Tackling chemical etching and its mechanisms of polyphenolic composites in various reactive low temperature plasmas

Abstract

Selective removal of the surface polymer layer from glass filled polyphenolic composites is important while using them for applications which demand good thermal and electrical insulation properties. Inductively coupled low temperature plasma discharge was verified to facilitate chemical etching to selectively remove the surface polymer from the glass filled polyphenolic composite. The etching rate inside the plasma was compared with commercially important feeding gases including H₂, N₂, O₂ and NH₃ as a function of discharge power. The protruding surface polymer was removed in volatile molecular fragments, whereas the glass additives stayed unaffected. The etching rate is shown to increase as a function of applied power inside all used gas discharges. Irrespective of the applied power, O₂ plasma displayed the highest etching rate. Furthermore, by combining optical emission spectroscopy (OES) and X-ray photoelectron spectroscopy (XPS) analyses, the work is extended to mechanistically present the possible chemical pathways which lead to creation of various functional groups on the surface during the plasma–surface interactions. This gives a new outlook towards the future applications of low temperature plasma processing in composite industry for surface modification and selective etching.

---

Introduction

Plasma is denoted as a complex mixture of high energy ions, radicals, excited atoms and electrons which was successfully implemented for fabrication of high performing micro/nanodevices including nanofibres, graphene sheets, thin coatings, quantum dots, etc.^1–5 Plasma processing is always associated with complex physical and chemical interactions in gaseous phase as well as at the plasma–surface interface. However, one of the major challenges is the understanding of chemical interactions and reaction pathways of different energetic plasma particles on the designated surfaces. This is extremely important when plasma is used for surface modification of delicate materials including polymeric substrates which are important for material fabrication in chemistry, chemical engineering, physics, biology and even medicine.^6–8 Surface modifications including functionalization and etching are preferred with low energy particles due to the high heat sensitivity of polymeric substrates. For treating such heat sensitive materials, weakly-ionized highly-dissociated plasmas in thermal non-equilibrium are more suitable, since they are able to restrict the thermal damages caused by extensive ion bombardment on the surface.⁹

The solidity of various materials towards the plasma interacting particles has strong correlation with the type and energy of the bonds in the molecular or solid material formation. The typical example for this is the etching of graphene sheets with H₂ plasma. Due to the higher reactivity of the C–C bonds at the edge, functionalization and followed carbon removal is faster from the edge of the graphene sheet compared to the basal planes.^10,11 These differences observed in the materials etching rates with respect to their chemical stability or structures are relevant for the polymeric substrates as well. Various polymers are etched at different rates depending upon their physical and chemical properties including the bond strength, crystallinity and the ability to cross link on the surface upon irradiation, etc.^12,13 This is the general outset for preferential plasma removal of one material over another.

The etching rates inside the plasma are typically higher for macromolecular carbon chains comparing to the inorganic materials (minerals, glass, etc.) or the elemental forms of carbon (graphite, carbon nanotubes, etc.). Various polymer materials are usually composited with these organic or inorganic materials in order to improve their material performance in various applications.^14–16 In many cases, preferential removal of the polymer matrix or the embedded fillers is utilized for upgrading the properties of such fabricated composites. For better accuracy and precision during such preferential plasma etching, clear understanding of the fundamental chemistry involved is essential. However, the field of plasma science is notably application driven which leads to lack of fundamental understanding of the elemental mechanisms involved in the process. More particularly, when it comes to plasma surface interactions, the basic chemical reactions involved are often on the side track of investigations. One of the main challenges for the plasma community is to control the efficiency of the etching process at atomic or molecular level, which requires a detailed understanding of the processes. Some of the recent articles tried to explore the chemical interactions between the plasma particles and polymeric substrates, however the research was limited to atmospheric pressure plasma treatments or limited number of feeding gases.¹⁷

The fabrication of various composite materials and their suitable surface modifications are essential to devise suitable materials for delicate applications such as improved adhesion, surface micro/nanostructuring, improvement of sensing properties of various semi-conducting materials, etc.^1,18–21 Glass filled polyphenolic composites which are important for engineering applications in aerospace and insulating devices.²² However, carbonisation of the very thin surface polymer layer when they are in contact with the electrical wires found to restrict their performance level in electrical insulation devices. The standard comparative tracking index test is the measure prescribed by International Electrotechnical Society safety regulations to assess the performance level of insulating materials. Conventional method to improve the tracking index values of polymer composites is associated with the incorporation of non-charring fillers such as melamine salts, which is not cost effective. In addition to that, even after the addition of such fillers, the thin surface polymer can result in test failure. An alternative to this, is to use cheap fillers such as glass fibres and remove the superficial polymer which confronts the test performance.²⁰ Compared to the conventional methods, plasma selective etching has the advantageous of fast processing, low production cost and easy scale up to industrial scale production. Herein we present a systematic study of plasma chemistry which influences the etching rate and its selectivity for the example case of glass filled polyphenolic resin. Moreover, we compare the etching rates of the composite resin for different plasma feeding gases including H₂, N₂, O₂ and NH₃. The fundamental chemical interactions and mechanisms during the treatment process are unravelled by combining various plasma diagnostics and surface analytical techniques.

Materials and methods

Plasma was generated with an inductively coupled RF discharge using AE Cesar plasma generator at a frequency of 13.56 MHz inside a borosilicate glass tube of 80 cm length (Fig. 1).²⁰ Before that the chamber was evacuated to a base pressure (<1 Pa) using a two-stage rotary pump, and then commercially available gases H₂, N₂, O₂ and NH₃ (Messer) were leaked individually into the system through mass-flow controller (Aera Transformer 785X by Hitachi Metals). The gas flows were controlled to keep the pressure inside the chamber at a constant value of 40 Pa. The plasma was generated at 3 different discharge input powers 100 W, 150 W and 250 W. The inductive coil and the reactor chamber were cooled with chilled water and forced air flow, respectively. The testing samples – commercially bought polyphenolic resins (phenol–formaldehyde novolac resin) filled with 60–80% of glass fibres (diameter < 8 μm and length > 10 μm) and glass spheres (diameter < 8 μm) were prepared, according to the manufacturer: Vyncolit Sumitomo Bakelite Co. Ltd., by injection moulding process. During the moulding process, the temperature inside the tool cavity was kept at 160–170 °C. The samples thus obtained were cured in three stages: (i) 2 h at 160 °C, (ii) 4 h at 180 °C and (iii) 4 h at 210 °C. For the treatments, the samples were cut to dimensions 1 cm × 1 cm × 5 mm (lbh) and placed on a glass substrate inside the plasma chamber. The etch rates were calculated from the mass lost after the plasma exposure at least with 3 samples for each treatment conditions. The treated samples were kept under inert atmosphere to avoid the possibility of aging effects, prior to various surface analysis. During the processing, the qualitative analysis of the plasma was carried out by Avantes AvaSpec-3648 optical emission spectrometer. The sample surface temperature was measured with an infrared pyrometer (Raytek Raynger MX4+) through the CaF₂ window with linear transmissivity in the IR region of interest. The surface characterization was done after treatments with X-ray photoelectron spectroscopy (XPS) using the PHI-TFA XPS instrument (Physical Electronics Inc). The sample morphology was imaged with a scanning electron microscope (SEM) FE-SEM Jeol JSM-7600F at electron beam energy of 15 kV.

Fig. 1 Schematic of the inductively coupled low pressure RF plasma reactor chamber used for the treatment process.

Results and discussion

The etching rates of the polymer–matrix composite were measured in terms of the mass lost after exposing it to various gas plasmas and is presented in Fig. 2. The etching was done selectively for polymer over inert glass fillers. The comparison of the etching rates revealed that the etching rates were maximal for O₂ plasma at all the 3 discharge powers used for the treatment process. Additionally, the etching rates showed an increasing trend with the increase in the discharge power. This was evidently due to the higher degree of dissociation of the feeding gas with the increase in applied discharge power.^23,24 While comparing the etching rates at lower discharge power (100 W), the etching rates inside H₂, N₂ and NH₃ were very close to each other, whereas the etching rate inside the O₂ plasma was comparatively higher. At 100 W discharge power, the observed etching inside the O₂ plasma was around 90% higher than that of H₂ plasma after 6 minutes of exposure. When the discharge power was increased to 250 W, this difference increased up to 3.5 times. This clearly indicated that the etching efficiency as a function of discharge power was more pronounced in the case of O₂ plasma. The plausible explanation for the lower etching rate inside other gas plasmas could be explained in terms of difference in the reactivity of various atomic and molecular species produced during discharge, which is described in the latter section of this article.^25,26

Fig. 2 Etching rates in various gas plasmas generated at pressure 40 Pa and different discharge powers (A) at 250 W (B) at 150 W and (C) at 100 W.

A sample floating inside the plasma immediately acquires a negative potential, and the positive ions could be accelerated onto the surface. One can argue the possibility of filler detachment from the surface due to ion bombardment during the plasma exposure, which could cause errors during the mass loss from the surface. However in our case, the ions are only weakly accelerated inside the plasma sheath, whereas the mass for the fillers is relatively higher. For this reason, there is low or no possibility of filler detachment by means of kinetic energy of ions colliding with the surface even at higher discharge powers. In addition, the reproducibility of plasma etching rates was confirmed by repeated measurements on multiple samples. While treating polymer samples, surface roughness measurements could be one suitable candidate for assessing the etching rate (see ESI†). In our case, this was not possible due to the variable filler size distribution.

As seen from the SEM images, the surface of the non-treated composite was completely covered with the polymer and the fillers were well embedded into the matrix (Fig. 3A). However, after exposing it to various gas plasmas, the polymer was selectively removed from the surface and the fillers got exposed. The SEM images of the samples treated with these plasmas at 250 W for 3 minutes were compared in Fig. 3B–E. This revealed that the surface of the samples treated in O₂ and NH₃ plasma (Fig. 3D and E) was predominantly polymer residue free. On the other hand, the polymer residue partially covered the embedded fillers on the surface of the samples treated with N₂ and H₂ plasma (Fig. 3B and C), respectively.

Fig. 3 SEM images of top surfaces of non-treated and various plasma treated samples at 250 W for 3 minutes (A) non-treated sample (B) treated in N₂ plasma (C) treated in H₂ plasma (D) treated in O₂ plasma and (E) treated in NH₃ plasma.

Unravelling the origin of etching, the energies of various bond types in the polymer chain play one of the most important roles which determinate the bond breaking mechanism and consequent removal. The cross linked polyphenol matrix is formed after the condensation reaction between the phenols and formaldehyde monomers at ortho or para sites. Thus the cross linked structure consist of phenolic rings connected to each other at ortho or para sites through benzylic carbon atoms. The bond energy of the aromatic C–C bond is roughly estimated to a value of 4.8 eV. The other possible bond breakings are at the benzylic C–C (3.9 eV) or benzylic C–H (3.8 eV) positions. Compared to the benzylic C–H, the aromatic C–H bond has much higher bond dissociation energy of ∼4.2 eV.²⁷

During the plasma–surface interactions, the primary bond cleavage is always expected at the position with the lowest bond energy. The extra stability is provided by the energy of aromatization, which retards the possibility of ring rupture compared to the stated competing reactions. Thus, one should expect the functionalization to be initiated at the benzylic position. Another major reaction that could be possible is also the quenching of the neutral atomic species by the functionalization of aromatic rings.²⁸ Namely, the incoming atomic species, generated by the dissociation of the feeding gas in the plasma, react with the aromatic ring in such a way that the proton from the aromatic ring is transferred to the incoming heteroatom causing a functionalization of aromatic ring. During such neutral atom quenching, the incoming plasma species preferably attack the ortho position due to their high electrophilic nature.²⁹ These two major initiation reactions between the incoming atom and the polymer substrate are represented in Scheme 1. However, further proceedings of the etching process is hard to explain due to the complexity of the competing reactions.

Scheme 1 (a) Plasma neutral atom substitutes the hydrogen at the benzylic position (b) neutral atom quenching by proton abstraction from the aromatic ring.

To gain more understanding on the chemical composition and the modifications of surface chemistry during plasma treatments, the XPS analyses were carried out. XPS survey spectrum (0–800 eV) revealed the elemental composition on the surface before and after exposure to different gas plasmas generated at various discharge powers. The case of 250 W treatment for 6 minutes is presented in Fig. 4. The non-treated sample surface contained mostly carbon (89.4%) and very little of oxygen (9.1%). After the exposure of polymer–matrix composites, the surface carbon content drastically decreased for all tested gas plasmas, whereas the concentration of the oxygen significantly increased. Additionally, the Si content was also increased after the plasma treatment. The Si signal originated from the exposed glass fillers, leading us to the confirmation that fillers are exposed and polymer is removed. After the polymer removal, notable amount of Na and traces of other minerals including Zn and Mg are also observed on the surface as additives to fillers. All these observations pointed out that the plasma treatment reduced the carbon content on the surface by the selective removal of the surface polymer. Another interesting observation is that the nitrogen content was slightly higher for the samples exposed to N₂ plasma rather than the one for NH₃ plasma.

Fig. 4 XPS survey spectrum with major elemental compositions for the non-treated sample and the ones treated in various gas plasmas for 6 minutes at an applied power of 250 W.

To compare the surface chemistry before and after the treatments, the deconvolution of C 1s peak was essential. The high resolution (HR) C 1s spectrum of the non-treated sample was fitted with 3 peaks at 284.8 (C1), 285.9 (C2) and 288.6 eV (C3) respectively (Fig. 5). The peak at lower binding energy was assigned to C–C/C–H bond types. Thus, the intensity of this peak C1 is connected to the surface polymer content. The peak C2 could be assigned to the C–OH bond originated from the phenolic functionality on the resin matrix, whereas the peak C3 at binding energy around 288.6 eV was assigned to the O–C=O functional groups. However, the concentration of O–C=O bonds is very low (∼2%), and probably originate from the degraded polymer during the high temperature curing process. Namely, the reports of such effects of curing temperature on the pyrolytic degradation of phenolic resins are found in literature and support this argument.³⁰

Fig. 5 Deconvolution of XPS HR C 1s spectrum of the non-treated polyphenolic sample.

When the samples are treated with different gas plasmas (like presented case for 250 W and 6 minutes) and compared, the changes in the surface chemistry arise after plasma treatments. The comparison of various plasma treatments revealed that the notable change was in the intensity of the C–C/C–H bond types (Fig. 6).

Fig. 6 XPS HR C 1s spectrum deconvolution of the samples treated at 250 W for 6 minutes (A) in N₂ plasma (B) in H₂ plasma (C) in O₂ plasma (D) in NH₃ plasma.

This was evident from the decrease in the intensity of the peak C1 at 284.7 eV. The decrease was more pronounced in the case of NH₃ and O₂ rather than of N₂ or H₂. In the case of NH₃ or N₂ plasma, C2 can be a grouping of C–N and C–O functional groups due to their comparable binding energy. The peak C3 in the case of O₂ plasma was assigned to the carboxylic functionalities. Moreover, the peak C3 (below 289 eV) was assigned as the amide linkage formed as a result of post oxidation after exposing to the external atmosphere.¹⁷ When the sample was exposed to H₂ plasma, such post oxidation gave rise to lower concentrations of carbon–oxygen bonds on the surface.^31,32

Other origins of oxygen functionalities could be from the reaction between the polymer and the oxygen from the desorbed water molecules or water formed as a by-product of polymer etching.

To explore more information about the plasma surface interactions, plasma characterization is needed. Optical emission spectroscopy (OES) analysis is a useful characterization technique which enables the detection of various excited species in plasma. Monitoring of the OES spectra of N₂ plasma in the beginning of the 6^th minute for various discharge powers revealed that the intensity of the bands increases consecutively with the increase in the discharge power. N₂ plasma emission bands mainly composed of N₂ first and second positive spectrum (N₂-FPS and N₂-SPS) (Fig. 7).

Fig. 7 OES measurements during the N₂ plasma treatment of the composite at the 6^th minute for various discharge powers.

The major lines gathered in the range 300 to 400 nm constitutes of N₂-SPS (C³Π_u → B³Π_g).³³ The N₂-SPS band at 337 nm: N₂ (C₃Π_u) is principally responsible for N₂-SPS, which is produced as a binary collision product of N₂ (A₃Σ_u⁺).³⁴ The existence of N₂-SPS can thus serve as the indication for N₂ (A₃Σ_u⁺), which has an energy state of 6.2 eV.³⁵ These species can very easily break various bonds present in the polymer chain. However, due to the ease of reactivity, primarily the bond breaking is expected at the benzylic position. Furthermore, the peak at 390 nm is assigned as a combination of N₂⁺, CN and CH fragments, where CN and CH species originate after the plasma–polymer interactions.³⁶ Other possible etch product is NH species which are typically observed at around 336 nm in their excited state. However, this peak is not clearly visible as it can be merged with broad N₂-SPS band at 337 nm. Another possible species inside the reactor chamber is atomic N species, which can separately appear as peaks at 745 nm and 869 nm. However in our case these emission lines are probably embedded inside the molecular emission bands of N₂-FPS and not seen separately due to lower resolution of recorded spectrum.^37,38 Moreover, the intensity of emission line from excited Na species (589 nm) is predominantly seen only when the discharge power is 250 W. Such species usually originate from the organic salts added to improve the processability of the polymer material.³⁹ Interestingly, this can serve as another indication of enhanced etching rate at higher discharge powers.

The represented OES spectra during the NH₃ plasma treatment revealed that the major peaks are the atomic lines corresponding to hydrogen species includes Hα, Hβ and Hγ at respective wavelengths 686, 486 and 434 nm (Fig. 8). The peaks corresponding to NH band at 336 nm (A³Π → X³Σ⁻) were also present at relatively high intensity. Additionally, the peaks corresponding to nitrogen molecular species in the N₂-SPS are present, whereas NH₂ species were hard to detect due to their faster quenching and lowered detection sensitivity, which is 200 times lower than that of the NH species.¹⁷

Fig. 8 OES measurements recorded during the NH₃ plasma treatment of the composite material at the beginning of the 6^th minute for various discharge powers.

The major interacting species inside NH₃ plasma are NH_x and H atoms, whereas in the case of N₂ plasma the excited N₂ and N₂⁺.^40,41 If NH₃ plasma is used, then there is a presence of H atoms. Moreover, it is well known that NH₃ plasma can induce simultaneous nitrogen incorporation and reduction. On the other hand, the major interacting species in N₂ plasma are supposed to be N₂ (A₃Σ_u⁺) and potentially dissociated N atoms, which induce polymer scissions. Due to the lower stability of the C–H bond at the benzylic position, primary attack of the plasma particles will probably occur at this position. The possible amination mechanism inside NH₃ plasma is represented in Scheme 2A. Either by the direct reaction of NH₂ radical at the benzylic position to form C–NH₂ bonds, releasing atomic H or by the reaction of NH₂ species with the dangling bonds generated by the influence of H atoms. Additionally, amination at the aromatic ring is also possible by means of NH₂ quenching as explained previously in Scheme 1. Another one of the possible mechanistic routes for amination and nitrile bond formation in N₂ plasma is speculated in Scheme 2B, where diatomic or atomic nitrogen species are involved. The UV emission from the excited plasma species are capable of inducing homolytic bond fissions on the polymer surface. However, the exact possible mechanism is hard to predict since plasma is composed of many species which can directly or synergistically influence chemical pathways. After the primary functionalization step, the process can go on through many complicated and competitive reactions. Here it is impossible also to neglect the possibility of interactions from complex radicals arisen from chain fragmentation.

Scheme 2 (A) Amination mechanism in NH₃ plasma and (B) formation of nitrile and amine functionalities by atomic and molecular nitrogen species.

Among various gas plasmas used, H₂ plasma had the lowest etching rate. The etching inside H₂ plasma is expected to proceed through C–C and C–H bond breakings through hydrogenation reaction. The OES spectrum presented in Fig. 9 revealed that the plasma contained atomic hydrogen observed from the emission lines at 434 nm (Hγ), 486 nm (Hβ) and 656 nm (Hα).^42,43 The molecular bands corresponding to H₂ Fulcher and CH were present in the range 590–640 nm and 387 nm, respectively. The intensity of emission from the atomic Na at 589 nm was pronounced only for 250 W. This peak was much weaker in the case of lower discharge powers. In addition to this, the intensities of the etch products originated from the fragmentation of the macromolecules (e.g. CH peak 387 nm, C2 Swan peak at 413 nm) which were the highest at 250 W.⁴⁴

Fig. 9 OES measurements recorded during the sample treatment inside H₂ plasma measured at the 6^th minute.

Going further, O₂ plasma was the most active etcher and displayed the highest etching rates. From the OES measurements it is observed that spectra are dominated with excited atomic oxygen as major etcher at 777 nm O (3p⁵P → 3s⁵S) and 844 nm O (3p³P → 3s³S) (Fig. 10). Other atomic line O (3p⁵P–4d⁵D⁰) was observed at 615 nm.⁴⁵ The spectra are populated with the emission lines from atomic hydrogen and molecular OH species (309 nm), which originate from the excited species released in to the system after the polymer etching as well as the dissociation of the desorbed water molecules. The emission lines observed from the excited CO species in the CO Angstrom region clearly indicated the chemical oxidation of the polymer.^20,45 The chemical etching of polymer material with O₂ plasma is well discussed in many articles, where the primary step is the functionalization of the exposed surface. After the surface is saturated with functional groups, the polymer starts to be removed as volatile neutral molecules from the surface including CO (ΔH⁰_f = −110.5 kJ mol⁻¹), CO₂ (ΔH⁰_f = −393.5 kJ mol⁻¹) or OH (ΔH⁰_f = 39 kJ mol⁻¹) and H₂O (ΔH⁰_f = −241.8 kJ mol⁻¹).

Fig. 10 OES measurements recorded during the sample treatment inside O₂ plasma measured at the 6^th minute.

The most relevant explanation for the higher etching rate inside O₂ plasma is expected due to the higher electronegativity of atomic O species generated in the system which has an electronegativity of 3.44. Thus very lower etching rates inside H₂ plasma can be explained on the basis of lower electronegativity of atomic H species (2.20). However, this argument is in contradiction with the lower etching rates inside N₂ gas discharge, where atomic nitrogen has electronegativity of 3.04. Inside various gas discharges, the concentrations of atomic species are the main factors that determines the polymer etching rates. Atomic species chemically react with the interacting polymer surface, forming volatile molecular fragments. Unfortunately, none of the currently available experimental techniques enable the quantification of various atomic species inside all the gas discharges used. On the other hand, ionic species typically recombine on the interacting surface with a probability 1, producing a local thermal spike. More power can typically increase the electron temperature in plasma. This means that cross-section for electron excitation of certain reactions will change, meaning that different reaction pathways will create different species or their densities. However this depends on the molecular type. Generally electrons with energies in the range of 1 to 100 eV will generate a higher dissociation as well as ionization. Following this, the higher etching rates are provided at higher discharge powers, especially in the case of NH₃ and O₂ plasmas where this is ascribed to the higher extend of heterogeneous surface recombination of the generated species.

In addition, the measured surface temperature is an important parameter among various factors which affects the etching rate of the polymer. The temperature was measured with the pyrometer and the time evolution for all types of plasma and all applied powers is presented in Fig. 11.

Fig. 11 Measured temperature on the sample surface during the treatment in various gas plasmas (A) at 250 W (B) at 150 W and (C) at 100 W.

During the plasma treatment, the surface temperature increases due to multiple reasons including exothermic carbon oxidation, neutral atom recombination on the surface, ion collision and relaxation of various plasma species. The surface heating rate as a result of plasma–surface interactions were the lowest for 100 W and highest for 250 W. This power–temperature dependence is explained by increased density of reactive plasma particles which act as etchers with increasing power. As these particles react with surface, the heating of surface becomes more extensive. An interesting feature is the highest temperature achieved is for the case of NH₃ plasma. Since this doesn't correspond fully to previous statement, it suggests that other processes are occurring on the surface which lead to more extensive heating. In the case of NH₃, it seems that the heating of the sample surface is due to heterogeneous surface recombination of the NH_x species. These groups accommodate (adsorb) on the sample surface where they recombine back into ammonia. On other hand, small deviations in O₂ plasma can be explained with exothermic polymer oxidation prior polymer removal which can similarly contribute to the surface heating. However, after the polymer removal the major surface heating is due to radical and ion recombination. When the polymer component is removed due to chemical etching, the recombination occurs predominantly on glass fillers. The fact that the sample temperature is highest for the case of ammonia plasma indicates the recombination of NH_x occurs at somehow higher probability than recombination of N, O and H atoms to parent molecules. The effect is particularly pronounced at the lowest power of 100 W where heating by heterogeneous surface recombination usually prevails over other mechanisms such as accommodation of positively charged ions and neutralization of charged particles. For the future, we intend to investigate such heating effects from the surface recombination by combining Langmuir and catalytic probe measurements.

Conclusions

The etching of polyphenolic–matrix composite in various gas plasmas at different discharge conditions is explored and differences are discussed. Among the various gases used, the etching rate was the highest with O₂ plasma. The increased etching rate at higher discharge powers was attributed to the increased density of excited and dissociated particles. The excited species inside various gas plasmas were monitored by OES analysis, which revealed the presence of dissociated atoms, other excited particles and molecules as well as ions. Various particles react with surfaces at different rates due to their potential energies. Impinging ionic species mostly recombine upon interaction with the surface and directly contribute to surface heating. On the other hand, metastables or neutral atoms can recombine on the surface with lower probabilities which are highly dependent on interacting material. The results of these interactions are then predominant formation of chemical bonds with the polymer.

During the treatment process, the polymer was selectively etched away while the glass fibres remained intact as observed by SEM images. Moreover by combining XPS analyses, OES measurements and information on strength of various bonds involved in the macromolecular chain, the surface reactions were mechanistically proposed and presented. The major etchers inside O₂ plasma are dissociated O atoms which give rise to oxidized carbon (CO or CO₂) and OH or H₂O. The etching rate was the highest for O₂ plasma plausibly due to the higher reaction rates of the O reactive species with the polymer. In the case of H₂ plasma, the proposed etching is occurring through dissociated H atoms which reduce the C–C and C–H bonds to volatile hydrocarbons. When using N₂ plasma, the major etchers included excited ionic and molecular species. Whilst in the case of NH₃ plasma, dissociation of the feeding gas provided atomic H and NH species as observed by OES measurements. However, there are other possible species including NH₂. The existence of NH₂ is speculated due to the presence of NH and Hγ since the major formation of NH₂ species occurs through the reaction between NH₂ and H in NH₃ discharges (NH₂ + H → NH + H₂). Due to various excitation and quenching mechanisms, all the species involved cannot be detected with emission spectroscopy. For better validation of our proposed results, more detailed analysis of plasma with other sophisticated techniques including laser induced fluorescence and absorption spectroscopic studies and probe measurements (Langmuir and catalytic) should be combined with molecular modelling.

As a result of plasma–surface interactions, the surface temperature was increased during the treatment which represented an important parameter that influenced directly or synergistically the etching rate. The observed surface temperature was the highest for NH₃ plasmas, although the etching rate was lower compared to O₂ plasma. This discrepancies were explained in terms of higher probability of radical recombination which contributes to surface temperature in addition to polymer oxidation and ion collisions. Whilst it is really difficult to distinguish the individual effects of various heating mechanisms. Thus to conclude, this work will enlighten the applications of cold plasma processing in composite technology and provide more understandings on the plasma–surface interactions. To gain more understanding of these mechanisms, our future works will be more focused on plasma characterization which enables to estimate the concentration of various atomic and ionic species inside the plasma, electron temperature and ion energy flux onto the sample surface. Thus to conclude, the etching rate and etching efficiency could be controlled by choosing the proper plasma feeding gas or discharge conditions.

Acknowledgements

Authors gratefully acknowledge Slovenian Research Agency (ARRS) for financial support (Project grant L2-6767). H. P. thanks Jozef Stefan International Postgraduate School for the PhD grant from innovative scheme. Authors acknowledge also R. Zaplotnik for assistance in mass spectrometry and A. Drnovšek for roughness measurements.

References

1. H. Puliyalil and U. Cvelbar, Nanomaterials, 2016, 6, 108,  DOI:10.3390/nano6060108.
2. M. Alexander, V. Roumen, S. Koen, V. Alexander, Z. Liang, T. Gustaaf Van, V. Annick and H. Chris Van, Nanotechnology, 2008, 19, 305604.
3. C. Yang, H. Bi, D. Wan, F. Huang, X. Xie and M. Jiang, J. Mater. Chem. A, 2013, 1, 770–775.
4. A. Michelmore, D. A. Steele, J. D. Whittle, J. W. Bradley and R. D. Short, RSC Adv., 2013, 3, 13540–13557.
5. Z. J. Han, I. Levchenko, S. Kumar, M. M. A. Yajadda, S. Yick, D. H. Seo, P. J. Martin, S. Peel, Z. Kuncic and K. Ostrikov, J. Phys. D: Appl. Phys., 2011, 44, 174019.
6. S. Huet, T. Belmonte, J. M. Thiébaut, S. Bockel and H. Michel, Thin Solid Films, 2005, 475, 63–67.
7. A. Sarani, A. Y. Nikiforov, N. De Geyter, R. Morent and C. Leys, Appl. Surf. Sci., 2011, 257, 8737–8741.
8. R. V. Goreham, A. Mierczynska, L. E. Smith, R. Sedev and K. Vasilev, RSC Adv., 2013, 3, 10309–10317.
9. K. Bazaka, M. V. Jacob, W. Chrzanowski and K. Ostrikov, RSC Adv., 2015, 5, 48739–48759.
10. L. Xie, L. Jiao and H. Dai, J. Am. Chem. Soc., 2010, 132, 14751–14753.
11. A. Davydova, E. Despiau-Pujo, G. Cunge and D. B. Graves, J. Phys. D: Appl. Phys., 2015, 48, 195202.
12. D. Hegemann, H. Brunner and C. Oehr, Nucl. Instrum. Methods Phys. Res., Sect. B, 2003, 208, 281–286.
13. M. B. Olde Riekerink, J. G. A. Terlingen, G. H. M. Engbers and J. Feijen, Langmuir, 1999, 15, 4847–4856.
14. S. Allauddin, M. K. Akhil Chandran, K. K. Jena, R. Narayan and K. V. S. N. Raju, Prog. Org. Coat., 2013, 76, 1402–1412.
15. A. Palanisamy, N. V. Salim, B. L. Fox, P. Jyotishkumar, T. Pradeep and N. Hameed, RSC Adv., 2016, 6, 55792–55799.
16. F. Malara, M. Manca, L. De Marco, P. Pareo and G. Gigli, ACS Appl. Mater. Interfaces, 2011, 3, 3625–3632.
17. Y.-W. Yang, J.-Y. Wu, C.-T. Liu, G.-C. Liao, K.-Y. Cheng, M.-H. Chiang and J.-S. Wu, Plasma Processes Polym., 2015, 12, 678–690.
18. E. Elias, S. Chandran, C. N. Chandran, F. G. Souza and S. Thomas, RSC Adv., 2016, 6, 21376–21386.
19. H. Puliyalil, P. Slobodian, M. Sedlacik, R. Benlikaya, P. Riha, K. Ostrikov and U. Cvelbar, Front. Chem. Sci. Eng., 2016, 10, 265–272.
20. H. Puliyalil, U. Cvelbar, G. Filipic, A. D. Petric, R. Zaplotnik, N. Recek, M. Mozetic and S. Thomas, RSC Adv., 2015, 5, 37853–37858.
21. R. Rajan, J. Riihivuori, E. Rainosalo, M. Skrifvars and P. Järvelä, J. Reinf. Plast. Compos., 2014, 33, 1416–1429.
22. A. Pizzi and C. C. Ibeh, in Handbook of Thermoset Plastics, ed. H. Dodiuk and S. H. Goodman, William Andrew Publishing, Boston, 3rd edn, 2014, pp. 13–44,  DOI:10.1016/B978-1-4557-3107-7.00002-6.
23. N. Škoro, N. Puač, S. Lazović, U. Cvelbar, G. Kokkoris and E. Gogolides, J. Phys. D: Appl. Phys., 2013, 46, 475206.
24. C. Canal, F. Gaboriau, A. Ricard, M. Mozetic, U. Cvelbar and A. Drenik, Plasma Chem. Plasma Process., 2007, 27, 404–413.
25. A. Vesel, I. Junkar, U. Cvelbar, J. Kovac and M. Mozetic, Surf. Interface Anal., 2008, 40, 1444–1453.
26. H. Nagai, S. Takashima, M. Hiramatsu, M. Hori and T. Goto, J. Appl. Phys., 2002, 91, 2615–2621.
27. D. F. McMillen and D. M. Golden, Annu. Rev. Phys. Chem., 1982, 33, 493–532.
28. G. N. Taylor and T. M. Wolf, Polym. Eng. Sci., 1980, 20, 1087–1092.
29. E. Zadok, B. Sialom and Y. Mazur, Angew. Chem., Int. Ed. Engl., 1980, 19, 1004–1005.
30. W. M. Jackson and R. T. Conley, J. Appl. Polym. Sci., 1964, 8, 2163–2193.
31. M. Pascu, D. Debarnot, S. Durand and F. Poncin-Epaillard, in Plasma Processes and Polymers, Wiley-VCH Verlag GmbH & Co. KGaA, 2005, ch. 13, pp. 157–176,  DOI:10.1002/3527605584.
32. Y. W. Park and N. Inagaki, Polymer, 2003, 44, 1569–1575.
33. R. M. v. d. Horst, T. Verreycken, E. M. v. Veldhuizen and P. J. Bruggeman, J. Phys. D: Appl. Phys., 2012, 45, 345201.
34. L. G. Piper, J. Chem. Phys., 1988, 88, 231–239.
35. T. H. Chung, Y. W. Lee, H. M. Joh and M. A. Song, AIP Adv., 2011, 1, 032136.
36. M. Camero, F. J. Gordillo-Vázquez, J. Ortiz and C. Gómez-Aleixandre, Plasma Sources Sci. Technol., 2004, 13, 121.
37. K. J. Clay, S. P. Speakman, G. A. J. Amaratunga and S. R. P. Silva, J. Appl. Phys., 1996, 79, 7227–7233.
38. P. Anne-Marie, R. André, D. Bénédicte, O. Emmanuel, L. Juliette and G. Mihai, J. Phys. D: Appl. Phys., 2005, 38, 1905.
39. M. A. Osman, A. Atallah and U. W. Suter, Polymer, 2004, 45, 1177–1183.
40. C. López-Santos, F. Yubero, J. Cotrino and A. R. González-Elipe, ACS Appl. Mater. Interfaces, 2010, 2, 980–990.
41. J. E. Klemberg-Sapieha, O. M. Küttel, L. Martinu and M. R. Wertheimer, J. Vac. Sci. Technol., 1991, 9, 2975–2981.
42. A. Vesel, J. Kovac, R. Zaplotnik, M. Modic and M. Mozetic, Appl. Surf. Sci., 2015, 357, 1325–1332.
43. M.-J. Wang, Y.-I. Chang and F. Poncin-Epaillard, Langmuir, 2003, 19, 8325–8330.
44. A. Chingsungnoen, J. I. B. Wilson, V. Amornkitbamrung, C. Thomas and T. Burinprakhon, Plasma Sources Sci. Technol., 2007, 16, 434.
45. H. Puliyalil, G. Filipič and U. Cvelbar, Plasma Processes Polym., 2016, 13, 737–743.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra15923k

---