minute pulses confirming

the p-type behavior of CNTs, including Pt-modified CNTs networked

films The sensor signal is recovered upon dry air when the chemical

test ambient is turned on, with a good dynamics Drift of the sensor

baseline has been observed due to thermal instability of sensing

coating in the microdevices This effect is more pronounced in the

CNTs modified with Pt loadings of 8 and 15 nm for CO2, and in the

unmodified CNT and modified with a Pt loading of 8 nm for CH4

Probably, this thermal drift could be attributed to non-nanotube

carbon (e.g., amorphous carbon) in the CNT networked films, as confirmed by SEM images These CNT layers have been used as-grown, and without any purification process (e.g., thermal annealing, acid treatment, and ultra-sonication) Maybe, these purification protocols could enhance the stability and perhaps also the sensitivity The calibration curves for the CO2and CH4greenhouse gasses have been shown in theFigs 5(b) and6(b) for all four CNTs-chemiresistors Generally, the Pt-modified CNT sensors exhibit higher gas sensitivity compared to un-functionalized CNT microsensors In the CO2case, the highest gas sensitivity has been measured for CNT sensor loaded with Pt

of 15 nm; while in the CH4case, the highest gas sensitivity has been achieved for CNTs sensor surface-modified with a Pt loading of 8 nm This enhanced gas sensitivity in the Pt-modified CNTs has been already reported by the authors[34]for other gasses (NO2, NH3, H2S, and CO) The improved sensitivity in the Pt-functionalized CNTs sensors can be explained by chemical gating of electron transfer promoted by Pt catalyst, with a consequent higher response

These results demonstrate that Pt-modified CNTs-sensors are interesting devices for monitoring greenhouse gasses (CO2and CH4),

at proof-of-concept level

3.4 Sensor characterization: domestic safety gasses (CO, C2H5OH) Carbon monoxide (CO) is very hazardous gas for health, especially in domestic environment due to gas leakage in the combustion systems

Fig 3 (a) Time response towards NO 2 gas of four CNTs-chemiresistors, at 120 °C Three

sensors are surface-modified with tuned loadings of Pt-catalyst at 8, 15 and 30 nm.

(b) Calibration curves, at 120 °C, towards NO 2 gas of four CNT-chemiresistors.

Fig 4 (a) Time response towards NH 3 gas of four CNT-chemiresistors, at 120 °C Three sensors are surface-modified with tuned loadings of Pt-catalyst at 8, 15 and 30 nm (b) Calibration curves, at 120 °C, towards NH 3 gas of four CNT-chemiresistors.

The Threshold Limit Value (TLV) for CO is 25 ppm, as reported by

American Conference of Governmental Industrial Hygienists (ACGIH)

[35] Therefore, new sensors with high performance at low cost are

required for domestic safety applications A strategy to reduce false

alarms in the domestic safety detectors is the usage of chemicalfilters to

cut the interfering gasses of the targeted analyte to be detected Usually,

a class of domestic interfering gasses is constituted by alcohols (ethanol,

methanol, iso-propanol, etc.), that can trigger a false response of the

sensor calibrated to targeted CO gas In particular, chemicalfilters based

on activated carbons are used to reduce the effect of interfering alcohols

on CO gas response

Fig 7(a) and (b) shows the time response of four

CNT-chemiresistors, unmodified and Pt-modified with tuned loadings of

8, 15 and 30 nm, towards CO gas in the range from 250 to 1000 ppm,

at the sensor temperature of 120 °C, with and without activated

carbonfilter The highest gas response has been measured for

Pt-modified CNTs having a loading of 15 nm: e.g., a resistance change of

120 and 110Ω has been recorded for CNT:Pt15 nm sensor to

1000 ppm CO for nofilter and with filter, respectively As observed,

the activated carbon filter reduces slightly the gas response and

deteriorates fairly the minimal gas detection from 300 to 250 ppm CO,

with and without carbonfilter, respectively However, the effect of

chemicalfilter on CO gas sensitivity for all four CNTs-based sensors

has been assessed in a broad range of CO gas concentration from 250

to 1000 ppm The results are shown in Fig 7(c) A comparative analysis of the calibration curves for all four sensors, with and with-out activated carbon filter, exhibits that the maximum filtering coefficient (FC) has been measured as low as 11% The FC has been

defined as 1−(ΔR1/ΔR2), where ΔR1 and ΔR2 are the gas responses

to a given CO concentration, with and without carbonfilter

In order to assess the effect of the carbonfilter on sensor response towards domestic interfering gasses, sensing characteristics of two typical sensors (CNT and CNT:Pt15 nm) operating at 120 °C have been measured for ethanol, in the range from 40 to 120 ppm, with and without chemical filter The results are shown in Fig 8 The gas response towards ethanol results to be lowered by the presence of the chemicalfilter by a filtering coefficient of 27% and 28% for CNT and CNT:Pt15 nm sensors, respectively Generally, the carbonfilter re-duces the gas response towards ethanol with afiltering coefficient higher than carbon monoxide (11%) Therefore, the chemicalfilter seems to be efficient to reduce the false alarm of the interfering ethanol against targeted CO gas Additionally, the chemical filter increases the response time and recovery time as well by lowering these sensing characteristics

Finally, these results demonstrate that the carbonfilter could be used to reduce the interfering effect of ethanol on CO detection for domestic safety to enhance the false alarm rate Although the chemicalfilter performance could be improved by other types of

Fig 5 (a) Time response towards CO 2 gas of four CNT-chemiresistors, at 120 °C Three

sensors are surface-modified with tuned loadings of Pt-catalyst as 8, 15 and 30 nm.

(b) Calibration curves, at 120 °C, towards CO 2 gas of four CNT-chemiresistors.

Fig 6 (a) Time response towards CH 4 gas of four CNT-chemiresistors, at 120 °C Three sensors are surface-modified with tuned loadings of Pt-catalyst at 8, 15 and 30 nm (b) Calibration curves, at 120 °C, towards CH 4 gas of four CNT-chemiresistors.

activated carbons with various porosity, mesh, size, composition and

solid solution

4 Discussion

The cross-sensitivity of the Pt-modified CNTs-sensors has been

studied by comparison of the mean sensitivity measured forfive targeted

gasses (NO2, NH3, CO2, CH4, and CO), at 120 °C The mean sensitivity Smis

defined as:

Sm=1

n∑n

i = 1 ΔR=R

ð Þi

where (ΔR/R)iis the percentage relative resistance change related to the i-measurand of gas concentration ci, weighted over a number n of exposures for the same gas The results achieved are shown inFig 9 The CNT-chemiresistors exhibit the highest sensitivity towards nitrogen dioxide (NO2) compared to other gasses (NH3, CO2, CH4, and CO) This enhanced sensitivity has been found also in a previous work of the authors[34] It can be attributed to a chemical affinity of the oxidizing NO2gas molecules with the carbon nanostructures, as largely reported in literature [8,15,16,18,19,26] In particular, the highest NO2sensitivity has been measured for CNTs modified with a loading of Pt of 8 nm, while the NO2gas sensitivity decreases with increasing Pt loadings of 15 and 30 nm This improved sensitivity is typical for the Pt clusters reduced size Generally, the higher sensitivity of the Pt-modified CNTs-sensors compared to unmodified CNTs is strongly related to catalytic spillover at nanoclusters surface This generates a lowering of the Pt work function by causing electron transfer from metal to nanotubes, with a consequent reduced conductivity in the p-type CNTs material Another competitive effect

of the electron transfer is based on capability of NO2electron-acceptor properties by causing an increase of hole density in the CNTs, with a

Fig 7 Time response towards CO gas of four CNT-chemiresistors, at 120 °C, (a) without

and (b) with activated carbon filter Three sensors are surface-modified with tuned

loadings of Pt-catalyst at 8, 15 and 30 nm (c) Comparison of the calibration curves, at

120 °C, towards CO gas of four CNT-chemiresistors, without and with activated carbon

filter.

Fig 8 Comparison of the sensor response, at 120 °C, towards interfering ethanol (C 2 H 5 OH) for (a) CNT-sensor and (b) Pt-modified CNT-chemiresistor loaded with

15 nm, without and with activated carbon filter.

consequent increased conductivity In the case of greater Pt loading of

30 nm, thefirst effect caused by injection of electrons from Pt metal

into CNTs seems to be dominant with a consequent reduced NO2

sensitivity compared to unmodified CNTs and Pt-modified CNTs

having lower loadings of 8 and 15 nm

The NO2gas sensitivity of the four CNT-sensors results to be higher

up to one order of magnitude compared to other tested gasses of NH3,

CO2, CH4, and CO Additionally, the chemical patterns in the

low-sensitivity range have been shown in the inset ofFig 9 In this case,

ammonia sensitivity is the highest compared to the other gasses (CO2,

CH4, and CO)

Also, CNT-sensor drift has been observed in the various transients

The baseline of a given CNT-sensor is affected by peculiar thermal effects

that can increase the rate of drift in the gas sensors if the non-nanotube

carbon material has not been completely removed from grown

CNTs-networks This can be a limitation for practical applications, mainly for

gas detection at sub-ppm level Material processing based on

pre-conditioning procedures (i.e., thermal annealing, acid treatment, and

surface modifications) can be adopted to enhance both the baseline drift

and long-term stability of the produced sensors This strategy has been

planned for future work

5 Conclusions

Carbon nanotubes have been grown by CVD technology directly onto

Co-catalyzed low-cost alumina substrates for toxic gasses (NO2and

NH3), greenhouse gasses (CO2and CH4), domestic safety gasses (CO and

C2H5OH) detection applications The CNTs are networked layers with a

maximum length up to 1–5 μm and diameter ranging from 5 to 40 nm

Surface-functionalizations with nanoclusters of sputtered Pt catalyst at

tuned loadings of 8, 15 and 30 nm are performed to remarkably enhance

gas sensitivity of the chemiresistive microsensors, operating at a

temperature of 120 °C A detection of 50 ppb NO2has been sensed at

120 °C by Pt-functionalized CNT-sensors The maximum NO2 gas

sensitivity has been measured by the Pt-modified CNT device with a

loading of 8 nm The cross-sensitivity of the Pt-modified CNT

micro-sensors has been studied as well The results demonstrate that the

highest NO2gas sensitivity has been achieved up to one higher order of

magnitude, if compared to other tested gasses (NH3, CO2, CH4, and CO) Additionally, the chemicalfilters based on activated carbon could be a good strategy to reduce false alarms caused by interfering alcohols (e.g., ethanol) to detect targeted CO gas in domestic environment for chemical safety Future work is planned for investigations on stability

of the CNT-based sensor signal by post-treatment and post-processing

of the carbon nanomaterial, and for improving the efficiency of chemical filters to reduce interfering effects on gas response for domestic safety and indoor air quality control In summary, these results demonstrate good sensing performance of the Pt-functionalized CNT-microsensors for selective NO2 gas detection at low temperature of 120 °C for environmental-air monitoring applications

References

[1] S Iijima, Nature 354 (1991) 56.

[2] S Iijima, T Ichihashi, Nature 363 (1993) 603.

[3] M.S Dresselhaus, G Dresselhaus, Ph Avouris, Topics in Applied Physics, vol 80, Springer-Verlag, Berlin, Germany, 2001.

[4] S.S Wong, E Joselevich, A.T Woolley, C.L Cheung, C.M Lieber, Nature 394 (1998) 52.

[5] C.H Hsu, C.F Chen, C.C Chen, S.Y Chan, Diamond Related Materials 14 (2005) 739.

[6] Ph Avouris, Acc Chem Res 35 (2002) 1026.

[7] Ph Avouris, J Appenzeller, R Martel, S Wind, Proc IEEE 91 (2003) 1772 [8] J Kong, N.R Franklin, C Zhou, M.G Chapline, S Peng, K Cho, H Dai, Science 287 (2000) 622.

[9] P.G Collins, K Bradley, M Ishigami, A Zettl, Science 287 (2000) 1801 [10] J Li, Y Lu, Q Ye, M Cinke, J Han, M Meyyappan, Nano Lett 3 (7) (2003) 929 [11] M Penza, G Cassano, P Aversa, A Cusano, A Cutolo, M Giordano, L Nicolais, Nanotechnology 16 (2005) 2536.

[12] Q Cao, J.A Rogers, Adv Mater 21 (2009) 29.

[13] O.K Varghese, P.D Kichambre, D Gong, K.G Ong, E.C Dickey, C.A Grimes, Sens Actuators B 81 (2001) 32.

[14] A Modi, N Koratkar, E Lass, B Wei, P.M Ajayan, Nature 424 (6945) (2003) 171 [15] M Penza, G Cassano, R Rossi, A Rizzo, M.A Signore, M Alvisi, N Lisi, E Serra, R Giorgi, Appl Phys Lett 90 (2007) 103101.

[16] L Valentini, I Armentano, J.M Kenny, C Cantalini, L Lozzi, S Santucci, Appl Phys Lett 82 (2003) 961.

[17] P Qi, O Vermesh, M Grecu, A Javey, Q Wang, H Dai, S Peng, K.J Cho, Nano Lett.

3 (2003) 347.

[18] C Cantalini, L Valentini, L Lozzi, I Armentano, J.M Kenny, S Cantucci, Sens Actuators B 93 (2003) 333.

[19] M Penza, R Rossi, M Alvisi, M.A Signore, E Serra, J Phys D Appl Phys 42 (2009) 072002.

[20] T Ueda, S Katsuki, K Takahashi, H.A Narges, T Ikegami, F Mitsugi, Diamond Related Materials 17 (2008) 1586.

[21] E.H Espinosa, R Ionescu, C Bittencourt, A Felten, R Erni, G Van Tendeloo, J.J Pieraux, E Llobet, Thin Solid Films 515 (2007) 8322.

[22] T Ueda, M.M.H Bhuiyan, H Norimatsu, S Katsuki, T Ikegami, F Mitsugi, Phys E

40 (7) (2008) 2272.

[23] K.H An, S.Y Jeong, H.R Hwang, U.H Lee, Adv Mater 16 (2004) 1005 [24] T Zhang, S Mubeen, N.V Myung, M.A Deshusses, Nanotechnology 19 (2008) 332001 [25] M Penza, R Rossi, M Alvisi, M.A Signore, E Serra, R Paolesse, A D'Amico, C Di Natale, Sens Actuators B 144 (2010) 387.

[26] M Penza, G Cassano, R Rossi, M Alvisi, A Rizzo, M.A Signore, Th Dikonimos, E Serra, R Giorgi, Appl Phys Lett 90 (2007) 173123.

[27] J Kong, M.G Chapline, H Dai, Adv Mater 13 (18) (2001) 1384.

[28] M.K Kumar, S Ramaprabhu, J Phys Chem B 110 (2006) 11291.

[29] A Star, V Joshi, S Skarupo, D Thomas, J.-C.P Gabriel, J Phys Chem B 110 (2006) 21014.

[30] W Wongwiriyapan, S Inoue, T Ito, R Shimazaki, T Maekawa, K Suzuki, H Ishikawa, S Honda, K Oura, M Katayama, Appl Phys Express 1 (2008) 014004 [31] R Leghrib, A Felten, F Demoisson, F Reniers, J.-J Pireaux, E Llobet, Carbon 48 (2010) 3477.

[32] M Penza, R Rossi, M Alvisi, M.A Signore, G Cassano, D Di Maio, R Pentassuglia,

E Piscopiello, E Serra, M Falconieri, Thin Solid Films 517 (2009) 6211 [33] M Penza, R Rossi, M Alvisi, E Serra, Nanotechnology 21 (2010) 105501 [34] M Penza, R Rossi, M Alvisi, G Cassano, M.A Signore, E Serra, R Giorgi, Sens Actuators B 135 (2008) 289.

[35] American Conference of Governmental Industrial Hygienists (ACGIH): Threshold Limit Values (TLV) and Biological Exposure Indices (BEI), 2001.

Fig 9 Chemical pattern of the mean sensitivity of four Pt-modified CNT-chemiresistors,

operated at 120 °C, for five targeted gasses The inset shows the mean sensitivity of the

low-sensitive chemical patterns towards four targeted gasses.