This paper reviews the progress that has been made in the use of surface analytical techniques such as XPS and ToF-SIMS to obtain chemical information from the buried interface of an organic coating or adhesives system. Such a task is nontrivial, as the interfacial region, at most a few nanometers wide, is buried between many tens or hundreds of micrometers of substrate and overlayer. Two methodologies are described for the determination of interface chemistry: the deposition of a very thin (ca. 2 nm) film of coating or adhesive on the substrate followed by the use of XPS or ToF-SIMS to "look" through the layer to extract chemical information specific to the substrate; and the use of chemical or mechanical sectioning to present the sample in such a way that the interface region of the system is submitted for analysis. In the former category, examples are given of the deposition of an organosilane on aluminum, which yields a covalent Al-O-Si bond, and the subsequent extension of covalency via a TDI urone (toluene diisocyante urone) curing agent and an epoxy resin into the bulk of the cured adhesive. The continuity of covalent bonding from the substrate to the bulk resin ensures good durability of such a system. Oxide stripping and ultra-low-angle microtomy techniques are used to provide examples of the development of unique interface chemistry and the elucidation of interface concentration gradients on a duplex organic coatings system. It is concluded that such approaches have much to offer the adhesion scientist in the search for the Holy Grail: the ability to reverse-engineer interface chemistry in order to confer specific properties.
In any attempt to understand the physico-chemical phenomena responsible for adhesion between a polymer phase, such as an adhesive or organic coating, and a solid substrate, there is a need to be able to access the interface or interphase between the two materials. This is easier said than done; the complexity of the situation is indicated in Fig. 1.1, in that significant thicknesses (many tens of micrometers at the very least) of both substrate and overlayer obscure the adhesion layer or boundary that exists between the two. In order to examine this critical region one must resort to the removal of large amounts of material by mechanical or chemical means, or construct model specimens in which the interphase region is more readily accessible by analytical methods such as XPS and ToF-SIMS. The simple expedient of forensic analysis of a failure interface produced by mechanical, chemical, or electrochemical means is unlikely to provide a specimen that will yield the required information. Although it may lead to a greater understanding of the mechanism of failure, fundamental aspects relating to adhesion phenomena will invariably be elusive.
The nature of interfacial bonding will, to a certain extent and in a very simplistic manner, influence joint strength. At this level the contributions to interfacial bonding are quite simply the number of bonds in place per unit area and the strength (more strictly the bond energy) of the various bond types. Thus it is possible to envisage the equivalence of the contribution to bond strength from many weak bonds (or rather less strong bonds) and fewer, stronger bonds, as indicated in Fig. 1.2 (a). This is all well and good, but when one starts to consider the resistance of the interface to aqueous exposure it is often found that strong bonds (such as covalent bonds) are much more resistant to hydrodynamic displacement than weaker bonds such as those of the van der Waals type (Fig. 1.2 b). Thus both the aeric density of bonding sites and a knowledge of bond type are essential if one is to understand the interfacial chemistry and its effect on adhesion and - more importantly - durability. In the longer term the aim is to be able to engineer specific chemistry at the interface which will provide the required level of performance from an adhesive joint or organic coating.
This paper will build on previous reviews which have sought to explore the manner in which surface analysis methods can be purposefully employed to understand adhesion phenomena, with an emphasis on the elucidation of interphase chemistry. The rationale behind such an approach is that it is this critical region of a polymer/metal or polymer/polymer couple that will influence the performance of the overall system, be it the durability of an adhesive joint or the corrosion protection performance of an organic coating.
In essence there are two potential ways in which the interphase region can be approached; either by the use of systems based on real adhesives or organic coatings to create a model interphase, or by the sectioning, by some means or other, to expose the interphase region prior to analysis by XPS or ToF-SIMS. In this paper the use of both approaches, which have been widely explored in the author's laboratory over the last three decades, will be described.
1.2 Development of a Model Interphase
This route to interphase analysis may also be described as the thin-film approach, inasmuch as the basic principle involves the deposition of extremely thin layers (< 2 nm) of the mobile phase onto a substrate. In the work described here the mobile phase is the organic component of the system, but the same approach is applicable to studies of the metallization of polymer substrates. It is also convenient to combine studies of polymer interactions with solid substrates with studies of the adsorption characteristics of the organic components themselves. Such an approach has much to offer in adhesion research and the basis of studies of adsorption from a liquid phase and its applicability in adhesion has been discussed in detail elsewhere so it will not be treated in depth here. A brief overview will, however, provide a background to this approach. The determination of gas-phase adsorption isotherms is a well-known methodology in surface chemistry; in this manner it is possible to describe adsorption as following Langmuir or other characteristic adsorption types. The conventional method of studying the adsorption of molecules from the liquid phase is to establish the depletion of the adsorbate molecule from the liquid phase. However, as first pointed out by Castle and Bailey, with the advent of surface analysis methods it is now very straightforward to monitor the actual uptake of the adsorbate on the solid surface by XPS or SIMS. The experiment itself is quite simple in that a set of coupons of the solid substrates are exposed to a series of dilute solutions of the candidate adsorbate. The uptake curve (the adsorption isotherm) is quite simply a plot of surface concentration of the adsorbate (determined by XPS or SIMS) versus the solution concentration. In the case of chemisorption of the adsorbate on the substrate the uptake curve will quickly reach a plateau, indicating that all the potential adsorption sites on the substrate are occupied by adsorbate molecules. An example of this is shown in Fig. 1.3, which indicates the adsorption of the diglycidyl ether of bisphenolA (DGEBA) on silane-treated aluminum. The measure of the uptake of the DGEBA is taken as the relative peak intensity of the mass 135 peak in the positive ToF-SIMS spectrum, which is very characteristic of the DGEBA molecule. If necessary the thickness of the overlayer at monolayer coverage (the plateau region) can be determined by XPS; similarly, the adsorption regime which best describes the adsorption characteristics (that is, whether adsorption conforms to the Langmuir, Temkin, or another type) may be determined by simple diagnostic tests, as described elsewhere. Although there are several possibilities, experience has shown that many of the adsorption phenomena of importance in adhesion are characterized by Langmuir adsorption, indicating an equivalence of adsorption sites (in terms of enthalpy of adsorption) on the solid substrate.
Although the adsorption isotherm provides us with much important evidence regarding the aeric density of bonding sites and the type of adsorption that occurs, it tells us little about the interfacial reactions responsible for adhesion. In order to achieve this goal it is necessary, as indicated earlier, to examine a thin layer of the adsorbate on the substrate. The choice of such a specimen can be made from an adsorption isotherm if chemisorption is known to occur (indicated by a curve of the form of Fig. 1.3). A specimen taken from the plateau region of such an uptake curve will yield the maximum number of interfacial bonds with the minimum overlayer thickness. It should thus, in principle, be possible to use XPS or SIMS to probe the interfacial chemistry directly. Although this is perhaps the optimum approach, it is sometimes not possible to obtain the entire uptake curve and in such cases a very dilute solution should be used.
An interesting study using the thin-film approach is provided by the adsorption of poly(methyl methacrylate) on a series of oxidized metal substrates. By careful examination of the XPS C1s spectra it was possible to relate small changes in the relative position of the methoxy and ester components to specific interactions between the polymer and the metal substrates. The type of bonding observed depends strongly on the acido-basic properties of the metal oxide. The adsorption isotherms from these systems were not simple to interpret, as polymer conformation changes as the solution concentration increased gave the erroneous appearance of multilayer deposition.
An elegant example of this type of approach is the recent study of the interaction of an organosilane adhesion promoter ([gamma]-glycidoxypropyltrimethoxysilane, GPS) on aluminum. The concept of a formal covalent bond between aluminum and the organosilane is not new and was first suggested back in the 1970s, but it is only unambiguously identifiable using high-resolution ToF-SIMS. The spectrum of Fig. 1.4 is a high-resolution mass spectrum of nominal mass m/z = 71. The intense peak labelled SiO[Al.sup.+] is indicative of the bonding scheme shown in Scheme 1.1. It should be noted that, in this case, the organosilane was applied to the metallic substrate as a primer (1% aqueous solution) and then cured at 93 °C for 30 min. Evidence is emerging from current work that when the organosilane is included in the formulation of a room temperature curing adhesive, such a reaction is not present at the interface. It seems that there is a subtle synergy between curing agent and organosilane, leading to an interpenetrating network. In a similar vein ToF-SIMS has also been used identify the interaction between an aminosilane and iron surfaces, but this interaction does seem to occur at ambient temperature.
In dealing with commercial systems the investigative scientist is faced with many potential problems but the most significant is that such systems will, in the main, be very complex formulations of many individual components, and for obvious reasons associated with commercial sensitivity, it is unlikely that those outside the manufacturing company will be privy to details regarding the formulation. A typical structural adhesive will have many components in the formulation, some of which are:
liquid epoxy resin
solid epoxy resin
fillers and additives
pigments and dyestuffs
Given this complexity, there is really no option but to investigate some components in isolation before going on to the fully formulated product.
In work on an aerospace structural adhesive, the adsorption characteristics of the liquid epoxy resin (DGEBA) and the curing agent (toluene diisocyante urone, TDI urone) were studied independently on bare aluminum and aluminum treated with GPS. Then the interactions of dilute solutions of the adhesive with the two substrates were studied, enabling a detailed model of the interfacial chemistry to be proposed. The adsorption isotherms for the TDI urone curing agent adsorbed from a dilute solution of the adhesive are presented in Fig. 1.5. It is clear that the adsorption on the bare aluminum surface appears to be twice that of the GPS-treated aluminum. If the intensity of the halved curve obtained from the bare aluminum is halved it is found that it is coincident with that from the GPS-treated substrate. In order to resolve this apparent conundrum it is helpful to consider the source of the ion used to provide the data of Fig. 1.5. The structure of the curing agent is shown in Scheme 1.2 and the ion at m/ z = 58 is assigned to the C[H.sub.3]-NH-C=[O.sup.+] ion that is readily generated at either end of the molecule. The reaction scheme that is thought to occur (Scheme 1.3) shows the manner in which a GPS molecule, bonded to an aluminum substrate, might interact with a curing agent molecule. The immobilization of the TDI urone by reaction with the oxirane ring of the GPS molecule will mean that it is less likely that the bonded end of the amine will yield the characteristic m/z = 58 fragment. In the case of the bare aluminum the curing agent will interact by way of acid-base interactions via, for example, the carbonyl group of the curing agent molecule; as the interaction is not so strong as the covalent bond described above, both ends of the molecule are available to yield the characteristic ion. Further evidence for such an interaction with the GPS-treated substrate is provided by way an ion at m/z = 277 which is formed by scission adjacent to the CH-OH group formed on opening of the oxirane ring.
This type of investigation provides important information regarding the type of bonding that forms at the interface and can be represented in schematic form: Scheme 1.4 indicates a formal covalent bonding from substrate, through adhesion promoter, curing agent, DGEBA, into the bulk of the cross-linked adhesive. Given the comments regarding the hydrodynamic stability of covalent bonds relative to secondary bonds exemplified by the schematic of Fig. 1.2, one would expect superior performance from such a system, and this is indeed the case.
1.3 The Buried Interface
Although the use of the thin-film method to provide a model interphase for analysis has much to commend it, the analysis of such a region formed between adhesive, coating, and substrate is perhaps more attractive in a number of situations. There are a number of options involving electron microscopy and surface analysis but in all cases the specimen preparation is the key to the optimum results. In the case of electron microscopy the most obvious expedient, the use a metallographic cross-section in conjunction with scanning electron microscopy (SEM) and energy-dispersive X-ray analysis (EDX) is ruled out for a number of reasons. Polishing a cross-section will invariably lead to smearing of the polymer phase which will reduce the level of information attainable, and although electron microscopy can be readily carried out the interaction volume of the electron beam with the sample will mean that analytical resolution will be of the order of one micrometer, not nearly good enough to supply an analysis of the interfacial region relevant to adhesion.
Excerpted from Adhesion Copyright © 2005 by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.