**5. Metal nanoparticlesdecorated CNTs**

Electronic, physical and chemical properties of metallic nanoclusters are usually sensitive to the changes in environment [81]. CNTs decorated with metallic nanoparticles (NPs) have been widely used to achieve selectivity and improve the sensitivity, response time and DLs for a variety of gas detections. Layer by layer, electrodeposition, chemical deposition, elec‐ trochemical deposition and sputtering are the methods used to prepare the metallic NP-CNTs composites discussed in this section.

SWCNTs films were modified with Pd NPs using sputtering method [82]. After apply differ‐ ent deposition times (40 – 160 s), it was found that 120 s was the optimum deposition time to obtain enhanced sensor response for 1% H2 in dry air at 50 o C. A typical response curve (igas/ iair vs. time) showed differences in response and recovery between the first and following H2 sensing cycles. FTIR studies were used to support and explain those differences and mecha‐ nisms of detection. The first cycle showed an overall larger electrical current in the presence of H2and then it reached a new steady state. When the atmosphere was change to dry air, the current did not go to its original value but remained in the steady state, which is consid‐ ered as an irreversible response. The explanation to this is that is atomized by the Pd NPs and spilled to the surface of the MWCNTs, occurring the chemical and irreversible reaction of hydrogenation of the carbonyl groups of the MWCNTs at the first cycle. The second cycle and following ones started at the steady state where the first cycle finished and the electrical current showed a decrease in the presence of H2 and when the atmosphere was changed to dry air, the electrical current recovered back to where the cycle started. This reversible be‐ havior is explained as physisorption of H2 molecules onto Pd/SWCNTs.

Pd/MWCNTs and Pt-Pd/MWCNTs composites were tested for the detection of H2 in a con‐ centration range of 20 ppm– 2% in N2 and 200 ppm – 2% in air [83]. Composites were pre‐ pared by growing CNTs yarns and then covered them with a layer of Pd NPs or sequentially deposited layers of Pd and Pt NPs, using a recently developed technique called self-fuelled electrodeposition (SFED). Exposure to 1% H2using N2 with 1% air as carrier gas. As with other Pd/CNTs-based sensors [82], an initial irreversible drop in resistance was ob‐ served, and after that, the sensor reached a steady state. A stable baseline was established just after a couple of exposure/recovery cycles. Pd-MWCNTs was not able to detect H2 con‐ centrations below 20 ppm but with the Pt-Pd/MWCNTs composites it was possible to detect concentrations as small as 5 ppm (0.0005%). The sensor saturated at 100 ppm and higher concentrations. When the same experiments were performed using air as carrier gas, it was found that the detection limit for both composites decreased.

rolysis step reduces the distance between the NPs and the SWCNTs and consequently re‐

Latest Advances in Modified/Functionalized Carbon Nanotube-Based Gas Sensors

http://dx.doi.org/10.5772/52173

353

Double wall carbon nanotubes (DWCNTs) have shown to have longer length, compared to SWNTs, provide a better percolation behavior, the possibility of modifying the outer layer without modifying the inner one and flexibility are some of their attractive charac‐ teristics. Considering those characteristics, Rumiche *et al.* evaluated Pd NPs/ (DWCNTs) composites as room temperature H2 sensor (Figure 5) [87]. Different amounts of DWNTs (15 and 20 µL) were deposited over silicon oxide substrates and decorated with 1, 3 and 6nm layers of Pd NPs. To evaluate their sensing properties, changes in resistance as function of time were recorded when exposed to 3%, 2%, 1%, 0.5%, 0.3%, 0.2%, 0.1%, and 0.05% of H2 in dry air. Samples containing 1nm Pd layer were unsuccessful detect‐ ing any of the tested concentrations. On the other hand, samples containing 3 and 6nm Pd layer showed overall similar performance in terms of increases of resistance and re‐ covery. When analyzing the results for the lowest tested H2 concentration (0.05%)it was found that the increase in resistance was comparable for samples with same Pd layer thickness and different DWCNTs content (e. g. 3nm thick Pd layer deposited over 15µL DWCNTs was comparable to of 3nm thick Pd layer over 20µL DWCNTs). However, the increase in Pd coating thickness produced a reduction in the response. The obtained re‐ sults confirmed that the combination of the amount of DWCNTs and the Pd-layer thick‐

Another H2 sensor based on N-doped MWCNTs electrochemically decorated with Au NPs (Au-NMWCNTs) was presented by Sadek and collaborators [88]. N-doped MWCNTs were chosen because they have enhanced surface reactivity and chemically active sites for the nu‐ cleation of Au NPs during the electrodeposition process. NPs of different sizes were ob‐ tained with variation of electrodeposition potential. Changes in resistance as function of time were used to evaluate the performance of the Au-NMWCNTs sensors when exposed to different H2 concentration between 0.06% and 1%. Sensitivity, response time and recovery time highly depended on the size of the AuNPs: the smallest the size the better the sensitivi‐

Penza *et al*. worked in the modification of SWCNTs with Pt, Ru and NiNPs to monitor toxic, landfill, and greenhouse gases (CO2, CH4, NH3, and NO2) [89, 90]. Pt NP layers with thick‐ ness of 8, 15 and 30 nm were sputtered over SWCNTs films and exposed to the different gas‐

that Pt-SWCNTs had better sensitivity than unmodified SWCNTs. It was also found that the sensitivity depended on the layer thickness. For instance, 8 nm Pt layer-SWCNTs showed highest sensitivity for NO2 and CH4 and 15 nm Pt layer-SWCNTs showed better sensitivity for NH3 and CO2. In a similar study, a sensor array containing Pt, Ru, and Ag NPs sputtered over SWCNTs with a thickness of 5nmwas able to detect and selectively discriminate be‐ tween landfill gases. Concentrations as low as 100 ppb of NO2 were selectively detected at a

C. Changes in resistance as function of time showed

duces the delay in electron transfer, allowing faster response times.

ness directly affects the sensitivity of the sensors.

ty and the shorter the response and recovery times.

es at an operation temperature of 120o

C.

temperature of 120 o

A Pd/MWCNTs flexible substrate H2 sensor was fabricated using the layer-by-layer techni‐ que [84]. For this, an Au IDE was sputtered over a polyester (PET) film, followed by the fab‐ rication of a poly (4-styrenesulfonic acid-*co*-maleic acid)/poly (allylamine hydrochloride) (PSSMA/PAH) bilayer film. Then, a MWCNTs layer was deposited over the PSSMA/PAH bilayer film and decorated with Pd NPs using chemical deposition. The sensors were ex‐ posed to H2 in a range of concentrations between 200 and 40000 ppm at room temperature and 53% RH. The rapid response and higher sensitivity of the Pd/MWCNTs when compared to plain MWCNTs is attributed to the well known catalytic effect of Pd NPs. Because the de‐ tection mechanism is the dissociation of the H2 molecules on the surface of the Pd NPs, the linear relation was found to be between the sensor's response and the square root of the con‐ centration. It also showed to be selective to H2 at concentrations higher than 1000 ppm (vs. NH3, CO, CH4 and others), to be highly reproducible, to have long-term stability and com‐ parable to similar sensors fabricated in rigid substrates.

Zilli *et al.* demonstrated that the H2sensing capacity of the Pd/MWCNTs nanocomposite is affected by the different stages of purification of the CNTs [85]. Pd NPs were chemically de‐ posited on pristine MWCNTs (Pd-CNT-P), gas-phase oxidized MWCNTs (Pd-CNT-O) and gas-phase oxidized/acid treated MWCNTs (Pd-CNT-A) and used as H2 sensing material. All three nanocomposites showed increase in resistance as function of time in the presence of 500µL (STP) of H2 and an immediate decrease when the H2 disappeared from the testing chamber. However, Pd-CNT-O showed to be more sensitive than Pd-CNT-P and Pd-CNT-A, which both showed similar responses. The reason for these results is that for CNT-P, cata‐ lytic Fe NPs, used for the CNT growth, are encapsulated inside the CNTs and in CNT-A, the Fe NPs were removed the during the acid treatment. After the gas-phase oxidation process (CNT-O), the Fe NPs become exposed and at the same time, oxygen-containing groups are formed in the surface of CNTs, which act as additional anchoring site for Pd NPs and thus making the Pd-CNT-O sample more sensitive to H2.

Pd/SWCNT composites were prepared by using a poly (amido amine) (PAMAM) dendrimer assisted synthesis, followed by a pyrolysis step to remove the dendrimers [86]. For H2 sens‐ ing experiments, the samples were deposited over Ti/Au electrodes and changes in resist‐ ance as function of time were measured. The Pd/SWCNTs samples showed to be more sensitive to 10, 000 ppm of H2 at room temperature, when compared to samples of chemical‐ ly reduced Pd on SWCNTs (without the presence of dendrimer) and Pd/PAMAM-SWCNTs. Moreover, these samples were able to detect all of the concentrations in a 10-100 ppm range. It was concluded that not only the dendrimers provided more nucleation sites for the Pd NPs and thus higher NPs density, but also that the removal of the dendrimers thru the py‐ rolysis step reduces the distance between the NPs and the SWCNTs and consequently re‐ duces the delay in electron transfer, allowing faster response times.

served, and after that, the sensor reached a steady state. A stable baseline was established just after a couple of exposure/recovery cycles. Pd-MWCNTs was not able to detect H2 con‐ centrations below 20 ppm but with the Pt-Pd/MWCNTs composites it was possible to detect concentrations as small as 5 ppm (0.0005%). The sensor saturated at 100 ppm and higher concentrations. When the same experiments were performed using air as carrier gas, it was

A Pd/MWCNTs flexible substrate H2 sensor was fabricated using the layer-by-layer techni‐ que [84]. For this, an Au IDE was sputtered over a polyester (PET) film, followed by the fab‐ rication of a poly (4-styrenesulfonic acid-*co*-maleic acid)/poly (allylamine hydrochloride) (PSSMA/PAH) bilayer film. Then, a MWCNTs layer was deposited over the PSSMA/PAH bilayer film and decorated with Pd NPs using chemical deposition. The sensors were ex‐ posed to H2 in a range of concentrations between 200 and 40000 ppm at room temperature and 53% RH. The rapid response and higher sensitivity of the Pd/MWCNTs when compared to plain MWCNTs is attributed to the well known catalytic effect of Pd NPs. Because the de‐ tection mechanism is the dissociation of the H2 molecules on the surface of the Pd NPs, the linear relation was found to be between the sensor's response and the square root of the con‐ centration. It also showed to be selective to H2 at concentrations higher than 1000 ppm (vs. NH3, CO, CH4 and others), to be highly reproducible, to have long-term stability and com‐

Zilli *et al.* demonstrated that the H2sensing capacity of the Pd/MWCNTs nanocomposite is affected by the different stages of purification of the CNTs [85]. Pd NPs were chemically de‐ posited on pristine MWCNTs (Pd-CNT-P), gas-phase oxidized MWCNTs (Pd-CNT-O) and gas-phase oxidized/acid treated MWCNTs (Pd-CNT-A) and used as H2 sensing material. All three nanocomposites showed increase in resistance as function of time in the presence of 500µL (STP) of H2 and an immediate decrease when the H2 disappeared from the testing chamber. However, Pd-CNT-O showed to be more sensitive than Pd-CNT-P and Pd-CNT-A, which both showed similar responses. The reason for these results is that for CNT-P, cata‐ lytic Fe NPs, used for the CNT growth, are encapsulated inside the CNTs and in CNT-A, the Fe NPs were removed the during the acid treatment. After the gas-phase oxidation process (CNT-O), the Fe NPs become exposed and at the same time, oxygen-containing groups are formed in the surface of CNTs, which act as additional anchoring site for Pd NPs and thus

Pd/SWCNT composites were prepared by using a poly (amido amine) (PAMAM) dendrimer assisted synthesis, followed by a pyrolysis step to remove the dendrimers [86]. For H2 sens‐ ing experiments, the samples were deposited over Ti/Au electrodes and changes in resist‐ ance as function of time were measured. The Pd/SWCNTs samples showed to be more sensitive to 10, 000 ppm of H2 at room temperature, when compared to samples of chemical‐ ly reduced Pd on SWCNTs (without the presence of dendrimer) and Pd/PAMAM-SWCNTs. Moreover, these samples were able to detect all of the concentrations in a 10-100 ppm range. It was concluded that not only the dendrimers provided more nucleation sites for the Pd NPs and thus higher NPs density, but also that the removal of the dendrimers thru the py‐

found that the detection limit for both composites decreased.

352 Syntheses and Applications of Carbon Nanotubes and Their Composites

parable to similar sensors fabricated in rigid substrates.

making the Pd-CNT-O sample more sensitive to H2.

Double wall carbon nanotubes (DWCNTs) have shown to have longer length, compared to SWNTs, provide a better percolation behavior, the possibility of modifying the outer layer without modifying the inner one and flexibility are some of their attractive charac‐ teristics. Considering those characteristics, Rumiche *et al.* evaluated Pd NPs/ (DWCNTs) composites as room temperature H2 sensor (Figure 5) [87]. Different amounts of DWNTs (15 and 20 µL) were deposited over silicon oxide substrates and decorated with 1, 3 and 6nm layers of Pd NPs. To evaluate their sensing properties, changes in resistance as function of time were recorded when exposed to 3%, 2%, 1%, 0.5%, 0.3%, 0.2%, 0.1%, and 0.05% of H2 in dry air. Samples containing 1nm Pd layer were unsuccessful detect‐ ing any of the tested concentrations. On the other hand, samples containing 3 and 6nm Pd layer showed overall similar performance in terms of increases of resistance and re‐ covery. When analyzing the results for the lowest tested H2 concentration (0.05%)it was found that the increase in resistance was comparable for samples with same Pd layer thickness and different DWCNTs content (e. g. 3nm thick Pd layer deposited over 15µL DWCNTs was comparable to of 3nm thick Pd layer over 20µL DWCNTs). However, the increase in Pd coating thickness produced a reduction in the response. The obtained re‐ sults confirmed that the combination of the amount of DWCNTs and the Pd-layer thick‐ ness directly affects the sensitivity of the sensors.

Another H2 sensor based on N-doped MWCNTs electrochemically decorated with Au NPs (Au-NMWCNTs) was presented by Sadek and collaborators [88]. N-doped MWCNTs were chosen because they have enhanced surface reactivity and chemically active sites for the nu‐ cleation of Au NPs during the electrodeposition process. NPs of different sizes were ob‐ tained with variation of electrodeposition potential. Changes in resistance as function of time were used to evaluate the performance of the Au-NMWCNTs sensors when exposed to different H2 concentration between 0.06% and 1%. Sensitivity, response time and recovery time highly depended on the size of the AuNPs: the smallest the size the better the sensitivi‐ ty and the shorter the response and recovery times.

Penza *et al*. worked in the modification of SWCNTs with Pt, Ru and NiNPs to monitor toxic, landfill, and greenhouse gases (CO2, CH4, NH3, and NO2) [89, 90]. Pt NP layers with thick‐ ness of 8, 15 and 30 nm were sputtered over SWCNTs films and exposed to the different gas‐ es at an operation temperature of 120o C. Changes in resistance as function of time showed that Pt-SWCNTs had better sensitivity than unmodified SWCNTs. It was also found that the sensitivity depended on the layer thickness. For instance, 8 nm Pt layer-SWCNTs showed highest sensitivity for NO2 and CH4 and 15 nm Pt layer-SWCNTs showed better sensitivity for NH3 and CO2. In a similar study, a sensor array containing Pt, Ru, and Ag NPs sputtered over SWCNTs with a thickness of 5nmwas able to detect and selectively discriminate be‐ tween landfill gases. Concentrations as low as 100 ppb of NO2 were selectively detected at a temperature of 120 o C.

sponse, followed by Rh/SWCNTs and F-SWCNTs, which were less sensitive. The other CNTs with different coating and loadings were insensitive to formaldehyde. However, when the array was exposed to formaldehyde at concentrations as low as 0.01 ppm (10ppb), Rh/SWCNTs sensor showed to be more sensitive and the presented an estimat‐ ed DL (by IUPAC definition) of 10ppb. The DL for pristine SWCNTs and F-SWCNTs sensors were 15ppb and 20ppb, respectively. The three sensors presented very fast re‐

Latest Advances in Modified/Functionalized Carbon Nanotube-Based Gas Sensors

Theoretical studies have been used to study the interaction between Pd and Pt NPs decora‐ tedCNTs and a wide variety of gases for different applications. Zhou *et al*. used the density functional theory (DFT) to study the adsorption and interaction of SO2, CH3OH, and CH4 with Pd-SWCNTs [91]. The replacement of a central C atom of the CNTs with a Pd atom causes structural deformations. As SO2 is adsorbed, there is a charge transfer from Pd-SWCNT to SO2. As for CH3OH, the appropriate adsorption conformation is thru the lone par of the oxygen of CH3OH and thus occurring an overall charge transfer from CH3OH to Pd-SWCNT. The interaction between Pd-SWCNT and CH4 is similar in that the charge transfer occurs from Pd-SWCNT to the gas. However, the interaction between Pd-SWCNTs and CH4

> Sensor Configuration

Pd MWCNT H2 Resistor 1% [82] Pd MWCNTs H2 Resistor 10000ppm\* [84] Pd MWCNTs H2 Resistor NS [85] Pd-Pt MWCNTs H2 Resistor 5ppm [83] Pd SWCNTs H2 Resistor 10ppm [86] Pd DWCNTs H2 Resistor 0.05 [87]

Au N-doped MWCNTs H2 Resistor 0.06%\* [88] Rh SWCNTs HCOH Resistor 10ppb [80]

Li and co workers investigated the adsorption of CO and NO on SWCNTs-decorated with Pd and Pt using first-principle calculations [92]. It was found that the electronic properties of SWCNTs change upon modification with Pd or Pd atoms. The semi-conductive band gap is decreased compared to pristine SWCNTs. The reason for the observed decrease in the band

DL Ref.

http://dx.doi.org/10.5772/52173

355

Resistor 100ppb NO2 [89 , 90]

is not as strong as the interaction between Pd-SWCNT and SO2 or even CH3OH.

CO2,CH4, NH3,NO2

**Table 3.** Metallic nanoparticles used to decorate SWCNTs for gas sensing applications

sponse (~18sec) and recovery time of approximately 1 minute.

NP NT Type Target

Pt, Ru, Ag SWCNTs

\*Tested concentration ^For concentrations >1%

**Figure 5.** A-D) Responses of the different Pd/DWCNTs composites when exposed to H2. (E) Calibration curves for the samples presented in figures (A-D). (From Rumiche *et al.* [87] Copyright © 2012, with permission from Elsevier. )

Lu *et al*. used pristine SWCNTs, fluorinated SWCNTs (F-SWCNTs) and rhodium doped SWCNTs (Rh-SWCNTs) and other various coatings and dopings on the SWCNTs for the room temperature detection of formaldehyde (HCOH) [80]. The measurements of changes in resistance as function of time when the array was exposed to the different concentrations of formaldehyde were used to analyze the sensor array performance. When exposed to 0.71 ppm formaldehyde, pristine SWCNTs showed the higher re‐ sponse, followed by Rh/SWCNTs and F-SWCNTs, which were less sensitive. The other CNTs with different coating and loadings were insensitive to formaldehyde. However, when the array was exposed to formaldehyde at concentrations as low as 0.01 ppm (10ppb), Rh/SWCNTs sensor showed to be more sensitive and the presented an estimat‐ ed DL (by IUPAC definition) of 10ppb. The DL for pristine SWCNTs and F-SWCNTs sensors were 15ppb and 20ppb, respectively. The three sensors presented very fast re‐ sponse (~18sec) and recovery time of approximately 1 minute.

Theoretical studies have been used to study the interaction between Pd and Pt NPs decora‐ tedCNTs and a wide variety of gases for different applications. Zhou *et al*. used the density functional theory (DFT) to study the adsorption and interaction of SO2, CH3OH, and CH4 with Pd-SWCNTs [91]. The replacement of a central C atom of the CNTs with a Pd atom causes structural deformations. As SO2 is adsorbed, there is a charge transfer from Pd-SWCNT to SO2. As for CH3OH, the appropriate adsorption conformation is thru the lone par of the oxygen of CH3OH and thus occurring an overall charge transfer from CH3OH to Pd-SWCNT. The interaction between Pd-SWCNT and CH4 is similar in that the charge transfer occurs from Pd-SWCNT to the gas. However, the interaction between Pd-SWCNTs and CH4 is not as strong as the interaction between Pd-SWCNT and SO2 or even CH3OH.


\*Tested concentration

**Figure 5.** A-D) Responses of the different Pd/DWCNTs composites when exposed to H2. (E) Calibration curves for the samples presented in figures (A-D). (From Rumiche *et al.* [87] Copyright © 2012, with permission from Elsevier. )

354 Syntheses and Applications of Carbon Nanotubes and Their Composites

Lu *et al*. used pristine SWCNTs, fluorinated SWCNTs (F-SWCNTs) and rhodium doped SWCNTs (Rh-SWCNTs) and other various coatings and dopings on the SWCNTs for the room temperature detection of formaldehyde (HCOH) [80]. The measurements of changes in resistance as function of time when the array was exposed to the different concentrations of formaldehyde were used to analyze the sensor array performance. When exposed to 0.71 ppm formaldehyde, pristine SWCNTs showed the higher re‐ ^For concentrations >1%

**Table 3.** Metallic nanoparticles used to decorate SWCNTs for gas sensing applications

Li and co workers investigated the adsorption of CO and NO on SWCNTs-decorated with Pd and Pt using first-principle calculations [92]. It was found that the electronic properties of SWCNTs change upon modification with Pd or Pd atoms. The semi-conductive band gap is decreased compared to pristine SWCNTs. The reason for the observed decrease in the band gap is due to charge transfer from the Pd and Pt atoms to the surface of the SWCNTs. Differ‐ ent from pristine SWCNT that show poor adsorption, Pd-SWCNTs and Pt-SWCNTs showed to chemisorb CO molecules as well as NO. However, Pt-SWCNTs showed bigger binding energy and charge transfer than Pd-SWCNTs. The formation of C-Pd, N-Pd, C-Pt, and N-Pt bonds demonstrate that the metal atoms provide additional adsorptions sites for gases and open the possibility to use both materials as sensors for the detection of CO and NO.
