Plasma functionalization mechanism to modify isocyanate groups on multiwalled carbon nanotubes

Carbon nanotubes (CNTs) have some great advantages for future applications because the nanotubes have excellent material properties, especially electrical and mechanical properties.¹ A space elevator is one of the fascinating applications with CNT mechanical properties.² The elevator is a building structure that connects the surface of the Earth to a counterweight located at 144 000 km height with a strong cable.^3,4 A station is located in between, providing us with some great opportunities, for example, an inexpensive launch of a communication satellite. Such a huge structure requires a cable with a low material density but enough tension to resist centrifugal force generated due to the Earth’s rotation. The tension must be more than 60 GPa,² while CNTs have a tension of ∼100 GPa,^5,6 which sounds exciting for realizing the structure.

Although there have been numerous theoretical reports about the elevator, CNT engineering remains challenging because CNTs have the material property of chemical inertness. One of the possible solutions to overcome inertness is the functionalization on the CNT surface,^7–10 which allows CNTs to increase chemical flexibility, but the technique has yet to reach mass production. A wet process with a strong acid is one of the techniques for CNT functionalization. This technique first allows one to modify carboxyl groups (–COOH) on the CNT surface,^11,12 and then, the groups are replaced with the desirable modified groups, such as amide groups¹³ and phenyl groups.¹⁴ The replaced groups enhance the CNT compatibility with other materials; for example, the functionalization improves CNT solvency in water.¹⁵ The wet process usually requires long hours to functionalize CNTs, so the functionalization with low-temperature plasma is sometimes considered. As already known, this kind of plasma contains a lot of energetic species in the gas phase, but electrons are the main species that are kinetically energized in the plasma. In such a plasma, the modification mostly occurs at the surface of CNTs while keeping a basic carbon structure.^16–18 Our previous works also showed amino group (–NH₂) functionalization on CNTs with hydrogen/nitrogen gas-mixture plasma¹⁹ and isocyanate (–NCO) group functionalization with nitrogen/oxygen or nitrogen/carbon dioxide plasma.^20–22 

The modification of functional groups enlarges the opportunities for the CNT application. For example, we have reported that the plasma-modified NCO group is useful as a composite filler to enhance the mechanical properties of the composite polyurethane in wear resistance.^23,24 One of these reports showed that the plasma-functionalized NCO group increased interfacial interaction between a polymer and the filler, suggesting that it can be a good idea to control the number of functionalization to make further mechanical improvements. Another recent report of us showed that the functionalization ratio becomes maximum when plasma has the densities of nitrogen species and oxygen species nearly equal in the gas phase; these species became the same densities when the plasma-forming gas was 90% nitrogen gas content in the nitrogen-oxygen gas mixture. Furthermore, our Raman spectroscopy measurements supported any functionalization (or maybe structural change) on CNTs under this particular plasma condition.²⁵ However, we have noticed that other species in plasma could contribute to the formation of NCO groups on CNTs, such as carbon monoxide (CO) and carbon nitride (CN). The gas species could be important because of its reactive characteristic and chemical compound to form NCO groups, although the contribution of carbon monoxide species to form NCO is not well understood yet. This report is a further analysis to find the relationship between the NCO functionalization on the CNT surface and the gas species in plasma, particularly, the effect of carbon monoxides.

Figure 1 shows the apparatus to make plasma functionalization onto multiwalled carbon nanotubes (MWCNTs; JEIO Co., Ltd.; 95% purity, 4% ash, CM-951; 10–30 nm diameter). This reactor generates capacitively coupled plasma (CCP) with a radio frequency of 13.56 MHz. A general plasma treatment was performed with 15 mg MWCNTs placed in a quartz Petri dish. The plasma was generated with the gas mixture of the following gases: nitrogen, oxygen, argon, and carbon dioxide. The total gas pressure was always 225 Pa under any plasma conditions in this article. The partial pressure of argon was 10 Pa (5% of the total pressure), and the partial pressures of the other gases were dependent on the measurement condition. The base pressure of this reactor is ∼15 Pa, and the residual gas is from ambient air. The residual gas contains water vapor, but the effect of water vapor from ambient air is not taken into account in this article. This is because the partial pressure of the vapor is small enough compared with the total pressure. The operating power for the plasma generation was constant at ∼50 W.

This article focuses on the relationship between the gas species in plasma and the NCO-modified groups functionalized on CNTs. The densities of the main species in the plasma were found from the optical emissions according to the plasma actinometry method.^26–28 Our previous article showed that nitrogen/carbon dioxide plasma (N₂/CO₂ plasma) and nitrogen/oxygen plasma (N₂/O₂ plasma) functionalize the isocyanate group (NCO groups). This article measured the density of species of atomic nitrogen (N), atomic oxygen (O), and CO molecules from the optical emission of plasma for further analysis between the gas phase species and the NCO functionalization. The measurements were performed with the use of two optical emission spectrometers (ASEQ, HR1; 0.4 nm resolution with 50 μm slit): the first spectrometer covers the emission of wavelength from 250 to 550 nm, and the second spectrometer covers the emission of wavelength from 650 to 950 nm.

We collected the optical emission at 30 mm from the plasma with a single lens connecting the optical fiber (2 × 2 400 μm core 50:50 multimode coupler; Thorlabs, TH400R5S2B). This specific fiber splits the incoming light emission into two fibers at the same ratio and delivers the two light emissions to the two spectrometers. This configuration was necessary to maintain a good optical emission spectrum measurement resolution through focusing wavelengths.

Figure 2 shows the typical optical emission spectra when we generated plasma with the gas mixture of N₂ and CO₂ from (a) 250 to 550 and (b) 730 to 850 nm range. Note that the partial pressures of N₂ and CO₂ were the same, and the plasma had 10 Pa of argon gas (<5% of the total pressure) as a tracer gas for the plasma actinometry. The use of two spectrometers is due to the purpose of deriving the N species density, O species density, and CO molecule density from the same plasma. In this article, these densities are all volume-averaged values. In addition, we assumed that the detector sensitivity is not too different in wavelength to wavelength in the single spectrometer. Furthermore, we kept the density derivation from the optical emission intensities measured within the same spectrometer.

Figure 3 shows the density evolutions of the main gas species as a function of nitrogen content ratio of the mixture gas. Here, we focused on the densities of the species, atomic nitrogen (N), atomic oxygen (O), and carbon monoxide (CO). The densities were calculated from the optical measurement with the plasma actinometry. Remember that the density derived here is that of the neutral species; the density of ions should be much lower according to the charge-neutrality principle in plasma.³¹ As seen in Figs. 3(a) and 3(b), the N species density had only a slight change under both N₂/CO₂ and N₂/O₂ plasma conditions at any nitrogen content ratio. This semiconstant trend was acceptable because the N specie density is a result of dissociation from the parent species with nitrogen atoms, such as from molecular nitrogen. The dissociation mainly occurs due to the collisions among the species in the plasma. This result was under the condition that the operating power was the same at any nitrogen content ratio, so the condition of electrons, such as their density and temperature, and the condition of other energetic species, such as metastable species, could be different at a ratio to ratio in the gas phase. Therefore, the measurement condition could have resulted in this semiconstant trend, as seen in the figure, although the molecular nitrogen ratio increased in the gas phase.

In contrast to the N species, the O specie density had a decreasing trend under both plasma conditions as the partial pressure of molecular nitrogen increased. In addition, the O specie density was always greater than the N specie density within the range between 0% and 90% of the nitrogen content ratio. Then, the O specie density, finally, became less than that of the N species above 90% of the nitrogen content. This means that the densities of N species and O species in the gas phase became almost the same when the nitrogen content became ∼90%. Furthermore, the two plasmas had the densities of the N and O species at ∼2–3 × 10²⁰ m⁻³ at the 90% nitrogen content.

Focusing on the CO density evolution, the result showed that both plasma conditions contain some CO species in the gas phase. In addition, the CO density had decreasing trends in both plasmas as a function of molecular nitrogen content. However, it should be noted that N₂/CO₂ plasma always had more CO density than N₂/O₂ plasma compared to the same molecular nitrogen content. These overall results indicate that the difference in the CO density is a good clue to analyzing the contributions of CO molecules to the NCO functionalization on the CNT surface, which will be discussed in Sec. III B.

This subsection showed the result of the functionalization ratio of the NCO groups on the plasma-treated MWCNTs. The functionalization ratio is defined as the ratio of the number of NCO groups over the carbon molecules in the CNTs, which is measured and calculated from the fluorescent intensity of the acridine yellow G (AYG, C₁₅H₁₅N₃⋅HCl), which was exposed to CNTs as an NCO probe. The details of this measurement were already reported elsewhere,^20,21 so please refer to our former works to avoid unnecessary duplication. Figure 4 shows the result from measuring the NCO functionalization ratio with fluorescent measurement.

As seen in the figure, the trend of the functionalization ratio looked almost the same for the two plasma conditions; the functionalization ratio gradually increased up to 90% nitrogen content, and then, the ratio started to decrease over 90% molecular nitrogen content. This result also showed that N₂/CO₂ plasma could slightly enhance the functionalization ratio better than N₂/O₂ plasma, consistent with our previous work reporting wear exams of the polyurethane composite with plasma-treated CNTs in the polyurethane composite comparison between N₂/CO₂ and N₂/O₂ plasmas.³³ 

One might be interested in other characterizations of the functionalized CNTs shown in this section, such as with FTIR and Raman spectroscopies. In fact, the characterizations with the spectroscopies were already reported in our previous articles,^23,25 and these measurements showed some evidence of NCO functionalization on CNT surfaces. Furthermore, our preliminary XPS analysis showed NCO functionalization on CNTs at the peaks of the binding energies on 288. 9(C1s), 400.1 (N1s), and 531.4 eV (O1s).³⁴ The analysis is currently undergoing, and it will be reported soon with details to confirm our observations in this article. This article shows the functionalization analysis with the fluorescent method because this analysis realizes the quantitative analysis of the NCO functionalization ratio on the plasma-treated CNTs. The quantization helps us to find the correlation between gas-state species in plasma and the functionalization measured with the fluorescent method.

As might already be noticed from the results shown in Secs. III A and III B, the nitrogen content of 90% is key to understanding the NCO functionalization process between the gas species in plasma and the CNT surface. The nitrogen content ratio appeared where the N species and the O species became almost the same densities, as seen in Fig. 3, and where the functionalization ratio became the maximum, as seen in Fig. 4. Furthermore, the functionalization ratio was almost the same at 90% nitrogen content for the two plasmas, ∼0.13%–0.15%, according to Fig. 4. It is true that Fig. 4 also shows that N₂/CO₂ plasma slightly increased the functionalization ratio in comparison with N₂/O₂ plasma, but the functionalization ratio evolution as a function of nitrogen content does not follow the trend of the CO molecule evolution. Furthermore, both N₂/CO₂ and N₂/O₂ plasma conditions had decreasing trends in the CO density trend; the CO density never became the maximum value at 90% of the nitrogen content. These overall results indicate that CO species in the gas phase are less contributive to functionalizing NCO groups on the MWCNTs; rather, almost the densities of both N and O species mainly determine the functionalization ratio.

One might be interested in the quenching effect; the nonradiating species could affect the NCO functionalization. Or, the other might be interested in how the wall condition affects functionalization. These effects could give some roles in CNT functionalization, and it is obvious that further diagnostics and analyses are required to have a good estimation of the effects of NCO functionalization. This article, rather, stays focused on the reports of the correlation between atomic gas species and functionalization in this article, so the investigation of these effects remains for our future work.

In addition, the isocyanate functional group has a C specie, which connects nitrogen and oxygen species with two double bonds (—N=C=O). This modified-group structure indicates that the C species in the isocyanate groups does not have direct connections to the C species in CNTs. This overall circumstance indicates that the C specie to form an NCO specie has to be supplied from atomic C specie in the gas phase. Unfortunately, the optical emission from C species was not clearly shown in our optical measurement shown in this article. Still, the optical emission observation with N₂/O₂ plasma [Fig. 3(b)] already showed CO molecules, which is an indication of the C species in the gas phase. (CO molecules could dissociate into C and O species.^35,36) The result in Figs. 3 and 4 revealed that CO molecules have a limited effect on the NCO functionalization, so CN species and/or NCO radical species have to be in the gas phase in plasma to form an NCO-modified group on the CNT surface. To consider the CN effect to form NCO groups on the surface, we observed the optical intensity of CN molecules at 388 nm, which is called the CN violet band $( B 2 Σ + − X 2 Σ + , Δ v = 0 ; 388.3 nm)$⁠.^37, Figure 5 is the evolution of the optical emission intensity of CN molecules as a function of nitrogen content. Note that the optical emission is identical enough at this observing wavelength to either the emission of the molecular nitrogen (N₂; $C 3 Π u− B 3 Π g$ system; 380.5 nm) or the emission of the ionized molecular nitrogen (⁠ $N 2 +; B 2 Σ u +− X 2 Σ g +$ system; 391.4 nm).³⁸ 

As seen in the figure, the molecular CN emission tended to increase as a function of nitrogen content in the plasma. This trend does not really follow the trend of the NCO functionalization ratio, as seen in Fig. 4. It is true that the optical emission intensity is not always proportional to the density of the observing species, but this qualitative analysis supports that CN species are less likely a main contributor to forming the NCO functional group on the CNT surface. Therefore, a series of observations and analyses found that NCO-modified groups form at the surface of CNTs by N, O, and C species individually coming to the surface. Otherwise, the NCO radicals form in the gas phase in the plasma, and then, the radicals functionalize the CNTs. Furthermore, the generation of the radicals is maximized when N and O species have similar densities and C species are in the gas.

This article reported the contributions of the gas species in low-temperature plasma for NCO functionalization on MWCNTs. For this analysis, we performed and compared two plasmas generated with an N₂/CO₂ gas mixture and an N₂/O₂ gas mixture. First of all, we made the density estimation with plasma actinometry for the two plasmas (Fig. 3). The measurement showed that N specie density increased, while O specie density decreased as a function of nitrogen gas content in the plasma. Here, the two specie’s densities became almost the same at 90% of the nitrogen content. Meanwhile, we made functionalization ratio measurements with the fluorescent method for the plasma-treated CNTs (Fig. 4). This measurement compared the CNTs treated with two plasmas as well, where the plasma was generated with the variation in the nitrogen content. The measurement showed that NCO groups became the maximum when both plasmas had 90% nitrogen content. Interestingly, the two plasmas gave almost the same functionalization ratio on CNTs even though the plasmas had significantly different CO densities. In addition, the density measurement showed CO had monotonic decreasing trends for the two plasmas, so it is straightforward to consider that the CO contribution should be limited to NCO functionalization on CNTs. At the end of this article, we analyzed the effect of CN species with the optical emission intensity at 388.3 nm (Fig. 5). The CN evolution trend did not follow the functionalization ratio trend as a function of nitrogen content, so CN species are less likely to be effective in forming NCO formation on CNTs. Therefore, this article concluded that plasma requires a similar density of N and O species in the gas phase for NCO functionalization. In addition, C species are supplied in the gas to form NCO species at the CNT surface. The NCO-modified groups possibly form in the gas phase or directly at the surface of CNTs by N, O, and C species coming from the gas in the plasma.

This research was financially supported, in part, by the research program of the Center for Low-Temperature Plasma Sciences, Nagoya University. In addition, Chubu University supported this research from the aspect of the facility. Furthermore, there are numerous efforts by some of the undergraduate students, J. Murakami, T. Muto, and X. Xun, in the Department of Electrical Engineering, Chubu University.

The authors have no conflicts to disclose.

Daisuke Ogawa: Conceptualization (lead); Data curation (lead); Formal analysis (lead); Funding acquisition (lead); Investigation (equal); Supervision (lead); Writing – original draft (lead). Keiji Nakamura: Data curation (equal); Formal analysis (equal); Project administration (equal); Resources (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.