Plasma polymerisation of an allyl organophosphate monomer by atmospheric pressure pulsed-PECVD: insights into the growth mechanisms

Abstract

Atmospheric plasma deposition of DiEthylAllyl Phosphate (DEAP) has been performed to study the behaviour of an allylic phosphate-based monomer in a nitrogen Atmospheric Pressure-Dielectric Barrier Discharge (AP-DBD). The deposition kinetics and chemical structures of the different coatings have been studied as a function of the duty cycle and the power dissipated in the discharge. It has been highlighted that it is possible to obtain different chemistries, from organic coatings in which the monomer structure is remaining to inorganic polyphosphate-based ones. Different mechanisms of deposition have been outlined and discussed taking into account the deposition kinetics, the coatings chemistries and the monomer reactivity in the gas phase.

---

Introduction

Phosphorus-based coatings are known to be widely used as corrosion protection layers¹ as well as to promote adhesion,² to generate amphoteric surfaces³ or to provide flame-retarding effects.^4–8 They have been also reported to be an emerging resource for bioengineered nanomaterials, biomedical devices and antimicrobial activities.^9–11 The deposition of phosphorus-containing coatings can be easily achieved by wet chemical processes such as electrochemical treatment,¹² sol–gel methods,^13,14 copolymerisation and chemical grafting.^15–17 Plasma-based techniques are generally considered as environmental friendly and safe processes as it does not need any solvents. Furthermore, it could provide conformal surface coatings with tailored physical and chemical properties on sensitive and complex substrates. Surprisingly and despite these advantages, only few papers report the use of plasma techniques to cure (through plasma-induced graft-polymerisation (PIGP))^18,19 or deposit (through low-pressure plasma-enhanced chemical vapour deposition (PECVD))^20,21 a coating from an organophosphorus monomer. Recently, a new PECVD process has been developed to deposit a phosphate-based polymer at atmospheric pressure.²² This is a single-step dry deposition technique which involves the reaction of a volatile organophosphorus monomer injected into an Atmospheric Pressure-Dielectric Barrier Discharge (AP-DBD). Dielectric Barrier Discharge (DBD) generates low-temperature plasmas, which are typically required for the treatment of thermolabile substrates (e.g. polymers). As most of plasma processes, with this direct AP-PECVD system based on a DBD, tuning plasma parameters can have a great influence on plasma polymer growth and then on the structural and chemical properties of the coatings. Among them, the pulsation of the discharge can play a major role. Hence, many works have been devoted to the analysis of the influence of pulsed plasma on the polymerisation processes for different types of monomers.^23–30 This pulsed-PECVD, where the power is modulated for durations from micro- to milliseconds, decreases the total power density delivered to the plasma and short plasma pulses (μs to ms) can activate vinyl-, allyl- or acrylic monomer molecules.²³ This pulsation, which consists in chopping the electrical excitation signal, can be split up into two different periods. During the plasma-on period, the monomer is fragmented depending on the plasma power and the resulted plasma polymer film is formed by random recombinations of the generated monomer fragments. Alternatively, during the plasma-off period, in absence of ions bombardment and photon-irradiation, the free radicals entrapped at the surface region of the growing film can now directly react with the vapour of monomer and initiate the polymerisation via reactive bonds (e.g. vinyl, allyl). Those radicals hence initiate a purely chemical radical chain reaction. The plasma polymers then expected consist in more chemically-conventional products than those encountered in the continuous-wave counterpart, where more fragmentation of precursor occurs. Thus, the t_ON in pulsed-PECVD can be considered as the initiation step and t_OFF as the propagation step.²⁶ The chemical composition and structure of the plasma polymer depends on the pulse-on/pulse-off ratio, the so-called duty cycle (DC). This paper relates the influence of pulse parameters and power density on the structure and growth mechanisms of phosphorus-containing thin films prepared by the atmospheric pressure plasma polymerisation of DiEthylAllyl Phosphate (DEAP).

Experimental section

The DBD configuration used in this study has already been described elsewhere.²² The discharge was produced between two plane-parallel high voltage electrodes (15 mm × 74 mm each) covered by an alumina dielectric barrier and a moving stage as the grounded electrode. Prior to any plasma treatment, the plasma chamber was pumped to a vacuum of 20 Pa and then filled to atmospheric pressure with the desired gas mixture (nitrogen in this case). This procedure was repeated three times to ensure high gas purity of the discharge. The gas gap between the high voltage electrodes and the substrate was maintained at 1 mm. The organophosphorus precursor (DEAP, C₇H₁₅O₄P) was introduced into the discharge using a bubbler system made of a cylinder and a frit. This latter was thermostatically controlled (298 K) and the gas pipes were slightly heated (308 K) to prevent the precursor from any condensing phenomena on the inner pipe surface. The concentration of precursor is close to 10 ppm. Total flow rate was maintained at 5 standard litres per minute (slm) for all the experiments while a slight pumping maintains the atmospheric pressure inside the reactor. Unless it is explicitly specified, the deposition process was performed in dynamic mode with a constant moving stage speed of 15 mm s⁻¹ and different exposition times depending on the desired coating thickness and mass. During the deposition process, the plasma was initiated using a Corona generator from SOFTAL electronic GmbH, generating a 10 kHz sinusoidal signal. Coatings were deposited using a modulated sinusoidal electrical excitation varying from a continuous wave (CW) to a pulsed wave (PW) with different ON-time and OFF-time pulses, labelled t_ON and t_OFF, respectively. The time-OFF was set to 30 ms for all the experiments. This time is sufficient to renew the gas mixture between the electrodes before each time-ON. The power dissipated in the discharge was calculated by analysing the corresponding experimental Lissajous figure during plasma ON-time. Duty cycle value is calculated as shown in Fig. 1 and the average power is proportional to the duty cycle value (e.g. average power is equivalent to the discharge power only when the duty cycle equals 100%).

Fig. 1 Oscillogram of the applied voltage and discharge current in a pulse period.

Optical emission spectra of the discharge were recorded in the 330–390 nm wavelength range thanks to an ARC SpectraPro-2500i spectrometer equipped with a CCD detector and a grating of 1800 lines per mm blazed at 300 nm.

The substrates to be coated were either 200 μm thick aluminium foils (Eurofoil-Luxembourg) for ATR-FTIR or 275 μm thick silicon wafers (intrinsic, double side polished, Siltronix) for the other analyses. Polycarbonate (Goodfellow) was also used for the XPS analysis of the sample synthesized with a duty cycle of 3%. Substrates were first cleaned by acetone and alcohol in an ultrasonic bath before each manipulation and plasma activated during 30 seconds with a continuous wave N₂ : O₂ (95 : 5 vol%) atmospheric pressure dielectric barrier discharge just before deposition.

DEAP was obtained from Sigma-Aldrich and used without further purification (98%). N₂ (99.999%) gas was obtained from Air Liquide.

The samples were weighted using a Sartorius ME-36S microbalance. Weighting measurements of the samples were performed three times before and after each deposition to monitor the mass deposition rate.

Coatings thicknesses were evaluated by means of spectroscopic ellipsometry (AutoSE, Horiba Scientific) at an angle of incidence of 70°, in a spectral range from 440 to 1000 nm. The used ellipsometric model assumes that the samples were made of a semi-infinite silicon substrate covered by a 2 nm thick silicon dioxide layer (native oxide layer) and the plasma thin film on the top. Plasma polymerised layers were assumed to be homogeneous, non-porous, isotropic, and were simulated using a dispersion law. Roughness was supposed to be negligible.

FT-IR analyses were performed either in Attenuated Total Reflectance (ATR) mode equipped with a Ge-ATR-crystal accessory or in transmission mode on a Bruker Hyperion 2000 spectrometer and using a liquid N₂ cooled MCT detector. ATR-FTIR was used over the 4000–1300 cm⁻¹ wavenumber range and transmission mode over the 1300–700 cm⁻¹ one. Spectra were acquired by averaging 500 scans with a spectral resolution of 4 cm⁻¹. The signal of spectra acquired in transmission mode was smoothed by a Savitzky–Golay filter considering 20 points at each individual point for the smoothing routine.

X-ray photoelectron spectroscopy (XPS) measurements were carried out on a Kratos AXIS Ultra DLD instrument equipped with a hemispherical energy analyser and a monochromatic Al K_α X-ray source (hν = 1486.6 eV) operated at 150 W. The pass energy was fixed at 160 eV for survey scans and 40 eV for core level spectra, with an analysis area of 300 × 700 μm. Electron flood gun was used to compensate for the charge. All spectra were calibrated using the carbon 1s peak aliphatic contribution at 285.0 eV. No etching step was performed prior to the analyses. The XPS spectra were processed with CasaXPS software and a Shirley-type background was subtracted from the spectra. Peaks were fitted with a Gaussian/Lorentzian (70–30%) line shape. For the plasma coating which presents both an organic and inorganic phase (DC = 3%), different charging effects for each phase were observed during the XPS measurements. This phenomenon induces different peak shifting in energy and results in the broadening of the curve, which leads to incorrect peak fitting. To fix this issue, polycarbonate substrate was systematically used to analyse these plasma films by XPS. Indeed, as an insulating substrate, polycarbonate will then charge the whole coating homogeneously and enables identical shift for all curves.

Time-of-Flight Secondary Ion Mass Spectrometry (TOFSIMS) analyses were carried out by mean of a TOFSIMS.⁵ (IonTOF) equipped with a bismuth liquid metal ion gun. Bi₃⁺ clusters were used for surface spectra acquisition with a primary ion dose kept constant at 1 × 10¹¹ ions per cm² to ensure static conditions. All experiments were performed by rastering the primary ion beam across an area of 500 × 500 μm² in the high current bunched mode to provide the highest achievable mass resolving power (R = m/Δm > 5000 at m/z = 100). Multivariate analysis of TOFSIMS spectra was performed through the principal component analysis (PCA) method using the Excel add-in Multibase (numericaldynamics). Before multivariate analysis, each spectrum was normalised to total ion count number. The integration limits were checked on each peak individually to ensure that the correct peak areas were being measured on all the spectra.

The ionization of DEAP vapours was achieved by plasma-based DART ion source (Direct Analysis in Real Time, IonSense). The identifications were performed by high resolution Orbitrap analyser (Thermo, mass resolving power: 240k at m/z = 400, relative error < 1 ppm).

Results

Thin films chemical characterisations

Herein, the influence of the duty cycle on the formation of phosphate-based plasma polymers using the DEAP as monomer is studied. As discussed previously, the time-ON is of particular interest as it is a key parameter which regulates the energy dissipated in the gas, the concentration and the nature of activated species, while the time-OFF enables the reaction/propagation of the free radicals.

The AP-DBD of DEAP, irrespective of the different duty cycles and power densities studied in this paper, led to the deposition of homogeneous and smooth coatings (Fig. S1†).

To understand the plasma deposition processes involved in the discharge and on the surface, the analysis of the chemistry and the structure of the different coatings were carried out (deposition conditions are reported in Table 1). Fig. 2 represents the FT-IR spectra of the associated coatings. The intensity of the absorbance was normalised to the coating thickness. Hence, it is possible to obtain reliable relative quantitative information. The spectrum of the pure liquid monomer is included as a reference. A list of the most intense absorption bands is given in Table 2.

Table 1 Deposition conditions for the study of coatings chemistry and deposition rate

Duty cycle (%)	Pulsation tON–tOFF (ms)	Power density (W cm^−2)	Electrical excitation (kHz)	[DEAP] (ppm)	Thickness^c (nm)
a Power density during tON for all conditions.b Considering a vapour pressure of 10 Pa (298 K) empirically calculated.c Thickness of the coatings used for FT-IR analyses.
3	1–30				46
9	3–30				62
25	10–30	1.0^a	10		50
33	15–30			10^b	64
40	20–30	(±0.1)	(±0.5)		52
50	30–30				57
100	N/A				70

---

Fig. 2 FT-IR spectra of DEAP and ppDEAP at different duty cycles for the same dissipated power (during t_ON) and for the same thickness using the deposition conditions reported in Table 1 (a) in the 1300–700 cm⁻¹ range and (b) in the 4000–1300 cm⁻¹ range.

Table 2 Absorption band assignments for FT-IR spectra

Absorption band (cm^−1)		Functional group/assignment	Vibration	References
DEAP (monomer)	ppDEAP
3087	—	=C–H2	Asymmetric stretching	32
—	3500–3000	O–H/N–H	Stretching	32 and 33
2985	—	C–H3	Asymmetric stretching	32
2911	—		Symmetric stretching
∼1430	—		Bending deformation
1370	—		Symmetric bending “umbrella”
1101	—		Bending “rocking”
2935	—	C–H2	Asymmetric stretching	32
2869	—		Symmetric stretching
—	1750–1700	C=O	Carboxyl/ketone stretching	32 and 33
	1700–1650		Amide stretching
—	∼1650	C=N	Imine/oxime stretching	32 and 33
—	∼1640	O–H (adsorbed H2O)	Symmetric bending “scissor”
1649	—	C=C	Stretching	32
—	∼1560	NH2	Scissors (primary amines)	32 and 33
1480	1510–1370	Et–O	Bending deformation
1370			Bending “wagging”
—	∼1400	C–N	Amide/amine stretching
—	1500–1230	P=N	Acyclic P=N stretching	34
1275	1270–1200	P=O	P=O stretching (unassociated)	32 and 34
—	1200–1170		P=O stretching (associated)
1166	1180–1155	C–H3 (in C–O)	Bending rocking	32
—	110–930	P–N(–C)	Asymmetric stretching
—	750–680		Symmetric stretching
—	1070	P–O (in P–OH)	Symmetric stretching
—	991		Asymmetric stretching
1032	—	P–O (in P–O–C)	Asymmetric stretching
864	—		Symmetric stretching
—	∼895	P–O–P	Asymmetric stretching

---

The DEAP monomer shows different peaks in the region around 3000 cm⁻¹. The C–H stretching groups for most saturated groups (CH₃, CH₂ and CH) are detected below 3000 cm⁻¹. The C–H stretching band above 3000 cm⁻¹ corresponds to unsaturated carbon atoms. The =CH₂ asymmetric stretching of allyl groups is visible at 3087 cm⁻¹. The stretching of the C=C double bond appears at 1649 cm⁻¹, which is typical of allyl or vinyl groups. This band is often a good indicator of the presence or absence of the allyl groups. However, as the double bonds are nearly symmetrically substituted in this case, the intensity of the C=C stretching band is very weak in the precursor's spectrum, meaning that it will not be detected in the less organised plasma polymer, where peak broadening occurs due to steric effects. The region 1500–1300 cm⁻¹ can be attributed to characteristic doublets to the C–H deformation vibrations of the ethoxy groups (e.g. ν_s.(CH₃)_umbrella = 1370 cm⁻¹). The peaks appearing at 1275 cm⁻¹ and 1166 cm⁻¹ are assigned to ν(P=O)_associated and CH₃ rocking vibration in phosphate esters, respectively. The strong sharp band which appears at 1032 cm⁻¹ is assigned to ν_as.(P–O–C) and correlated to an additional band at 864 cm⁻¹ corresponding to the ν_s.(P–O–C) vibration. As displayed on Fig. 2a, the comparison between the monomer and the plasma-polymer spectra clearly shows an evolution of the absorption band of the P–O(–C) bond, which tends to shift towards higher wavenumbers along with the increase of the t_ON. Besides, the asymmetric stretching vibration band assigned to P–O–P (∼900 cm⁻¹) clearly appears for the highest duty cycles. As this band is characteristic of inorganic polyphosphate groups,^22,31,34 it is probable that the hypsochromic shift from 1032 cm⁻¹ to 1070 cm⁻¹ is related to a stronger inorganic character of the coatings and corresponds to the evolution from P–O (in P–O–C) to P–O (in P–OH). The specific evolution of ν(P–O–C) and ν_as.(P–O–P) stretching vibrations obviously shows the drop of the P–O–C concentration by increasing the average power until reaching a steady state (DC = 33%). Meanwhile, the absorption band assigned to the P–O–P vibration appears for a t_ON around 15 ms (DC = 33%) and reaches a steady state for duty cycles higher than 40%. P–O–C band is always observed showing that a pure inorganic polyphosphate is never obtained.

Fig. 2b shows that the absorption of the C–H bonds present in the monomer spectrum just below 3000 cm⁻¹, and evidenced with dashed lines, is more and more reduced/overlapped in the plasma polymer spectra with the increase of the t_ON. The apparition of the broad band between 2800 cm⁻¹ and 3500 cm⁻¹ corresponds to the stretching vibration of O–H. The shoulders observed in the same region are attributed to N–H stretching vibration. In addition, nitrogen-containing groups (such as amine/amide/imine/oxime) and carbonyl compounds can also be noticed thanks to the broad band around 1690 cm⁻¹.³³ The chemical structure seems to change around a duty cycle of 25% (disappearance of CH₃ groups, shift of P–O–C band and appearance of N–H shoulders). Hence, the pulsation of the discharge acts on the structure of the coating by promoting either its organic (low DC) or inorganic (high DC) character. The organic environment is shown to be constituted mainly by C–H, C–O, C–N and/or C=N groups and P–O–C units, whereas the inorganic character of the coating is due to cross-linked polyphosphates. The relative atomic concentrations obtained by XPS are depicted in Fig. 3 (i.e. C 1s, P 2p, N 1s and O 1s) and allow discriminating the different quantities of elements present in the coatings.

Fig. 3 Evolution of the relative atomic concentration as a function of the t_ON duration.

Here again, the relative concentration in carbon tends to drop with the increase of the t_ON duration, meanwhile the concentration of oxygen increases. For DC = 3%, the relative concentrations of each elements are closed to those of the precursor, which would indicate a high retention of the monomer structure. This is in correlation with the previous FT-IR results, confirming the influence of the pulsation of the discharge on the organic/inorganic character of the coating. Besides, it can be added that the relative concentration of nitrogen increases with the duration of time-ON, which is based on the principle that the incorporation of nitrogen in the coating increases with the initiation time of the plasma. Concerning the quantity of phosphorus, working in CW condition (i.e. DC = 100%) enables to raise its concentration probably due to the loss of the ethoxy groups (loss of carbon amount) and the formation of polyphosphates (relative increase of phosphorus atomic quantity). As the chemistry of the coatings seems to change through its organic parts, the analysis of the C 1s de-convolution (Fig. S2†) of the high resolution XPS is of great interest (Fig. 4).

Fig. 4 Evolution of the carbon functional groups concentration obtained by reconstruction of XPS C 1s core shell spectrum for different time-ON period with the same dissipated power (∼1.0 W cm⁻²).

First, it can be reported that the contribution attributed to the functional groups C–C/C–H drops as the duty cycle increases. On the contrary, ketone/amide groups contribution increases with the average power, i.e. when the fragmentation of the monomer in the discharge becomes higher. Moreover, carboxyl groups can be identified only for duty cycles starting from 40% (t_ON > 20 ms). Finally, a contribution attributed to alkoxy groups (C–O), and potentially imines/oximes groups (C=N), appears in every coatings. This may also be a recombination from highly-fragmented monomers as well as partial oxidation of the organic parts (at least at the extreme surface).

To complete the information about the chemistry of the different coatings, it is interesting to have an insight into plasma polymer structure. All samples obtained at duty cycles ranging from 3% to 100% were analysed by TOFSIMS. The negative ion mass spectra (from m/z = 0 to 200) obtained for each plasma films were then submitted to Principal Component Analysis (PCA) in order to better correlate the duty cycle and molecular fragment evolutions without arbitrary selection of peaks. PCA is a multivariate analysis method, which enables a statistical study of the dependence (covariance) between different variables. The interest of applying this method to TOFSIMS data is to reduce the dimension of the data set, which initially consists of a huge number of interrelated ions variables, into basic principal components (PCs) aiming at identifying the PCs accounting for much of the variability of the data set. Fig. 5 depicts the evolution of the scores on the Principal Component 1 (PC1) vs. the Principal Component 2 (PC2). Firstly, it should be made clear that several acquisitions were achieved at different areas of each sample for all the conditions in order to get an idea on the data dispersion. Fig. 5 indicates that the acquisitions of each condition are gathered together, indicating a good homogeneity of the coatings. When looking at the evolution of data on PC1 axis, it also shows that the most significant variations on molecular fragments are directly related to the duty cycle, i.e. to the modification of the organic/inorganic character of the deposits. As observed from Table S1,† the peaks obtained for high duty cycles are characteristic of inorganic clusters of phosphates (e.g. (PO₃H)_nOH⁻), while the peaks obtained for low duty cycles correspond to organophosphate fragments (C_xH_yPO₄⁻).

Fig. 5 Score plot obtained from PCA statistical processing of the negative mode ToF-SIMS data presenting PC1 vs. PC2.

To a lesser extent, Fig. 5 also shows variations in spectra according to PC2, which are related to signals arising from contaminations ((SiO₂)_nOH, Cl⁻…). By this Principal Component Analysis taking into account variations of hundreds of peaks, even if no clear evidence of the plasma polymerisation through the allyl group of the monomer have been highlighted throughout the investigated conditions, the organic/inorganic character of the plasma polymers can be clearly associated with the duty cycle through the t_ON's duration. Besides, the influence of the plasma dissipated power in function of the duty cycle has also been studied to go further in the understanding of the deposition mechanism (Fig. 6).

Fig. 6 Evolution of the FT-IR spectra of the ppDEAP thin films as a function of the power density and the duty cycle.

The analysis of these results indicates that the duty cycle is not the only parameter that gives the coatings its organic/inorganic character and that, within the investigated range, the discharge power during the ON-time also plays an important role. As an example, the curve corresponding to the DC of 33% and to the power density of 1.5 W cm⁻² is very close to the curve for which DC = 100% at a power of 0.5 W cm⁻². Within the investigated range, the power density has an important effect on the chemistry of the coatings for intermediate DC (e.g. 33%) but has low effects for extreme DC, where the coatings chemistry is mainly controlled by the DC. Fig. 7 summarizes the influence of the power density and duty cycle on the chemical composition of the coatings.

Fig. 7 Influence of the interconnectivity between the duty cycle and the power density on the chemistry of the coatings.

Basically, it can be concluded that increasing the power density leads to the rise of electron density and then monomer dissociation rate, and that the higher the duty cycle (t_ON), the higher the rate of collisions between monomers and electrons, also leading to an increase of monomer dissociation rate. Hence, by tuning these two parameters, it is possible to synthesize different kind of chemistries and structures. However, it is important to keep in mind that, in our case and due to technological limitations, the range of variation is not the same between the power density and the duty cycle. Indeed, the discharge power varies by a factor of 3 (between 0.5 W cm⁻² and 1.5 W cm⁻²) when the duty cycle can be varied on two orders of magnitude (between DC = 3% and DC = 100%). Furthermore, it is not possible to obtain a strong organic character using continuous-wave operation, even at the minimum discharge power (i.e. 0.5 W cm⁻²) or to perform a strong inorganic deposit at low duty cycle (i.e. 3%), even by using the maximum discharge power (i.e. 1.5 W cm⁻²).

Discussion

Firstly, the kinetic of deposition is discussed. Fig. 8 displays the mass deposition per pulse obtained in the conditions reported in Table 1 as a function of the time-ON duration. The mass deposition rate increases with the t_ON, rising from 1.3 ng per cm² per pulse with a t_ON of 1 ms (DC = 3%) up to 5.3 ng per cm² per pulse with a t_ON of 30 ms (DC = 50%). However, this increase is not linear and proportional to the time-ON. A first trend can be observed for lower time-ON (t_ON < 5 ms). In that case, the deposition rate increases faster than the t_ON, probably due to an additional radical growth mechanism that should occur during t_OFF, which enable to deposit further matter per pulse (i.e. t_ON + t_OFF). For higher duty cycles, the mass deposition rate per pulse is lower than the increase of the t_ON, meaning that the t_OFF is negligible in the growth mechanism.

Fig. 8 Evolution of mass deposition per pulse of DEAP-based thin films and of CN emission intensity, using the conditions reported in Table 1.

Then, for the range of low duty cycles, the increase of the t_ON improves the activation of the monomer by creating more free-radicals at the surface which can later react with monomers during time-OFF. For higher duty cycles, the growth rate is limited by the deposition during the time-ON period and by an etching phenomenon, which explain why the deposition rate evolution is lower than the increase of the t_ON.

The transition from the first growth mechanism (classical t_ON deposition followed by radical growth mechanism during t_OFF) to the second one (deposition only during t_ON) is clearly related to the t_ON duration, with a transition point around 4 ms for a power density of 1.0 W cm⁻². This evolution can be related to the different dissociation rate of the monomer occurring in the gas phase of the plasma. Therefore, emission of CN is monitored by optical emission spectroscopy (Fig. 8). Indeed, as no CN groups are present in the initial molecule, they result from the recombination ensuing from the fragmentation of the precursor in the nitrogen plasma as well as from surface interactions between nitrogen plasma and carbonaceous moieties. As observed in Fig. 8, emission intensity of CN increases with the t_ON, which indicates that the rate of dissociation of the precursor is favored for higher duty cycles. This evolution in fragmentation rate according to the t_ON explains the differences of chemistry between the coatings and then the evolution of the growth rate. For low duty cycles, the monomer is low-fragmented or just activated (low intensity of CN emission) and the deposited coating exhibits retention of the monomer structure. For higher duty cycles, the monomer is more and more fragmented (higher intensity of CN emission) and the thin film reveals a polyphosphate structure which is not detected at a DC of 3%. Besides, it is also necessary to consider the contribution of the etching at the surface of the coatings during the t_ON. A part of the intensity of CN emission can result from this etching phenomenon. To evaluate this surface modification mechanism, a complementary study was carried out on the different coated substrates to estimate the part of the CN emission coming from the monomer dissociation and the other one coming from the etching of the deposited coating. For an organic coating (duty cycle of 3%, t_ON = 1 ms), the intensity of CN emission is also detected when pure nitrogen plasma is ignited onto the organic coating (Fig. S3†).

These results seem to show that different growth mechanisms take place and evolve according to the t_ON duration, which cause different dissociation mechanism of the monomer and lead to different deposition kinetics. Different types of fragments are created during t_ON depending on its duration, which influences the growth rate. This evolution of fragmentation rate can be explained by the residence time of the monomer into the discharge. Fig. 9a and b presents the FTIR chemical mapping of coatings realised in static mode with a duty cycle of 3% (t_ON = 1 ms) and 100% over a surface area of 30 mm × 6 mm. The area of the asymmetric stretching vibration band of P–O–P is integrated and normalised to the stretching vibration peak of P=O in order to display the evolution of the polyphosphate (i.e. inorganic part) through the coatings.

Fig. 9 FTIR mapping of the coatings realised at a power density of 1 W cm⁻² and duty cycles of 3% (a) and 100% (b). The asymmetric stretching vibration of P–O–P area is integrated and normalised to ν(P=O) for the chemical mapping of polyphosphates. The band assigned to ν(P=O) is integrated to evaluate the thickness rate evolution of the coatings for the static deposition at DC = 3% (c) and DC = 100% (d). Their corresponding photographs, (e) and (f) respectively, are added.

For a DC of 3%, the coating does not exhibit any polyphosphate backbone (Fig. 9a) as the ratio of peak intensity is low under and outside the electrode area (blue color). Meanwhile the coating realised at a DC of 100% clearly evidences a gradient of the coating chemistry (Fig. 9b), from blue (no inorganic part) to red (strongly inorganic due to polyphosphate networks). This can be explained by the residence time of the monomer in the discharge. Actually, the monomer does not have time to be highly fragmented during 1 ms of discharge (low CN emission intensity), which leads to an organic coating (no P–O–P detected, Fig. 9a). For the continuous wave condition (i.e. DC = 100%), the monomer reacts with the discharge all along its way under the high voltage electrodes. The monomer is more and more fragmented with its residence time in the discharge, and lead to a chemical gradient from organic to inorganic character (Fig. 9b).

In that case, the reactive species created in nitrogen have enough time and are sufficiently energetic to enable the fragmentation of weak chemical bonds (e.g. C–O bonds, as shown by MS/MS measurements, Fig. S4†) and to etch the surface. It is interesting to notice that below 4 ms of residence time through the discharge (which correspond to the transition point observed in Fig. 8) the coating presents an organic character. Afterwards the polyphosphates appear and increase in amount due to smaller molecular fragments which will condensate to form the inorganic polymer. The polyphosphate concentration evolution indicates that gas phase reactions play an important role in the monomer dissociation in addition of surface reactions. The evolution of the thickness in static mode is presented for duty cycles of 3% (Fig. 9c) and 100% (Fig. 9d) with its corresponding photographs (e and f, respectively).

These letters indicate that the deposition rate is higher at the entrance of the plasma zone. It rapidly goes through a maximum before decreasing slowly because of the depletion of the reactive species density. At DC = 3%, the presence of a very thin deposit out of the high-voltage electrode can be noticed, confirming that a free-radical mechanism should take place. At DC = 100%, no deposit is observed out of the high voltage electrodes indicating that the short lifetime species (ions, electrons) are needed to bring energy for condensations reactions (or other reactions) of polyphosphate fragments and/or surface activation to enable the grafting onto the surface.

To sum up, at a power density of 1.0 W cm⁻², the chemical characterisations of the films have shown two different characters depending on the t_ON duration and therefore on the duty cycle. For low time-ON period (DC < 33%, 1.0 W cm⁻²), the high content of carbon (q.v. FTIR and XPS analyses) indicates a deposition mechanism mainly through low fragmentation or activation of the monomer during the time-ON and few etching (both displayed by few emission of CN). This range of DC is characterised by a low density of reactive species compared to monomer concentration and time-ON mainly allows activating slightly the monomer by low fragmentation rate. During the time-OFF, the plasma polymerisation may be dominated by more chemically-conventional products (free-radicals) than those encountered in the time-ON, which are mainly created from direct electronic collisions, ion collisions, UV radiations and random recombinations. The plasma-OFF period is characterised by reactions of active species created in t_ON with available precursor. Crosslinking mechanisms and deposition kinetics are thus different and depend on the free radicals generated. The radical propagation or recombination of organic fragments on their activated sites may occur (retention of aliphatic contribution) to grow the coating. The “organic” character of the coatings can be explained by the soft activation of unmodified and slightly oxidised/nitrogenised (low content of amine/amide/oxime/aldehyde/ketone) monomer molecules, leading to recombination through activated bonds (e.g. allyl group).

A possible growth mechanism occurring during t_ON and t_OFF is proposed (Fig. 10). For higher t_ON duration (33% < DC < 100%, 1.0 W cm⁻²), the organic parts of the monomer are mostly lost leading to a condensation of the phosphate groups into polyphosphates. Remaining (nitrogenised) organic elements detected in the coatings are the molecular fragments created during the first periods of the time-ON (<4 ms) and embedded in the polyphosphate skeleton. Afterwards, the (nitrogenised) organic fragments can be oxidised, leading to higher content of amide, imine/oxime and carboxyl groups. The higher the duty cycle, the more oxidised are the organic fragments incorporated into the polyphosphate structure. This is evidenced with the increase of the ketone and carboxylic acid groups detected in the coating for the highest duty cycles. A significant ratio of monomer molecules are decomposed, leading to a high crosslinking ratio due to the formation of polyphosphates. The interaction time between energetic species of the discharge and the precursor increases and produces a higher rate of precursor dissociation as well as stronger etching of the surface.

Fig. 10 Simplified reaction pathway proposed for DEAP plasma polymerisation in pulsed-AP-DBD during t_ON and t_OFF.

It has been pointed out that the duty cycle has an influence on the deposition mechanisms of DEAP and then on the physical chemistry properties of the coatings. It is possible to tune the reactivity of the discharge by modifying the residence time of the monomer in the discharge, which will lead to different dissociation rates and then different type of fragments can be created. Depending of the obtained fragments, different growth kinetic and chemistry mechanisms can occur during t_ON and t_OFF. However, the reactivity of the discharge can also be modified by playing on the dissipated power. As this parameter is directly related to the t_ON, the plasma power can move the boundary (i.e. the transition point) between the two deposition mechanisms. By increasing or decreasing the power density, the transition zone will be lowered or raised, respectively. This is related to the binding energy of the molecules which are broken faster when the power density is higher and therefore the residence time can be lowered and vice versa.

Conclusions

This work investigates the interest of chopping an AC electrical signal in an AP-DBD process in order to better control the chemistry and deposition mechanism of P-containing coatings. It has been demonstrated that the chemistry of the coatings can be adjusted from organic phosphate-containing thin films to inorganic polyphosphate-based deposits. This evolution is in correlation with the duty cycle modification and/or the power dissipated in the discharge. For low duty cycles, the coatings exhibit an organic character with retention of the monomer structure, while for high duty cycles the thin films are based on a polyphosphate backbone. These parameters regulate the fragmentation rate of DEAP in the discharge to obtain different mechanisms of deposition of phosphate-containing plasma polymers. Kinetics and chemical structures can be tuned enabling to go further in the comprehension of the diethylallylphosphate behaviour in an AP-PECVD process working in pure nitrogen. During the time-ON, it is important to consider the discharge power in terms of creation of reactive species as well as the duration of the t_ON (dissociation rate of the monomer related to its residence time in the discharge). Some etching phenomena have also been reported to play a role in the coating chemistry. During time-OFF, the reactions can be controlled by the created reactive species (when t_ON is low and not too energetic) and the time which is left to let them react.

Acknowledgements

This work was supported by the National Research Fund, Luxembourg (FNR CORE project 10/MS/788816). We gratefully thank Dr Elodie Lecoq, Dr Simon Bulou, Dr Martina Modic and Dr Maryline Moreno for their skilful support and Julien Bardon for his fruitful discussions. Mathieu Gerard is also acknowledged for mechanical engineering.

Notes and references

1. O. A. Lam, G. David, Y. Hervaud and B. Boutevin, J. Polym. Sci., Part A: Polym. Chem., 2009, 47, 5090.
2. C. Bressy-Brondino, B. Boutevin, Y. Hervaud and M. Gaboyard, J. Appl. Polym. Sci., 2002, 83, 2277.
3. E. Akdogan and M. Mutlu, Colloids Surf., B, 2012, 89, 289.
4. E. D. Weil, J. Fire Sci., 2011, 29, 259.
5. M. Jimenez, S. Duquesne and S. Bourbigot, Thermochim. Acta, 2006, 449, 16.
6. S. Bourbigot and S. Duquesne, J. Mater. Chem., 2007, 17, 2283.
7. J. R. Ebdon, D. Price, B. J. Hunt, P. Joseph, F. Gao, G. J. Milnes and L. K. Cunliffe, Polym. Degrad. Stab., 2000, 69, 267.
8. A. R. Horrocks, Polym. Degrad. Stab., 2011, 96, 377.
9. J. Galezowska and E. Gumienna-Kontecka, Coord. Chem. Rev., 2012, 256, 105.
10. A. Muñoz-Bonilla and M. Fernández-García, Prog. Polym. Sci., 2012, 37, 281.
11. C. Queffélec, M. Petit, P. Janvier, D. A. Knight and B. Bujoli, Chem. Rev., 2012, 112, 3777.
12. J. A. Rotole and P. M. A. Sherwood, Chem. Mater., 2001, 13, 3933.
13. P. H. Mutin, G. Guerrero and A. Vioux, J. Mater. Chem., 2005, 15, 3761.
14. S. Karatas, N. Kayaman-Apohan, O. Turunc and A. Güngör, Polym. Adv. Technol., 2011, 22, 567.
15. G. David, E. Ortega, K. Chougrani, A. Manseri and B. Boutevin, React. Funct. Polym., 2011, 71, 599.
16. A. Kannan, N. Choudhury and N. Dutta, J. Electroanal. Chem., 2010, 641, 28.
17. P. R. Davies and N. G. Newton, Appl. Surf. Sci., 2001, 181, 296.
18. M. J. Tsafack and J. Levalois-Grützmacher, Surf. Coat. Technol., 2006, 201, 2599.
19. S. Zanini, C. Riccardi, M. Orlandi, C. Colombo and F. Croccolo, Polym. Degrad. Stab., 2008, 93, 1158.
20. M. J. Danilich and R. E. Marchant, Plasmas Polym., 2002, 7, 127.
21. J.-C. Lin, Y.-F. Chen and C.-Y. Chen, Biomaterials, 1999, 20, 1439.
22. F. Hilt, D. Duday, N. Gherardi, G. Frache, J. Bardon and P. Choquet, Plasma Processes Polym., 2013, 10, 556.
23. J. Friedrich, Plasma Processes Polym., 2011, 8, 783.
24. A. Choukourov, H. Biederman, I. Kholodkov, D. Slavinska, M. Trchova and A. Hollander, J. Appl. Polym. Sci., 2004, 92, 979.
25. N. D. Boscher, P. Choquet, D. Duday and S. Verdier, Plasma Processes Polym., 2010, 7, 163.
26. M. E. Alf, A. Asatekin, M. C. Barr, S. H. Baxamusa, H. Chelawat, G. Ozaydin-Ince, C. D. Petruczok, R. Sreenivasan, W. E. Tenhaeff, N. J. Trujillo, S. Vaddiraju, J. Xu and K. K. Gleason, Adv. Mater., 2010, 22, 1993.
27. N. D. Boscher, D. Duday, P. Heier, K. Heinze, F. Hilt and P. Choquet, Plasma Processes Polym., 2013, 10, 336.
28. C. L. Rinsch, X. Chen, V. Panchalingam, R. C. Eberhart, J.-H. Wang and R. B. Timmons, Langmuir, 1996, 12, 2995.
29. N. M. Mackie, N. F. Dalleska, D. G. Castner and E. R. Fischer, Chem. Mater., 1997, 9, 349.
30. D. O. H. Teare, C. G. Spanos, P. Ridley, E. J. Kinmond, V. Roucoules, J. P. S. Badyal, S. A. Brewer, S. Coulson and C. Willis, Chem. Mater., 2002, 14, 4566.
31. G. Crobu, A. Rossi, F. Mangolini and N. D. Spencer, Anal. Bioanal. Chem., 2012, 403, 1415.
32. G. Socrates, Infrared and Raman Characteristic Group Frequencies, Wiley, Chichester, Sx, England, 2001.
33. E. Lecoq, D. Duday, S. Bulou, G. Frache, F. Hilt, R. Maurau and P. Choquet, Plasma Processes Polym., 2013, 10, 250.
34. M. Nocun, J. Non-Cryst. Solids, 2004, 333, 90.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra11625a

---