Investigations of CO₂, CH₄ and N₂ physisorption in single-walled silicon carbon nanotubes using GCMC simulation

Abstract

In this paper, we report N₂, CH₄ and CO₂ adsorption in single-walled silicon carbon nanotubes (SiCNTs) using grand canonical Monte-Carlo and calculate the isosteric heat of gas adsorption. The results demonstrate that at ambient temperature and high pressure, gas adsorption in these nanotubes is in the order of CO₂ > CH₄ > N₂ and nanotubes’ order is (10,10) < (20,20) < (40,40), while this order of adsorptivity of nanotubes will be inverted for N₂ when the pressure is very low. Then, we fit our simulation results to Langmuir and Langmuir–Freundlich equations to illustrate the mechanism of gas adsorptivity. The fitting exhibits that the simulation data obtained are very close to Langmuir–Freundlich behavior, which emphasizes that the dominant adsorptivity has occurred in multi-layer adsorption. Moreover, the comparison between our simulation results and other reports, which studied these gases’ adsorption on different nanoporous materials, experimentally and theoretically, is presented to illustrate that SiCNTs still have the best gas adsorptivity ability at ambient temperature and low/high pressure.

Introduction

The useful discovery of carbon nanotubes by Iijima [1] has opened up the possibility of synthesis of nanometer-scale, one-dimensional materials including nanowires, nanoparticles, and nanotubes [2, 3]. Silicon carbon nanotube (SiCNT) was first synthesized in 2001 [2]; furthermore, it has been improved to grow for several kinds of one-dimensional nanostructures as a tube furnace in 2002 [3] and it is the one of the most important nanotubes with many excellent properties, especially; as promising membrane for gas adsorption applications [4].

In the past decade, much of nanoscientific investigations has focused on finding new materials in nanoscale with particular properties such as high-ability natural gas adsorption capacity, owing to their supplies of energy that satisfy industrial, governmental and business agencies [5, 6].

From the environmental, an economical benefit and the predominant cause of greenhouse effect and urgent issue suppressing global warming [7, 8], a large of worldwide attentions have been reported on adsorptivity properties of carbon dioxide (CO₂), methane (CH₄) and nitrogen (N₂) (as natural gas components) on various nanoporous materials like nanotubes to reducing CO₂ global emission and remove it from natural gas and pollutant air which as mention above, they have been considered as alternative resources of energy [5, 6].

In recent years, many researchers have studied CO₂, CH₄ and N₂ adsorption, theoretically and experimentally. That in follow we are going to examplify number of them and then compare their results with our data:

Mishra et al. [9] have reported CO₂, CH₄ and N₂ adsorption properties of a zinc-based metal organic framework, which is commonly known as ZnDABCO. Yi et al. [10] also studied these gas adsorption equilibrium isotherms on microwave-activated carbon, experimentally. Likewise, Zhang et al. [8] have presented experimental and computational studies to evaluate the adsorption-based separation of CO₂ from flue gas and natural gas using micro-porous metal organic framework Cu-TDPAT as a sorbent material. In another work, Karra et al. [11] performed atomistic grand canonical Monte-Carlo (GCMC) simulations to understand pore size, heat of adsorption, open metal sites, electrostatics, and ligand functionalization contributing to adsorption of CO₂, CO, and N₂ in MOFs: IRMOF-1, IRMOF-3, Cu-BTC, Zn₂[bdc]₂[dabco].

Billemont et al. [12] have studied using experiments and molecular simulations, the adsorption of CO₂ and CH₄ in nanoporous carbons in the presence of water. They observed a small volume of water; the shape of the adsorption isotherms remained similar; both the molecular simulations and experiments showed a slight decrease in the CO₂ and CH₄ adsorption amounts [12].

In 2007, Huang et al. investigated adsorption of CO₂ and CH₄ using GCMC on single-walled carbon nanotubes (SWCNTs) [5], and they proposed Harris and Yung’s EPM2 [13] and the spherical models for CO₂ as different potential models.

In this paper, applying GCMC simulation, we study adsorption of CO₂, CH₄ and N₂ to calculate their adsorptivity properties in the single-walled silicon carbon nanotubes (SWSiCNTs) purly.

GCMC computational details

Following our previous work [14, 15], the GCMC simulation was adopted here for SWSiCNTs with (10,10), (20,20) and (40,40) chirality and 50 nm tube length with 17.857, 35.714 and 71.429 Å diameters, respectively, in order to consider CH₄, N₂ and CO₂ physisorption from 294 to 343 K and 1 to 21,000 kPa range of temperature and pressure, respectively. The Si–C bond length is 1.87 Å [16].

In this research, all gas molecules are treated as one-site the simplest 12-6 Lennard–Jones (LJ) potential to calculate their interactions with together and with Si and C atoms of SWSiCNT and the LJ parameters between different atoms are calculated by Lorentz–Berthelot rules. We chose classical 12-6 LJ representation according to UFF parameterization for Si and C atoms of SWSiCNTs [17], whereas for CO₂, we used the LJ potential parameters from Ref. [5], and for CH₄ and N₂ from Ref. [18], which all LJ parameters here, are listed in Table 1.

Table 1 The list of 12-6 LJ potential parameters of gas molecules and atoms of SWSiCNT that have used in this work

The standard GCMC simulation is a stochastic method that can compute the adsorption isotherms of various gases in each solid framework such as nanotubes, with having a constant volume (V) in equilibrium and an infinite reservoir of gases imposing its chemical potential (μ) at fixed temperature (T) [19]. In this work we used the Music code for all GCMC calculations [20]. The last half of 15 million trials was used for calculating the ensemble averages in gas adsorption simulations. The periodic boundary condition (PBC) is imposed in tube length direction. Further GCMC details are available in our previous publications [14, 15].

It should be noted in the results and discussion section, at first we present the CO₂, CH₄ and N₂ adsorption isotherms and isosteric heats which are calculated and then the gas adsorption fitting to Langmuir and Langmuir–Freundlich equations will be exhibited and in the last subsection of results, we have been compared the gas adsorptivity of gases with other published experimental and simulation data.

Results and discussion

This work investigates gas adsorption properties of SWSiCNTs using GCMC.

The gases are CH₄, CO₂ and N₂. Therefore, our results are presented in some following sections.

CO₂, CH₄ and N₂ adsorptivities

At first, we show CO₂, CH₄ and N₂ adsorptivities in (10,10), (20,20) and (40,40) SWSiCNTs at 300 and 343 K at varying pressure ranges 1–21 MPa. Figures 1, 2 and 3 present the obtained data of gases in (10,10), (20,20) and (40,40) SWSiCNTs, respectively.

Fig. 1

CH₄, CO₂ and N₂ adsorption in (10,10) SWSiCNT at 300 and 343 K

Fig. 2

CH₄, CO₂ and N₂ adsorption in (20,20) SWSiCNT at 300 and 343 K

Fig. 3

CH₄, CO₂ and N₂ adsorption in (40,40) SWSiCNT at 300 and 343 K

A result obtained from Fig. 1 shows that the gases have less adsorption, especially at higher temperature. According to Fig. 1, maximum adsorption in all gases takes place almost at 10 MPa (moderate pressure) and range of adsorptivity of gases together is low, about between 7.5 and 9.5 mol/kg. This means that (10,10) SWSiCNT has low gas adsorptivity and it seems that CH₄ and CO₂ have better adsorption, which relates to amount of adsorption as shown in Fig. 1.

From Fig. 3, we observe that the gases have more adsorptivity in (20,20) SWSiCNT than in (10,10) SWSiCNT. However, the adsorption is decreased as a function of increasing the temperature. According to Fig. 3, maximum of adsorption in all gases occurred almost at 15 MPa pressure, and range of adsorptivity of gases together was about between 13 and 19 mol/kg. This means that the (20,20) SWSiCNT has more gas adsorptivity rather than the (10,10) SWSiCNT, and CH₄ and CO₂ have better adsorptivity. Moreover, CO₂ has better amount of adsorption than CH₄ especially at low and moderate pressure (see Fig. 3).

Figure 5 also indicates that the gases have more adsorptivity in (40,40) SWSiCNT than in (10,10) and (20,20) SWSiCNTs. The gas adsorptions are decreased with increasing the temperature. Although, according to Figs. 1, 3 and 5, the maximum gas adsorptions occur in the (40,40) SWSiCNT, the range of adsorptivity of gases together in the latter nanotube is about between 10 and 45 mol/kg. This means that the (40,40) SWSiCNT has more gas adsorptivity rather than the (10,10) and (20,20) SWSiCNTs. Furthermore, it is clear that CO₂ has better adsorptivity for this reason: the most adsorptivity we can see in Fig. 5. Figure 5 shows that the (40,40) SWSiCNT can adsorb CO₂ at ambient temperature at low and moderate pressure (<10 MPa). In addition, CH₄ is adsorbed better than N₂. Consequently, the large diameter SWSiCNT can adsorb CO₂ much more than CH₄ and N₂, while small diameter SWSiCNT can adsorb CO₂ the same as CH₄. The important point is that CO₂ has more ɛ/k _B rather than CH₄ and N₂ (Table 1), so, its interaction with atoms of SWSiCNT nanotubes is more considerable.

Therefore, more adsorptivity of CO₂ is expected energetically. It means that natural adsorptivity directly depends on the atomic interactions.

Adsorption fitting to Langmuir and Langmuir–Freundlich equations

In the next investigation, we have fitted adsorption data to Langmuir ($\mathit{\theta }=\frac{{\mathit{\theta }}_{\text{m}}kP}{1+kP}$) and Langmuir–Freundlich ($\mathit{\theta }=\frac{{\mathit{\theta }}_{\text{m}}k{P}^n}{1+k{P}^n}$) equations, where θ is the amount adsorbed, P is the equilibrium pressure, and θ _m, k, b, n are isotherm parameters. Figures 4 and 5 present CH₄, CO₂ and N₂ plots at 300 and 343 K related to (10,10) and (40,40) SWSiCNTs, respectively. It is obvious that both figures show that the data of adsorption correspond to the Langmuir–Freundlich equation. These observations emphasize that the Langmuir–Freundlich is the best fitting that illustrates the adsorption of gases imply with varying energetic inhomogeneity in both (10,10) as smaller diameter and (40,40) as larger diameter SWSiCNTs [14, 15]. However, the gases adsorption adaptation to the Langmuir–Freundlich model can distinguish that multi-layer adsorptivity is very feasible [14, 15]. Because the pressure is increased and temperature is lowered, additional layers are formed. This has led to the modern concept of multilayer adsorption. In Fig. 6, we can see that even in the (10,10) SWSiCNT, the secondary layer of gas adsorption is taking place. So, Langmuir’s model of uni-molecular adsorption is valid here.

Fig. 4

The data of adsorption fitted on Langmuir (dash lines) and Langmuir–Freundlich (color lines) equations in (10,10) SWSiCNT at 300 and 343 K. a CH₄, b CO₂, c N₂

Fig. 5

The data of adsorption fitted on Langmuir (dash lines) and Langmuir–Freundlich (color lines) equations in (40,40) SWSiCNT at 300 and 343 K. a CH₄, b CO₂, c N₂

Fig. 6

The cross-section of (10,10) and (40,40) SWSiCNTs while CO₂, CH₄ and N₂ are adsorbed inside them at 300 K and 21 kPa. a CH₄, b CO₂, c N₂. “1” and “2” are assigned to (10,10) and (40,40) SWSiCNTs, respectively

The comparison of gases adsorption

In this subsection, we compare the CO₂, CH₄ and N₂ adsorptions in SWSiCNTs and other nanoporous materials. Therefore, we first focus on CO₂ adsorption.

The comparison of CO₂ adsorption

Mishra et al. [9] have reported CO₂ adsorption property of ZnDABCO at 294 and 350 K, experimentally. Figure 7a that is related to 294 K, shows that ZnDABCO has more CO₂ adsorptivity than the (10,10) SWSiCNT, but it has less than the (20,20) and (40,40) SWSiCNTs. However, CO₂ adsorption properties of ZnDABCO and (20,20) SWSiCNT are very close. Figure 7b also is related to 350 K, and shows that ZnDABCO has CO₂ adsorptivity close to the (10,10) SWSiCNT, but it is less than the (20,20) and (40,40) SWSiCNTs. As a result, it seems, in relatively high temperature, SWSiCNT has good CO₂ adsorption than inorganic nanoporous materials. In fact, many inorganic nanoporous materials such as MOFs are destroyed in high temperature, while nanotubes, especially SiCNTs, have the most thermal stability [21].

Fig. 7

a CO₂ adsorption property of ZnDABCO reported from Ref. [9] and our results at 294 K. b CO₂ adsorption property of ZnDABCO reported from Ref. [9] and our results at 350 K. c CO₂ adsorption equilibrium isotherm from Refs. [8, 10] and our data at 298 K and range 1–100 kPa. d Adsorption simulation of CO₂ from Ref. [11] that is about CO₂ adsorption in Cu-BTC and our data at 298 K. e Adsorption simulation of CO₂ from Ref. [12] experimental and simulation results, Ref. [5] simulation data and our data at range of 1,000–25,000 kPa

Figure 7c also presents the CO₂ adsorption equilibrium isotherms reported from Refs. [8, 10] and our data at 298 K and range 1–100 kPa. This figure again indicates that the (10,10) and (20,20) SWSiCNTs have more CO₂ adsorptivity than activated carbon and Cu-TDPAT. It should be noted that our obtained simulation data values are above the simulation data values of Cu-TDPAT [8].

Karra et al. [11] reported atomistic GCMC simulations to realize the adsorption of CO₂, CO, and N₂ in IRMOF-1, IRMOF-3, Cu-BTC, Zn₂[bdc]₂[dabco] at 298 K and pressure lower than 1,200 kPa. Figure 7d presents their simulation data (Ref. [11]) and our simulation results.

Another study is performed by Billemont et al. that they have been used both experimental and simulation approaches to indicate the adsorption of CO₂ on the Filtrasorb F400 activated carbon (experimental absolute adsorption) and on CS1000A and CS1000AF (simulated adsorption) as nanoporous carbons in the presence of water at 318.15 K [12]. They show that CS1000AF has more adsorption of CO₂ than CS1000A theoretically. Then, we compare our data with simulation data of CO₂ adsorptivity on CS1000A and also Filtrasorb F400. In the section that is related to adsorption of CH₄ comparison, we will remind about Billemont et al. (Ref. [12]) research again. Figure 7e presents this comparison.

Huang et al. [5] using GCMC have investigated the adsorption of CO₂ (applying Harris and Yung’s EPM2 and spherical models for CO₂ as different potential models) and CH₄ at 343 K. Their obtained data of simulation are shown in Fig. 7-e. Our results that we consider the spherical model for CO₂ at 300 and 343 K with comparing to Billemont et al. of simulation and experimental indication (Ref. [12]), illustrate that (20,20) SWSiCNT has more CO₂ adsorptivity than nanoporous carbon, but (10,10) SWSiCNT is less CO₂ adsorptivity than nanoporous carbon. In addition, the results of Huang et al. (Ref. [5]) have shown CO₂ adsorptivity close to the (10,10) SWSiCNT in the desired pressure range and at 343 K. Therefore, the (20,20) SWSiCNT can adsorb CO₂ more than nanoporous carbon and SWCNTs at the same thermodynamic conditions. As a result, Billemont et al. of simulation and experimental research (Ref. [12]) has been done at 318.15 K, and our investigation is at 300 and 343 K, and we know the adsorption decreases with increasing temperature, so, we can realize that (20,20) SWSiCNT even at 343 K has more CO₂ adsorptivity than nanoporous carbon at 318.15 K.

Consequently, as final investigation of the adsorption of CO₂, we can find that SWSiCNT with diameter large enough (diameter larger than (10,10) tube) is proper to adsorb CO₂ as compared to some MOFs, nanoporous carbon and CNTs.

The comparison of CH₄ adsorption

In our work, also, we compare CH₄ adsorptivity in SWSiCNTs with other nanoporous materials such as Cu-TDPAT and ZnDABCO as MOF, activated carbon and CNTs. At first, we report our obtained simulation data at 294 K and pressure under 3,000 kPa, with Mishra et al. (Ref. [9]) as experimental study that it has been considered the adsorption of CH₄ on a zinc based metal organic framework: ZnDABCO. Figure 8a illustrates the adsorption isotherms of CH₄ in this case. In Fig. 8a, we can find that CH₄ adsorptivity even in the (10,10) SWSiCNT is more than in ZnDABCO; however, our data are from simulation, and Mishra et al.’s (Ref. [9]) work is an experimental study. Furthermore, Fig. 8a, shows that CH₄ adsorptivity behavior is the same, in both ZnDABCO and SWSiCNT.

Fig. 8

a CH₄ adsorption property of ZnDABCO reported from Ref. [9] and our results at 294 K. b CH₄ adsorption equilibrium isotherm from Refs. [8, 10] and our data at 298 K and range of 1–100 kPa. c Adsorption simulation of CH₄ from Ref. [12] experimental and simulation results, Ref. [5] simulation data and our data at range of 1,000–25,000 kPa

Figure 8b also compares Yi et al.’s (Ref. [10]) experimental study and Zhang et al.’s (Ref. [8]) experimental and simulation works with our simulation data at 298 K and pressure range of 0–100 kPa, while Yi et al. (Ref. [10]) represented their data on activated carbon, and Zhang et al. (Ref. [8]) reported their data on Cu-TDPAT. Figure 8b shows that considering GCMC simulation on adsorptivity of CH₄ in Cu-TDPAT and SWSiCNT, we can realize that SWSiCNT, even with (10,10) chirality is more CH₄ adsorbable than Cu-TDPAT at same thermodynamic conditions. Also, this figure emphasizes that experimental data from Ref. [10] on activated carbon are very close to experimental and simulation data from Ref. [8] on Cu-TDPAT. This is a very gratifying challenge in gas adsorption investigations, because reports contain both simulation and experimental studies on two different nanoporous materials. Simulation data and experimental data from Ref. [8] provide similar estimates of the adsorption of gases on MOFs. Also, our obtained data exhibit that SWSiCNT is a more CH₄ adsorbable nanoporous material. Therefore, we understand that this type of analysis (simulation calculations) can be accurate, while the achievement of this method can be significant, and results obtained from the simulation calculations are credible in this field.

In addition, we compare CH₄ adsorption at 300 and 343 K and pressure range 0–25,000 kPa with Ref. [12] as experimental (on Filtrasorb F400 activated carbon at 318.15 K) and simulation (on CS1000A and CS1000AF activated carbons) investigations and with Ref. [5] as simulation study (on CNTs). Figure 8c exhibits that GCMC simulation data of CH₄ on (10,10) SWSiCNT is more than GCMC simulation data of CH₄ on CNTs (Ref. [5]). In fact, CNTs have about 1 mol/kg CH₄ adsorbed value at 343 K (considering both EPM2 and spherical models of CH₄) while (10,10) SWSiCNT has about 7–9 mol/kg CH₄ adsorbed value at 343 K (considering spherical model of CH₄ in our work). However, simulation investigation on CS1000A and CS1000AF activated carbons at 318.15 K from Ref. [12] is very close to our data of (10,10) SWSiCNT at 300 and 343 K, but (20,20) SWSiCNT, even at 343 K (that is more than 318.15 K) shows more CH₄ adsorptivity.

Finally, with considering to Fig. 8a, b and c, we can result that SWSiCNT with large enough diameter is more CH₄ adsorbable at room temperature, and also, GCMC simulation can predict suitable the CH₄ adsorption behavior, this approach can also be reliable in terms of accuracy.

The comparison of N₂ adsorption

In last comparison, we represent our N₂ adsorption from GCMC simulations with experiments reports and with few existing theoretical works.

Figure 9a exhibits our data and Mishra et al. (Ref. [9]) experimental report at 294 K under pressure of 3,000 kPa. Mishra et al. [9] present their temperature investigations and they have been resulted that ZnDABCO has a little more N₂ adsorptivity at 294 K than 350 K. Also, our GCMC simulation data of N₂ is more than Mishra et al. (Ref. [9]) reports. In Fig. 9a, we can realize that N₂ cannot more adsorbable in one type of materiel at different thermodynamics conditions. It means that with varying the thermodynamic conditions, for instance in ZnDABCO, according to Ref. [9] as experimental work, we can see that the increase of pressure cannot significantly affect nitrogen uptake, and even though with decreasing the temperature, not significant changing occur in N₂ uptake. Also, we can observe from our simulation calculations that the increase of pressure cannot have considerable effect on nitrogen uptake in (10,10) and (20,20) SWSiCNTs. Here, like CO₂ and CH₄ adsorptions compared in previous subsections, we find that N₂ adsorption behavior from simulation data is similar to experimental study.

Fig. 9

a N₂ adsorption property of ZnDABCO reported from Ref. [9] and our results at 294 K. b N₂ adsorption equilibrium isotherm from Ref. [8, 10] and our data at 298 K and range of 1–100 kPa. c Adsorption simulation of N₂ from Ref. [12] experimental and simulation results, Ref. [5] simulation data and our data at range of 1,000–25,000 kPa

In this paper, Fig. 9b shows interesting results. In Sect. 3.1 related to reports of gas adsorption in our work, we can see that N₂ adsorptions, at 300 K and range 1,000–21,000 kPa, are 5–30 mol/kg in (10,10), (20,20) and (40,40) SWSiCNTs (please see Figs. 1, 2, 3). But Fig. 9b exhibits our simulation of N₂ adsorption at 298 K at pressure range 0–12 kPa (very low pressure). Figures 1, 2 and 3 show that at pressure range 1,000–21,000 kPa, N₂ adsorption in SWSiCNTs is in the order of: (10,10) < (20,20) < (40,40), but at 0–12 kPa (according Fig. 9b), we can observe that N₂ adsorption in SWSiCNTs is in the order of: (10,10) > (20,20). Also, Yi et al. [10] have been studied as experimental work on activated carbon shows N₂ adsorption similar to (20,20) SWSiCNT at 298 K. This observation of N₂ adsorption can see in Fig. 9c which this latter figure also, exhibits that: (1) N₂ adsorption in SWSiCNTs are order as: (10,10) > (20,20) under 100 kPa similar to results of Fig. 8b, (2) our GCMC simulation data of N₂ adsorption in (20,20) SWSiCNT is more than Cu-TDPAT (Ref. [8]) and (3) experimental and simulation data of N₂ adsorption in Ref. [8] investigation is very close to together.

Then, from Fig. 9a, b and c we can understand that N₂ adsorption at least under 100 kPa (about 1 atm) is differ than over than 1,000 kPa (10 atm) in SWSiCNTs so that, at low pressure (about 1 atm and ambient pressure) SWSiCNT with small diameter is more N₂ adsorbable than other SWSiCNT, contrariwise, SWSiCNT with large diameter is more N₂ adsorbable at pressure over 1,000 kPa (10 atm). Also, N₂ can adsorb lower than other gases because of their weakness interaction potentials with atoms of adsorbent (here SWSiCNT) and this lower adsorptivity is proved even with experimental observations.

Conclusion

This work deals with CO₂, CH₄ and N₂ physisorption in SWSiCNTs by using GCMC approach. The adsorption isotherms of gases are prepared, and interpretation of these quantities illustrates that CO₂ has more adsorption in the pure state at ambient temperate and high pressure, and with increasing the diameter of nanotube, we observe, the increase of gas adsorption, while at very low pressure this result is inverted. In addition, we compare our simulation results with other experimental and theoretical investigations, which consider these gas adsorptions on different nanoporous materials. The observations show that SWSiCNTs are the best candidates for gas adsorption.