Praktikum zur Materialanalytik Energy Dispersive X-ray Spectroscopy B513 Stand: 19.10.2016 Contents 1 Introduction... 2 2. Fundamental Physics and Notation... 3 2.1. Alignments of the microscope... 3 2.2. Generation of characteristic X-rays... 3 2.3 Moseley s law... 5 2.4 Detection of X-rays... 5 2.5 Area of analysis... 6 3 The experiment... 7 4 Literature... 7
1 Introduction Basically, a transmission electron microscope is an instrument usable for extracting diffraction patterns similar to an XRD machine, spectroscopic information as well as for magnifying extremely small structures. It is the combination of all three possibilities which make such an instrument a universal tool for the investigation of various materials. Diffraction patterns contain information on the symmetry of the atomic arrangement of the crystals investigated and the interatomic spacing of individual parts of the when operating in selected area electron diffraction (SAED) mode. The micrographs may be used for the determination of layer morphology and the defect structure when experiments are carried out under well-defined conditions, e.g. by high resolution transmission electron microscopy (HRTEM) or by bright field imaging which is analogous to classical light microscopy. Furthermore, different spectroscopic information can be extracted; of interest in this lab course are characteristic X- rays which we can record with an Energy Dispersive X-ray (EDX) detector. Figure 1: Setup of the JEM2100 JEOL microscope 2
2. Fundamental Physics and Notation 2.1. Alignments of the microscope In order to record the best possible images with the highest possible resolution the microscope has to be properly aligned before every use. This is necessary because not all components of the microscope are exactly on the optical axis, apertures have been used and are misaligned or slight changes in the environment influenced the microscope. These alignments comprise of centering the electron beam, all apertures in use (at least the condenser aperture, optionally also the diffraction and objective aperture) and correcting the pivot points when tilting the electron beam. On top of that, instrumental aberrations like astigmatism have to be corrected. Other aberrations, like the chromatic and spherical aberration can only be corrected in some special microscopes with dedicated correctors for every aberration. 2.2. Generation of characteristic X-rays When high velocity electrons hit the sample they interact with it in various ways as depicted in Fig. 2. Figure 2: Overview of different electron beam-matter interactions. Aside from all these theoretically information-bearing signals also heat and ballistic damage is created in the sample. [1] These interactions include direct collisions between the incident beam and shell electrons of the sample s atoms, so that shell electrons are removed from their orbital (Fig. 3; simply displayed as shells according to the Bohr model). Removal of deep core-level electrons and fill-up of these vacant states called transitions by outer-shell electrons causes the emission of characteristic X-rays. They are referred to as such, because their energy is characteristic, i.e. it is equal to the energy difference of the involved energy levels. Since these energy differences are specific for the single elements, the X-rays can be used to identify the elemental make-up. Next to the emission of this characteristic radiation also the continuous spectrum of Bremsstrahlung is generated by deflection and deceleration of electrons in the sample; however 3
this radiation is almost white in nature and does not yield any information about the chemical composition. Figure 3: Electronic shell model and indications for the nomenclature of electron transitions. [2] The nomenclature of characteristic X-rays has to include the origin and destination of the electron transition. The destination, i.e. the vacant state to be filled up, is simply addressed by the letter of its shell (K, L, M, ) and the origin of the electron transition by an index for the next higher shell (α) or next-next higher shell (β) and so on. Since electrons even within common shells can differ in their energy level due to bonding effects, an additional numeric index can be added, however, since the energy resolution of an EDX system is typically worse than 5 ev and the energy differences are typically smaller than that, these differences shall not be discussed here. As a simple example the transition of an L-electron to a vacant state in the K-shell is called Kα (and actually consists of Kα1 and Kα2) Besides the emission of characteristic X-rays the transition can also cause the emission of Auger electrons, but with increasing atomic number the fluorescence yield increases to nearly 100%, i.e. the probability for the emission of Auger electrons in heavy atoms is almost zero. Figure 4: Fluorescence yields for K and L shells for 5 Z 110. The plotted curve for the L shell represents an average of L1, L2, and L3 effective yields. [3] 4
2.3 Moseley s law The foundation for determining the elements is Moseley s law which shows the correlation between atomic number Z and the energy E of the emitted X-rays for the transition from the energy level n2 to level n1: E = h c λ = h c R (Z σ) 2 ( 1 n 1 2 1 n 2 2 ) Here, h denotes Planck s constant, c the speed of light and R is Rydberg s constant. The screening constant σ takes into account the presence of the other shell electrons. The law strictly only holds for small and medium atomic numbers. With increasing main quantum number n the energy of radiation drastically decreases. For Kα radiation (n1 = 1, n2 = 1 and σ = 1) the formula simplifies to: E = h c R (Z 1) 2 3 4 Thus, if we can measure the energy of the emitted X-rays we can deduce the atomic composition. 2.4 Detection of X-rays Since one cannot directly detect and measure X-rays, the task of an EDX detector is to first convert the radiation into electric signals, i.e. voltage peaks and finally display them as a spectrum. The main components of an EDX detector are collimator, window, detector crystal and a bunch of electronics. The collimator makes sure to reduce stray radiation falling onto the actual detector as much as possible. It is made from a highly absorbing material (i.e. for X-rays and electrons) and is basically a tube or a set of these pointing from the source of the X-rays to be detected (the sample) to the detector. The window acts as a barrier between the highly sensitive detector crystal beneath it and the vacuum chamber of the microscope where contaminants might be introduced from the atmosphere or released from the sample. Since the detector crystal is typically cooled down with liquid nitrogen (77 K) it would otherwise lead to pronounced condensation of contaminants on the crystal. Since the window is blocking the path of the X-rays it has be made from a very light material which does not absorb X-rays in a significant amount, e.g. polyimide or beryllium. However, no matter how light the window, it will always absorb some X-rays, especially low energy radiation, that s why some modern and state-of-the-art detectors work without liquid nitrogen cooling so they do not need a (permanent) window. On the downside, though, no cooling also introduces a lot of thermal noise in the crystal. The detector crystal itself is typically made from silicon. Its task is to convert the X-rays into electric charges. This is accomplished by generation of electron-hole pairs. For the generation of one electron-hole pair in silicon an energy of 3.86 ev is required. Thus the number of electron-hole pairs is proportional to the energy of the absorbed X-ray quant. For example a calcium Kα quant with an energy of 3691 ev creates roughly 956 electron-hole pairs and thus a charge of around 1.53 10-16 C. The design of a modern silicon drift detector (SDD) is depicted in Fig. 4: an n-doped anode is surrounded by circular areas of p-doped silicon. Upon application of a voltage the n-doped areas of these pn-junctions will be depleted of electrons (the majority charge carrier) and the space charge region increases in size. Electron-hole pairs created by incident X-rays cannot recombine but are separated: the circular electrodes create a potential gradient through which 5
the electrons drift to the middle anode and the holes to the circular electrodes. The anode is connected the gate of the integrated field effect transistor (FET) which converts the charge signal into a voltage. Figure 5: Schematic design of a silicon drift detector. [4] As mentioned before modern SDDs are more accurate, more effective, quicker and cheaper than older designs based on high-contact area silicon detectors. Also, the electronic thermal noise is much lower so that cooling with a Peltier element to around 253 K is sufficient. The signal generated by the FET is then amplified and the previously mentioned calcium Kα quant creates a voltage peak of around 1.53 mv. This method of detection, however, demands that the detection of one X-ray is finished before the next one can be analyzed, therefore a deadtime is required in which no other X-ray can be detected. The deadtime correction is called pile-up suppression: if during processing of one voltage pulse a second one is detected both of them are disregarded and the deadtime is increased. This also prolongs the overall time of analysis. In the end the individual voltage pulses are displayed according to the energy they represent and summed up to create an EDX spectrum. With the help of experimentally determined scattering and absorption cross-sections for individual elements and a number of numerical methods a quantification of the results can be done. For instance, the peaks have to be fitted by appropriate models (e.g. Gaussian), overlapping peaks in the spectrum have to be deconvolved and the area under the peaks has to be calculated. The detection limit for EDX is around 0.1 wt% for elements with Z 10. For lighter elements, however, the detection limit is much worse and is somewhere around the region of a few wt%. 2.5 Area of analysis Aside from the sheer detection of elements, it is a common task to also localize them. Thus, if a spatially and temporal continuous illumination of the sample was chosen (like it is done in the standard bright field mode) one would not be able to localize the elements, because X-rays would be generated at every location of the illuminated area at the same time. Instead, however, the TEM can be operated in a scanning mode very similar to a scanning electron microscope (SEM), accordingly this mode is called scanning TEM (STEM). In STEM mode, just like in an SEM, the electron beam is scanned over the sample and several generated signals can be captured by different detectors to create an image (e.g. unscattered electrons: STEM bright field (BF); scattered electrons STEM high angle annular dark field (HAADF)) and among them we can of course record the characteristic X-rays generated at each point. An image of the elemental distribution is commonly referred to as an elemental map. Some of these 6
signals can be recorded simultaneously, so that we can create an image of the sample, e.g. by STEM-HAADF and record the EDX elemental map at the same time. 3 The experiment In this lab course you are going to analyze an unknown sample and determine both the quantitative elemental content as well as their lateral distribution. 4 Literature [1] Taken from: http://www.gatan.com/techniques/eels [2] Energy-dispersive X-ray microanalysis: An introduction by Noran Instruments (cime.epfl.ch/webdav/site/cime2/shared/files/teaching/edx/introduction%20to%20eds.pdf ) [3] M. O. Krause, Atomic Radiative and Radiationless Yields for K and L Shells, J. Phys. Chem. Ref. Data 8, 307 (1979). [4] Taken from: http://www.azom.com/article.aspx?articleid=11973 7