Engineered Interfaces in Fiber Reinforced Composites Part 2 pptx

In this mechanism of adhesion, a bond is formed between a chemical group on the fiber surface and another compatible chemical group in the matrix, the formation of which results from usu

14 Engineered interjaces in fiber reinforced composites

reinforcements like glass, silica, and alumina, but are less effective with alkaline surfaces like magnesium, asbestos, and calcium carbonate (Plueddemann, 1974)

of a graphitic-like structure from the fiber surface particularly at low levels of treatment; and the second part is chemical bonding at the acidic sites However, much further work is still needed to verify this hypothesis

In this mechanism of adhesion, a bond is formed between a chemical group on the fiber surface and another compatible chemical group in the matrix, the formation of which results from usual thermally activated chemical reactions For example, a

silane group in an aqueous solution of a silane coupling agent reacts with a hydroxyl group of the glass fiber surface, while a group like vinyl on the other end will react with the epoxide group in the matrix The chemical compositions of the bulk fiber and of the surface for several widely used fiber systems are given in Table 2.2 It is interesting to note that except for glass fibers, the chemical composition of the surface does not resemble that of the bulk fiber, and oxygen is common to all fiber surfaces Further details regarding the types of surface treatments commonly applied to a variety of organic and inorganic fibers and their effects on the properties

of the interfaces and bulk composites are given in Chapter 5

2.2.5 Reaction bonding

Other than in polymer matrix composites, the chemical reaction between elements

of constituents takes place in different ways Reaction occurs to form a new compound(s) at the interface region in MMCs, particularly those manufactured by a molten metal infiltration process Reaction involves transfer of atoms from one or both of the constituents to the reaction site near the interface and these transfer processes are diffusion controlled Depending on the composite constituents, the atoms of the fiber surface diffuse through the reaction site, (for example, in the boron fiber-titanium matrix system, this causes a significant volume contraction due

to void formation in the center of the fiber or at the fiber-compound interface (Blackburn et al., 1966)), or the matrix atoms diffuse through the reaction product

Continued reaction to form a new compound at the interface region is generally harmful to the mechanical properties of composites

Chapter 2 Characterization of interfaces

F, Fe, Na impurities (inner core), borate

B (outer core)

C (outer core), 0, N

Boron (B/W core) W 2 B ~ , WB4 B z 0 3 as methyl B-OH, B-0-B

Silicon carbide Si, W (inner core), Si, C Si-0-Si, Si-OH

(SiC,/W core)

"After Scolar (1974)

Special cases of reaction bonding include the exchange reaction bond and the oxide bond The exchange reaction bond occurs when a second element in the constituents begins to exchange lattice sites with the elements in the reaction product

in thermodynamic equilibrium (Rudy, 1969) A good example of an exchange reaction is one that takes place between a titanium-aluminum alloy with boron fibers The boride compound is initially formed at the interface region in an early stage of the process composed of both elements This is followed by an exchange reaction between the titanium in the matrix and the aluminum in the boride The exchange reaction causes the composition of the matrix adjacent to the compound

to suffer a loss of titanium, which is now embedded in the compound This eventually slows down the overall reaction rate

The oxide bond occurs between the oxide films present in the matching surfaces of fiber and matrix The reaction bond makes a major contribution to the final bond strength of the interface for some MMCs, depending on the fiber-matrix combination (which determines the diffusivity of elements from one constituent to another) and the processing conditions (particularly temperature and exposure time) A general scheme for the classification of interfaces in MMCs can be made based on the chemical reaction occurring between fiber and matrix according to Metcalfe (1974) Table 2.3 gives examples of each type In class I, the fiber and matrix are mutually non-reactive and insoluble with each other; in class 11, the fiber

and matrix are mutually non-reactive but soluble in each other; and in class 111, the fiber and matrix react to form compound(s) at the interface There are no clear-cut definitions between the different classes, but the grouping provides a systematic division to evaluate their characteristics For pseudoclass 1 composites that include B-AI, stainless steel-A1 and Sic-A1 systems, hardly any interaction occurs in solid state diffusion bonding, but a reaction does occur when the A1 matrix is melted for liquid infiltration

In general, in most CMCs, chemical reaction hardly occurs between fiber (or

whisker) and matrix However, an extremely thin amorphous film can be formed,

16 Engineered interfaces in jiber reinforced composites

Table 2.3

Classification of fiber-metal matrix composite systemsa

2.2.6 Mechanical bonding

Mechanical bonds involve solely mechanical interlocking at the fiber surface Mechanical anchoring promoted by surface oxidation treatments, which produce a large number of pits, corrugations and large surface area of the carbon fiber, is known to be a significant mechanism of bonding in carbon fiber-polymer matrix

composites (see Chapter 5) The strength of this type of interface is unlikely to be

very high in transverse tension unless there are a large number of re-entrant angles

on the fiber surface, but the strength in longitudinal shear may be significant depending on the degree of roughness

In addition to the simple geometrical aspects of mechanical bonding, there are many different types of internal stresses present in composite materials that arise from shrinkage of the matrix material and the differential thermal expansion between fiber and matrix upon cooling from the processing temperature Among these stresses, the residual clamping stress acting normal to the fiber direction renders a synergistic benefit on top of the mechanical anchoring discussed above These mechanisms provide major bonding at the interface of many CMCs and play

a decisive role in controlling their fracture resistance and R-curve behavior Further details of these residual stresses are discussed in Chapter 7

Chapter 2 Characterization of interfaces 17

2.3 Physico-chemical characterization of interfaces

2.3.1 Introduction

Composite interfaces exist in a variety of forms of differing materials A

convenient way to characterize composite interfaces embedded within the bulk material is to analyze the surfaces of the composite constituents before they are combined together, or the surfaces created by fracture Surface layers represent only

a small portion of the total volume of bulk material The structure and composition

of the local surface often differ from the bulk material, yet they can provide critical information in predicting the overall properties and performance The basic unknown parameters in physico-chemical surface analysis are the chemical composition, depth, purity and the distribution of specific constituents and their atomic/microscopic structures, which constitute the interfaces Many factors such as process variables, contaminants, surface treatments and exposure to environmental conditions must be considered in the analysis

When a solid surface is irradiated with a beam of photons, electrons or ions, species are generated in various combinations An analytical method for surface characterization consists of using a particular type of probe beam and detecting a particular type of generated species In spectroscopy, the intensity or efficiency of the phenomenon of species generation is studied as a function of the energy of the species generated at a constant probe beam energy, or vice versa Most spectro- scopic techniques are capable of analyzing surface composition, and some also allow

an estimation of the chemical state of the atoms However, it may be difficult to isolate the contributions of each surface layer of the material being probed to these properties Since most surface analysis techniques probe only the top dozen atomic layers, it is important not to contaminate this region For this reason and particularly to reduce gas adsorption, a vacuum always has to be used in conjunction with these techniques The emergence of ultrahigh vacuum systems of less than loT6 Pa (or 7.5 x Torr), due to rapid technological advances in recent years, has accelerated the development of sophisticated techniques utilizing electrons, atoms and ions Amongst the currently available characterization techniques, the most useful ones for composite interfaces are: infrared (IR) and Fourier transform infrared (FTIR) spectroscopy, laser Raman spectroscopy, X-ray

photoelectron spectroscopy (XPS), Auger electron spectroscopy (AES), secondary

ion mass spectroscopy (SIMS), ion scattering spectroscopy (ISS), solid state nuclear magnetic resonance (NMR) spectroscopy, wide-angle X-ray scattering (WAXS),

small-angle X-ray scattering (SAXS) and the measurement of the contact angle A

selected list of these techniques is presented in Table 2.4 along with their atomic

processes and the information they provide Each technique has its own complexity, definite applications and limitations Often the information sought cannot be provided by a single technique This has resulted in the design of equipment that utilizes two or more techniques and obtains different sets of data from the same surface of the sample (e.g ISSjSIMS two-in-one and XPS/AES/SIMS three-in-one equipment) Adamson (1982), Lee (1989), Castle and Watts (1988) and Ishida (1994)

18

have presented excellent reviews of most of these techniques, with Ishida (1994)

being particulalry informative for characterization of composite materials

In addition to surface analytical techniques, microscopy, such as scanning

electron microscopy (SEM), transmission electron microscopy (TEM), scanning

tunneling microscopy (STM) and atomic force microscopy (AFM), also provide

invaluable information regarding the surface morphology, physico-chemical inter-

action at the fiber-matrix interface region, surface depth profile and concentration

of elements It is beyond the scope of this book to present details of all these

microscopic techniques

2.3.2 Infrared and Fourier transform infrared spectroscopy

IR spectroscopy, one of the few surface analytical techniques not requiring a

vacuum, provides a large amount of molecular information The absorption versus

frequency characteristics are obtained when a beam of IR radiation is transmitted

through a specimen IR is absorbed when a dipole vibrates naturally at the same

frequency as the absorber, and the pattern of vibration is unique for a given

molecule Therefore, the components or groups of atoms that are absorbed into the

IR at specific frequencies can be determined, allowing identification of the molecular

structure

The FTIR technique uses a moving mirror in an interferometer to produce an

optical transformation of the IR signal as shown in Fig 2.6 During this operation,

the source radiation is split into two: one half is reflected into the fixed mirror and

the other half transmitted to the moving mirror If the mirrors are placed equidistant

from the beam splitter, their beams will be in phase and reinforce each other In

contrast, the beams that are out of phase interfere destructively An interferogram is

produced from the equations involving the wavelength of the radiation, and a

Fourier analysis is conducted to determine the relation between the intensity and

frequency FTIR can be used to analyze gases, liquids and solids with minimal

preparation and little time This technique has been extensively applied to the study

Fixed mirror

-

Movable

mirror-

Unmodulated incident

Splitter

1 Detector Fig 2.6 Schematic diagram of an interferometry used in the FTIR spectroscopy After Lee (1989)

backscattered from the upper surface of the specimen The electrons in the specimen can also be excited and emitted from the upper surface which are called secondary electrons Both backscatterd and secondary electrons carry the morphological information from the specimen surface The microscope collects these electrons and transmits the signals

to a cathode ray tube where the signals are scanned synchronously providing morphological information on the specimen surface

Environmental SEMs are a special type of SEM that work under controlled environmental conditions and require no conductive coating

on the specimen with the pressure in the sample chamber only 1 or 2 orders magnitude lower than the atmosphere

TEM is composed of comprehensive electron optics, a projection system, and a high-vacuum environment When a portion of high voltage primary electrons is transmitted through an ultrathin sample, they can be unscattered and scattered to carry the microstructural information of the specimen The microscopes collect the electrons with a comprehensive detection system and project the microstructural images onto a fluorescent screen The ultimate voltage for a TEM can generally be from I O to 1000 keV, depending on the requirement of resolving power and specimcn thickness

The STM, like other scanning probe microscopes, relies on the scanning of

a sharp tip over a sample surface When the tip and sample are very close

so that the electron clouds of tip and sample atoms overlap, a tunneling current can be established through voltage differences applied between the

two electrodes When a raster scan is made, the relative height coordinate z

as a function of the raster coordinate x and y reflects the surface

topography of the sample The STM is limited to conducting materials a s

it is based on the flow of electrons

In AFM, a sharp tip integrated with a soft spring (cantilever) deflects as a result of the local interaction forces present between the apex of the tip and

the sample The deflection of this cantilever can be monitored at its rear

by a distance sensor The forces existing between tip and sample, when they are close, can be van der Waals, electrostatic or magnetic force Atomic-scale friction, elasticity and surface forces can also be measured AFM can be employed for both conductive and non-conductive specimens, without having to apply a high vacuum, presenting a major advantage over STM

20 Engineered interfaces in Jiber reinforced composites

Infrared (IR) and Fourier

transform infrared (FTIR)

spectroscopy

Raman spectroscopy (RS)

The sample surface is bombarded with an incident high energy electron beam, and the action of this beam produces electron changes in the target

atoms; the net result is the ejection of Auger electrons, which are the

characteristics of the element Because of the small depth and small spot size of analysis, this process is most often used for chemical analysis of microscopic surface features

When a sample maintained in a high vacuum is irradiated with soft X-rays, photoionization occurs, and the kinetic energy of the ejected photoelectrons is measured Output data and information related to the

number of electrons that are detected as a function of energy are generated

Interaction of the soft X-ray photon with sample surface results in ionization from the core and valence electron energy levels of the surface elements

The sample surface is bombarded with a beam of around 1 keV ions of

some gas such as argon and neon The action of the beam sputters atoms from the surface in the form of secondary ions, which are detected and analyzed to produce a characterization of the elemental nature of the surface The depth of the analysis is usually less than a nanometer, making this process the most suitable for analyzing extremely thin films

I n ISS, like in SIMS, gas ions such as helium or neon are bombarded on

the sample surface at a fixed angle of incident The ISS spectrum normally

consists of a single peak of backscattered inelastic ion intensity at an energy loss

that is characteristic of the mass of surface atom From the pattern of scattered

ion yield versus the primary ion energy, information about elements present on

the sample surface can be obtained at ppm level

The absorption versus frequency characteristics are obtained when a beam

of IR radiation is transmitted through a specimen The absorption or emission of radiation is related to changes in the energy states of the

material interacting with the radiation In the IR region (between 800 nm

and 250 pm in wavelength), absorption causes changes in rotational or

vibrational energy states The components or groups of atoms that absorb

in the IR a t specific frequencies are determined, providing information about the molecular structure The FTIR technique employs a moving mirror to produce an optical transformation of the IR signal, with the beam intensity after the interferometer becoming sinusoidal FTIR has been

extensively used for the study of adsorption on polymer surfaces, chemical modification and irradiation of polymers on the fibersurfaces

The collision between a photon of energy and a molecule results in two different types of light scattering: the first is Rnyleigh scattcring and the second is Raman scattering The Raman effect is an inelastic collision where the photon gains energy from or loses energy to the molecule that corresponds to the vibrational energy of the molecule Surface-enhanced Raman spectroscopy has been successfully used to obtain information about adsorption of polymers onto metal surfaces, polymer-polymer interaction and interdiffusion, surface segregation, stress transfer at the fiber-matrix interface, and surface structure of materials

Chapter 2 Characterization of interfaces 21

"After Adamson (1982), Lee (1989) and Ishida (1994)

of adsorption on surfaces of polymers (Lee, 1991) and of chemical modification and

irradiation of polymers on the fiber surfaces, including silane treated glass fibers (Ishida and Koenig, 1980; Garton and Daly, 1985; Grap et al., 1985; Miller and

Ishida, 1986; Liao, 1989; DeLong et al., 1990) Fig 2.7 shows typical IR spectra of

glass fiber-epoxy matrix composites with and without an amino silane coating on the fiber

2.3.3 Laser Raman spectroscopy

Laser Raman spectroscopy uses a light scattering process where a specimen is irradiated monochromatically with a laser The visible light that has passed into the specimen causes the photons of the same wavelength to be scattered elastically, while

22 Engineered interfaces in jiber reinforced composites

it causes the light of slightly longer or shorter wavelengths to be scattered inelastically The inelastic proportion of the photons imparts energy to the molecules, which are collected for analysis An interesting feature of the Raman spectroscopy is that certain functional groups or elements scatter incident radiation

at characteristic frequency shifts The vibrational frequency of the group or element

is the amount of shift from the exciting radiation Functional groups with high polarizability on vibration can be best analyzed with Raman spectroscopy

Raman and IR spectroscopies are complementary to each other because of their

different selection rules Raman scattering occurs when the electric field of light induces a dipole moment by changing the polarizability of the molecules In Raman spectroscopy the intensity of a band is linearly related to the concentration of the species IR spectroscopy, on the other hand, requires an intrinsic dipole moment to exist for charge with molecular vibration The concentration of the absorbing species is proportional to the logarithm of the ratio of the incident and transmitted intensities in the latter technique

As the laser beam can be focused to a small diameter, the Raman technique can

be used to analyze materials as small as one micron in diameter This technique has been often used with high performance fibers for composite applications in recent years This technique is proven to be a powerful tool to probe the deformation behavior of high molecular polymer fibers (e.g aramid and polyphenylene benzobisthiazole (PBT) fibers) at the molecular level (Robinson

et al., 1986; Day et al., 1987) This work stems from the principle established earlier by Tuinstra and Koenig (1970) that the peak frequencies of the Raman- active bands of certain fibers are sensitive to the level of applied stress or strain The rate of frequency shift is found to be proportional to the fiber modulus, which

is a direct reflection of the high degree of stress experienced by the longitudinally oriented polymer chains in the stiff fibers

In the case of carbon fibers, two bands are obtained: a strong band at about

1580 cm-' and a weak band at about 1360 cm-', which correspond to the Ezs and

AI, modes of graphite (Tuinstra and Koenig, 1970) The intensity of the Raman- active band, AI^ mode, increases with decreasing crystalline size (Robinson

et al., 1987), indicating that the strain-induced shifts are due to the deformation

of crystallites close to the surfaces of the fibers The ratio of the intensities of the two modes, Z(Alg)/Z(Ezg), has been used to give an indirect measure of the crystalline size in carbon fibers (Tuinstra and Koenig, 1970) Table 2.5 gives these ratios and the corresponding average crystal diameter, La, in the graphite plane, as determined

by X-ray techniques Typical examples of strain dependence of the Raman frequencies is shown in Fig 2.8 for two different carbon fibers, and the corresponding plots of the shifted Raman frequency are plotted as a function of the applied strain in Fig 2.9

Enabled by the high resolution of spectra, which is enhanced by the use of spatial filter assembly having a small (200 pm) pin hole, the principle of the strain-induced band shift in Raman spectra has been further extended to the measurement of residual thermal shrinkage stresses in model composites (Young et al., 1989; Filiou

et al., 1992) The strain mapping technique within the fibers is employed to study the

Chapter 2 Characterization of interfaces 23

Table 2.5

Intensity ratio of Raman bands I(AI,)/I(E2J and the corresponding apparent crystal diameter, La, for

various carbon fibers"

Thornel 10 Union Carbide

Raman Frequency (ern-')

Fig 2.8 Laser Raman spectra obtained (a) for a polyacrylonitrile (PAN)-based HMS4 carbon fiber, and

(b) for a pitch-based P75S carbon fiber After Robinson al (1987)

24 Engineered interfaces in jiber reinforced composites

in an epoxy matrix is shown in Fig 2.10 for different levels of applied strain The

IFSS is calculated based on the force balance between fiber axial direction and interface shear

2.3.4 X-ray photoelectron spectroscopy

XPS, also known as electron spectroscopy for chemical analysis (ESCA), is a

unique, non-destructive analytical technique that provides information regarding the chemical nature of the top 2-10 nm of the solid surface with outstanding sensitivity and resolution In XPS, the solid surfaces are subjected to a beam of almost monochromatic X-ray radiation of known energy in a high vacuum environment (4 x 10-9-1 x lop8 Torr) Electrons are emitted from the inner orbital with kinetic energies characteristic of the parent atoms The intensities of the kinetic energy are analyzed and the characteristic binding energies are used to determine the

chemical composition The total absorbed X-ray photon energy, hv, is given by the sum of the kinetic energy, E K , and the electron binding energy, EB

Once the kinetic energy is measured with an electron spectrometer for a given X-ray photon energy, the binding energy characteristic of the parent atoms can be directly determined The electron binding energy represents the work expended to remove an electron from a core level of the inner orbital to the Fermi level in its removal from the atom Peaks in the plots of electron intensity versus binding energy correspond

to the core energy levels that are characteristic of a given element

Chapter 2 Characterization of interfaces 25

Fig 2.10 Fiber strain and interfacial shear stress (IFSS) profiles along the fiber length for a heat-treated

Kevlar 49 fiber-poxy resin composite At applied strains of (a) 0.60% (b) 1.90% and (c) 2.5% After

Galiotis (1993a,b)

26 Engineered interfaces in jiber reinforced composites

Table 2.6

XPS analysis, elemental composition of carbon fibers"

"After Cazeneuve et al (1990)

In XPS, only large areas can be analyzed because X-rays are difficult to focus with sufficient intensities on a small target area Signals from small regions of a heterogeneous solid surface are usually weak and difficult to isolate For these reasons, X P S is not well suited to depth profiling One significant recent advance is the development of the X-ray monochromator, which collects some of the X-rays from a conventional source and refocuses them on the sample This allows a small sample area to be illuminated and analyzed with X-rays, resulting in an increased ability to distinguish different chemical states Another innovation is the addition of

a parallel detection system, which has the abiIity to collect simultaneously all the points of a special range, substantially increasing the speed and sensitivity of the instrument The conventional unit, which contains a single exit slit, is able to collect only a single point

Applications of XPS for composite interface studies include the quantitative assessment of the local concentration of chemical elements and functional groups that are required to evaluate the contributions of chemical bonding at the fiber- matrix interface region in polymer matrix composites (Yip and Liu, 1990; Baillie

et al., 1991; Nakahara et al., 1991; Shimizu et al., 1992; Kim et al., 1992; Wang and Jones, 1994) Fig 2.1 1 shows examples of XPS spectra obtained for carbon fibers with and without surface sizing The corresponding elemental compositions of these fibers are given in Table 2.6 The main difference between the sized and unsized carbon fibers is the quantity of nitrogen (Le 5.3% and 0.8% in unsized and sized fibers, respectively), which is considered to originate from the residue of a polyacrylonitrile (PAN) precursor or from the surface treatment at the end of the manufacturing process (Cazeneuve et al., 1990) To identify functional groups present on the fiber surface, the small chemical shifts are analyzed to obtain information of oxidation states and the overlapping peaks are deconvoluted (Kim

et al., 1992) This means that the larger the chemical shifts the easier the identification of functional groups However, certain functional groups can be difficult to distinguish, e.g carboxylic acids, esters, alcohols, and aldehydes, which

all contain a carbonyl oxygen and as a result have overlapping C1, spectra

2.3.5 Auger electron spectroscopy

AES is similar to XPS in its function, but it has unparalleled high sensitivity and spatial resolution (of approximately 30-50 nm) Both AES and XPS involve the identification of elements by measurement of ejected electron energies Fig 2.12

Chapter 2 Characterization of interfaces

Binding energy (eV)

Fig 2.1 I Spectra of (a) unsized and (b) sized T300 carbon fibers which are obtained from XPS After

Cazeneuve et al (1990)

compares the reactions in XPS, AES, SIMS and ISS, and the latter two techniques

will be discussed in the following sections In AES, it is possible to focus an electron

beam laterally to identify features less than 0.5pm in diameter and into a monolayer

in thickness In addition, by simultaneous use of analytical and sputter etching, it

may provide composition profiles However, the AES electron beam is highly

concentrated with high flux density and beam energy, which can damage the

polymer surface causing pyrolysis during measurement This makes it difficult to

employ AES technique on a thin film In this regard, X P S is a more delicate

technique as the power required is an order of magnitude lower than in AES

28 Engineered interfaces in fiber reinforced composites

SIMS and ISS Ion

Excitation

XPS Auger Electron Elec ‘&on

Fig 2.12 A comparison of XPS, AES, SIMS and ISS reactions After Lee (1989)

In AES, an energetic beam of electrons strikes the atoms of the sample in a vacuum and electrons with binding energies less than the incident beam energy may

be ejected from the inner atomic level, creating a single ionized excited atom This irradiation causes ejection of orbital electrons from the sample and the resulting excited atom either emits an X-ray (fluorescence) or an electron is ejected from the atom (Auger process) This vacancy is filled by de-excitation of electrons from other electron energy states The energy released can be transferred to an electron in any atom If this latter electron has a lower binding energy than the energy from the de- excitation, then it will be ejected with its energy related to the energy level of the separation in the atoms Auger electrons are the result of de-excitation processes of these vacancies and electrons from other shells and re-emission of an electron to carry away excess energy The electrons emitted have a short mean free path, and thus all Auger electrons are from the first few atomic surface layers The kinetic energies of the free electrons are detected and they reflect the variations in binding energies of the levels involved in the process

The Auger electron spectra shown in Fig 2.13 contain peaks corresponding to the intensity of Auger electrons as a function of kinetic energy These electrons are emitted following the creation of a core hole in the electron shells by radiation of an incident electron beam The kinetic energy is independent of the energy of the incident beam, and the intensity of an Auger peak relates to the concentration of

atoms or ions in the volume being analyzed As in XPS, changes in chemical and

oxidation states are reflected by the shifts in the peak position Whether or not the chemical state can be recognized depends on the width of the Auger peak A very wide peak cannot be used to provide information on the chemical state The intensity of a peak or the peak area is a complex function of the angle of incidence and the current of the primary beam, the inelastic mean free path of the escaping electron, the local angle of the detected electrons, etc It is essential to understand these factors to conduct proper composition analysis