Marcus P. Corrosion mechanisms in theory and practice

Подождите немного. Документ загружается.

chamber is incorporated as part of the siloxane network. There is a significant amount

of dimethylsilyl groups in the polymer, which can be formed only by cleavage of the

Si—C bond. Thus, it can be concluded that the polymerization mechanism

is dominated mainly by the formation and recombination of silyl radicals. The

abstraction of methyl groups is confirmed by the atomic concentrations measured by

XPS. The measured concentrations correspond to Si

4.1

0.9

compared with

in the case of the monomer [117]. The oxygen in the polymer is almost

exclusively bound to silicon, as measured by IR spectroscopy and XPS.

Morphology of Ultrathin Plasma Polymer Films

Despite the fact that plasma polymerization leads to rigid and laterally homogeneous

films, a varying number of pinholes is observed in most cases [118]. Pinholes are

crucial defects when the plasma polymer serves as a coating for reactive metals.

In this case, electrolyte can penetrate through these channels to the uncovered

metal surface and give rise to localized forms of corrosion. Not surprisingly,

the ultrathin plasma polymers contain several pinholes or intrinsic channels that

can be located by immersion of the coated sample in a copper sulfate solution

512 Rohwerder et al.

Figure 30 IRRA spectrum of an about 13-nm-thick plasma polymer deposited from a

mixture of argon and hexamethyldisilane (p

= 25 Pa, p

HMDS

= 10 Pa). For comparison,

the gas-phase spectrum of hexamethyldisilane is also shown. Indicated are the peaks that

are specific for the polymer. All peaks are assigned in Table 3 [115].

[c(CuSO

) = 10

–3

mol/e] [115]. Copper was deposited in the pinholes while the rest

of the surface was effectively isolated from the electrolyte even by films of a few

nanometers thickness. Van Ooij et al. [104] studied the topology of plasma polymers

deposited on stainless steel substrates from a DC plasma by means of STM. The

topology was dependent on the plasma parameters. Films deposited at low pressure

(6 Pa) were significantly smoother than those deposited at higher pressures (66 Pa).

The lateral homogeneity of ultrathin plasma polymers was confirmed by

atomic force microscopy measurements on flame-annealed gold substrates [117].

Flame annealed Au was chosen because of its large (about 200 × 200 nm)

monoatomic flat terraces. The measurement was done in the tapping mode to protect

the plasma polymer surface from damage. Figure 31b shows, as an example, the

structure of an about 3-nm-thick hexamethyldisilazane (HMDSZ) plasma polymer.

The film structure can be revealed by a comparison with the bare flame-annealed Au

surface (Fig. 31a). The shallow features of the plasma polymer show an average

diameter of about 20 nm and a height of about 0.5 nm. The structure of the

HMDS-PP resembles that of HMDSZ-PP and is published elsewhere [117]. By

calculating the roughness (RMS value) of the plasma polymers on monoatomic flat

terraces, the microstructure of the plasma polymers can be revealed as a function of

the nature of the monomer and the film thickness (see Table 2). For comparison, a

self-assembling monolayer of octadecylmercaptan (ODM) on flame-annealed

gold was prepared (adsorption from ethanolic solution at open circuit potential,

Δt

ads

= 12 h) and measured in the same way.

The plasma polymer–coated surface shows increased roughness in comparison

with the bare gold surface, whereas the ODM monolayer leads only to a slight

increase in roughness. The lower RMS value of the 1-nm-thick HMDSZ-plasma

polymer in comparison with the thicker films might be due to incomplete coverage

of the measured area so that the bare gold surface contributes to the overall value.

However, it appears that a constant roughness value is measured for a thickness

above 3 nm.

Corrosion Prevention by Adsorbed Monolayers 513

Table 1 Assignment of the IR Peaks of Gaseous HMDS and the HMDS Plasma Polymer

514 Rohwerder et al.

Figure 31 (a)AFM picture of a flame-annealed Au surface. The measurement was done

in the tapping mode. The brightness increases with the height. (b) AFM picture of a

flame-annealed Au surface coated with an about 3-nm-thick hexamethyldisilazane

[(CH

)

Si—NH—Si(CH

)

, HMDSZ] plasma polymer that was deposited at p

HMDSZ

= 5 Pa.

The brightness increases with the height [105].

Interaction between the Polymerizing Plasma and the Oxide Surface

By using an in situ FTIR cell, Grundmeier and Stratmann [117] could reveal

interactions at the iron oxide/plasma polymer interface. Prior to the plasma polymer

deposition, the iron surface was plasma oxidized. The in situ IRRA spectrum of the

plasma-oxidized surface now served as reference spectrum for the subsequent

measurements. A very short (1–2s) plasma polymer deposition was then performed

and the in situ IRRAspectra were measured again. This procedure was repeated twice

on the same sample, resulting in three layers. The first absorption spectrum was

subtracted from the second one and the second from the third. This procedure leads

to three spectra that represent the structure of each of the three plasma polymer

layers (see Fig. 32). All spectra were normalized by the thickness of the respective

layer, d, as measured by the QCM. The IRRA spectrum of the first layer (thickness

~ 0.7 nm) shows two peaks at 1200 and 955 cm

–1

, which are not found in the spec-

trum of the third layer. The spectrum of the second layer (thickness ~ 1nm) contains

the same two peaks but the intensities are much lower than in the first layer. In addition,

the Si—O—Si peak occurs at 1065 cm

–1

in the first layer and shifts to 1056 cm

–1

the second and to 1042 cm

–1

in the third layer. The higher frequency of the Si—O—

Si peak at the interface might be a consequence of the change in the angle of the

siloxane group induced by polar interactions between the oxygen of the polymer and

hydrogen on the oxide surface. Any change in the bonding angle means that a

higher energy of the system leads to an increase in the vibration frequency of the

Si—O—Si asymmetric stretching, as observed for an SiO

layer on silicon [119].

A comparison of the thickness of the layers with the results concerning the

microstructure of the films leads to the conclusion that almost complete coverage

of the surface by the plasma polymer is reached when the film thickness is about

2 nm. Thus, the special peaks in the first and second layers might be assigned to

the interaction of the deposited layer with the oxidized iron surface. The third

layer is deposited onto a surface that is almost completely covered by the

HMDS-plasma polymer film and therefore cannot interact with the oxide.

According to the literature, the peak at about 955cm

–1

can be assigned to a

metal-oxygen-silicon bond and the peak at 1200 cm

–1

is most probably due to a

Si—O—CH

or a metal—O—CH

group [116]. Silanol groups would also lead

to absorption in this wavelength region. The IR peak of silanols, however, was

observed at 920 cm

–1

for the chemisorption of polysiloxanes on iron using the

Corrosion Prevention by Adsorbed Monolayers 515

Table 2 Calculated RMS Values for Different Plasma Polymer Layers and a Self-

Assembled octadecylmercaptan (ODM) Film of Flame-Annealed Gold [105]

same reflection assembly [120]. This value is about 35 cm

–1

lower in wave number

than the peak observed in Figure 32. Van Ooij et al. [104] studied the influence of

the substrate on the plasma polymer formation using stainless steel as substrate.

Films were deposited from a DC plasma. The authors observed especially for films

deposited at high pressures a dependence of the film structure on the film thickness

and thus proposed a two-stage mechanism. While the film adjacent to the substrate

resembles polydimethylsiloxane, it becomes more random for higher film thickness.

Applications of Plasma Polymers as Corrosion-Resistant Layers

on Reactive Metals

Delamination of Ultrathin Plasma Polymers

Plasma polymers may be applied without subsequent painting. In this case, corrosion

may start in local defects in the plasma polymer and spreading of these defects is a

critical degradation process. Moreover, plasma polymers may be used as model

ultrathin polymers coatings. Grundmeier et al. [121] studied the delamination of

ultrathin plasma polymers by using a scanning Kelvin probe and surface analytical

methods.

516 Rohwerder et al.

Figure 32 In situ IRRA difference spectra revealing the structure of the subsequently

deposited three layers of HMDS plasma polymer on plasma-oxidized iron. The polymer

was deposited at a partial pressure of p

HMDS

= 5Pa. The value d means the thickness of

each deposited plasma polymer layer [117].

The delamination rate was determined with a scanning Kelvin probe. A small

amount of fine sodium chloride was introduced into a circular deepening in the

middle of the polished and ethanol-cleaned iron sample. After the sample was

introduced into the plasma reactor, it was cleaned and activated in one step by an

oxygen plasma, leading to a carbon-free and highly oxidized iron surface. In the

next step an ultrathin plasma polymer of hexamethyldisilane was deposited on the

cleaned substrate, leading to a well-defined metal-polymer interface. The thickness

of the deposited polymer was controlled by the in situ measurement of the resonance

frequency of the quartz crystal and was about 5 nm, so that the film thickness is

in the range of the escape depth of the photoelectrons.

This special kind of defect preparation has the advantage that water

condenses slowly after the insertion of the sample into the scanning Kelvin probe

at 93% relative humidity. The growing droplet is fixed in the deepening due to the

hydrophobic nature of the plasma film (water contact angle 92 ± 2°). Thus, the

electrolyte does not spread over the delamination zone. Because a saturated

sodium chloride solution is formed within the deepening, no drying of the defect

is observed. A Cr/Ni wire was used as the vibrating reference electrode and

provided a local resolution of about 100 μm. X-ray photoelectron spectra were

obtained using a Physical Electronics ESCA 5600 XP spectrometer.

In Figure 33 the potential distribution over the circular defect and the

surrounding intact plasma polymer–coated area as measured with the scanning

Kelvin probe is shown with increasing corrosion time. Obviously, negative

potentials are measured within the defect and very positive potentials are observed

in the intact region. With increasing corrosion time the delamination front marked

by the sharp decrease of the corrosion potential shifts into the former intact area

and reveals that undermining occurs. Line scans shown in Figure 34 reveal the

activation of the defect and the progress of the delamination front.

An interesting question refers to the comparability of the delamination of

ultrathin polymers to that of macroscopic thick coatings. Therefore the sample was

inserted into the XP spectrometer and the delaminated area was analyzed locally

with high-energy resolution (0.6 eV). Figure 35 presents the line scans of the Na 1s

and the Cl 2p peaks starting at the border of the defect and moving toward the intact

area. The local resolution was 400 μm and the total length of the line scan was 2000 μm.

The line scan reveals that no chloride is found in the delaminated area and that the

concentration of sodium ions decreases toward the delamination frontier. The final

length of the undermined region measured by the scanning Kelvin probe and the

distribution of sodium ions as measured by means of electron spectroscopy for

chemical analysis (ESCA) are in good accordance.

It can be concluded from the former results that the delamination of the

ultrathin polymers behaves similarly to that of the model coating. It is now an

advantage of the very small thickness that the metal-polymer interface can be

characterized by XP spectroscopy without the separation between polymer and metal

substrate, which otherwise leads to significant distortions caused by organic

contamination, water absorption, and oxidation at ambient atmosphere. It was found

that the ultrathin solid films are oxidized and transformed to a gel-like structure due

to the osmotic swelling of the film in the delaminated region [121].

Corrosion Prevention by Adsorbed Monolayers 517

518 Rohwerder et al.

Figure 33 Potential distribution over the circular defect and the surrounding intact

plasma polymer–coated area as measured with the scanning Kelvin probe is shown with

increasing corrosion time [121].

Corrosion Prevention by Adsorbed Monolayers 519

Figure 34 Scanning Kelvin probe line scans shown across the circular defect in Figure 33,

revealing the activation of the defect and the progress of the delamination front [121].

Influence of Ultrathin Plasma Polymers on Delamination Kinetics

of Polymer-Coated Steel and Galvanized Steel

As stated before, an ultrathin plasma polymer usually requires a stabilizing top coat.

Iron substrates were plasma modified by an oxygen plasma treatment and by the

deposition of an ultrathin plasma polymer as described before. Grundmeier and

Stratmann [115] investigated the influence of an interfacial plasma polymer on the

delamination kinetics of a clear coat on iron. The sample was coated with a

water-based, one-component, polyurethane acrylate primer. An artificial defect with

a reservoir for the electrolyte (0.5 M NaCl) and a linear defect border was prepared

within the coated substrate. To reveal the influence of the plasma deposition on the

delamination kinetics the sample was prepared with one half of the substrate not

modified and the other half plasma modified. Masking of the unmodified region was

achieved using a thin aluminum foil placed on the desired area. This technique

guarantees that both areas have the same properties before the plasma treatment and

are coated with the primer in the same way. Any observed differences should,

therefore, be due only to the plasma deposition.

Figure 36 shows a 3D plot of a scanning Kelvin probe measurement of an iron

sample that was coated with a 5-nm-thick HMDS plasma polymer on one half of the

sample surface prior to the deposition of the primer [115]. The water contact angle

on the HMDS plasma polymer was 92 ± 2°. This means that the surface energy

decreased with respect to a plasma-oxidized surface, which shows a water contact

angle of less than 10°. Nevertheless, wetting of the plasma polymer surface by the

primer was sufficient to lead to homogeneous spreading and adhesion of the primer.

The area modified by the plasma polymer shows little undermining even after

1265 min, whereas the area coated only by the primer shows fast delamination.

A comparison of the delamination kinetics is shown in Figure 37. The progress of the

steepest slope of the potential profile is plotted versus the time after the activation of

the defect. Comparison of the corresponding kinetics leads to the conclusion that the

ultrathin plasma polymer gives rise to a significant decrease in the delamination rate.

Plasma polymers with a special surface structure suitable to bond to an epoxy

amine primer were used as interfacial coupling layers on iron and galvanized steel

and led to even better results [122]. However, in all cases the system always

showed cathodic delamination at the polymer/metal oxide interface, indicating

the importance of oxygen reduction on the oxide surface. For verification, after the

delamination of the sample, the delaminated polymer was pulled off and the

underside of the polymer and the iron surface were investigated by XPS to reveal

whether the system delaminated at the plasma polymer/metal interface or at the

520 Rohwerder et al.

Figure 35 Small spot ESCA line scans of the Na

and the Cl

peaks starting at the

border of the defect (see Figs. 33 and 34) and moving toward the intact area (spatial

resolution 400 μm, total length of the line scan 2000 μm) [121].

Corrosion Prevention by Adsorbed Monolayers 521

Figure 36 Potential profile of a delaminated primer-coated iron specimen that was half

side plasma polymerized and afterward coated by a 100-μm-thick primer (x = 0: defect

border, y < 0: not plasma modified, y > 0: plasma plasma polymerized). The thickness of

the HMDS plasma polymer was 5 nm [105].