Scanning Electron Microscopy (SEM) was used to determine the morphology of the carbon nanotubes through a scanning electron microscope, model JEOL 5900 LV, at 25 kV. The samples were mounted on an aluminium holder. In all cases, the images were obtained with a backscattered electron detector. Energy X-ray Dispersion Spectroscopy (EDS) determined the semi-quantitative elemental chemical compositions of the samples with an EDAX-4 spectrometer.

Transmission Electron Microscopy (TEM) is the most important and reliable technique for correctly identifying the nature and the form of carbon nanomaterials in academic research and in industry. The information provided by TEM directly reveals both the surface and the intrinsic structure of the nanotube [23] . The observation for this research was carried out with a JEOL JEM-2010 TEM operating at 400 kV.

Compounds were identified by comparing the diffraction patterns to the Joint Committee on Powder Diffraction Standards (JCPDS) using conventional methods An X-ray diffraction analysis was performed for CNTs. The samples were placed in a holder of Lucite and then on the goniometer of the diffractometer (Siemens D-5000 diffractometer with a copper anode X-ray tube with λ = 1.543 Å). Kα radiation was selected with a diffracted beam monochromator at 25 kV and 2θ steps of 0.02˚ for 50 min to acquire X-ray patterns from 4˚ to 70˚, with high enough intensities to achieve the lines to identify in angle 2θ the mineral components of the analysed material. Compounds were identified by comparing the diffraction patterns to the Joint Committee on Powder Diffraction Standards (JCPDS) using conventional methods [24] .

5 mg of the samples were placed directly into platinum crucible, and the analysis was carried out under N₂ flow at the heating rate of 10˚C min⁻¹, with the temperature ranging from room temperature up to 950˚C using a TGA-TDA 51 TA Instruments Thermo gravimetric Analyser.

Specific surface area (SSA, m²g⁻¹) were determined for both dry and degassed natural carbon nanotubes using a multipoint N₂ adsorption/desorption Brunauer- Emmett-Teller (BET) method at room temperature in a Micrometrics BEL Japan INC surface area analyser.

Raman spectroscopy is one of the most powerful tools for characterization of CNTs. Without sample preparation, a fast and nondestructive analysis is possible. All allotropic forms of carbon are active in Raman spectroscopy [25] , and the position, width, and relative intensity of bands are modified according to the carbon forms [26] . The Raman analysis was carried out using a Horiba JobinYvon spectrometer, HR Lab Ram 800. The laser was used to excite the sample, 532 nm. All the measurements were performed at room temperature.

A study in 2009, realized by the Center for Sustainable Transport in Mexico, found that in the Toluca city, the car brands: Nissan-Tsuru 1997, 1999, 2008, Chevrolet-Chevy 1998, Chevrolet-Corsa 2007, have more presence and 87% of the vehicles are personal and private and four-fifths of them run every day. For this reason in it investigation, the vehicles were randomly selected and were used five cars of these brands. The experiments were carried out in the verificenter

TO-913 Macrover Auto SAQ. S.A. de C.V., located in Ignacio Comonfort Avenue, Santa Ana Tlapaltitlán, Toluca, México. The car emissions of hydrocarbons (HC), oxides of nitrogen (NO_x), carbon monoxide (CO) and carbon dioxide (CO₂), were verified according to program.

In this study, the concentrations of air pollutants after passing through the CNTs filter were calculated using the following equation [22] :

RemovalefficiencyofCNTfilter = ( 1 − Cwith-CNTs Cwithout-CNTs × 1 00 % )

where C_with-CNTs is the flow of air pollutants after passing through CNT filters, and C_without-CNTs is the flow of air pollutants passing through membrane supports without CNTs loading.

1) Scanning Electron Microscopy before sorption process

The carbon nanotubes analysed in this work were colour black. The SEM image 20,000 magnification for CNTs is shown in Figure 2(a), in this can be observed that the bulk of the material is mainly constituted by fibbers; the material has a variety in diameter from 25 to 130 nm and 15 to 50 μm in length.

In conjunction with electron microscopy, elemental chemical composition was determined. The identification of surface feature was performed by qualitative EDS analyses. The EDS results are: for the CNTSs were found: C (91.29%), O (6.23%) and Fe (2.48%). In this cases iron nanoparticles were used as the catalyst, for this the presence or iron in this percentage.

2) Scanning Electron Microscopy after sorption process

Below are the micrographs of the CNTS used in sorption of gaseous pollutant of exhaust from mobile sources, where it is observed that the morphology of the CNTS representative is unchanged, Figure 2(b), also in chemical composition, where in the initial percentage of unchanged: 82.70% C, 15.66% O, 1.24% Fe, however was found the presence of silicon (Si), although at a low concentration of 0.38%.

3) Transmission Electron Microscopy analysis

These studies confirm the collected material consists of multi-walled carbon nanotubes (MWCNTs), with the dominant tube diameter in the range 72.5 nm with length 50 μm and the inter-tube spacing is of 0.34 nm, Figure 3.

According to the Joint Committee on Powder Diffraction Standard (JCPDS, card

01-075-2078), the main mineral compound present in CNTs is graphite. The intensity of the strongest peak to 26˚ for graphite is considered to represent the crystalline compound, and the intensity between 43˚ and 46˚ represents the background due to amorphous components. This behaviour indicates that the material does not suffer modification of its structure, Figure 4.

Thermal analysis methods are useful for identifying relative changes due to processing. It was used to study the degree of CNTs purification [25] . The loss of weight were from 25 to 800˚C of 3.78%, from 800 to 870˚C of 0.46%, from 870 to 900˚C of 2.34% and 900 to 1000˚C of 3.71%. In agreement to the results, it is confirmed that the CNTs do not present a high loss of weight (10.5%), Figure 5.

The total surface area (TSA) determined from classical BET method by physisorption of nitrogen at 78 K is presented. The CNTs showed that specific surface area is of 153 m²/g, the total pore volume of 0.9534 cm³ g⁻¹ and an average pore diameter of 24.931 nm, classifying to the CNTs as mesoporous material (2 to 50 nm) according to the IUPAC classification. The data were analysed using the Langmuir adsorption model. It was concluded that CNTs had greater adsorption capacity, Figure 6.

The Raman spectrum of CNTs sample is indicated in Figure 7. As shown, the band at 1580 cm⁻¹ (G band) is due to tangential stretching of the carbon atoms in a two dimensional hexagonal lattice as a graphite sheet, this band reveals the presence of multi-walled nanotubes in the sample. The peak centered at around

1349 cm⁻¹ (D-band) indicates disturbances on the hybridization of sp² carbon atoms and has its overtone at 2681 cm⁻¹, is mainly due to the presence of amorphous carbon and defects in the sample [27] . These two positions are close to 1346 cm⁻¹ and 1578 cm⁻¹ which have been reported in the literature as features for multi-walled nanotubes concentric hexagonal lattice [28] .

Table 1 and Table 2 show the vehicle models that were tested, presents the average removal efficiencies obtained, there can be seen that reduction of air pollutants using the sorption device packaged with the CNTS it is different for each vehicle.

For the tests conducted Corsa 2007, the flow was maintained at 80 mL/min, the temperature varied between 40˚C and 50˚C. For 1998 Chevy, the flow was 110 mL/min and the temperature at 70˚C, to Nissan models 1997 and 1999, the flow varied between 85 and 90 mL/min, the temperature was 85˚C, 2008 model presents a flow of 40 mL/min, and the temperature was 40˚C.

The maximum permissible emission of pollutants applies to vehicles that are

evaluated under this test, are given by the Mexican Official Standards NOM- 041-SEMARNAT-1999 [29] and NOM-050-SEMARNAT-1993 [30] that apply to gasoline vehicles.

As shown in Table 1, the summary Comparison of Average Emission Results, was significantly higher variability was observed for the removal efficiency of the MWNT filters when collecting CO and NO₂. Nonetheless, the average removal efficiency of the CNT filters were observed for all air pollutants, but the removal efficiency was less. For the first test, the emissions were taken when vehicles were accelerated to 4 km/h.

Table 2 shows summary Comparison of Average Emission Results for the test fast idle (40 km/h) for different vehicles, shows the removal efficiencies of MWNT filters when collecting four air pollutants, CO, HC, CO₂, NO_x. However the removal efficiency for the CNT, was found significantly higher for NOx in all vehicles.

Nissan vehicles presented variability in its results; this is because they are different models. However the 2008 model had higher removal efficiency of the contaminants. The vehicle Corsa 2007 Chevy 1998 showed higher removal efficiency for NO_x.

After the removal, the CNTs showed that decrease 16% in BET surface area, it was 128 m² g⁻¹. The total pore volume of 0.9999 cm³ g⁻¹ and an average pore diameter of 31.06 nm, according to the IUPAC classification as mesoporous material (2 to 50 nm). The data were analysed using the Langmuir adsorption model. It was concluded that CNTs do not showed changes in adsorption capacity. Nitrogen adsorption isotherm of the CNTS, after sorption, follows the behavior of type IV and V of the physical adsorption isotherms IUPAC.