Marcus P. Corrosion mechanisms in theory and practice

Подождите немного. Документ загружается.

It is clear from the situation just described that the latter situation is ideal for a

stable chemical modification of the metal substrate and that molecules should be

chosen that are able to interact strongly with the metallic substrate. However, there

are some characteristic differences between corrosion inhibitors and molecular

adhesion promoters: whereas for corrosion inhibition the composition and structure

of the metal surface are defined by the corrosive medium, the surface properties can

be changed and adjusted to the structure of the adhesion promoter. Inhibitors must be

soluble in the electrolyte (e.g., water) and can be applied only for well-defined

reaction conditions. Adhesion promoters, however, may be applied from aqueous or

nonaqueous solvents or even from the gas phase and the reaction conditions can be

optimized for the given substrate. Therefore, some characteristic molecular features

of inhibitors such as heteroatoms S, P, and O should be incorporated into the

structure of the adhesion promoter; however, the molecule itself should show

minimum solubility in water and the possibility to bind a polymer onto the adhesion

promoter.

In this chapter, only very simple molecules will be discussed. They are

composed of one reactive center such as —SH, —Si(OCH

)

, or PO(OH)

, which

should be able to bind to the metal surface [20,21], and a long aliphatic chain (e.g.,

—C

), which allows ordering of the individual molecules and the formation of a

dense packing on top of the substrate (self-organization) (see Fig. 2). We will discuss

482 Rohwerder et al.

Figure 2 (a) Reaction of silanes with hydroxylated surface in humid atmospheres.

(b) Reaction of mercaptans with metal surfaces.

how these modified surfaces may be prepared, what kind of structure is observed,

and how stable the modification is in aggressive electrolytes.

The preparation technique must fulfill certain requirements: the surface

properties of the substrate, e.g., the density of chemisorbed OH groups, must be

well defined and the organic monomer must be allowed to bind to the reactive

centers of the surface without destroying the defined surface structure; the

monomer itself should form a dense structure with a high degree of ordering so

that the substrate surface is not accessible to water molecules.

Another method for the modification of metal surfaces by ultrathin organic

films is plasma polymerization. Plasma polymerization, as a process technology

for corrosion-resistant thin-film deposition, has been explored during the last 20

years. Plasma polymers can be deposited from an electric discharge containing

organic or metal-organic molecules [44,45]. A glow discharge is formed by expos-

ing a gaseous monomer at reduced pressure to an electric field. The monomer is

fragmented in the discharge and the reactive intermediates generated polymerize

on a substrate according to a special reaction mechanism [46,47]. The resulting

films can be highly cross-linked and, depending on process parameters, show

more inorganic or more organic properties; moreover, adhesion is excellent to

most metal surfaces, i.e., the process is less specific to certain metals than is the

case with molecular self-assembly, and deposition of ultrathin films is fast. The

main disadvantage is that up to recently the process required a very low residual

pressure; i.e., vacuum equipment was needed. Lately, plasma polymers have also

been prepared under atmospheric pressure conditions.

ORGANIC MONOLAYER FILMS

Corrosion Protection by Self-Assembled Films

As outlined above, in order to improve the stability of the polymer/metal interface it

is of utmost importance to find ways to prepare interfaces that have better ability to

inhibit oxygen reduction and are less vulnerable to the products of oxygen reduction.

One way is to use monolayers of bifunctional molecules as adhesion promoters.

Ideally, such a molecule should form a tight chemical bond to the metal or metal

oxide surface with its head group and to the polymer with its tails groups. The

monolayer should be as dense as possible with as few defects as possible, for

optimum stability and inhibition capability. Also, for technical application the

formation of such monolayers should be quick, i.e., be finished within a few seconds.

The following paragraph will focus on electrochemical aspects of the self-

organization and the resulting effect on the final defect structure. The discussion will

distinguish between oxide-covered and oxide-free surfaces. The protective impact of

the films will also be discussed.

In recent years the process of molecular self-assembly on solid surfaces, i.e.,

the adsorption and self-organized formation of highly ordered monolayers from

monomers in solution, has received increasing interest, especially the self-assembly

of thiols on gold [48]. As could be shown, thiol monolayers proved to be excellent

inhibitors of oxygen reduction and moreover are not easily destroyed by the

radicals set free during the oxygen reduction [49]. First tests on iron also gave

promising results [50,51] although the preparation is not easy on this substrate.

Corrosion Prevention by Adsorbed Monolayers 483

Because thiol molecules do not adsorb on iron oxide, the iron has to be electrochem-

ically polarized to cathodic potentials in order to get rid of the oxide layer and then

to keep it free of oxide during self-organization. The SA of BTA on copper was

investigated by Magnussen and Behm [52]. Now, even though self-organization is

the subject of hundreds of publications, up to very recently nothing has been known

about the influence of the electrode potential and surface charge on the self-assembly

process. For bare metal surfaces the surface charge is controlled by the electrode

potential, for oxide-covered samples by the pH of the solution.

Self-Assembly on Oxide-Free Metal Surfaces

Thiol Self-Assembly on Gold The well-studied system thiol/Au is ideal for

investigating the effect of the electrode potential on the kinetics of self-assembly

and on the resulting defect structure.

A direct way to monitor the self-assembly in situ is to measure the decrease

in capacitance of the immersed sample (see Fig. 3). The double layer at the

metal/electrolyte interface is pushed apart by the growing monolayer, which

causes a decrease of capacitance. Because an electrode’s capacitance depends on

the electrode potential, it is useful to normalize curves obtained for adsorption at

different potentials, e.g., by referring the change in capacitance ΔC at the time t to

the final capacity change ΔC

max

. Such normalized curves for the adsorption at –400

and –800 mV versus a special Ag/AgCl reference electrode [19,53] in ethanol

and typical nanoscopic structures of the films at different stages of the thiol

self-assembly are shown in Figure 4. A detailed analysis of the curves yields the

distinction of three characteristic potential ranges for thiol self-assembly [19,53].

An intermediate potential range from –400 to + 200 mV versus the Ag/AgCl

reference electrode where the rate of self-assembly shows only a

comparatively weak dependence on the electrode potential; scanning

tunneling microscopy (STM) images of the completed films show basically

the same features already known from the literature [54–62].

A cathodic potential range where the rate of self-assembly decreases significantly

with increasingly cathodic potentials; STM images of the completed films

show that the average domain size also increases, so that for adsorption at

–800 or –900 mV the thiol domains fill whole terraces, even if the lateral size

of these exceeds 100 or even 200 nm (Fig. 4).

An anodic potential range where there is a slight decrease in the rate of

self-assembly with increasingly anodic potentials.

Whereas the film is finished to 95% within 10 s at potentials of –400 to

+200 mV, yielding a film composed of domains with a lateral size of 10–20 nm,

it takes thousands of seconds at very cathodic potentials for the domains to cover

the surface, but the domains are much larger. A detailed analysis shows that the

thiol adsorption is at least a two-step process: physisorption followed by

chemisorption [53a]:

484 Rohwerder et al.

Corrosion Prevention by Adsorbed Monolayers 485

Figure 3 If the capacity of the gold surface is measured during thiol adsorption, a decrease in capacitance can be observed due to the

pushing apart of the double layer by the adsorbing thiol.

It is the electrode potential dependence of the latter step that is responsible for the

extreme slow-down in adsorption velocity at cathodic potentials.

Assuming a symmetry factor β = 0.5 (because of the symmetry of the system)

and a complete charge transfer, it can be shown from the measured capacity curves

that about 36% of the overall potential difference between the electrode and the

solution occurs over the alkane chains; i.e., the longer the alkane chain, the smaller

the slowdown effect at cathodic potentials. This could be shown for the adsorption

of octadecylthiol as compared with decylthiol [53b].

Thiol on Iron For technical application this proves to be a problem, because

adsorption times on the order of thousands of seconds are out of the question. In

order to circumvent this problem, another coating technique has to be applied: the

sample (the working electrode in the figure) is polarized to cathodic potentials to

486 Rohwerder et al.

Figure 4 The normalized capacity curves clearly show that thiol (here decanethiol)

self-assembly is much faster at intermediate potentials (here –400 mV) than at cathodic

potentials (–800 mV). Typical STM images of the film structure are shown for different

stages of self-assembly at the two potentials. At –400 mV, thiol self-assembly is fast and

occurs via small individual domains (1), which sets the foundation for the final domain

structure (2). At –800 mV, self-assembly is very slow and occurs via domains that are

interconnected by molecular rows of chemisorbed thiol (3), and thus very large domains

result in the final film (4).

reduce the oxide layer and to prevent oxide formation during the coating process

and then pulled through a thiol film, floating on top of the aqueous electrolyte

[50,51]. In this way about 10-nm-thick multilayer films can be prepared, which

show excellent blocking of oxygen reduction (see Fig. 5). It is quite obvious that

the oxygen reduction is drastically limited on the modified surface and a much

higher cathodic overpotential is necessary to reduce oxygen at significant rates. If

oxygen is reduced, however, the film is destroyed by the radicals formed during

oxygen reduction.

For adhesion promotion only the first monolayer will be effective. Because it

proved not to be possible to prevent iron surface reoxidation completely while

pulling the sample through the floating thiol film, and thus parts of the sample

are not covered by chemisorbed thiols but by oxide, it is necessary to get rid of

such defects in a second preparation step in order to improve the quality of the

chemisorbed monolayer beneath the multilayer film. Immersing the as-prepared

sample in aqueous solution and cycling the potential from the cathodic limit,

where hydrogen evolution is beginning, through the potential range where iron

is oxidized (peak in curves 1 and 2), and back for several times results in a

healing of the disordered film (see Fig. 6), as can be seen from the decrease of

the oxidation peak with increasing number of cycles. With each sweep into the

potential range where iron oxide is reduced, bare iron is exposed at the defect

sites to the thiol in the multilayer, and in the subsequent anodic sweep the thiol

can at least partly be chemisorbed before reoxidation of the remaining bare iron

surface sets in. Finally, all the surface is covered by chemisorbed thiol mole-

cules. Then the modified sample is pulled out of the solution and the polymer

coating is applied. Alternatively, the sample may be polarized at negative –800 mV.

But then the healing process takes much longer (see Fig. 7a) because the thiol

chemisorption at cathodic potentials is very slow (see earlier). Most of the

multilayer thiol is dissolved into the polymer so that the first chemisorbed

monolayer should perform its function as an adhesion-promoting layer.

Corrosion Prevention by Adsorbed Monolayers 487

Figure 5 Rate of oxygen reduction on iron (c) and iron modified by one monolayer of

n-decylmercaptan ().

488 Rohwerder et al.

Figure 6 (Upper left) Model of a thiol multilayer on iron with remnants of oxide. After several cycles of the electrode potential

(CV) in O

-free electrolyte (right), the interface between thiol and iron is healed (lower left).

Scanning Kelvin probe mappings of a thus modified iron sample coated with a

polymer show more than three times slower delamination kinetics. Still, a more

pronounced effect is desirable for future applications.

Closer investigation shows that the bond between the first thiol monolayer

and the metal surface is destroyed by reoxidation of the iron surface beneath the

multilayer film within tens of seconds. Of course, this results in diminished ability

of the layers to promote adhesion compared with what would be expected from an

intact monolayer. This reoxidation is faster the thinner the multilayer film. For

a monolayer film the reoxidation takes about 10s. That is an impressive factor of

to 10

slower than for bare iron, achieved by an only 1-nm-thick monolayer

film, as can be seen from Figure 7b.

Corrosion Prevention by Adsorbed Monolayers 489

Figure 7 (a) Current observed during healing at –800 mV of a decanethiol monolayer

compared with bare iron. (b) Normalized oxygen peak vs. oxygen exposure for

thiol-covered iron in comparison with bare iron.

A first step to find ways to improve this technique is a understanding of

the reoxidation process, which is at the center of current research. Important

parameters are chain length and functional groups of the thiol molecules. If the

choice of more suitable thiols together with a more refined modification technique

resulted in a thiol layer that survives even one order of magnitude longer in air,

this could be an important breakthrough, because then the polymer coating could

be applied before most of the thiol film is destroyed.

Self-Assembly on Stable Oxide Surface

Self-Assembly of Phosphonate Films Whereas in the case of iron the adhesion

layers have to be adsorbed directly on the metal because the oxides are unstable, the

situation is different for metals that form stable oxides. Here it is better to have the

films formed on the stable oxide. On aluminum oxides, for example, phosphonates,

X—R—PO

2–

, form stable and well-ordered monolayers. In X—R—PO

2–

, X

designates the functional end group, R the alkane chain, and PO

2–

the head group.

With technical samples, high-resolution STM is not applicable to obtain

information about the molecular order. Other methods such as Fourier transform

infrared (FTIR) spectroscopy and X-ray photoelectron spectroscopy (XPS) have

to be applied.

The adsorption of phosphonates on aluminum surfaces is an acid-base reaction.

The driving force is the formation of a surface salt [63]. Figure 8 shows FTIR

spectra supporting this theory. The similarity in the spectra of a Zn-biphosphonate

multilayer on gold and of an self-assembly (SA) film of biphosphonic acid on

aluminum oxide clearly supports this theory. The peaks at wave numbers 1100 and

950 cm

–1

are attributed to vibrational modes of the RPO

2–

anion.

490 Rohwerder et al.

Figure 8 Infrared adsorbance spectra of a phosphonate film on an aluminum oxide

surface and of Zn-biphosphonate multilayer on gold (see figure on the left).

The following adsorption scheme is supposed:

Corrosion Prevention by Adsorbed Monolayers 491

This requires the presence of OH groups on the aluminum oxide surface. Different

levels of OH on the oxide surface can be adjusted by gas-phase adsorption of

different amounts of oxygen and water on aluminum in a special ultrahigh vacuum

(UHV) chamber, which also allows electrochemical experiments and adsorption

from solution without exposing the samples to air. After the OH ratio in the oxide

layer was determined with XPS, the sample was exposed to the phosphonate

solution and then the amount of adsorbed phosphonates was determined with XPS.

Figure 9 shows that for low OH content in the oxide surface no phosphonate

adsorption occurs.

Because aluminum is only rarely used as the pure material, it is of crucial

importance for a successful application of the protecting and adhesion-promoting

monolayer films that not only the aluminum surface is covered by the molecules but

also the inclusions, such as Al

Fe, as they are typical of most aluminum alloys. Even

Figure 9 Intensity of the phosphorus 2s photoelectron signal as a function of the OH

–

content in the oxide surface.