3-1. Molecular modeling calculations
A model for CNT is indicated in figure 2 with different views. CNT is optimized at B3LYP/3-21g** level of the theory then the HOMO/LUMO band gap is calculated as indicated in Fig. 3a. The molecular electrostatic potential (ESP) contour is calculated at the same level of theory and plotted in Fig. 3b. TiO2 is supposed to interact with the surface of CNT following two schemes, the first form is adsorbed state as shown in Fig. 4a. Then HOMO/LUMO band gap energy is calculated as shown in Fig. 5b., while Fig. 4c shows the calculated molecular ESP.
Fig. 5a presents the second form of interaction between 7 TiO2 and the surface of CNT which is through the complex form. Fig. 5b shows the HOMO/LUMO and Fig. 5c shows the molecular ESP. Before discussing the band gap and ESP, it is important to follow up on the physical changes in the surface of CNT as a result of TiO2 surface interaction either through adsorb and/or complex state. Table. 1 presents the B3LYP/3-21g** Mulliken atomic charges, TDM, and HOMO/LUMO band gap energy. The Mulliken charges are charges based on the local electron density (charge density) that depends strongly on the basis set and function because it is sensitive to the probability density. In this work, the function and basis set is the same for the studied structures (B3LYP/3-21g**). The Mulliken atomic charges are followed on one of the carbon atoms with which TiO2 is going to interact. The value of the charge is calculated before and after the interaction between CNT and TiO2. Mulliken atomic charges as listed in table 1 is 0.028287 for CNT, then changed as TiO2 interacted with CNT to be -0.084450 corresponding to adsorb state and -0.226259 corresponding to complex state. This is an indication for changing the charge upon the carbon atom as a result of the interaction with TiO2 in either adsorb state or complex state.
Table 1
Calculated B3LYP/3-21g** Mulliken atomic charges (MAC), total dipole moment (TDM) as Debye, HOMO/LUMO band gap energy (∆E) as eV, for Carbon nanotube, Carbon nanotube with 7TiO2 as complex state and Carbon nanotube with 7TiO2 as adsorb state.
Structures
|
MAC
|
TDM
|
∆E
|
X
|
Y
|
Z
|
Total
|
Carbon nanotube (CNT)
|
0.028287
|
0.0759
|
0.5694
|
-0.0085
|
0.5745
|
0.245
|
CNT/7TiO2 as adsorb state
|
-0.084450
|
-1.7088
|
1.9695
|
10.4649
|
10.7849
|
0.543
|
CNT/7TiO2 as complex state
|
-0.226259
|
-1.6525
|
-0.2346
|
11.9222
|
12.0385
|
0.633
|
Regarding the TDM with x, y and z distribution for CNT, the TDM is 0.5745 Debye. Changing the charge upon carbon is expected to make redistribution process which is already regarded as a result of the interaction, which is, in turn, changing the values of the dipole moment in x, y, and z. This is the reason for changing the TDM to 10.4649 Debye for adsorbing state and 12.0385 Debye for the complex state.
The correlation between the Mulliken atomic charges and TDM could be also tried for HOMO/LUMO bandgap energy. As listed in table 1, the bandgap is changed from 0.245 eV to 0.543 eV then 0.633 eV ongoing from CNT to adsorb state to complex state. Changing the surface properties of CNT as a result of interaction with TiO2 dedicates its surface for many applications. Another important physical parameter that describes the surface of a given structure is the molecular ESP which is plotted for the studied structures as indicated in Fig. 4c, 4-c and Fig. 5c respectively. The contour of ESP is a useful tool necessary for describing the surface of a given structure as it determines the sites for both electrophilic and/or nucleophilic reactions. In other words, the ESP could be very effective to determine the ability to form hydrogen bonding (Politzer et al. (1981)). This ESP contour is describing the charge distributions for the studied CNT and CNT/7TiO2 throughout colors. It was stated earlier that, the colors are ranging from negative to positive on going from red and yellow to blue. In other words, the following decreasing order is describing the increase in potential namely red < orange < yellow < green < blue (Politzer et al. (1985); Politzer et al. (1996)).
The electrophilic reactivity is always related to red then electrophilic reactivity is correlated with blue. Fig. 3b presents the ESP contour of CNT indicating that there is a uniform distribution of ESP. Regarding the CNT/7TiO2 adsorb state, the electrophilic active regions (negative regions) are localized on the left side as indicated in figure 4-c. This is also confirmed by the distribution of HOMO/LUMO as shown in figure 4-b. While ESP for CNT/7TiO2 complex state is localized uniformly up and down inside the CNT as indicated in figure 5-d. This is also confirmed by the distribution of HOMO/LUMO as shown in figure 5-b.
3-2. Comparison with experimental results
For verifications of the model the CNT is prepared by CVD using ferrocene and Xylene as catalyst and hydrocarbon source respectively, and then characterized tested with different characterization techniques. X- ray is one of the most powerful and nondestructive technics for crystal structure identification. Fig. 6 shows the XRD pattern of MWCNTs decorated with TiO2. The pattern shows three characteristic diffraction peaks at 2θ = 25.5o, 37.9o 47.79o, 55 o and 62 o. These diffraction peaks are corresponding to (101), (0 0 4), (2 0 0), (211) and (204) crystal planes of the pure anatase TiO2 phase (JCPDS Card No. 86-1157) (Kalaiarasi et al. (2018), Morales et al. (2012), Wang et al. (2014)). It was noticed that the XRD peaks are broad which confirm the nano size of the TiO2 particles. No diffraction peaks of CNT has been detected in the CNT/7TiO2 nanocomposite, which can be due to that the main characteristic diffraction peak of CNT (002) at 26.2° is perhaps shaded by the (101) peak at 25.3° of anatase TiO2 .
No other diffraction peaks belonging to any other impurities have been recognized indicating a pure anats phase. These results of this morphology suggest that the model which describes the interaction between CNT surface and TiO2 could be through the surface as indicated earlier in the molecular modeling part.
As indicated in Fig. 7(a,b), the HR-TEM images indicate the morphology and structure for CNT/TiO2. They revealed that the as-prepared TiO2 nanoparticles looks like spherical dark black dots and appear like agglomerated particles over the surface of CNTs. The amount of loaded TiO2 nanoparticles is not so high and distributed over all the external surface of CNTs. the TiO2 nanoparticles grown out the surface of the nanotubes indicating the weak force between TiO2 nanoparticles and CNTs. Moreover the HR-TEM revealed that the CNTs are of multiwall type with thick wall and narrow cavity. It Fig. 8 presents the TGA as thermal analysis which is well known as an effective tool to describe the amount of nano-metal oxide on the surface of CNTs which is sometimes known as the loading ratio. TGA was performed in the air from room temperature up to 800 oC at a scanning rate of 10 oC/min. The TGA thermogram exhibited to mass changes at 372 oC and 503 oC. The first mass change is a shoulder peak at 372 oC which may be attributed to the complete oxidation of the remaining non-oxidized Ti ions. This may be an indication that not all the metals are complexed and may exist as adsorb state. The second mass loss is due to the CNTs consumption (Chen et al. (2015)). The amount of loaded TiO2 nano particles has been estimated from TGA thermogram to be approximately 75%. The TGA results confirmed the both of XRD and HR-TEM results.
Table. 2 presents the FTIR spectrum for CNT/TiO2. The band at 3437 cm-1 could be assigned for the OH stretching from carboxyl groups, which can be ascribed to the oscillation of carboxyl groups (O=C-OH and C-OH). It is also the band at which the water and/or the moisture could appear during the measurements. The two CH- bands at 2920 cm−1 and 2850 cm−1 are assigned for symmetric and asymmetric bands respectively of CH2. These two bands are an indication of the fact that the methylene structure of carbon nanotubes is not destroyed. While the band at 2358 cm-1 can be associated with the OH stretching from strongly hydrogen-bonded -COOH group (Hussain et al. (2011)). The band at 1728 cm−1 is assigned for carboxyl groups and that at 1160 cm−1 is assigned for carbonyl group; both of which are formed in the modification stage (Saleh et al. (2012)). The band at 1580 cm−1 is assigned for the stretching of the carbon nanotube backbone, while the band at 1397 cm−1 is the O-H deformation of the C–OH group. Finally, the band at 660 cm−1 is the characteristic band for TiO2 (Sun et al. (2011)). For compassion with theoretical IR, only two bands are chosen; one for CNT and the other for the interaction between a metal oxide and CNT as indicated in Table. 3. As listed in table 2, the experimental band for stretching of CNTs was at 1580 cm-1 while that of TiO2 interacted with CNT was at 660 cm-1. The calculated B3LYP-3-21g** values for the same bands were 1517 cm-1 and 652 cm-1 respectively. Gathering the accuracy of the method together with the computational time, one can conclude that the studied level of theory is suitable for studying the CNT/ TiO2 regarding the computational time.
Table 2
FTIR band frequencies and their assignment for CNT/TiO2 nanocomposites.
Frequencies
|
Assignment
|
3437
|
O-H stretching of the carboxyl group coupled with that of the hydroxyl group
|
2920
|
Symmetric stretching of CH2
|
2850
|
Asymmetric stretching of CH2
|
2370
|
OH stretching from strongly hydrogen-bonded -COOH group
|
1728
|
Carboxyl group
|
1580
|
Stretching of the carbon nanotube backbone
|
1397
|
O-H deformation of C–OH group
|
1160
|
Carbonyl group
|
660
|
MO
|
Table. 3
FTIR spectrum of CNT/TiO2 as compared with the calculated DFT:B3LYP/3-21g** spectrum.
FTIR
|
Cal
|
Assignment
|
1580
|
1517
|
Stretching of the carbon nanotube backbone
|
660
|
652
|
MO
|