JUSTIN I. WALKER ALL RIGHTS RESERVED

Transcription

1 2008 JUSTIN I. WALKER ALL RIGHTS RESERVED

2 SPECTROSCOPIC ANALYSIS OF MATERIALS FOR ORTHOPAEDIC AND ENERGY CONVERSION APPLICATIONS A Thesis Presented to The Graduate Faculty of The University of Akron In Partial Fulfillment of the Requirements for the Degree Master of Science Justin I. Walker December, 2008

3 SPECTROSCOPIC ANALYSIS OF MATERIALS FOR ORTHOPAEDIC AND ENERGY CONVERSION APPLICATIONS Justin I. Walker Thesis Approved: Accepted: Advisor Rex D. Ramsier Dean of the College Ronald F. Levant Committee Member Edward A. Evans Committee Chair Robert R. Mallik Dean of the Graduate School George R. Newkome Date Department Chair Robert R. Mallik ii

4 ABSTRACT X-ray photoelectron spectroscopy (XPS) and electron dispersive X-ray spectroscopy (EDS) are two complimentary techniques that are used to obtain data about the chemical and electronic states of materials. The information that XPS acquires is from the surfaces of samples, while EDS gets information from the bulk. Additionally, scanning electron microscopy (SEM) is a method to get more qualitative information about substances, by taking high magnification images. Many materials can be analyzed using these methods, but two are detailed in this thesis: electrospun aluminum oxide doped zinc oxide (AOZO) nanofibers and cobalt chromium molybdenum (CoCrMo) metal powder particulates. AOZO nanofibers are on the forefront of material research. XPS was utilized to discover how the amount of aluminum dopant, used in fabricating the fibers, affected the fibers conductivity. CoCrMo is used to make prosthetic implants and the elemental structure is a significant factor in determining biocompatibility. XPS and EDS data was compared to ascertain discrepancies in composition, and how that affected biocompatibility. iii

5 ACKNOWLEDGEMENTS Charles Barr Edward Bender George Chase James Ehrman Edward Evans Rajendra Kamal Yong-Cheol Kang Mark Kovacik Adria Lotus Robert Mallik Lisa Park Juyun Park Soo-Jin Park Tom Quick Rex Ramsier iv

6 TABLE OF CONTENTS Page LIST OF FIGURES... vi CHAPTER I. BACKGROUND...1 II. CHARACTERIZATION METHODS...4 X-ray Photoelectron Spectroscopy...4 Scanning Electron Microscopy...11 Electron Dispersive X-ray Spectroscopy...21 III. APPLICATIONS...23 Characterization of Alumina Doped Zinc Oxide Nanofibers...23 Characterization of CoCrMo Metal Powders...33 IV. SUMMARY...41 REFERENCES...43 APPENDICES...46 APPENDIX A. STANDARD OPERATING PROCEDURE FOR THE VG ESCALAB MK II...47 APPENDIX B. STANDARD OPERATING PROCEDURE FOR THE FEI QUANTA 200 ENVIRONMENTAL SCANNING ELECTRON MICROSCOP WITH EDAX...53 v

7 LIST OF FIGURES Figure Page 1 An illustration of an X-ray producing a photoelectron Example XPS survey spectrum, obtained using a VG ESCALAB MK II The main constituents of an XPS system The typical construction of an X-ray source Photoelectron path The electron multiplier Example SEM image A simplified illustration of an SEM The electron multiplier Side view of an esem Illustration of an EDS instrument Elecrospinning instrument High resolution X-ray photoelectron spectra of AOZO nanofibers Example deconvoluted XP spectra Electrical conductivities of AOZO nanofibers Representative SEM image of AOZO nanofibers Energy dispersive X-ray spectra for AOZO nanofibers SEM images of CoCrMo particulate powders...35 vi

8 19 ED spectra for CoCrMo particulate powders XP spectra for CoCrMo particulate powders Table of bulk composition percentages of CoCrMo powders Table of surface composition percentages of CoCrMo powders Cell activity after CoCrMo exposure A top-down schematic of the XPS chamber layout...49 vii

9 CHAPTER I BACKGROUND Various characterization methods are used to determine a wide array of information about a multitude of materials. X-ray photoelectron spectroscopy (XPS) is among the most common techniques for obtaining chemical and electronic information from materials surfaces. Scanning electron microscopy (SEM) is used to obtain high resolution images of samples. Also, energy dispersive X-ray spectroscopy (EDS) is another characterization technique, which is closely connected to SEM instruments, that provides chemical information about the bulk of materials. XPS is an important spectroscopic technique used in surface science. The surface is the site of interaction for most systems. XPS has become a standard due to its surface sensitivity and tendency to minimize degradation of materials. An XPS instrument produces X-rays that are directed toward a sample, which then emits photoelectrons. Only the photoelectrons that are within about 10 nanometers of the surface are able to leave the sample and make it to an analyzer. SEM instruments produce detailed images, with nanometer spatial resolutions, of materials by bombarding samples with a beam of electrons, then analyzing the 1

10 backscattered and secondary electrons which are emitted from the sample. A magnification of up to 250,000 times is possible using this method, which is far beyond the capability of any optical microscope. X-rays are also emitted using the same technique, which can be analyzed as well. EDS is a spectroscopic technique that can be performed simply by adding an attachment to an SEM instrument, in order to collect the emitted X-rays. These X-rays are from deeper within the sample, from the first few microns, giving elemental information about the bulk of materials. When combined with XPS, this method can provide more complete chemical profiles of materials. The materials that can be studied by these methods include many condensed matter materials that can be made to fit into the instruments. Naturally there are certain other limitations, which are discussed later, but most substances can be analyzed. Each method has its own caveats when used on any given class of materials. This thesis details studies of two substances: aluminum oxide doped zinc oxide nanofibers and CoCrMo metal powder particulates. Aluminum oxide doped zinc oxide (AOZO) nanofibers are a type of material still in development that have a variety of possible applications. The link between the conductivity of AOZO nanofibers and aluminum concentration, using chemical data collected by XPS, is presented in a subsequent chapter. The nanofibers discussed in that chapter were created using electrospinning, which is also explained. Cobalt chromium molybdenum (CoCrMo) metal powders are one class of materials used in the creation of prosthetic implants. These materials are used for their biocompatibility, or their ability to be accepted by the body while minimizing chances of 2

11 infection, inflammation, or outright rejection by the body. This thesis presents a study comparing the surface composition of these powders with their bulk composition, in order test the link between surface elemental makeup and cell viability. 3

12 CHAPTER II CHARACTERIZATION METHODS X-ray Photoelectron Spectroscopy XPS is a characterization technique used to find the elemental composition of materials' surfaces along with the chemical and electronic state of the atoms in those materials. When XPS was first developed in the mid-1960s, it was referred to as ESCA: Electron Spectroscopy for Chemical Analysis. It can be used to detect elements from lithium to lawrencium [1]. XP spectra are recorded by bombarding a sample with a focused beam of X-rays normally from an aluminum or magnesium source. The count of the number of electrons that make it to a detector, which is proportion to those that are ejected from the sample, is plotted versus the binding energy of those electrons. The emitted electrons' binding energy is described by the equation: E binding = E photon - E kinetic Φ sp, where E photon is the incident X-ray energy, E kinetic is the measured kinetic energy of the emitted electron, and Φ sp is the spectrometer's work function. Φ sp is determined by factors such as the geometry and materials of the instrument and is unique to each spectrometer [1]. 4

13 This technique is surface sensitive, ejecting photoelectrons from roughly the first 10 nanometers of the surface, due to the short free mean path of electrons in normal materials [2]. XPS requires an ultra-high vacuum (10-9 millibar) so that the electrons do not have to interact with atmosphere on the way to the analyzer, and so that the surfaces of the samples stay clean. Since the sample has to be placed in such a high vacuum, certain materials cannot be analyzed with this technique. For instance, liquid samples are not usually analyzed because they would evaporate upon entering the machine without special treatment. Figure 1: An illustration of an X-ray producing a photoelectron. In this case, the X-ray ejects an electron from the 1s level of a boron atom. These electrons produce the boron s characteristic 189 ev peak in XP spectra. When electrons are ejected from the core of atoms by the X-rays, they can be replaced by higher level electrons. The energy lost from the electron when it falls from one level to another will most likely be transferred to another electron, which is ejected from the atom. These electrons from this Auger effect also show up in the XP spectra, but do not usually interfere with analysis. These peaks are labeled by A(XYZ), where A is the element, X is the level of the electron ejected by the X-ray, Y is the fallen electron's original level, and Z is the Auger electron's level [1]. There are 3 general consequences of photon-electron collision, which depend on the initial energy of the photons. For energies above 2 MeV, pair production occurs, where an elementary particle and its antiparticle are created. When the photon has an 5

14 Relative Intensity (arb. units) units) energy between 20 kev and 2 MeV, Compton Scattering occurs, and the electron gains some kinetic energy from the photon. The photo ionization regime involves collisions between photons that have energy less than 20 kev [3]. XPS falls in to this regime, since it typically employs X-rays with energies of ev. Generally, XP spectra involve plotting ejected electron count versus the binding energy of those electrons. Each element has a distinctive set of XPS peaks at a particular binding energy. These peaks correspond to electron levels in the sample's atoms. The areas under the peaks, once divided by a sensitivity factor unique to each peak, directly relate to the amount of each element. These factors are determined experimentally and are used to normalize the intensities of different elemental peaks. The sensitivity factors arise from photoionization probabilities and are well known [4]. Silicon Carbide XP Spectrum C KVV O 1s C 1s 0.8 O KLL 0.7 Si 2s Si 2p Carbon 60.8% ± 25.3% Oxygen 11.3% ± 2.8% Silicon 27.9% ± 7.0% Binding Energy (ev) Figure 2: Example XPS survey spectrum, obtained using a VG ESCALAB MK II. This spectrum for silicon carbide has elemental characteristic peaks, as well as Auger peaks. 6

15 The XP spectra contain more than just these characteristic elemental peaks. As stated before, Auger lines are common in XP spectra, along with X-ray satellites, and continuous backgrounds. X-ray satellites arise from non-monochromatic sources, which manifest as lines below the characteristic peak. The most notable of these are the α 3 and α 4 satellites that total 10 percent of the peak and appear 9.8 ev and 11.8 ev below it, respectively, for an aluminum anode. The background arises from inelastic collisions of photoelectrons and Auger electrons with atoms, ions, or electrons in the sample [1]. An XPS system consists of the following components: an ultra high vacuum (UHV) stainless steel chamber with UHV pumps, an X-ray source, an electron collection lens, an electron energy analyzer covered in Mu-metal magnetic field shielding, an electron detector system, sample mounts, a sample stage, and a set of stage manipulators. Figure 3: The main constituents of an XPS system. The UHV chamber is kept at a high vacuum with pumps, so that the X-rays can make the way to the sample from the X-ray source. The photoelectrons ejected from the sample travel through the hemispherical analyzer to the detector, with minimal scattering. 7

16 The ultra high vacuum system centers on a stainless steel chamber, which is used for shielding as well as its structural integrity. Ultra high vacuum pumps, such as ion pumps, diffusion pumps, and titanium sublimation pumps, are used to evacuate the chamber. These pumps are backed by turbomolecular pumps and roughing pumps. Generally, a non-monochromatic X-ray source with a millimeter diameter beam is used. This naturally leads to a low spatial resolution. Electrons are accelerated from a filament to an anode coated with magnesium or aluminum by a 10 kv potential. This anode is usually just 1-5 centimeters from the sample. The X-rays lines produced from the collision with the magnesium or aluminum are at ev or ev respectively. These non-monochromatic X-ray sources also produce Bremsstrahlung X- rays, which is caused by slowing electrons and makes up part of the XP spectra background [1]. Figure 4: The typical construction of an X-ray source. The water connections allow deionized water to be pumped through the anode, to prevent overheating. Taken after [5]. 8

17 Electrons ejected from the material pass through an electron collection lens on the way to the electron analyzer. Usually, the analyzer is constructed of two concentric hemispheres. These hemispheres are electrically isolated and kept at controlled voltages, forming an electrostatic field, which allows only electrons with certain energies to pass completely through. Figure 5: Photoelectron path. The electrons ejected from the sample travel up an UHV chamber, past electromagnetic lenses and apertures to the Hemispherical Analyzer. The voltage difference between the two hemispheres determines which electrons are allowed through to the detector. 9

18 The electrons that do make it through are funneled into the electron multiplier. The multiplier is a spiral glass tube with a cone funnel entrance and a conductive coating on the inside, which has a high voltage applied to it. Electrons hitting the cone or the inside of the spiral create more secondary electrons in an avalanche effect. The electrons are accelerated to the end of the spiral by the high voltage, causing a current. The current pulses are again amplified then counted by circuit, which is the data that is extracted from the system [6]. Figure 6: The electron multiplier. The funnel at the opening directs electrons into the spiraling glass tube. High voltage connections are at each end of the glass tube. The quantitative accuracy of this method depends on many factors: signal to noise ratio, peak sensitivity, the accuracy of sensitivity factors, correction for electron transmission function, correction for electron mean free path energy dependence, surface homogeneity, and sample degradation. Considering all these factors, the method is credited with 10-20% uncertainty for calculating atomic percentages, under routine conditions [1]. 10

19 There are some other limitations associated with XPS. To start, the elemental detection is limited to 1,000-10,000 parts-per-million [1]. Also, the technique is more sensitive to core level electrons than outer shell electrons. And, as stated before, X-ray beams are at least several millimeters in diameter, so this limits the ultimate spatial resolution [7]. The sample sizes for older instruments are generally small, on the order of 1x1 to 3x3 centimeters. Scanning Electron Microscopy Scanning electron microscopy is a characterization technique used to view detailed images of materials surfaces. This technique involves bombarding the sample with electrons in order to form an image, which leads to much higher spatial resolution than is possible with standard optical microscopes. Electron microscopes were first developed in the 1930s in Germany, but the first commercial SEMs were not available until 1965 [8]. SEM images are obtained by focusing a fine (nanometer-sized) beam of electrons on the surface of a material. The beam diameter is several nanometers wide, and once it reaches the sample it is referred to as the electron spot size. This bombardment releases secondary electrons, backscattered electrons, characteristic X-rays, and light from cathodoluminescence. The secondary electrons are usually collected to form the raster scan images of surfaces. The intensity of the signal registered by the SEM relies on shape, composition, volume, and crystal orientation within the material [9]. 11

20 Figure 7: Example SEM image. This is an image of powdered metal particles, which are discussed in the next chapter. SEMs can produce images that magnify samples surfaces by 15 to 250,000 times, or more. In simplest terms, the magnification is determined by the ratio of the raster dimensions to the display dimensions. Since the display dimension can be considered constant, diminishing the raster dimension increases the magnification. So, unlike conventional optical microscopes, the objective lens power has no effect on magnification [9]. SEMs are able to obtain greater spatial resolutions than optical microscopes due to the de Broglie wavelength of electrons, given by λ = h/mυ, where λ is the electrons wavelength, h is Planck s constant, m is the electrons mass, and υ is the electrons 12

21 velocity. Typically, the wavelength of electrons is much smaller than the wavelength of photons, due to their larger momentum. The apparent brightness of the image depends on the number of secondary electrons that make it to the detector. If the beam is incident upon the material normal to the surface, then the region from which electrons escape from the material is uniform. As the angle of incidence increases, the region from which the electrons escape enlarges on one side of the beam, which causes steep areas of the resulting image to appear much brighter, and the image gains a distinct, three-dimensional look [9]. There are several other image-related parameters, beyond the magnification and brightness, which are important in scanning electron microscopy. The depth of focus, or the vertical distance that is clear on the image, is 100 times the beam diameter or more. The spatial resolution is determined by the electron spot size, which depends on electron wavelength and lens system in the instrument. SEMs can image nanometer-sized objects, while X-ray analysis is limited to micron-sized objects [9]. The most common method to produce images on a standard SEM system is secondary electron imaging. The electrons impinging on the sample penetrate to micron depth, creating the interaction volume. This interaction volume is where electrons are repeatedly absorbed and scattered within the sample. However, the secondary electrons are only emitted from the first few nanometers from the surface [10]. They are called 'secondary' electrons because they arise due to another form of radiation, referred to as the primary radiation. The primary radiation can be in the form of electrons, ions, or photons with enough energy to cause ionization. The secondary electrons, which have less than 50 ev when they emerge, are counted by an Everhart-Thornley detector [11]. 13

22 Another method for producing SEM images involves backscattered electrons. The detectors that are required for these types of images differ from the detectors required for secondary electron images because the electrons are deflected to different areas. These electrons have a kinetic energy greater than 50 ev, and are electrons that are reflected from the interaction volume. The image formed using this method depends on the position and energy sensitivity of the detector, and the angle of incidence [9]. Backscattered electron images can be divided into three subcategories. Highdeflection backscattered electron images are created from collecting electrons from a sample that have been deflected by more the ninety degrees. Low-deflection backscattered electron images are created from collecting electrons that have been deflected less than ninety degrees. These images have better resolution than highdeflection images. Low-loss backscattered electron images are created by collecting electrons that have lost less than 1 ev, while being backscattered. This is achieved using oblique angles of incidence and energy filters. Low-loss images have better resolution than both high-deflection and low-deflection images [9]. There are other types of images that can be produced by scanning electron microscopes. One type of image is a specimen-current image, which involves collecting the current that runs from sample to ground. Another type of image is the induced-signal image. This is used for semiconductors and dielectrics in order to view surface topography, crystal defects, and p-n junctions. Even more types of images can be created, as the SEM can be a versatile characterization tool [9]. Scanning electron microscopes share some components and design similarities with XPS instruments. Like XPS systems, SEMs must be operated under vacuum. Low 14

23 pressure is needed to limit corrosion and ion bombardment of the emitting surface in the electron gun, in addition to allowing the electrons to travel freely in the instrument. Usually, gases in SEMs are pumped out by oil diffusion pumps that are backed by rotary mechanical pumps. This lowers the pressure in the instrument to around 10-5 millibar. Using cold traps and airlocks, the pressure can be lowered to 10-9 millibar. To operate a field emission electron gun, which is addressed below, an ion pump is necessary [9]. Figure 8: A simplified illustration of an SEM. Most of the functions in an SEM are controlled by a computer, in modern SEMs. The scanning coils raster the electron beam and the electron spot size, while the column blanking coils act as a gate for the electron beam. Taken after [9]. Naturally, instead of having an X-ray gun as a source, an SEM possesses an electron gun. Electrons are emitted thermionically from a tungsten filament and accelerated toward an anode. Tungsten is ideal for this application because it has the 15

24 lowest vapor pressure and highest melting point of all metals. A current of 1 microamp is supplied to the electron source for X-ray microanalysis, and a current of 1 to 10 picoamps for a 10 nanometer electron spot size. The pressure in an SEM with a tungsten filament must be less than 5 x 10-5 millibar, and at such a pressure the filament will last tens of hours [9]. Another electron source is the rod-shaped lanthanum hexaboride (LaB 6 ) cathodes. These sources have a three nanometer electron spot size, and X-ray analysis is also possible. The vacuum must be slightly better, about 10-6 millibar, which requires a vacuum system upgrade for tungsten filament systems. Under operating conditions, it has a lower temperature and higher current density that tungsten [12]. Optimally, the rod must have a sharp tip, and the heating coil must be precisely placed. The heat from the rod can corrode supporting metals, unless they are properly designed. These sources last hundreds of hours [9]. Electrons can also be created using a field emission gun. This is the brightest of the sources, operating at 0.1 nanoamps. However, it requires the lowest pressure, 10-9 to millibar. Also, field emission guns are poor for X-ray analysis and only last for a few hours [9]. Regardless of what source is used, the electrons will be channeled through magnetic lenses. The final lens is normally a pinhole lens, which has a smaller hole on the side closest to the specimen. This minimizes the magnetic field at the specimen in order to maximize secondary electron escape. The focal length of the final lens is less than 1 centimeter [9]. 16

25 Once electrons have been ejected from the sample, they must be collected. The detector is composed of a scintillator within an open-faced metal box. The open face is covered by a grid held at a large, positive potential, in order to accelerate secondary electrons toward the detector. To collect backscattered electrons, a grounded grid is used. Scintillators are materials that give off light after ionization events, which can be made of a variety of materials. The light emitted from the scintillator is channeled through a light pipe to strike a photomultiplier. This detector design is especially useful because it is effectively noise free [9]. Figure 9: SEM secondary electron detector. Positive voltage, on the order of hundreds of volts, is applied to the metal box of the electron detector, so that the electrons leaving the sample are accelerated toward it. Taken after [9]. The pulses from the photomultiplier are then processed by electronics and a computer. The standard method of display is an intensity-modulated image, in which the 17

26 brightest parts of images correspond to areas of greatest signal. These areas are where the number of scattered electrons is the highest, which, as mentioned before, are the steepest parts of the sample [9]. Samples in an SEM must be conductive so that electrons striking the sample have a path to ground. Samples that are not naturally conductive can be coated with a conductive coating. This coating can be graphite, gold, platinum, or tungsten, and is deposited by high vacuum evaporation or low vacuum sputter coating. This conductive layer also increases contrast, even for already conductive materials. It is possible to obtain images of nonconductive samples without conductive coating, if an environmental scanning electron microscope (esem) or a field emission gun is used [9]. The goal of raising the pressure at which SEM could be conducted was primarily pursued through the 1970s and 1980s. Raising the pressure in an SEM makes it possible to image nonconductive samples, as well as samples that would otherwise evaporate to be studied was the year that the first commercial esem instruments were debuted [13]. Once key patents expired, Philips and FEI added systems to the market. The basic design of an esem is the same as an SEM, but they differ in important ways. Through differential pumping, the sample chamber is kept at high pressure, while the chamber housing the electron source is kept at high vacuum. These two chambers are separated by pressure limiting apertures. Most of the gas that makes it through the first aperture, into an intermediate pressure chamber, is pumped out. Whatever gas that makes it through to the low pressure electron gun chamber can be pumped out at a sufficient rate for even field emission guns to function [13]. 18

27 The electron beam degrades as it travels from the source to the sample. The vacuum in the electron gun chamber preserves the electron beam, but as the electrons travel through the intermediate chamber they start to scatter. While the beam travels through the sample chamber, a significant number of electrons are scattered by the gas. Before the electron beam has been so diffused that a discernable signal cannot be obtained, a measureable amount of electrons strike the sample. At this point, the diffused electrons are evenly spread out, so they only contribute a manageable background signal to the image [13]. There is a limit to the distance the electron beam can travel in the sample chamber, for a given source voltage and gas, and still create a useful image. This distance must be short enough that the equation, P specimen *d 1 pascal meter, holds true. P specimen is the pressure in the specimen chamber, d is the distance the beam travels in the chamber, and it is assumed that 5 kv is applied to the electron beam [13]. Naturally, the detector that is used in a standard SEM will not work in the high pressure of an esem due to electrical discharge of the high voltage supplied to it. Instead, a gaseous detection device is used, which only employs several hundred volts or less. In such a device, a uniform electric field is created by two parallel plates, in order to accelerate the electrons from the sample toward an anode. Thermal diffusion also moves these electrons radially. On the way to the anode, the electrons collide with gas molecules giving off even more electrons. At a short distance, this process reaches 100% efficiency, even at low voltage [14]. A scintillator/photomultiplier dectection scheme, like that in a standard SEM, is also used in the gaseous detection device [15]. The gaseous detection device can also include a backscattered electron detector [16]. 19

28 As mentioned, an esem has the ability to image nonconductive materials, unlike a standard SEM. This is facilitated by the high conductivity of the gas within the specimen chamber. This conductivity arises from the ionization caused by the incident electrons and the subsequent secondary electrons and backscattered electrons [17,18]. A nonconductive material that is imaged in an esem can still be coated with a conductive material to achieve higher quality images. Doing so would counteract the benefits of using an esem, but it would negate the need for to SEMs for two types of materials. Figure 10: Side view of an esem. The differential pumping system, made possible by the pressure limiting apertures, allows esems to image materials that could not be imaged in a conventional SEM. The anode and cathode create the uniform electric field that is used to direct the secondary electrons toward the detector. 20

29 Energy Dispersive X-ray Spectroscopy Energy dispersive X-ray spectroscopy is a characterization technique used to perform elemental analysis on materials. EDS was developed in late 1960s and is a technique similar to wavelength dispersive X-ray spectroscopy (WDS), but with a few key differences [19]. ED spectra result from the counts of scattered X-rays collected at varying energies, while WD spectra are made from the counts of diffracted X-rays at varying wavelengths. Notably, EDS has a small size to collection angle ratio compared to WDS [20]. EDS is usually performed in an SEM instrument, since the required components are very similar. As in SEM imaging, electrons are impingent upon and penetrate into the bulk of an EDS sample. Then, X-rays from inner-shell ionizations are ejected from the sample and enter a detector that is under high vacuum. This detector is separated from the sample by gold-coated beryllium film window, which is tens of nanometers thick. The window cannot be exposed to atmosphere so an airlock is used [20]. When X-rays enter the detector, electron-hole pairs are created in a lithium-doped silicon crystal. The charge pulses that are created from these pairs are proportional to incoming X-ray energy. These charge pulses are the information that is collected by the system and used to create spectra. A liquid nitrogen system is used to cool the detector in order to reduce electronic noise [20]. The spectrum resulting from EDS analysis consists of Gaussian characteristic peaks from the X-ray emissions that occur when excited atoms return to ground state. 21

30 There is also a background signal from Brehsstrahlung radiation, which results from electrons being slowed by inelastic collisions [20]. One important parameter of EDS is the X-ray take-off angle, which is the angle between plane of sample and detector axis. Studies have shown the optimal angle is 20 degrees [21]. Knowing this exact value is critical for proper analytical calculations. An angle that is too low leads to large path-length corrections, while angles that are too large lower collection efficiency because the detector must be retracted [20]. Other important parameters include the solid angle of collection and the detector resolution. The solid angle of collection is defined as the area of the detector divided by the distance from the detector to the sample. The detector resolution determined by analyzing the manganese K α peak, which has a full width half maximum (FWHM) of 5.89 kev. During routine EDS analysis, a resolution of 150 ev is typical, and is limited by the instruments electronics and statistical X-ray distribution [20]. Figure 11: Illustration of an EDS instrument. The field effect transistor (FET) and preamp are some of the initial electronics used to analyze the charge pulses from the Si(Li) detector. Taken after [20]. 22

31 CHAPTER III APPLICATIONS Characterization of Alumina Doped Zinc Oxide Nanofibers Zinc oxide is a thoroughly studied material with a hexagonal wurtzite crystal structure with both covalent and ionic bonds [22]. It is a semiconducting oxide with a wide band gap, 3.37 ev, which has gained interest recently, especially for nanometer scale applications [23-25]. Zinc oxide is highly reactive to air and its electrical properties are considerably altered by the adsorption of oxygen, carbon dioxide, hydrocarbons, sulfur-containing compounds, and water. In order to tune the optical, electrical, and mechanical properties of zinc oxide crystals and polycrystalline films, they have been doped with elements of the alkali metals such as Li and with Ga, In, N, Al, Sn, P, etc. from groups IIIB to VIIB [26]. Nanostructures made of zinc oxide doped with alumina are of special interest. Materials of this type have been investigated for transparent conducting oxide for applications such as conducting window material for solar cells. Additionally, it has been researched for use in light emitting diodes and laser diodes emitting in the ultraviolet 23

32 spectrum, electrophotography, gas sensing, and thermoelectronics. Many techniques have been used to create doped and undoped zinc oxide films, such as pulsed laser deposition, DC magnetron sputtering, RF magnetron sputtering, metal organic chemical vapor deposition, molecular beam epitaxy, metal organic vapor phase epitaxy, reactive deposition, spray pyrolysis, and sol gel on various substrates [27]. Fabrication of the nanofibers in this study involved sol-gel methods followed by electrospinning. Sol-gel is a wet-chemistry process that is usually used to create metal oxides, which begins with a solution (sol) and generates an integrated network (gel). Four solutions were created with varying levels of aluminum oxide to create the nanofibers. The nanofibers are denoted as undoped zinc oxide and AOZO1 through AOZO3, with the aluminum oxide content increasing from AOZO1 to AOZO3. The electrospinning technique for processing polymers has existed since the 1930s, and has been refined ever since [28,29]. An electrospinning instrument consists of a high voltage power supply, a collection device, and a prepared polymer solution with a delivery system. This delivery system can be a pipette with a conducting wire or a syringe with a syringe pump. The pipette is a gravity feed system with the conducting wire placed in the solution, while the syringe provides a constant feed system with a voltage applied to the tip of the syringe. The collection device is either grounded or oppositely charged with respect to the feed system, depending on the desired rate of solution ejection [30]. In order to perform electrospinning, the collection device is first prepared and positioned a precise distance from the feed system. Then, the stationary feed system is filled with a polymer solution. Next, tens of thousands of volts is applied across the feed 24

33 system and collection device. As the now charged solution is forced from the end of the feed system, it forms a Taylor cone. This Taylor cone arises from the interaction of the solution's surface tension with electrostatic forces between the charged particles. As the charged jet approaches the collection device, the solvent evaporates from the solution, forming a microscopic fiber [30]. Figure 12: Electrospinning instrument. The collection device can be as simple as a metal plate, but is often more complicated to allow either large amounts of nanofibers to be fabricated or special products such as nanofiber yarns to be made. The creation of the polymeric precursor solutions is an involved process. First, a zinc oxide precursor solution was made by dissolving zinc acetate (Zn(CH 3 COO) 2 ) in water at a weight ratio of 1:4 (Zn(CH 3 COO) 2 :water). Next, polyvinylpyrrolidone (PVP) solution (PVP:C 2 H 5 OH) with a 1:6 weight ratio was added to equal parts of the aqueous zinc acetate solution. An aluminum oxide precursor solution was made by mixing 5 25

34 grams of aluminum acetate (Al(CH 3 COO) 3 ) with 5 grams of water and 5 grams of ethanol, while the whole solution was magnetically stirred for about 24 hours without heating. A PVP solution, composed of 3 grams of PVP added to 50 milliliters of C 2 H 5 OH, was added to the aluminum oxide precursor solution at a weight ratio 1.7:1 (PVP solution:al(ch 3 COO) 3 solution). The PVP/Al(CH 3 COO) 3 composite solution was added to the PVP/Zn(CH 3 COO) 2 composite solution according to the amount of doping desired and the solution was magnetically stirred for 1 hour before electrospinning [27]. After that, the polymeric precursor solutions were electrospun at the constant flow rate of 2 microliters per minute using a voltage of 20 kv, with a distance from the anode to the fiber collector fixed at about 20 centimeters. The fibers were then brought to a temperature of 873 Kelvin at rate of 10 Kelvin per minute, and then maintained at 873 Kelvin for 5 hours to create ceramic nanofibers [27]. XPS was performed using a VG ESCALAB MK II to examine the chemical properties at the surface of the AOZO nanofibers. The pressure inside the XPS chamber was less than millibar before analysis was performed. A twin anode X-ray source with a magnesium Kα ( ev) and aluminum Kα ( ev), and a concentric hemispherical analyzer (CHA) were in the instrument. The aluminum Kα source was used to create the XP spectra. In order to obtain spectra, the instrument operated with an anode voltage of 9 kv, an electron multiplier voltage of 2850 ev, an anode current of 20 milliamps, a filament current of 4.2 amps, a pass energy of 20 ev, a dwell time of 100 milliseconds, and an energy step size of 0.02 ev in constant analyzer energy (CAE) mode for high resolution scans [27]. 26

35 Since the aluminum 2p XPS peak overlaps with zinc 2p satellite peaks, the aluminum 2s area was chosen for analysis. These peaks were shifted due to charging effects, so corrections were made using the carbon 1s peak of drift carbon at ev as a reference. The increased concentrations of aluminum oxide dopant lead to changes in the chemical signatures found in the XP spectra. The high energy zinc 2p features increased in intensity as aluminum oxide was added, while the oxygen 1s and aluminum 2s peak maxima shifted to higher binding energies. These peak shifts imply a correlation between the addition of the aluminum oxide dopant with changes in the chemical states of zinc, as well as aluminum and oxygen, in addition to higher oxidation states occuring as the amount of aluminum increased [27]. (a) Zn 2p, X1 (b) O 1s, X5 (c) Al 2s, X13 Zn2p 1/2 Zn2p 3/2 AOZO3 cps (a.u.) AOZO3 AOZO2 AOZO3 AOZO2 AOZO2 AOZO1 AOZO1 AOZO1 Undoped ZnO Undoped ZnO Undoped ZnO Binding Energy (ev) Figure 13: High resolution X-ray photoelectron spectra of AOZO nanofibers. From bottom to top, AOZO0 to AOZO3 nanofiber spectra of the Zn 2p region (a), O 1s region (b), and Al 2s region (c). The high energy zinc 2p peak shifted due to the combination of changing chemical state, an overlapping aluminum 2p peak that increased as alumina was added, and charging affects. Figure from [27]. Deconvolution of the XPS peaks in the three different areas of interest (zinc 2p, oxygen 1s, and aluminum 2s) was performed using XPSPEAK shareware (ver. 4.1). The 27

36 FWHMs used for peak deconvolution were , , and for zinc 2p, aluminum 2s, and oxygen 1s, respectively. Two states were found for zinc during the deconvolution, corresponding to metallic zinc at a binding energy of ev and to zinc oxide at around ev. As the amount of aluminum oxide dopant was increased, the ratio of metallic zinc to zinc oxide decreased from 3.06 to 1.13 [27]. (a) Zn 2p 3/2, X1 (b) O 1s, X5 (c) Al 2s, X25 AlO x ZnO Zn 0 AlO y cps (a.u.) AOZO3 raw sum Zn 0 background AOZO3 surf OH ZnO AOZO3 background Al 0 ZnO AlO y AlO x AOZO1 AOZO1 AOZO Binding Energy (ev) Figure 14: Example deconvoluted XP spectra. This figure shows the zinc 2p 3/2 area (a), the oxygen 1s area (b), and the aluminum 2s region (c) for AOZO1 (bottom) and AOZO3 (top). Metallic zinc at a binding energy of ev and zinc oxide near ev are show in (a). The binding energies for oxygen are near ev for ZnO, ev for low binding energy AlO x, ev for surface hydroxide, and ev for high binding energy AlO y, shown in (b). In (c) metallic aluminum (Al 0 ) at ev, low binding energy AlO x at ev, and high binding energy AlO y at ev, where x is less than y. The deconvoluted zinc 2p 3/2 is shown in (a) instead of showing of the entire zinc 2p area because the spin-orbit splitting constant of zinc 2p is so wide (22.97 ev). Figure from [27]. Detailed chemical analysis was also done on the oxygen 1s peak. Constituent peaks for four different chemical states of oxygen were found near ev for ZnO, ev for low binding energy AlO x, at ev for surface hydroxide, and ev for high 28

37 binding energy AlO y, where x is smaller than y. The ratio of the oxygen in zinc oxide to aluminum oxide decreased from 2.29 to 1.35 as the aluminum content increased from 1.75 to 3.33 atomic percent. The three aluminum constituents deconvoluted were metallic aluminum at ev, low binding energy aluminum oxide at ev, and high binding energy aluminum oxide at ev. The ratio of aluminum/zinc rose from to in the metallic and oxide states of the metals, as the aluminum content increased. Conversely, the ratio of metallic aluminum to AlO x +AlO y decreased from 1.10 to 0.32, and that of metallic zinc to zinc oxide decreased from 3.06 to The lowered content of metallic aluminum and the simultaneous increase in aluminum oxide seems to have caused the decrease in conductivity [27]. Figure 15: Electrical conductivities of AOZO nanofibers. The nanofibers were calcined at 873 Kelvin for 5 hours. Figure from [27]. 29

38 The electrical properties of the AOZO nanofibers were studied by making I-V measurements using a Keithley 2410 sourcemeter. The AOZO nanofibers were first crushed and formed into disks that were 6 millimeters in diameter and 0.75 millimeters thick. A 150 nanometer layer of metallic nickel was deposited on the flat surfaces of the disks by plasma enhanced physical vapor deposition (PEPVD). Then, the coated AOZO disks were attached to glass slides. Electrical contacts with silver wires were attached to the flat surfaces and a potential difference of ± 50 volts was applied to the wires, and measurements were made at room temperature. The resistance R obtained from the I-V measurement, length L, and cross sectional area A of the disks were used to obtain conductivities σ of the AOZO materials using σ = R/LA. Measurements revealed that the aluminum content affects the conductivity of the AOZO nanofibers in a non-monotonic manner [27]. A Philips XL-30 Environmental Scanning Electron Microscope equipped with a Noran X-ray detection system was used to take SEM images of the nanofibers. In order to aquire the images, the fibers were silver-coated with a S150B Sputter Coater, then imaged using an accelerating voltage of 20 kv. The average diameter of these fibers, which was found to be about 100 nanometers, was just small enough to be considered a nanomaterial [27]. It is important to note that the silver-coating increases the diameter of the fibers, so the diameter of the actual fibers will be smaller that the measured diameters. There are unique motivations and caveats associated with acquiring SEM images of these types of fibers. Obtaining SEM images for these types of materials is important to verify that that are actually fibrous, and have a diameter between 1 and 100 nanometers. Such materials are desirable for their high surface area to volume ratios, 30

39 while being light weight considering their mechanical strength. Also, the actual resolutions of images of these materials are far lower than theoretical SEM limits, primarily because the fibers tend to move slightly, while the instrument is operational. Also, working with these fibers in any vacuum instrument was difficult, due to the poor adhesion to the sample stubs. Care had to be taken to ensure that the fibers were not simply drawn out of the machine by the vacuum pumps. Figure16: Representative SEM image of AOZO nanofibers. This is sample AOZO1 exhibiting an average fiber diameter 100 nm after being calcined at 873 K for 5 hrs [27]. An FEI QUANTA 200 Environmental Scanning Electron Microscope equipped with an EDAX X-ray detection system was used to acquire EDS data. The instrument maintained a 60 pascal pressure, a voltage of kv, and a take of angle of 35.18, 34.61, 35.51, and was used for undoped zinc oxide then AOZO1 through AOZO3, respectively. 1 µm 31

40 Energy (kev) Figure 17: Energy dispersive X-ray spectra for AOZO nanofibers. Undoped zinc oxide then AOZO1 through AOZO3 from left to right. Notice the increase of aluminum from left to right. The zinc and oxygen peaks do not show such obvious trends. The bulk content of aluminum was close to that found on the surface of the nanofibers. The atomic percent of aluminum was 3.07%, 5.82%, and 6.53% in AOZO1 through AOZO3, respectively, while the resulting aluminum to zinc ratio rose as 0.063, 0.147, and While these percentages and ratios are higher than was found on the surface, that information alone is not sufficient to conclude that more aluminum was actually inside the fibers. However, it is reassuring to find that the aluminum content in the bulk was similar to the surface content, and that the same trends in composition are apparent. This study indicated that the incorporation of 1% by weight of aluminum dopant increased the conductivity of zinc oxide nanofibers by two orders of magnitude. 32

41 However, XPS data indicated that further addition of aluminum lowered the conductivity due to increased formation of aluminum oxide. Since AOZO nanofibers could be used in catalysis, photocatalysis, photonics, electronics, and sensor devices, this information is important. Characterization of CoCrMo Metal Powders The methods and techniques used to manufacture and clean prosthetic implants can change the elemental composition at the material s surface [31-33]. It is this surface composition that is important, when considering how these materials will react with biological interfaces. Maximizing biocompatibility is an important aspect of developing joint replacement prostheses materials. Until now, most studies concerning biocompatibility with regard to elemental composition have been performed on the bulk form of these materials, while minor attention has been given to the surface. Similarly, metal particulate powders that are used to simulate prosthetic implant wear debris in cell cultures have had little surface elemental composition characterization [34]. One of the contributing sources of inflammation and bone loss near prosthetic implants has been shown to be metal wear debris that was able to be engulfed by cells. Previous studies investigating cell cultures exposed to metal particulates have shown inconsistent results; with some reporting extreme toxicity to cells and others show complete biological inertness, for what was thought to be the same material. This prompted an investigation into the discrepancies in elemental composition between the bulk alloy compositions to surface compositions of cobalt-chrome-molybdenum 33

42 (CoCrMo) American Society for Testing and Materials (ASTM) F75 particulate powders. XP and ED spectra of CoCrMo powders from three different sources were compared to cell culture studies of the same powders to investigate the relation between cellular viability and surface composition relation [34]. The three powdered metals, denoted as CoCrMo-I, CoCrMo-II, and CoCrMo-III, were all grade F75 by American Society for Testing and Materials standards. Each of the powders was presumed to be fabricated exactly the same way, and to have exactly the same elemental composition. Two of the powders were from the same vendor, acquired on two different occasions, while the third was from another vendor (CoCrMo-I) [34]. Like the nanofiber study, an FEI QUANTA 200 Environmental Scanning Electron Microscope equipped with an EDAX X-ray detection system was used to acquire EDS data, while using a voltage of kv at a pressure of 60 pascal. Unlike the nanofiber study, this instrument was also used to acquire the SEM images with a spot size of 4.0 nanometers, and a sample to detector depth of approximately millimeters for various magnifications. The SEM images revealed marked differences between the three supposedly identical powders. CoCrMo-I and CoCrMo-III had similar geometry (spherical), but differed significantly in average size, while CoCrMo-II had a completely different geometry (irregular). Recall that two of these powders were from the same source and were supposed to be identical. However, CoCrMo-I and CoCrMo-II differ in such a way that suggests that they might have been manufactured or processed in two different manners, since they were obtained at two separate dates [34]. Again, the resolutions 34

43 obtained in these images were far less than ultimate limits, but in line with practical limits of the instrument and material. Figure 18: SEM images of CoCrMo particulate powders. Clockwise from top left: CoCrMo-I, CoCrMo-II, CoCrMo-III at about 2,000x, CoCrMo-III at about 15,000x. Similar figures are found in [34]. The ED spectra that were acquired agreed much more closely with expectations for these powders than the SEM images. The ED spectra showed that metal composition in the bulk of all three powders were nearly identical. Additionally, these data collected from the EDAX system indicated that the metallic powders adhered well to ASTM standards for F75 and those specified by their sources. Variations were evident in the overall content of the samples, but these variations were in the non-metallic elements found within the powders [34]. 35

44 Energy (kev) Figure 19: ED spectra for CoCrMo particulate powders. Again, XPS was performed using a VG ESCALAB MK II to examine the surface composition at the surfaces of the powders. The pressure inside the XPS chamber was less than millibar before analysis was performed. The aluminum Kα source was used to create the X-rays. The instrument operated with an anode voltage of 9 kv, an electron multiplier voltage of 2850 ev, an anode current of 20 milliamps, a filament current of 4.2 amps, a pass energy of 20 ev, a dwell time of 100 milliseconds, and an energy step size of 0.5 ev in constant analyzer energy (CAE) mode. 36

45 Figure 20: XP spectra for CoCrMo particulate powders. CoCrMo-I, CoCrMo-II, and CoCrMo-III survey spectra shown in (a), (b), and (c). Similar figures also found in [34]. The gathered XP spectra indicated that surface compositions of the CoCrMo metal particulate powders varied with respect to each other, and with respect to their bulk compositions. CoCrMo-II was the only powder that exhibited little discrepancy between surface and the bulk composition. Conversely, CoCrMo-I and CoCrMo-III spectra indicated that surface segregation had occurred. At the surface, CoCrMo-I showed an increase in the relative quantity of molybdenum, manganese, and silicon, resulting in a decrease in the amount of chromium and cobalt. The CoCrMo-III powder had increased concentrations of chromium, molybdenum, silicon, manganese, and indium, but a significant decrease in the cobalt concentration. The measured difference in elemental 37

46 composition suggests that a factor in the fabrication or processing of these materials causes a discrepancy between the bulk and surface compositions. Table of Bulk Compostion Percentages of CoCrMo Powders CoCrMo-I CoCrMo-II CoCrMo-III F75 CoCrMo Cobalt (Co) 62% 62% 62% % Chromium (Cr) 34% 32% 34% 27-30% Molybdenum (Mo) 3% 4% 3% 5-7% Silicon (Si) 1% 2% 1% <1% Manganese (Mn) <1% Figure 21: Table of bulk composition percentages of CoCrMo powders. The values in the F75 column are the standards as published by the ASTM. A similar table can be found in reference [34]. Table of Surface Compostion Percentages of CoCrMo Powders CoCrMo-I CoCrMo-II CoCrMo-III F75 CoCrMo Cobalt (Co) 30% 69% 24% % Chromium (Cr) 30% 28% 42% 27-30% Molybdenum (Mo) 5% 3% 7% 5-7% Silicon (Si) 27% - 12% <1% Manganese (Mn) 8% - 8% <1% Indium (In) - - 7% - Figure 22: Table of surface composition percentages of CoCrMo powders. The values in the F75 column are the standards as published by the ASTM. A similar table can be found in reference [34]. In addition to characterizing the powders, a cell culture study was performed on knee joint tissue, from four donors. The cells were transferred to 25 centimeter 2 flasks, where they were left to grow to a full single layer (4 x 10 6 cells per flask). After that, the cells were exposed to varying doses of CoCrMo-I and CoCrMo-II, which were sterilized first. Control flasks for each set of donor cells were created, where no metal was added, along with flasks containing gram and 0.04 gram doses of each powder. Five days following the creation of these alloy-contaminated cultures, cell viability counts were made for each flask. These counts were normalized using the control flasks, before 38

47 finding averages and standard deviations for each of the four sets of cells. Dyes were used to determine which cells were still viable in the flasks [34]. Figure 23: Cell activity after CoCrMo exposure. Acridine orange staining was used to compare lysosomal (asterisks) and nuclear (arrows) activity following exposure to CoCrMo-I and CoCrMo-II. (A) 15 minutes after the exposure to the 0.04 gram dose of CoCrMo-I, the cells show very active nuclei (intense green color) and intact lysosomal activity (intense red color). (B) 15 minutes after the exposure to the gram dose of CoCrMo-II, the cells show less active nuclei (duller green) and lysosomal leakage (greenorange color). (C) 75 minutes after the exposure to the 0.04 gram dose of CoCrMo-I, the nuclei are apparently very active and lysosomes seem intact and active. (D) 75 minutes after exposure to the 0.004gram dose of CoCrMo-II, the cells show no non-reactive nuclei, lysosomal leakage, or cell membrane dissolution (large holes throughout the membrane). (E) 24 hours after exposure to the 0.04 gram dose of CoCrMo-I, the nuclei look to be extremely active and lysosomes seem active and intact. (F) 24 hours after the exposure to the gram dose of CoCrMo-II, the cells appear to be near death and show little nuclear activity and incomplete cell membranes Figure taken from [34]. Exposing the cells to the different samples resulted in noticeably different viability results. At the gram doses, CoCrMo-I powder caused a normalized reduction of 11% in viability, while the CoCrMo-II powder led to an 86% reduction in viability. For the 0.04 gram doses, the cells exposed to CoCrMo-I powder had a 30% 39