Particle-induced x-ray emission (PIXE) using hydrogen and a plurality of kinds of focused ion beams

Abstract

To provide practical implementation of particle-induced X-ray emission (PIXE) for a focused ion beam device or for a dual beam device having both a focused ion beam and a scanning microscope function.SOLUTION: An analytical method includes the steps of: guiding ions including a mixture of a proton and a non-hydrogen ion onto a specimen, on condition that kinetic energy of the ion of the mixture is 50 kiloelectron volt (keV) or less; and detecting and measuring an X-ray emitted from the specimen in response to collision of the proton and the non-hydrogen ion with the top of the specimen.SELECTED DRAWING: Figure 1A

Description

[Technical Field]

[0001] The present invention relates to focused ion beam apparatus and methods for using such apparatus. More specifically, the present invention relates to elemental analysis of a sample by a focused ion beam comprising a mixture of protons and other ion species using a particle-induced X-ray emission method, and by a focused ion beam apparatus comprising both a standalone focused ion beam apparatus and a dual-beam apparatus that also includes scanning electron microscopy capabilities. [Background technology]

[0002] Focused ion beam (FIB) instruments are essential for sample preparation, nanofabrication, and materials analysis. Standalone FIB instruments are known, but they are typically used as components of dual-beam or triple-beam microscopes, in combination with scanning electron microscope (SEM) components[1], and more recently, with femtosecond laser ablation system components.[2] FIB instruments occupy a position between small-scale electron beam processing and large-scale bulk processing in semiconductor fabrication and processing applications, as the focused beam enables mask-free processing of small-sized feature shapes that are inaccessible with bulk processing such as reactive ion etching (RIE). In such applications, FIB instruments offer significantly higher throughput than that provided by electron beam processing such as electron beam lithography and focused electron beam induced processing (FEBIP). [3,4] For this reason, FIB instruments are typically used in industrial applications such as semiconductor failure analysis where cross-sectioning, tomography, or TEM thin section preparation is required on a relatively small scale and high throughput is essential for cost reasons. Focused ion beam instruments are also a research and development site for applications such as the preparation of biological samples for cryo-TEM analysis, the preparation of samples for atomic probe tomography, or the fabrication of optical components such as immersion lenses. Thus, FIB instruments are ubiquitous and can be found in numerous factories and laboratories around the world. [3,5,6,7]

[0003] FIB devices are similar to scanning electron microscopes and can be used in a similar manner. In either operation of such a device, a densely focused beam of charged particles is scanned or rastered across the surface of the sample under consideration. The impact of the beam's energy on the sample surface induces either backscattering of particles from each respective focal point or emission of secondary charged particles (ions or electrons). The particles backscattered or emitted from each point can be detected by a particle detector, which provides a measure of the relative quantity of particles detected from each point. The number of such particles that can be detected, and therefore the signal intensity, may depend on the topography and composition of the sample. In this way, a map of the sample's topography and / or composition can be developed. However, while SEMs utilize a focused electron beam to image the sample, FIBs utilize a focused beam of ions instead. Due to this difference in beam composition, FIB devices produce images with somewhat lower resolution than SEM images and, in contrast to SEM devices, can cause damage to the sample through the sputtering process. Nevertheless, the sputtering process can be advantageously used in micro-etching or micro-milling applications. Therefore, many dual-beam systems include both a FIB component or subsystem and a SEM component or subsystem. In such dual-beam systems, the SEM component can be used to monitor the micro-etching or micro-milling procedures performed by the FIB component through its imaging capabilities.

[0004] From another perspective, given that the primary purpose of both FIBs and particle accelerators is to generate and accelerate ions for imaging, fabrication, and ion beam analysis purposes, FIBs can be considered small, relatively low-energy particle accelerators. The emergence of ion microprobes (modules joined to particle accelerator beamlines to generate focused beams capable of micrometer-level resolution) has further blurred the line between FIBs and particle accelerators. [8] However, compared to FIB instruments, particle accelerators are somewhat rarer, with only about 30,000 accelerators in operation as of 2014, ranging from small benchtop particle accelerators to large hadron colliders, varying considerably in their intended use and application, and their costs can reach billions of dollars. [9,10]

[0005] One of the main functions of many particle accelerator facilities is ion beam analysis. Beam analysis (IBA) is a set of techniques that involve the analysis of the interaction between high-energy particles and a sample, and the signals that result from these interactions. These techniques include:

[11] ● Rutherford backscattering spectroscopy Backscattering spectroscopy (RBS): Analysis of primary ions dynamically scattered from a sample. ● Ion beam induced charge (IBIC) involves the generation of hole pairs by a primary ion beam and subsequent analysis of charge transport within the sample. ● Nuclear reaction analysis (NRA): Analysis of charged particles released as a result of nuclear reactions following the interaction between primary ions and target nuclei. ● Particle-induced gamma-ray emission ray emission (PIGE): Analysis of gamma rays emitted following the interaction between primary ions and target nuclei. ● Particle-induced X-ray emission Emission (PIXE): Analysis of characteristic X-rays emitted as a result of ionization of a target atom by primary ions.

[0006] Of particular interest to this disclosure is the particle-induced X-ray emission (PIXE) technique. This technique is ideal for trace element analysis, and in particular, with a sensitivity of 100–500 ppm, it can determine the sample composition with a sensitivity of less than 1 ppm (parts per million) compared to similar X-ray spectroscopy techniques such as energy-dispersive spectroscopy performed with scanning electron microscopy (SEM-EDS) [12,13].

[14] PIXE analysis is typically performed by detecting the emitted X-rays while an ion beam is collided with the sample surface. Theoretically, the beam can contain any type of ion. However, in practice, very light ions such as protons are used because of their high velocity under a given accelerating potential, since the X-ray production cross section (XRPCS) is proportional to the ion velocity. Heavier ions can also be used, but this requires a very high accelerating potential to match the proton velocity

[15] , which can result in significant sample damage

[16] .

[0007] Typically, PIXE analysis techniques use several megaelectron volts (mega PIXE is performed at an incident particle energy of electron-volt (MeV) because the X-ray production cross-section increases significantly at such high energies

[17] . The optimal energy range for PIXE is known to be 3 MeV, where the X-ray production cross-section is maximized while the background contribution remains sufficiently low.

[18] However, these high primary ion energies have so far limited the use of PIXE techniques to particle accelerator facilities, which, as mentioned above, are limited in terms of availability and cost. As a result, PIXE techniques may currently be somewhat inaccessible to the average laboratory user.

[0008] Low-energy PIXE (LE-PIXE) is performed at much lower energies, e.g., below 1 MeV. Despite the lower particle energies required, LE-PIXE techniques are still typically performed using decelerated accelerator beamlines or modified ion implanters. Thus, the general use of LE-PIXE faces many of the same problems as PIXE in terms of cost and availability. [19,20] Performing PIXE at lower energies offers several advantages. In particular, LE-PIXE offers higher sensitivity to light elements compared to higher-energy PIXE. This improved sensitivity to light elements is mainly due to two main factors: a very low bremsstrahlung emission signal and a low secondary fluorescence yield at low energies.

[19] Moriya et al. demonstrated significantly higher sensitivity to lighter elements by excitation with 180 keV protons compared to 2 MeV protons, with a PK of 50 for 180 keV protons and 0.9 for 2 MeV protons. α The signal-to-noise ratio for X-ray emission lines is described. These authors concluded that sensitivity for all elements with atomic number Z less than or equal to 18 is superior for excitation energies of 150 keV compared to 2 MeV, which is due to lower background radiation for 150 keV protons compared to 2 MeV protons.

[21] Light elements that have only single X-ray transitions at lower energies, such as Be, B, C, N, and O, are often obscured by background signals such as bremsstrahlung. Higher energy PIXE also requires a thick Mylar window to block backscattered ions that may have energies up to the primary beam energy. Transmission of low-energy X-rays through such a thick window is insufficient due to X-ray absorption by the window, and therefore measurement of X-ray signals produced by light elements is impossible. Thus, the PIXE technique is often limited to elements with masses greater than Al. [20,22]

[0009] Despite the potential advantages of light element analysis described above, the drawbacks of LE-PIXE and very-low-energy PIXE (VLE-PIXE), such as the significant decrease in X-ray signals at such low energies, low sensitivity, and inability to effectively detect characteristic peaks above 1 keV, often outweigh the advantages. As a result, LE-PIXE has been studied in limited numbers, and VLE-PIXE has been studied very little.

[20] Consequently, alternative characterization methods such as SEM-EDS are typically used, or researchers rely on PIXE at energies above 1 MeV. To date, as described in the literature, only a few LE-PIXE systems utilizing decelerating accelerator beamlines or ion implantation systems have been developed. [22,23,24,25,26] Consequently, experimental data at acceleration energies below approximately 140 keV are largely absent, making the analysis of ionization mechanisms at such low energies extremely difficult. Lapicki sought extended experimental data at these energies to help clarify existing models, and cited the use of LE-PIXE as one of the motivational factors for extending these models.

[27]

[0010] The earlier claim

[28] of the observation of PIXE in a FIB microscope utilizing Ga+ ions has been disproven. The result is an artifact caused by backscattered ions originating from a charged insulated sample. These ions are accelerated away from the charged sample and, upon collision with a ground surface (e.g., a pole piece), generate secondary electrons known as type III secondary electrons. These secondary electrons are then accelerated towards the positively charged sample, generating characteristic X-rays. The electrostatic potential generated by such an insulated sample has been shown to be on the order of tens of kV.

[29] The signal is Ga to the sample +Another indicator not derived from ion collisions is the presence of a strong bremsstrahlung background, which is typical of SEM-EDS spectra but is known to be absent in PIXE spectra at such low energies. The generation and analysis of X-rays by re-acceleration of secondary electrons toward a charged sample are sometimes called charge-induced X-rays (CHIX).

[30]

[0011] The ability to perform PIXE on focused ion beam instruments represents a significant step in the advancement of PIXE techniques, opening up the availability of PIXE to a variety of additional instruments, providing a complementary technique to SEM-EDS, and thereby giving significantly improved sensitivity to trace elements. This specification introduces the term very low energy PIXE (VLE-PIXE), which describes PIXE performed below 50 keV, the energy range available to standard commercially available FIB instruments. Performing PIXE in this very low energy range comes at the cost of significantly reduced X-ray generation due to the mechanisms involved in the ionization of target atoms. Indeed, PIXE performed at such low energies using only protons was considered impossible due to the expected extremely low X-ray generation cross-section.

[29] As a result, efforts to develop PIXE at such low energies have essentially been stagnate. Therefore, the ability to perform VLE-PIXE on FIB microscopes with the sensitivity available for PIXE performed at much higher energies represents a significant breakthrough in PIXE analysis. Therefore, the development of methods to enhance X-ray generation by PIXE at ultra-low energies remains necessary. [Overview of the Initiative]

[0012] Considering the above background, the inventors addressed the above need in the art through the development of a beam doping mechanism, thereby demonstrating that the addition of small amounts of heavy ions such as Ar or Xe to a hydrogen beam results in a dramatic improvement in X-ray generation of the PIXE technique in the energy range below 50 keV. The inventors discovered and ultimately demonstrated that even when only a small percentage of heavy ions such as Ar or Xe are added to a hydrogen beam used to generate protons, and the protons are then focused onto a sample in a focused ion beam (FIB) apparatus, the performance of the PIXE technique exceeds conventional expectations by several orders of magnitude. The PIXE signal increases with the percentage of heavy ions constituting the beam, reaching a maximum at approximately 80% heavy ions. When the PIXE signal is collected during FIB milling, this leads to the possibility of endpoint detection and real-time elemental mapping that can be utilized in 3D tomography workflows.

[0013] According to the first aspect of this instruction, the analysis method is: The method involves guiding and focusing a beam of ions containing a mixture of protons and non-hydrogen ions onto a sample, wherein the kinetic energy of the ions in the mixture is 50 kiloelectron volts (keV) or less, and the method involves guiding and focusing the beam. This includes detecting and measuring X-rays emitted from a sample in response to collisions of protons and non-hydrogen ions onto the sample.

[0014] According to the second aspect of this instruction, the method for milling a region of the sample surface is: A beam of ions containing a mixture of protons and non-hydrogen ions is guided and focused onto a region, wherein the kinetic energy of the ions in the mixture is 50 kiloelectron volts (keV) or less, the ion beam is raster-scanned across the region, and collisions of non-hydrogen ions cause sputtering on the surface of a sample within the region, guided and focused. This includes detecting and measuring X-rays emitted from a sample in response to collisions of protons and non-hydrogen ions onto the sample. [Brief explanation of the drawing]

[0015] The above and various other aspects of the present invention will become apparent from the following description with reference to the accompanying drawings, which are given only as examples and are not necessarily drawn to a specific scale. [Figure 1A] This is a schematic diagram of a first apparatus in which the method described herein may be implemented, the apparatus comprising a focused ion beam (FIB) column, a plasma ion source adapted to generate ions from a mixture of gases and supply the ions to the FIB column, and an X-ray detector. [Figure 1B] This is a schematic diagram of a dual-beam apparatus in which the method described herein may be implemented, the apparatus comprising a FIB column, a scanning electron microscope column, a plasma ion source adapted to generate ions from a mixture of gases and supply the ions to the FIB column, and an X-ray detector. [Figure 2] This is a schematic diagram of a known inductively coupled plasma ion source for generating ions and supplying them to a FIB column. [Figure 3] This is a schematic diagram of the sample-facing ends of the SEM and FIB columns of a dual-beam scanning microscope and / or milling apparatus, illustrating the separation of an ion beam containing multiple ion species into multiple beamlets, each beamlet containing a subset of ion species due to the interaction between the ions and the immersion magnetic field. [Figure 4] This plot shows the measured lateral deviation of the H3+ ion component of the hydrogen beam as a function of the relative magnetic field strength of the immersion magnetic field through which the ions pass after exiting the FIB column. [Figure 5] This is a plot of the measured currents of separated beamlets generated from a 650 pA hydrogen beam produced from a 200 W inductive plasma ion source. [Figure 6A] This plots the change in the percentage of measured current intensities of various separated beamlets as a function of radio-frequency (RF) power applied to the inductively coupled plasma ion source. [Figure 6B]Figure 6A plots the integrated measurement currents of all beamlets as a function of the RF power applied to the inductively coupled plasma ion source. [Figure 6C] This plot shows the change in the proportion of each of the multiple ion species that make up a multi-ion beam, as a function of the total beam current. [Figure 7] Al sputtered with 1 μm aluminum

[0016]

number

[0017] The following description is provided to enable those skilled in the art to create and use the present invention, and is provided in the context of a particular use and its requirements. Various modifications to the described embodiments will be readily apparent to those skilled in the art, and the general principles herein may be applied to other embodiments. Accordingly, the present invention is not intended to be limited to the shown embodiments and examples, but rather to provide the broadest possible scope according to the shown and described features and principles. For a more detailed and complete understanding of the features of the present invention, please refer to Figures 1A to 18 in conjunction with the following description.

[0018] In the description of the invention as described herein, unless otherwise implicitly or explicitly understood or stated, a singular term encompasses its plural equivalents, and a plural term encompasses its singular equivalent. Furthermore, unless otherwise implicitly or explicitly understood or stated, it is understood that for any given component or embodiment described herein, any of the possible candidates or substitutes listed for that component may generally be used individually or in combination with each other. Moreover, it should be understood that the figures shown herein are not necessarily drawn to scale, and some elements may be drawn simply to illustrate the invention. Reference numerals may also be repeated in various figures to indicate corresponding or similar elements. In addition, unless otherwise implicitly or explicitly understood or stated, any enumeration of candidates or substitutes is for illustrative purposes only and not limiting.

[0019] Apparatus Embodiment Figures 1A and 1B schematically illustrate two examples of apparatuses according to this teaching, on which the method according to this teaching can be implemented. In both apparatus 1a and apparatus 1b (Figure 1B), a FIB column 6 housed in a vacuum chamber 13 receives a mixture of ions, including at least hydrogen ions (i.e., protons) 2 and ions heavier than protons 3, from an inductively coupled plasma ion source 4, which includes a coiled electrode 5 to which a radio frequency (RF) voltage waveform is applied during operation. The ions are received from a gas inlet pipe 11 and generated from a gas containing a mixture of hydrogen, including hydrogen molecules 2m, and at least one other non-hydrogen gas containing molecules 3m. These different gases are supplied in a purified form and can be mixed in appropriate proportions in a gas mixing manifold (not shown) before being introduced into the gas inlet pipe 11. Alternatively, the gases may be supplied in a pre-mixed form, thereby eliminating the need for a gas mixing manifold and any metering valves required for proportional mixing. When an appropriate RF voltage is applied to the coiled electrode 5, the plasma is ignited in the ion source 4 in a known manner.

[0020] Continuing with Figures 1A and 1B, ions 2 and 3 are focused onto beam 7 and guided along the length of the FIB column 6 toward the beam focus on the surface of sample 8 by electric and / or magnetic fields applied to a series of ion optics 15, which may include ion lenses and ion guides. The ion optics accelerate the ions toward the sample, causing the ions in the ion beam to collide with the sample surface at kinetic energies of less than 50 keV. As will be described in more detail below, when both proton 2 and the heavier ion 3 collide with the same focus on sample 8, a measurable amount of X-ray photons 9 are generated, which are derived in known ways by emitting inner-shell electrons from the atoms of the sample and filling the resulting electron holes with electrons from higher-energy electron shells. The resulting X-rays are detected by an energy-dispersive X-ray detector 10 and recorded as an energy-dispersive spectrum. In this way, the elemental composition of the sample at the beam focus can be determined. Apparatus 1b further includes a scanning electron microscope (SEM) column 12 in a vacuum chamber, used to guide and focus the electron beam onto the same focal point or, via rasterization of the electron beam, onto a region of the sample surrounding the focal point of the ion beam 7. An image of the region of the sample may be created in a known manner using the detection of secondary or backscattered electrons by an electron detector (not shown).

[0021] Selection of an ion source Liquid metal ion source The main factor distinguishing focused ion beam (FIB) microscope types is the ion source. This supports the morphology and function of the microscope, as well as its intended use in processing and analysis. The most common type of ion source in focused ion beam microscopes is the liquid metal ion source (LMIS). The most commonly used metal for these applications is gallium (Ga), mainly due to its low melting point (29.8°C) and high mass, and therefore sputter yield. [4] When a large negative potential is then applied between the needle tip and the extraction electrode, the extraction bias is balanced by the surface tension of the Ga liquid, constricting the liquid to a fixed point known as a Taylor cone, and the cusp forms a tip radius of about 5 nm at the tip of the Taylor cone. [5] In high vacuum, this potential is equivalent to the Ga from the cusp. + This is sufficient to generate ion field emission. [31,32,33] Modern LMIS-based FIB instruments have a beam spot size of about 10 nm. [7] Because LMIS can only be operated with a liquid metal source, it cannot generate light ions such as protons necessary to generate X-rays by the VLE-PIXE technique.

[0022] Gas field ionization source Gas field ionization source The Source (GFIS) utilizes a metal needle whose final radius at the tip is exactly three atoms.

[34] A gas at a very low partial pressure is introduced into the needle, where gas molecules rapidly adsorb to the tip surface. A strong electric field is then applied to the tip, resulting in field emission from only the three gas atoms adsorbed at the final radius of the tip.

[35] While typically working with helium gas, these ion sources can also generate heavier ions such as neon with higher sputter yields for faster FIB processing.

[35] Regarding lighter ions such as protons, Moritani et al. and Matsubara et al. demonstrated the generation of protons in a GFIS source. However, the generated protons are H2 + and H3 +It is produced in very small proportions relative to the ions. [36,37,38]

[0023] A typical FIB operates with a fixed current, modifying the beam current by changing the aperture size, whereas the current of a GFIS source is modified by changing the pressure of the gas surrounding the tip. Higher pressures allow for faster replenishment of gas atoms on the tip, resulting in faster extraction rates and higher currents. However, as the gas pressure increases, the potential barrier for field ionization also increases, resulting in a maximum operating current of about 100 pA. [39,40] The small number of protons produced by this ion source, coupled with the very small current available to the GFIS, makes this ion source unsuitable for VLE-PIXE analysis. Furthermore, a GFIS can only host one source species at a time and cannot be used for doped beam VLE-PIXE as required by the present invention.

[0024] Inductively coupled plasma ion source Plasma is defined as a fully or partially ionized gas consisting of electrons and ions, and is identified as a readily available ion source for use in FIB systems. Plasma sources are one of several types, named after the technique used to generate the plasma. One major advantage of plasma ion sources is their ability to accommodate multiple ion species, a requirement of doped beam VLE-PIXE. Inductively coupled plasma (ICP) can be generated by using an antenna located outside the plasma chamber (e.g., coiled electrode 5 in Figures 1A-1B). The main advantage of the ICP ion source configuration is that the plasma itself does not come into contact with the antenna, which significantly increases the antenna's lifetime and expands the range of possible plasmas to include those containing oxidizing gases. ICP is formed by applying RF waves to a helical or coiled conductive antenna separated from the gas by the use of a dielectric insulator, typically high-quality quartz with low dielectric loss and good resistance to high coil temperatures.

[0025] At relatively low RF power, the RF field from the antenna penetrates the dielectric, providing a field strong enough to ionize the gas. This is known as the "E mode" because the primary excitation mechanism is the electric field of the RF coil, similar to that of a DC discharge. At higher RF power, the energy is sufficient to accelerate electrons in the sheath near the edge of the dielectric, also known as the skin layer. This is known as the "H mode" because a true inductive coupling is formed between the antenna and the plasma, thereby generating a very high-density plasma. An ideal ICP acts like a transformer, with the primary winding being the antenna and the induced electric field in the plasma forming a single-winding secondary coil. ICPs are used in applications requiring very high-density plasmas, are suitable for use as a high-brightness ion source, and are ideal for use as an ion source for VLE-PIXE.

[0026] Comparison of ion sources Table 1 shows a selection of measured ion beam source metrics from the literature for the most common FIB sources: LMIS, GFIS, and ICP. Underlined species indicate the source used for the listed metrics. Based on this table, it can be seen that, in general, instruments capable of high resolution are limited in terms of the current they can supply, and vice versa. In VLE-PIXE, due to the low efficiency of the technique at such low energies, ICP is an essential development for delivering very high effective current to the sample by an ICP ion source. The ability to generate protons by utilizing hydrogen precursors is also an essential development required for VLE-PIXE in FIB microscopy, and the ability to simultaneously host multiple source species is important for doped beam VLE-PIXE experiments.

[0027] [Table 1]

[0028] Configuration of the selected plasma source Based on the above comparison of ion sources, the inventors have found that known FIB apparatuses equipped with an ICP ion source and a gas mixing apparatus, as described herein by reference, as PFIB apparatuses, such as those described in U.S. Patent Nos. 8,076,650 and 8,822,913, are optimal for VL-PIXE operation. Figure 2 is a schematic diagram of an ion source 100 used in such a PFIB apparatus. A predetermined gas mixture is supplied by a gas mixing system 220. The gas mixture is supplied from an external gas supply line 104 through a gas filter 106 to a plasma chamber 102 in a source tube 103, and then to a capillary tube 108 having a flow limiter 110. Energy is supplied from an RF power supply 113 to the plasma chamber 102 by an antenna coil 114, and ions are extracted by an extraction electrode 120 through a source electrode aperture 116 in a source electrode 118. The segmented Faraday shield 121 reduces capacitive coupling between the coil 114 and the plasma in the chamber 102, thereby reducing energy diffusion of extracted ions within the chamber 102. The power supply 113 preferably drives the antenna 114 in a "balanced" manner, i.e., the electrical phase shift across the antenna is tuned to reduce modulation of the plasma potential, as described in U.S. Patent No. 7,670,455 incorporated herein by reference. The balanced antenna preferably provides a null point in the radio frequency energy field within the plasma, which reduces energy diffusion of ions extracted from the plasma chamber 102.

[0029] Gas conductance to and from the plasma chamber 102 passes through a flow limiter 110 in a capillary tube (located at the top of the source tube 103) and an aperture 116 (typically less than 1 / 4 mm in diameter) in the source electrode 118. A pump 122, connected to a gas supply line 104 via a valve 123, removes gas from the plasma chamber 102 through the capillary tube 108 and the gas supply line 104. An ion column pump (not shown) extracts gas from the plasma chamber 102 through the source electrode aperture 116.

[0030] The gas mixing system 220 receives gas from multiple gas sources, such as gas storage units 130A, 130B, 130C, and 130D, and supplies the gas to the gas supply line 104 through the corresponding valves 131A to 131D. The valves 131A to 131D are adjusted to provide the gas to the plasma chamber 106 by supplying the desired gas mixture to the gas supply line 104. Multiple types of gas can be supplied to the plasma chamber simultaneously by opening multiple valves 131A to 131D at the same time. Preferably, the valves 131A to 131D are metering valves that control the ratio of gas to the gas inlet 104.

[0031] The beam voltage source 132 supplies a high voltage to the plasma in the chamber 102, and the extraction voltage source 134 supplies voltage to the extraction electrode 120. The extracted ions or electrons are focused by the focusing electrode 136. Further details of the focusing column and sample chamber are not shown.

[0032] To remove gas from inside the plasma chamber, the gas supply line 104 is pumped as shown to remove gas from the source tube above the flow limiter 110 in the capillary tube 108. The volume of the FIB system below the source electrode 118 can also be appropriately pumped using a main chamber vacuum pump (not shown).

[0033] Both the source electrode aperture 116 and the flow limiter 110 have small diameters and correspondingly very low gas conductance, making it impossible to rapidly pump out the inside of the source tube 103. This is particularly disadvantageous for generating FIB systems where it may be desirable to perform continuous process steps using different ion species. Firstly, it may take much longer to evacuate the first process gas from the source tube 103 before the base pressure becomes low enough to introduce a second process gas. Insufficient gas purging can lead to plasma contamination by ionization. U.S. Patent No. 8,633,452, incorporated herein by reference, describes a plasma chamber design that rapidly changes the gas within a plasma source by providing an alternative path for the gas to enter and exit the vacuum chamber.

[0034] The mass filter 202 is located below the ion source. The mass filter 202 is preferably an E×B filter, but other types of mass filters may also be used. The mass filter 202 includes an electrode 204 that provides an electric field and magnets (not shown) positioned above and below the plane of the paper to provide a crossing magnetic field. A connector 206 provides an electrical connection to the electrode 204 and a mechanical connection for adjusting the position of the electrode 204. The electric field is adjustable to select the mass of ions that pass through the filter without deflection and through the aperture in the beam path. Ions with masses other than the selected mass are deflected and do not pass through the aperture. While the mass filter 202 is shown schematically, it may comprise a more complex mass filter or a composite filter having two or more regions of E×B fields separated along the beam axis.

[0035] Gas delivery system As shown in Figure 2, the emergence of the combination of the plasma ion source 100 and the gas mixing system 220 made it possible to apply up to four parallel source species. Typically, Xe, Ar, O2, and N2 are supplied from gas storage units 130A-130D used in conventional FIB operation. Gases Xe and Ar are generally selected due to their high mass, making them ideal for rapid sample processing (e.g., etching, milling) due to their high sputter yield and non-reactive nature, thereby reducing the impact on the chemical and electronic properties of the sample. Gases O2 and N2 are generally selected due to their availability and the potential for interesting applications due to the chemically active nature of these species. Although the sputter yield of these species is much lower than that of Xe and Ar, they are still not negligible and can be used for processing and preparing materials by FIB milling. One example is the use of an O2 beam for processing organic materials, as the chemical properties of the oxygen species allow for volatilization of the sputtered species and prevent redeposition of carbon-based materials.

[47]

[0036] The VLE-PIXE technique taught herein relies on the generation of lighter ion species, such as protons, in a beam that also contains heavier ion species. The coupled plasma ion source 100 shown in Figure 2 provides the capability to generate such a coupled ion beam. Traditionally, hydrogen has been unsuitable for material preparation and processing under conventional FIB operation due to its very low mass and therefore very low sputter yield. However, the inventors have recognized that the true potential of hydrogen-containing ion beams lies instead in the analysis of materials, which has many possible applications, such as VLE-PIXE. Several previous attempts have been made with H + Ions have long been introduced into FIBs using gas field ionization sources (GFIS). [36,37,38,39,40] However, the very low currents available, especially the proton current, make material analysis with such instruments difficult.

[0037] To perform VLE - PIXE as described herein, it is desirable that the current from the protons be a significant amount, because the low XRPCs at such low energies can be compensated somewhat by increasing the number of protons incident there within a given time frame. For performing doped - beam VLE - PIXE, the ability to support multiple ion species simultaneously, as provided by the combined plasma ion source 100 and gas mixing system 220, is advantageous for success. However, as described above, hydrogen is not suitable for material preparation and processing under conventional FIB operation because of its very small mass and very low sputtering yield. Since the currently described method requires hydrogen, a standard supply oxygen cylinder was simply replaced with a cylinder of high - purity hydrogen. The removal of the oxygen cylinder was also for safety purposes, as the combination of oxygen and hydrogen on the same system created an explosion hazard.

[0038] According to the apparatus of the present teachings, the hydrogen gas required for VLE - PIXE analysis is transferred to the inlet of the plasma chamber through a metering orifice that ultimately controls the pressure within the plasma chamber. The hydrogen gas and another, heavier gas are transferred through a series of shut - off valves to a gas delivery manifold that supplies the resulting gas mixture to the metering orifice. The back - pressure on this orifice, provided by the gas delivery system, determines the flow rate through the orifice and, for the particular apparatus of the inventors, is optimal in the range of 2 - 4×10 3 mbar, as measured just before the orifice on a positive - pressure gauge (piezo - based gauge). The evacuation rate of the plasma chamber is fixed by the pumping speed of a turbo - molecular pump (not shown), thereby creating a pressure drop between the metering orifice and the exit aperture of the plasma chamber. Although the exact pressure within the plasma chamber is unknown, using the particular apparatus of the inventors, it has been demonstrated that a measured pressure of 6 - 7×10 -6 mbar on a cold - cathode vacuum gauge after the plasma chamber is optimal for hydrogen plasma ignition.

[0039] Operation of the plasma source Once a stable gas pressure is established within the plasma chamber, the plasma is ignited. This ignition process typically occurs automatically following a gas species switch. The following are typical RF plasma operating parameters determined using our apparatus, and are not intended to be limiting. ● The RF impedance matching capacitor is set to the pre-ignition value. ● The RF generator provides RF waves at the desired power and frequency. ● The plasma igniter discharges, which ignites the plasma inside the chamber. ● Plasma detection logic is applied, which relies on the increase in reflected RF power during plasma ignition. ● The RF impedance matching capacitor moves to the post-ignition value and automatically tunes to minimize reflected power. Once ignited, the plasma is stabilized for a period of 15–30 minutes before use to achieve the most stable beam. Operation of the plasma source during beam doping typically involved extinguishing the plasma, establishing a new precursor gas mixture, and reigniting the plasma source. Some manual RF capacitor tuning was also required, as the optimal tuning value for the plasma system is established for a single gas species.

[0040] Throughout the beam doping experiments, it was observed that the addition of dopant species to the source plasma could have unexpected effects on the plasma composition, and therefore the beam composition. As will be further explained below (see, for example, Figure 10), changing the partial pressure of Ar in the plasma source can result in a proportion of hydrogen molecular species in the beam that differs somewhat from the proportion expected based on a pure hydrogen beam. This is fairly typical of multi-gas plasmas, as Bai et al. explained, that plasma parameters can change significantly along with the mixing ratio in inductively coupled plasmas.

[48]

[0041] Characterization of ion beams Understanding the molecular and isotopic composition of a focused ion beam is fundamental to verifying the successful implementation of the VLE-PIXE technique, as well as to understanding the beam-sample interaction and the resulting physical processes. In particular, in doped beam VLE-PIXE experiments, the plasma is H + Ions (i.e., protons), and H2 + and H3 + Ions and a certain proportion of dopant species are generated, the ratio of which depends on several factors such as plasma power, pressure, and the partial pressure of the gaseous precursor. Characterizing the beam before it is incident on the sample is the best way to determine the experimental conditions. This section teaches the techniques used to characterize the beam composition.

[0042] Beam Chemistry Several reaction pathways exist within the plasma that lead to the formation of various primary hydrogen ions.

[49] The ratio of these species depends heavily on plasma conditions such as pressure, power, electron density, chamber size, and electron temperature. Fukumasa et al. found that when the source pressure is 1 × 10⁻⁶ -4 mbar to 1 x 10 -3 When changed to mbar, all other parameters remain constant, H + H2 + and H3 + It has been demonstrated that the molecular ratio of these species can be dramatically shifted to 70% / 15% / 15%~10% / 30% / 60%.

[49] For this reason, the determination of the ratio of hydrogen molecular species should be done by the precise characterization of the current of each individual hydrogen species in the sample. To perform such a procedure, a single ion beam composed of the above species should be separated into individual beamlets, and the ion beam current of each beamlet should be characterized. A method for doing so is described in the following subsections.

[0043] Splitting an ion beam into "beamlets" Magnetic immersion lens of an electron column (for example, SEM column 12 in Figure 3) The immersion lens (MIL) acts as the final lens for the electron beam and utilizes a strong magnetic field to help guide the electrons collected by the detector located within the column. [50, 51] This magnetic field is known to cause ion deflection as ions move from the FIB column to the sample as a result of the Lorentz force applied to positively charged ions, as schematically shown in Figure 3. Importantly, this Lorentz force depends on the mass-to-charge ratio of the ions, with a greater force applied to lighter ions than to heavier ions.

[0044] The strength of the magnetic field generated by the magnetic immersion lens (MIL) can be adjusted by increasing or decreasing the current flowing through the MIL, which can be applied in both positive and negative polarity. This procedure is typically used to adjust the focus of the electron beam on the sample. However, the magnetic immersion lens contains a ferromagnetic core that plays a role in enhancing the magnetic field generated by the coil. As a result, when the current applied to the lens coil is removed, a residual magnetic field still exists. Therefore, a current must be applied to the lens coil in such a way that it acts to compensate for the residual magnetic field, resulting in an effective magnetic field free state. This is typically applied during FIB operation of a conventional dual-beam apparatus, such as apparatus 1b in Figure 1B, where the ion beam moves freely onto the sample without deflection.

[0045] However, changing the MIL current creates the possibility of intentionally inducing deflection of ions in the beam. The magnetic field is aligned so that the deflection is only in the X direction relative to the beam scanning field of view. This results in the separation of ion beam components by mass / charge ratio, with the heaviest components falling closest to the beam axis and the lightest components falling at a certain distance. Figure 3 is a schematic diagram of the magnetic field immersion lens 21 and the resulting immersion magnetic field 23 near the sample-adjacent ends of SEM column 12 and FIB column 6, showing the separation of the ion beam 7 into multiple deflected beamlets. The degree of deflection can be described using the Lorentz equation and therefore can be used to identify various separated species from the location of burn marks generated on the surface of the sample 8.

[0046] Measured transverse deflection H3 of ions from a position without a magnetic field. + This is shown in Figure 4. From these values, the magnetic field strength can be determined by solving the Lorentz equation. This information can then be used to predict the transverse deflection of several common mass components for a given relative magnetic field strength. The calculated deflections were found to be very accurate when compared to the spot burn images of the separated beams. Conversely, the masses of the species can be calculated based on their transverse deflection, which are identified using the spot burn images of the separated beams.

[0047] Following mass calculations, several considerations must be taken to isolate the most likely molecular or isotope candidates. One such consideration is whether the majority of ions generated from the source gas into the plasma are ionized on their own. This determination is important because mass identification is based on the mass-to-charge ratio (m / z) rather than simply the ionic mass. Thus, a double-charged ion could be misidentified as an ion having half its actual mass. The performance of an early version of the previously reported Helios Hydra PFIB source showed a double-ionization rate of only 0.75% for an ICP source operating with Xe at 300 W RF power.

[46] This power is significantly higher than the 37 W typically used with Xe in commercial versions of this ICP source, and therefore the rate of double-ionized Xe is expected to be even lower than 0.75%. Based on its most abundant isotope, the bicharged Xe with an average m / z of 66 Thomson isotopes can also be clearly distinguished from any other possible beam species, considering that other candidate species are transition metals such as single-charged copper, zinc, or gallium, which should not be present in the beam.

[0048] It has been shown that Ar can generate a double ionized species at a typical operating power of 200 W. 2+ Spot also, Ne + This could be due to Ar 2+ It can be clearly identified as. Research on nitrogen ICP sources with a maximum power of 400W shows that N2 2+ Species generation is undetectable, and dissociated N + The species demonstrated that it is far more likely.

[52] Nitrogen species N2 2+ Also, N + It overlaps with and is indistinguishable. Research on oxygen ICP sources with a maximum power of 300W also involves double ionized O2 2+ Unable to detect O and O +This demonstrates that dissociation into species is far more likely to occur.

[53] It is impossible to double ionize an H2 molecule, because there are no electrons left to facilitate bonding, which results in two free H + Because it is equivalent to an ion.

[0049] Another process that must be considered is the possibility of chemical reactions occurring in the plasma source. While gaseous Ar is typically considered a non-reactive noble gas, it can form molecules with various other species in the plasma. These molecules, known as polyatomic interference, are particularly prevalent in ICP mass spectrometry, where the Ar carrier gas can form molecules with sample species. These interfering molecules are, 52 Cr, 56 Fe, 75 As, and 80 Mistakenly identified as Se, 40 Ar 12 C, 40 Ar 16 O, 40 Ar 35 Cl, and 40 Ar 40 This can lead to inaccurate identification of species such as Ar.

[54] These polyatoms fall between the primary m / z peak of Ar at 40Th and the primary peak of Xe at 132Th. Since no other ions are expected between these two masses, this simplifies the identification of Ar polyatoms.

[0050] Measurement of beam component current Once the beam components are identified, each beamlet species can be quantified. For these measurements, a standard Faraday cup design is ineffective because it records the current of all charged particles hitting the top surface of the cup, thus measuring only the total beam current. A special Faraday cup must be used so that particles impacting the top surface of the cup are rejected, and only those entering through the aperture are measured. Therefore, a special Faraday cup, known as a "separation" Faraday cup, was designed for this purpose.

[0051] Figure 18 is a schematic cross-sectional view of the isolation Faraday cup 400 according to this instruction, illustrating the operating principle of the apparatus. The Faraday cup 400 comprises a first electrode 401 and a second electrode 403 separated from the first electrode 401 and separated from the first electrode by one or more electrical insulating spacers 402. For example, the apparatus may generally be cylindrical (or any other shape), in which case the apparatus may comprise a single ring-shaped spacer 402, and the first electrode may be in the form of a circular plate. The first electrode comprises apertures having a diameter smaller than the spacing between individual isolated beam components (i.e., beamlets), five of which are shown as 7a, 7b, 7c, 7d, and 7e. Thus, ions from only one such beamlet (e.g., beamlet 7a) can pass through the aperture at any time. Ions from other beamlets are collided with the surface of the first electrode 401. Since the first electrode 401 is grounded via the conductor 407, the current originating from the ions of the beam component striking the first electrode flows directly to ground. Only the charge of the beam component (e.g., individual beamlets) entering the aperture 404 is collected by the isolation Faraday cup 400 and enters the charge collector cavity 405. The current entering the cavity through a single beam component passing through the aperture 404 can be measured by a potentiometer (not shown), such as a pico-ammeter electrically coupled to the second electrode 403 by the conductor 408. Similar designs have been previously implemented for measuring the shape of an ion beam.

[76]

[0052] The charge collector cavity 405 of the second electrode 403 does not necessarily have to be provided in the specific shape shown in Figure 18. More generally, the cavity may be shaped to improve the collection efficiency and accuracy of the device by directing ions and / or secondary electrons to the inner walls of the cavity rather than directing them upward toward the cavity opening. Thus, the cavity may include additional cavity segments and additional cavity walls that are not necessarily parallel to any of the cavity walls shown in Figure 18.

[0053] To measure the current associated with each individual component of the ion beam, the separated beamlets may be scanned across the surface of the isolation Faraday cup 400 and across the aperture 404. Arrow 406 schematically illustrates the relative motion between the beamlet set and the isolation Faraday cup 400. The scanning process can be achieved by using a built-in scanning pattern generator of a commercially available FIB or SEM microscope. Alternatively, the isolation Faraday cup may be programmed to move in small steps by a movable stage assembly while the ion beamlets remain stationary.

[0054] Figure 5 shows beamlet current measurements for a 200W hydrogen plasma, where the current of each beamlet can be converted to a percentage of the total beam composition. In this scan, species H + H2 + and H3 + N2 + , O2 + Ar + and Xe + It can be identified in the same way as the contribution from species. The composition ratio of this beam is shown in Table 2.

[0055] It is known that the RF power used to generate the source plasma can alter the composition of the hydrogen ion beam due to changes in plasma density and electron temperature that affect the reaction pathways within the source plasma.

[49] Therefore, as shown in Figure 6A, the beam composition was measured as a function of the plasma source RF power to establish the optimal RF power for proton production. Traces 31, 32, 33, 34 and 35 in Figure 6A relate to the measured proportions of each species.

[0056] [Table 2] H3 + H2 + , N2 + +O2 + Xe + and H +As shown in Figure 6B, increasing the plasma RF power increases the H in the beam. + An increase in the ion proportion is observed, and correspondingly, the total beam current increases. Therefore, by operating the plasma source at the highest possible RF power, both the maximum beam current and the maximum proportion of protons in the hydrogen beam can be achieved. However, increasing the plasma power also raises the temperature of the plasma chamber, which can result in permanent damage to the plasma source. For this reason, 200W of power was considered the maximum safe operating RF power.

[0057] H in the beam + Because the proportion of ions (protons) is relatively low, a high current must be used to generate a sufficient X-ray signal for analysis during VLE-PIXE measurements. Therefore, compositional current measurements were performed to determine whether increasing the beam current affects the proportion of species in the beam. The results of these measurements are shown in Figure 6C. Traces 41, 42, 43, 44, 45, and 46 in Figure 6C represent species H3, respectively. + H2 + H + , N2 + +O2 + Xe + and Ar + This relates to the measured proportion. These measurements demonstrate that only small changes in beam composition occur over changes of several orders of magnitude in beam current. For this reason, beamlet characterization can be effectively performed at lower currents with smaller beam spot sizes, resulting in better separation between beamlets and reduced damage to the isolation Faraday cup, especially when characterizing doped beams that have higher sputter yields due to the presence of heavier ion species.

[0058] Beam purity control Achieving a high-purity hydrogen beam presented a significant challenge due to residual contamination remaining in the FIB source during source type switching, such as N2 from air pollution. + and O2 + , as well as Ar due to residual contamination from previous ion source species and previous beam doping experiments+ and Xe + Several types of pollutants were commonly observed in the beam.

[0059] The gas delivery system and plasma source are pumped using the PFIB vacuum system during gas source switching, but the effectiveness of contamination reduction by this pumping depends on the pumping time before the introduction of the next gaseous source species. In this situation, the pumping rate is limited by the gas flow rate, not by the pump, and is determined by the conductance of the system.

[55] Several factors, such as the long, narrow tubing used in the gas delivery line, the metering orifice at the inlet of the plasma chamber, and the aperture separating the differential pumping zone in the FIB column, limit the conductance of the vacuum system, and as a result, if a sufficiently long pumping cycle is not used, contaminating gas molecules remain in the system. In applications requiring high-purity hydrogen beams, the gas delivery system and plasma source chamber have been evacuated by vacuum pumping for extended periods, typically overnight. An overnight vacuum pumping cycle has been found to reduce the Ar value to less than 1% of the beam composition.

[0060] To further reduce the number of contaminant species, additional efforts are required, such as introducing the desired gas species through multiple purge cycles and pumping the system several times to allow scrubbing of residual gases. However, multiple purge cycles waste a considerable amount of high-purity source gas, are costly, and may require frequent replacement of source gas cylinders. For critical experiments where the absence of other beam species is desired, a technique may be used to split the ion beam into separate beamlets, directing different beam components (i.e., different ion species) to different regions on the sample. This technique has been advantageously used, for example, to isolate X-ray signals generated on a sample region by specific ion species.

[0061] Utilizing shorter vacuum pump cycles allows small amounts of Ar to remain in the beam, which is typically undesirable when a high-purity hydrogen beam is required. However, it has been found that short vacuum pump cycles provide a simple way to achieve low-concentration doped hydrogen beams without requiring additional gas mixing in the source gas delivery system. Low-concentration doped (less than 10%) beams could be easily achieved using modified pump cycles (e.g., as proposed in the following sections) by starting with a source composed of the desired dopant species and switching to a hydrogen beam, thereby stopping the delivery of Ar gas and changing to the delivery of hydrogen gas, so that a significant proportion of the dopant species remained in the plasma source and gas delivery lines. Although this method did not provide much control over the proportion of dopant species, post-characterization of the beam composition was typically performed, and the measured beam composition could be explained in any experimental data. To achieve higher-concentration doped beams, it was necessary to mix the gases in the plasma chamber.

[0062] Optimized beam chemistry control parameters The following are presented as typical operating parameters determined using our apparatus, and are not intended to be limiting.

[0063] Source Plasma The gas supply regulator is 2-4 x 10 -3 The plasma chamber outlet pressure should be adjusted to achieve mbar. -3 It should be in the mbar range. Plasma power should be set to 200W or higher if there is no risk of damage.

[0064] Beam splitting Beam splitting should be used to isolate individual components. For beam splitting, the immersion lens coils of adjacent SEM column currents should be adjusted to correct the intensity of the immersion magnetic field. A coil current of approximately 2.04 ampere-turns corresponds to the focused beam, and a coil current of 0 ampere-turns for 30 keV in eucentricity corresponds to the lightest species (H + ) and the heaviest species (Xe) at approximately 550 μm + This corresponds to the separation between ) and ).

[0065] High-purity hydrogen beam The focused ion beam gas delivery line and plasma chamber should ideally be vacuum-pumped overnight to remove any residual gas species. For very pure beams, the desired gas should be introduced into the system, then pumped out over a long period of time, and then reintroduced for plasma ignition.

[0066] Low-concentration doped hydrogen beam Starting from the source of the desired dopant species, the beam should be switched to hydrogen in a modified pumping cycle. With a primary vacuum pump delay of 10 seconds and a turbomolecular pump delay of 10 seconds, approximately 10% of the dopant species will be yielded, with the remainder being hydrogen.

[0067] Minimizing beam scattering The chamber of the focused ion beam apparatus should be pumped evacuated for at least two hours after reaching the target vacuum pressure before starting the experiment. The pressure in the lower FIB column, generally maintained by an ion getter pump, should be monitored and equilibrated before starting the experiment to avoid changes in beam current or composition during the experiment.

[0068] Verification of VLE-PIXE signal generation As outlined below, experiments were conducted to verify that the detected X-ray signals were indeed generated as a result of proton collisions with the sample. Furthermore, the role of spurious signals such as stray electrons, neutralized ions, and backscattered ions generated by the microscope and related instruments was analyzed, as were other spectral artifacts, including those originating from the EDS detector itself. Optimized operating parameters to reduce the influence of these spurious signals on VLE-PIXE measurements are described.

[0069] Proton contribution to the signal It is well known that for a given acceleration energy, the X-ray production section (XRPCS) is maximum for protons compared to heavier ions, due to their low mass and consequently high speed. Since the XRPCS decreases rapidly as the ionic mass increases, for example, with a He of 2amu mass... + It is true that even ions cannot generate enough X-rays for analysis.

[56] For this reason, the VLE-PIXE signal is expected to be generated by protons. However, as mentioned above, doped beams from inductively coupled plasma ion sources contain various ion species. It is necessary to experimentally verify that the proton component of the beam is responsible for generating most or all of the intensity of the X-ray spectrum observed when such a beam is guided onto a sample.

[0070] As previously described herein, an immersion magnetic field generated by a magnetic immersion lens placed in close proximity to the ion beam can spatially separate the ion beam into its individual mass components. The separated ion components of the beam can then be individually guided toward a sample while measuring the resulting X-ray emission. In particular, experiments were conducted in which the separated beam components were guided toward a binary sample containing a 1 μm aluminum layer sputter-coated onto half of a single-crystal Cu(100) substrate. By moving the separated beamlets across the interfaces between different materials on the binary sample, the X-ray signal contribution from each beam type could be determined.

[0071] Prior to the above experiment, a 2nA preset, 30keV hydrogen beam was measured before the experiment. The beam showed no detectable Ar signal and contained 10.84% ​​H + , 11.94% H2 + 53.14% H3 + , 5.69% O2 + and N2 + , and 18.37% Xe + It was shown that it consists of H + The beam was positioned so that only the component fell onto the Al film, and the rest fell onto the Cu substrate. The X-ray spectrum was captured for 20 minutes. Next, H + Components and H2 + The stage on which the sample was mounted was moved so that both components fell onto the Al film. The stage movement was repeated for the remaining beam components, and the spectra for each condition were compared. To isolate the contributions from each, the spectra from each successive irradiation were subtracted from the previous one. The Al peak with the highest intensity was Al

[0072]

number

[0073]

number

[0074] As shown in Figure 7, most of the X-ray signal from Al is due to H to the sample. + It is produced from collisions of ions (i.e., protons). + Despite being a higher proportion than ions, H2 + and H3 +It is also observed that the contribution of molecules to the signal can be negligible. The use of a magnetic field immersion lens for beam separation in this experiment also has the additional advantage that any stray electrons resulting from ion collisions with residual gas molecules will be deflected away from the sample.

[0075] The signal generated by protons alone is the primary contributor to the spectrum, but it is still much weaker than the signal generated when the hydrogen beam is doped with heavier ion species. This is evident in the doped beam spectra of NIST standard materials. α Compared to the line (for example, Figure 14A), the Al K at 1.486 keV in the X-ray spectrum generated by protons alone. α This can be determined by the complete absence of the line, but the concentration of Al in the latter sample is remarkably low (see, for example, Table 4). Al K in NIST samples α The presence of the line is, Al

[0076]

number

[0077] Interference from backscattered ions At acceleration voltages greater than 24 keV, a broad and strong background signal was found to appear at low X-ray energies and extend down to energies of 3–4 keV. Initially, this was thought to be bremsstrahlung radiation resulting from primary ion collisions with the sample surface, but proton collisions at such low energies are not expected to generate detectable bremsstrahlung radiation. In addition, the background was shown to clearly scale in intensity with increasing target masses: Mg (Z=12), Cu (Z=29), and Au (Z=79), a phenomenon not typically observed to such an extent in bremsstrahlung radiation.

[0078] To confirm the nature of this background, two spectra were collected on the same single-crystal Cu sample using a windowed EDS detector and a windowless EDS detector at 8.1 keV, below the background threshold. It was found that background was present when the spectrum was captured with the windowless detector, but not when the windowed detector was used. This result demonstrates that the window plays a role in blocking the background signal.

[0079] It was concluded that this broad background signal is due to ions that are backscattered from the sample and collide with the EDS detector surface, thereby generating the signal. When a windowed detector is used, this signal does not appear up to an energy of 24 keV due to the transmission threshold of backscattered ions through the window. However, when a windowless detector is used, backscattered ions collide freely with the detector surface. The fact that a thin detector window can block all backscattered ions up to an energy of 24 keV while still allowing transmission of X-rays at very low energies is a significant advantage of VLE-PIXE over LE-PIXE and PIXE techniques. VLE-PIXE spectra were typically taken at energies below 24 keV to avoid the influence of backscattered ion background. However, spectra may be taken above 30 keV up to at least 50 keV if there is no risk of interference between the backscattered ion background and the analyte peak.

[0080] Dependence of signal on overlap of coupled proton / heavy ion beams As described above in this specification, the ion orbits can be differentially deflected according to the ion's mass-to-charge value by passing the ions through an immersion magnetic field as they move from the FIB column to the sample. As a result, the beam can be separated into individual mass components, as shown in Figures 8A and 8B. The magnetic field strength can be adjusted so that the beam components completely overlap (Figure 8A) or separate from each other as the magnetic field strength increases (Figure 8B). In the case of VLE-PIXE, the inventors' initial expectation was that the spatial separation of protons and dopant species (e.g., Figure 8B) would have no significant effect on the VLE-PIXE signal intensity. What was actually observed was that the X-ray signal decreased rapidly as the protons and heavy ion species were gradually separated. This is shown in Figure 8C, where the VLE-PIXE spectrum of the NIST standard material SRM 654b (further discussed below in the subsection on trace element analysis of this specification; see Table 4) is shown as a function of relative magnetic field strength (correlated with beam separation). Specifically, Figure 8C shows that the intensity of several prominent X-ray peaks changes as the magnetic field strength increases. The specific peaks shown in this plot were selected so that they remain present in all spectra. These results demonstrate that collisions of proton and heavy ion dopant species at the same location on the sample result in an X-ray signal greater than the sum of the hydrogen and heavy ion dopant signals. Scanning electron microscope images of the burn marks of the irradiation spots shown in Figures 8A and 8B demonstrate the change in beam component overlap with increasing magnetic field strength.

[0081] Furthermore, the data in Figure 8C shows that the decrease in the VLE-PIXE signal is not uniform across the entire X-ray spectrum. Instead, this decrease is mainly observed in the portion of the spectrum containing peaks with X-ray energies greater than approximately 1 keV. The intensity of the peaks in this low-energy region is relatively invariant with respect to the magnetic field strength. For example, as seen in Figure 8C, Ti

[0082]

number

[0083] It should also be noted that from the data plotted in FIG. 8C, the maximum X-ray intensity having a magnetic field strength is offset by about 20% from the zero relative magnetic field position. This offset is thought to be due to an offset in the magnetic field strength setting required to cancel the residual magnetic field of the magnetic field immersion lens coil, which becomes the reference point of zero relative magnetic field strength when adjusted. The zero relative magnetic field strength position is set by manually changing the intensity of the magnetic field immersion lens coil until all beam component secondary electron images visually converge. Stronger H2 + and H3 + species for H + The very weak intensity of the beam makes it very difficult to visually align the images generated by H + ions. Therefore, visually aligning the beam images may not result in a complete overlap of the H + species and the remaining beam components at the zero relative magnetic field strength point.

[0084] To conclude with confidence that the peak intensity behavior shown in FIG. 8C (specifically, the decrease in X-ray intensity of the X-ray peaks associated with energy levels above about 1 keV with an increase in beamlet separation) is accurately explained in terms of the complex interaction of both protons and dopant ions with the sample, the inventors needed to rule out possible non-sample-related alternative explanations. Such alternative explanations include (a) changing the total beam current with an increase in the strength of the magnetic field used to separate the ion beam into beamlets, and (b) the influence of stray electrons generated from gas scattering within the vacuum chamber and the FIB column.

[0085] To test the effect of magnetic field strength on the total beam current, this current was measured at several values ​​of relative magnetic field strength by integrating the individual current measurements of all component beamlets. From these experiments, it was observed that the total beam current fluctuates with changes in magnetic field strength, but this fluctuation is the opposite of the behavior observed in Figure 8C regarding the change in X-ray intensity due to magnetic field strength.

[0086] To further determine the effect of such changes in beam current on the VLE-PIXE spectrum, several VLE-PIXE spectra of SRM 654b were acquired as a function of beam current by more than one order of magnitude. Based on these experiments, considering the decrease in total X-ray count between the strongest spectrum in Figure 8C and the spectrum at the maximum magnetic field strength, it was determined that a decrease of more than 90% in the total beam current was required to observe the measured change in the VLE-PIXE signal (Figure 8C). Since changes of less than 20% of the total current were observed across the entire range of magnetic field strengths during the acquisition of a series of SRM 654b spectra, changes in current due to the magnetic field cannot be the cause of the observed effect. Furthermore, it was found that changes in the VLE-PIXE signal observed with respect to the total ion current during the acquisition of a series of SRM 654b spectra affected the entire spectrum equally. This result is in contrast to the general observed change in magnetic field strength generated by the magnetic field immersion lens, which only affects the spectral region above approximately 1 keV. This provides additional evidence that changes in beam current cannot be the cause of the decrease in the VLE-PIXE signal due to the changing magnetic field.

[0087] Finally, another hypothesis that the X-ray intensity behavior observed in Figure 8C is related to the acceleration of stray electrons on the sample was ruled out by our general observation that while electrons can cause enhancement of the VLE-PIXE signal, they do so at the expense of an increase in the bremsstrahlung background. Such an increase in the bremsstrahlung background is not observed in the VLE-PIXE spectrum. Therefore, it can be concluded that the enhancement of the VLE-PIXE signal induced by beam doping cannot be based on increased electron generation in the lower column, for example, due to enhanced scattering from gas molecules by dopant heavy ion species.

[0088] Influence of heavy ion species identity on VLE-PIXE performance As established above, the effects of proton and heavy ion dopant species on the same region of the sample are essential for generating an enhanced X-ray signal. To establish the effect of dopant ion species on X-ray generation, beams were prepared using hydrogen doped with over 99% hydrogen, as well as hydrogen doped with N, Ar, and Xe. The effect of increasing the immersion magnetic field generated by the magnetic immersion lens was investigated for each beam composition. To establish a hydrogen beam with over 99% hydrogen, the FIB column was evacuated overnight to remove residual gas species. The gas mixing procedure described above was used to establish low-concentration doped beams. The compositions of these beams before and after measurement are shown in Table 3 below.

[0089] The intensity of the magnetic field increase for all spectra again follows a similar trend to that shown in Figure 8C, increasing first and then decreasing. However, the response of the undoped (over 99% hydrogen) VLE-PIXE spectrum to increasing magnetic field strength shows only a moderate decrease in peak intensity in the portion of the EDS spectrum above 1 keV. This response indicates that some residual contamination still exists despite the best efforts to establish a very clean beam, and demonstrates that even trace amounts of contamination are sufficient to enhance X-ray generation. The undoped and nitrogen spectra are very similar, except for a noticeably enhanced peak at about 0.392 keV, which may be attributable to emissions from a nitrogen emitter. The similar changes with respect to the magnetic field for both of these spectra suggest that nitrogen has a negligible effect on the enhancement of the VLE-PIXE spectrum, and that any enhancement to the separated beam may be attributable to residual Ar or Xe contaminants. The Ar-doped and Xe-doped spectra show significant peak enhancement across the entire spectrum above 1 keV compared to the completely separated beam. Compared to the Xe-doped spectrum, a strong peak at an X-ray energy of approximately 0.22 keV is observed in the Ar-doped spectrum, attributable to emission from the Ar emitter. However, the Xe-doped spectrum shows no additional peaks and is therefore determined to be a preferred species for the analysis of low-energy peaks in regions where the influence of Ar or N may obscure the sample characteristic peaks.

[0090] [Table 3]

[0091] The enhancement factors for each dopant species relative to the undoped hydrogen beam are shown in Figure 9. As shown in this figure, a significant increase in X-ray generation is demonstrated by the introduction of Ar and Xe species, while the effect of N species was negligible. The low spectral enhancement of nitrogen may be due to the simpler nature of the N electronic structure compared to Ar and Xe, resulting in less significant changes in the electronic structure of the target atom due to the formation of quasimolecules. The higher velocity of N species, as well as its smaller atomic radius, may also result in quasimolecules with shorter lifetimes compared to Ar or Xe. We have calculated that the collision time for N is approximately 0.58 times that of Ar and approximately 0.37 times that of Xe, which also reduces the probability of simultaneous interaction with the sample. All of these factors result in the reduced performance of N compared to Ar and Xe, as reflected in Figure 9.

[0092] The performance increase between the Ar dopant species and the Xe dopant species was small. This effect may be due to the fact that the closest approach distance for Xe is much larger than that for Ar, defined as the minimum internuclear distance achieved during collision: 5.76 × 10⁻⁶ for Ar and Xe, respectively. -11 m and 5.18 × 10 -10 This metric, calculated as m, affects the energy level shift associated with quasi-molecular formation, with smaller approach distances resulting in larger shifts in projectile and target energy levels. For N, this is 8.71 × 10⁻⁶. -12 Despite the extremely short closest approach distance for m compared to Ar and Xe, the simple nature of the N-electron structure and short-lived quasi-molecules counteracts the effect of the shorter closest approach distance.

[0093] The effect of the proportion of Dopant species on VLE-PIXE performance Experiments were conducted to determine the effect of the proportion of dopant species on VLE-PIXE performance. These experiments were performed using Ar as the dopant, even though Xe is preferable due to the absence of additional peaks observed in the Ar-doped spectrum. Xenon was not used due to limitations of the Xe source plasma. However, Xe is expected to yield similar results to Ar, as demonstrated by the increased performance compared to a hydrogen beam with over 99% purity, as shown in Figure 9. In the first part of the experiment, VLE-PIXE spectra were acquired using a NIST standard material SRM 654b (Table 4) with an Ar beam with over 99% purity. Contamination from previous source species meant that some amount of hydrogen was still present, which could result in some X-ray generation. With the plasma source gas supply line completely filled with Ar, the hydrogen beam was switched on and a small amount of hydrogen gas was introduced. The beam composition was measured and the VLE-PIXE spectrum was captured. The source plasma was then switched off, and a small portion of the gas supply line was evacuated using the plasma source vacuum system. The plasma source was ignited again, and the initial Ar content was diluted with the remaining hydrogen. This was repeated several times to obtain numerous VLE-PIXE spectra against the partial pressure of argon in hydrogen. The beam composition percentages for each vacuum pumping cycle are shown in Figure 10, and it is shown that the Ar content decreased from 99% to 1.7% during the measurement process.

[0094] It should be noted that the experimental method described in the paragraph above generally results in significant contamination of the gas delivery line to the hydrogen gas source cylinder. Following this measurement, complete fluidization and pumping of the line are recommended. For commercial applications where a desired ratio of hydrogen to dopant species is required, ideally, a pre-mixed gas source is used so that the beam composition remains fixed throughout the experiment and can be reproduced in a more predictable manner. For more flexible beam mixing, mass flow controllers (MFCs) can be used to precisely determine the mixing ratio of each beam species.

[0095] The increase in the total signal intensity is demonstrated as the proportion of Ar increases from 1% to approximately 80% as shown in Figure 11. Subsequently, as the Ar content approaches 99% and the hydrogen content in the beam approaches 0, the VLE-PIXE signal rapidly decreases. This result demonstrates that a certain proportion of hydrogen is required to generate the VLE-PIXE signal. Even the addition of a small proportion of heavy ion species to the hydrogen beam results in a significant enhancement of the VLE-PIXE signal. Conversely, adding a small proportion of hydrogen to the heavy ion species beam can result in a heavy ion beam that can also generate characteristic X-ray signals. This opens up the possibility of technologies such as an endpointing method that allows the X-ray signal to be analyzed during FIB milling by the heavy ion beam and enables precise control over the sample delamination. Also, this enables the combination of both the milling step and the characterization step into one, potentially shortening the experimental time required for this technique and providing an alternative to serial tomography.

[0096] Determination of the spectral intensity contribution of isolated heavy ions Since the X-ray signal intensity is observed to increase with the addition of a small proportion of dopant heavy ions to an ion beam mainly composed of hydrogen, it is important to determine the extent to which the increased signal intensity is due to only the dopant. Therefore, to determine the effect on X-ray generation by the heavy ion species alone, VLE-PIXE spectra were acquired using a beam where the majority of the beam was composed of the dopant species N, Ar, and Xe. In the first experiment, a high purity (over 99%) nitrogen beam was established by vacuum pumping the FIB plasma chamber overnight. The beam composition before the experiment was measured by the method described above. These measurements determined that the majority of the species in the beam could be attributed to N2 + ions and N + ions. There may have been trace amounts of other beam species such as hydrogen species present, but they were below the detection threshold of the measurement.

[0097] Measurements of X-rays emitted from the sample indicate that a considerable amount of X-rays are generated when the sample is irradiated with an ion beam composed almost entirely of nitrogen. The X-ray spectrum of the sample shows titanium

[0098]

number

[0099]

number

[57] . In addition, the appearance of a broad and strong peak at an energy of 0.39 keV is likely attributable to molecular orbital emission resulting from vacancy transfer from N projectile ions to the target atom. This molecular orbital emission was first observed by Saris and Macdonald [58,59].

[0100] Similar experiments were conducted using an Ar beam, where a relatively high-purity beam was also established. The same Ti, V, Al, and Si peaks observed in the high-purity nitrogen beam experiment were also observed in the X-ray spectrum induced by the high-purity argon beam experiment. Furthermore, a strong X-ray peak was observed at an energy of approximately 0.22 keV, which corresponds to Ar L at 0.2217 eV. η Ar at the peak and 0.2201 eV

[0101]

number

[0102] Finally, we performed a similar experiment using a high-purity Xe beam. Note that the RF power of the Xe plasma at 37W is significantly lower than the 200W used for other species. Since the ionization potential of other beam species is much higher, it is highly likely that this Xe beam is composed almost entirely of Xe, and this hypothesis supports the idea that Xe + This is corroborated by analyzing the beam composition prior to the analysis in which only the species were detected. Weak X-ray peaks attributable to Ti, Al, Si, and V are observed. The X-ray spectrum also indicates that Xe functions well as a dopant for the hydrogen beam, and is considered an optimal dopant species compared to N or Ar, as a result of a significant reduction in spectral background. However, it should be noted that the use of Xe may result in greater sample damage due to the high sputter yield. In addition, limitations on the RF power that can be applied to the Xe plasma restrict the use of Xe as a dopant at high partial pressures of either hydrogen or Xe, where higher RF power is required for the effective ionization of hydrogen molecules and proton formation.

[0103] To determine whether the signal intensity of the spectrum observed under irradiation with a doped hydrogen beam is actually greater than the sum of the spectra produced by separate irradiation with each component, the strongest spectra from high-purity hydrogen beam experiments and high-purity Xe beams were added proportionally to generate an artificially produced Xe-doped hydrogen spectrum. The results show that when individual pure H and pure Xe spectra are added together and adjusted to the respective proportions in the beam, the resulting intensity is significantly lower than the VLE-PIXE spectrum achieved when the hydrogen beam is doped with a small proportion of Xe, thereby confirming the hypothesis that separate proton and heavy ion beams must overlap to obtain a high-quality VLE-PIXE spectrum, and that any detailed theory of the X-ray emission mechanism must explain the interaction between both species and the sample.

[0104] Optimized doped beam VLE-PIXE parameters The following are typical device operating parameters determined using our apparatus, and are not intended to be limiting.

[0105] Beam configuration ● The immersion magnetic field of the SEM column, if present, can be adjusted to achieve the highest possible X-ray signal, which can be offset by approximately 20 percent relative to the lens setting that produces the focused beam.

[0106] Dopant species ●Xe is ideally used as a dopant species because it does not produce the additional molecular orbital peaks that would result from this species. ●Ar can be used in situations where high RF power is required, as high RF power Xe plasma can cause permanent damage to the plasma source.

[0107] Heavy seeds doped into a hydrogen beam ● Evacuum the plasma chamber and FIB column overnight. ● Switch the ion source input to the desired dopant species for a period of 30 minutes or more. ● Modify the vacuum pump feeding cycle to account for a 10-second PVP delay and a 10-second TMP delay. ● Switch the ion source input to hydrogen. This will yield approximately 10% dopant species and the remainder hydrogen, which is ideal for trace element analysis.

[0108] Hydrogen-doped heavy seed beam ●Ideally, a premixed gas source is used so that a beam composition of 80% heavy ions and the remaining hydrogen species is achieved. Alternatively, the plasma source gas delivery line can be filled with low-pressure hydrogen, and the heavy ion plasma is turned on to introduce the remaining heavy ions. ● For more flexible gas mixing, a micro-flow controller can be used to control the proportion of gas precursor species.

[0109] Applications of VLE-PIXE EDS mapping A key capability of micro and nanoscale analytical techniques such as SEM-EDS and micro-PIXE is their ability to spatially resolve elemental components within a sample. This offers a significant advantage over bulk analytical techniques, as the distribution of elements within a sample can be established with a resolution limited by the instrument used.

[0110] Examples of VLE-PIXE elemental maps collected with a Xe-doped hydrogen beam of approximately 50 nA (approximately 6% Xe and the remainder hydrogen) are shown in Figures 12B–12C. Figure 12A is a redrawn version of a MoS2 thin section exfoliated on a Si(100) wafer, observed using FIB secondary electron (SE) contrast imaging. Using only the contrast image, it was not possible to initially identify several smaller thin sections above the main thin section. However, the elemental maps of Mo and S shown in Figures 12B–12C identify the presence and identity of multiple thin sections and correlate well with the SE image. These maps were captured in an automated manner using the provided Oxford Instruments AZtec® software package and corresponding external scanning generator (these were installed as standard on the FIB microscope). Thus, such experiments can be performed essentially without any modifications to the instrument.

[0111] Several challenges related to VLE-PIXE mapping capability need to be overcome. In particular, the high current required for VLE-PIXE measurements results in significantly reduced resolution compared to optimal imaging conditions for FIB. However, generational advancements in PFIB instruments improve high-current resolution. Also, due to the relatively low signal generated even by the doped hydrogen beam, the map acquisition time is quite significant, reaching about one hour for the maps shown in Figures 12B-12C. Improving detector sensitivity helps to reduce map acquisition time. Furthermore, increasing the proportion of dopant species can increase the generated X-ray signal, as mentioned above. However, improvement in signal intensity is achieved at the expense of increased sample damage due to sputtering.

[0112] The presence of dopant species in the ion beam used for VLE-PIXE mapping inevitably leads to sputtering of the underlying sample, meaning the sample is continuously modified during map acquisition. This phenomenon must be considered during doped-beam VLE-PIXE mapping, but it can still be important in certain situations, such as when each map may arise from subsequent layers of the sample. These maps can then be stacked to establish a three-dimensional elemental map, a technique known as tomography. Current tomography operations are typically performed sequentially, by first removing layers of material and then performing sample analysis. However, the use of doped-beam VLE-PIXE mapping can replace the sequential tomography procedure with a single-step method of simultaneous layer removal and analysis.

[0113] VLE-PIXE mapping may also be used in conjunction with an end-pointing procedure, thereby analyzing the X-ray signal during milling. This procedure allows milling to be precisely stopped at the interface between two layers due to the appearance of a new characteristic X-ray peak corresponding to the start of the next layer, and can be performed in real time, in contrast to other methods that may require continuous delayering and analysis.

[0114] Trace element analysis Energy-dispersive spectroscopy performed using a scanning electron microscope (SEM-EDS) is a technique similar to PIXE, both of which rely on particle collisions for core ionization, resulting in the emission of X-rays characteristic of the element being analyzed. While PIXE uses ions, typically protons, for ionization, SEM-EDS utilizes electrons for the same purpose.

[60]

[0115] The X-ray generation cross-section for electrons is an order of magnitude larger than that for protons in the energy range used by VLE-PIXE (i.e., below 50 keV). Therefore, X-ray generation by electrons is expected to be far more effective. [61,62] However, SEM-EDS also has a significant drawback: the appearance of strong bremsstrahlung radiation, which can obscure low-intensity X-ray peaks. The bremsstrahlung background effectively reduces the sensitivity of SEM-EDS to an LOD of about 100–500 ppm (parts per million) compared to the Lower Limit of Detection (LOD) of PIXE in the MeV range of around 1 ppm or below 1 ppm.

[14]

[13] The LOD for VLE-PIXE is expected to be several orders of magnitude higher than that for PIXE due to the decrease in XRPCS for protons in the VLE-PIXE energy range.

[63]

[0116] Direct comparison between PIXE and SEM-EDS is often difficult due to significant differences in the equipment between the two techniques. However, the implementation of VLE-PIXE on a dual-beam PFIB / SEM microscope allows both techniques to be performed sequentially using the same detector, X-ray window, software, and vacuum conditions, thereby enabling a more direct comparison. To compare the two techniques, we performed analyses of two standard reference materials (SRMs) provided by the National Institute of Standards and Technology (NIST). Both of these materials have well-tuned and certified compositions containing a range of major, minor, and trace elements. Below, the VLE-PIXE results are directly compared with SEM-EDS, a similar technique commonly used for elemental analysis in SEM, and the relative sensitivity of the two techniques is evaluated as described below.

[0117] The first sample measured was NIST SRM 654b, which is a Ti-based alloy containing a range of bulk and trace elements. The certified composition is shown in Table 4, with the remainder of the composition being Ti (not specified). The concentrations in Table 4 are indicated by asterisks ( * Unless otherwise indicated by an asterisk (), the values ​​are expressed as a percentage in mass fractions. * If marked with ), the concentration is given in mg / kg. The values ​​shown in parentheses are for informational purposes only.

[0118] The beam composition measurements prior to analysis are shown graphically in Figure 13 as a function of beamlet separation and expressed as a percentage of the total beam current in the table embedded within Figure 13. These measurements established the presence of a hydrogen beam doped with approximately 6.9% Xe, as described in this specification.

[0119] [Table 4]

[0120] As shown, a higher proportion of dopant species enhances X-ray signal generation, but this comes at the cost of increased sample damage due to sputtering by heavier ion species. Therefore, to maintain sample integrity, a small proportion of dopant species was selected for these experiments. To account for the relatively low X-ray generation with smaller proportions of dopant species, the capture times for SEM-EDS and doped beam VLE-PIXE were chosen to be 20 minutes and 4 hours, respectively, so that both spectra were acquired with a similar number of total X-ray counts (approximately 10M counts).

[0121] The Xe-doped VLE-PIXE spectrum at an accelerating voltage of 24 keV is shown in Figure 14A, compared to the SEM-EDS spectrum at an accelerating voltage of 5 keV. A 24 keV ion beam was used to minimize transmission of backscattered ions through the X-ray detector window, and a 5 keV electron beam was selected to enhance sensitivity to peaks in the X-ray energy range below 2.5 keV, an effect known as the overpotential ratio.

[64] A comparable comparison in the X-ray energy range of 4–10 keV is shown in Figure 14B, where a 30 keV electron beam was used to enhance sensitivity to X-ray peaks in this range. The SEM-EDS spectrum continues up to a value of 30 keV, but only the bremsstrahlung background is observed in the SEM-EDS spectrum at energies greater than 10 keV. Therefore, the spectrum has been truncated for clarity.

[0122] Qualitatively, the VLE-PIXE spectrum can be characterized by the near-complete absence of bremsstrahlung background, in contrast to the broad and strong bremsstrahlung background typical of the SEM-EDS spectrum. An example of this bremsstrahlung background is labeled in Figure 14A. As a result, several peaks that cannot be identified in the SEM-EDS spectrum are present in the VLE-PIXE spectrum, leading to the identification of two additional subcomponents, Ni and Cu. Ni is represented by the Ni Lα peak, as well as the Ni K α Peak and K β Based on the presence of peaks, it can be identified in the VLE-PIXE spectrum. Cu is Cu L α It can be identified in the VLE-PIXE spectrum based on the peaks.

[0123] Additional peak, Fe L α and Fe K β The identification of this peak can be used to confirm the presence of Fe in the sample, which is an important feature in the case of overlapping peaks. Peaks corresponding to elements not outlined in the analysis, for example, PK at 2.013 keV. αA peak was also observed. However, the lack of identification makes the identification of this element speculative. At an X-ray energy of 3 keV, there are several spectral features that do not match any known X-ray characteristic peaks and are therefore considered to be detector artifacts. For this reason, this portion of the spectrum has been removed for clarity.

[0124] To evaluate the effectiveness of the doped beam VLE-PIXE technique, a second sample of NIST reference material SRM 1242 was measured using the same procedure as described above. This sample was selected so that titanium, the main component of SRM 654b, was absent in SRM 1242, and cobalt and tungsten, the main components of SRM 1242, were absent in SRM 654b. This selection of the additional sample also helped to eliminate the possibility that the X-ray spectrum was due to some artifact of the system, as both samples were likely to produce identical X-ray spectra under such circumstances. The complete composition of SRM 1242 is shown in Table 5. The concentrations in Table 5 are expressed as percentages as mass fractions. Values ​​shown in parentheses are for informational purposes only.

[0125] The beam composition measurement before analysis is shown as a function of beamlet separation in Figure 15, and is expressed as a percentage of the total beam current in the table embedded in that figure. Similar to the NIST 654b sample, a hydrogen beam was established with a dopant ratio of 6.2%Xe. A comparison of the SEM-EDS and VLE-PIXE spectra for the second NIST sample is shown in Figures 16A and 16B. Here again, a 5 keV electron beam was used for peak comparison in the low-energy region, and a 30 keV electron beam was used in the high-energy region.

[0126] [Table 5]

[0127] Comparing the VLE-PIXE spectra of SRM 1242 and SRM 654b reveals several peaks specific to each sample, identified as both major and minor components of each sample. In the spectrum of the SRM 1242 sample, tungsten M-series peaks and Co K and L-series peaks are clearly identifiable, none of which were present in the SRM 654b spectrum. (Cr K) α The peak intensity also increased significantly compared to that in the first NIST sample, corresponding to the higher concentration of Cr in this second sample. The heavy element tungsten is particularly interesting due to the presence of outer shell M and N transitions, demonstrating the full range of K, L, M, and N X-ray transitions present in a single sample.

[0128] The emergence of new peaks corresponding to W and Co demonstrates the effectiveness of this technique and eliminates the possibility of X-ray signals resulting from instrument artifacts. The same unidentified spectral features at approximately 3 keV are still present, however, they have been cropped again in Figures 16A-16B to better identify spectral lines in the lower energy region.

[0129] In the characterization of SRM 1242, several peaks that should not be present were observed, specifically Ti and Al K. α A peak appeared. As previously stated herein, the Ti peak can arise from particles colliding with the X-ray detector window support grid or with the resputtered material coating the X-ray detector window. The fact that this Ti peak is present in both the VLE-PIXE and SEM-EDS spectra suggests that this is indeed the case and that this peak should be ignored. The presence of these obvious artifact peaks suggests that as the electron energy increases, some backscattered electrons bypass the electron trap and Ti K α This suggests that it may start generating a signal.

[0130] The signal-to-noise ratio (SNR) and limit of detection (LOD) were calculated for the elements identified in the spectra shown in Figures 14A–14B and 16A–16B, using a standard threshold of 3σ

[65] . The full set of values ​​is listed in Table 6 for the SRM 654b sample and in Table 7 for the SRM 1242 sample, with the better of the two techniques underlined. These results are also summarized in Figure 17, where the minimum LODs for each element measured using doped beam VLE-PIXE and SEM-EDS are compared. Table 5. NIST SRM 1242 certified concentration values.

[0131] [Table 6]

[0132] For most elements, the LOD is lower for doped beam VLE-PIXE compared to SEM-EDS, demonstrating higher sensitivity for doped beam VLE-PIXE. This result is mainly due to the absence of bremsstrahlung background generated during ion collisions compared to electron collisions. The reduced background allows for the identification of elemental peaks at lower concentrations than would be possible in the presence of background. Further subcomponents and trace components were identified. However, due to peak overlap, it was not possible to determine the SNR and LOD of these subcomponents and trace components. One example is phosphorus K. α Peak and Tungsten M γ This is an overlap with the peak.

[0133] The maximum improvement in LOD for doped beam VLE-PIXE compared to SEM-EDS was 14.66 times for Si, and the average improvement was 4.30 times for elements detected by both techniques (n=10). The minimum LOD for Xe-doped VLE-PIXE was 2.78 ppm for Cu on the SRM 1242 sample, compared to the minimum LOD for EDS, which was 132.86 ppm for Si on the SRM 654b sample. These LOD values ​​are:

[0134] [Table 7] The EDS values ​​were 100–500 ppm

[14] and the PIXE values ​​were around 1 ppm or less, which are consistent with the literature cited for PIXE performed at MeV energy.

[13] These results demonstrate that the addition of a small amount (about 6%) of Xe to the hydrogen beam can increase the sensitivity of VLE-PIXE to a level close to that of PIXE performed at MeV energy. El Ghawi et al. noted that the LOD of a 250 keV proton should be several orders of magnitude higher than that of a 2 MeV proton

[63] , and that the LOD for 30 keV should also be expected to be high.

[0135] Although the inventors have shown that VLE-PIXE is more sensitive than SEM-EDS, doped beam VLE-PIXE still has several drawbacks that must be considered when choosing between the two techniques. ●The large XRPCS of electrons compared to protons at this energy significantly reduces the experimental time required for SEM-EDS compared to doped beam VLE-PIXE. ● The spatial resolution of SEM-EDS is considerably better than that of doped beam VLE-PIXE due to its much smaller electron probe size. ● The ions used in doped beam VLE-PIXE cause sample damage due to proton and heavy ion implantation, as well as heavy ion sputtering. Therefore, doped beam VLE-PIXE and SEM-EDS can be considered complementary techniques. SEM-EDS is useful in situations where rapid identification of bulk and minor components or high spatial resolution is required, while VLE-PIXE finds applications where the identification of trace elements is required at the expense of experimental time.

[0136] LE and PIXE data were not available for the specific SRM samples measured in these experiments. Table 8 lists a set of conventional PIXE LOD values ​​from several sources for comparison with doped beam VLE-PIXE. This list is not exhaustive and is simply intended to provide a comparison of VLE-PIXE techniques. The listed LOD values ​​are optimal LOD values ​​for given experimental conditions.

[0137] Typical LODs of 1–10 ppm are consistently given for PIXEs performed at energies above 1 MeV. However, information on LODs achieved by LE-PIXEs is very limited. Nouli et al. demonstrated an LOD of 850 ppm for P for LE-PIXEs at 600 keV.

[71] Zahraman et al. demonstrated that the measured SNR increased as the energy decreased to 750 keV, resulting in a decrease in LOD.

[72] They attributed this to a decrease in background signal with decreasing energy, which helps to better identify the underlying PIXE signal. Overall, the table of results in this literature demonstrates that the best LOD of a doped beam VLE-PIXE calculated above at 2.78 ppm exhibits similar performance to PIXEs performed at much higher energies.

[0138] Our experimental results demonstrate that the sensitivity of VLE-PIXE performed using a low-doped Xe beam is comparable to that of PIXE performed at much higher energies. Compared to SEM-EDS, the sensitivity of doped beam VLE-PIXE was found to be superior in almost all cases. However, this sensitivity was achieved at the cost of significantly increased experimental time due to very low X-ray generation. Therefore, doped beam VLE-PIXE can be considered a complementary technique to SEM-EDS, which can be used for the identification of elements obscured by the bremsstrahlung background typical of SEM-EDS spectra.

[0139] [Table 8]

[0140] Endpointing and real-time tomography Potential commercial applications of the doped beam VLE-PIXE technique may be real-time elemental mapping processes or end-pointing methods. As shown above, small amounts of hydrogen can be added to a heavy ion species beam, such as Ar or Xe, to generate a strong VLE-PIXE signal. If this VLE-PIXE signal is monitored during the milling process on a layered sample, analysis of changes in characteristic peaks during milling can determine where one layer of the sample ends and the next layer begins. Such “end-pointing” techniques may have important applications in the semiconductor industry.

[73] In addition, the potential chemical effects induced by a beam of a mixture of hydrogen and heavy ions can provide some degree of control over another important process in sample delayering, semiconductor fabrication, and failure analysis.

[74] This method can be extended by providing ion beams containing a mixture of three or more ion species, such as a Xe beam for inducing fast sputtering, a nitrogen beam for inducing chemical effects, and a hydrogen beam for generating analytical X-ray signals.

[0141] Real-time elemental mapping is a more advanced version of end-pointing techniques in which an X-ray elemental map is collected during sample milling. X-ray elemental mapping allows for the localization of elemental composition across the entire sample. When the beam is scanned across the sample during X-ray map acquisition, H + H2 + and H3 + The proportion of heavy ion species with a higher sputtering yield removes some material so that each subsequent X-ray map originates from a different layer of the sample. Each map forms a "slice" of the material that, when combined, establishes a three-dimensional map of the material's elemental composition.

[0142] Typically, such three-dimensional mapping is performed sequentially, with layers of material being removed from the sample and the exposed layers being analyzed by electron backscatter diffraction or alternative methods such as EDS. This technique is known as serial sectioning tomography and is known to be slow because it requires repeated switching between FIB milling and sample analysis methods. [7] By using the VLE-PIXE method, which employs a beam mixed with hydrogen and heavy ion species, the milling and analysis steps can be combined into a single operation, and the mapping results can be captured in real time. This has the potential to significantly increase the throughput of tomography operations.

[0143] conclusion Our experimental investigations have demonstrated the implementation and application of the doped beam VLE-PIXE technique using a commercially available focused ion beam microscope. This technique offers trace element analysis with performance comparable to PIXE performed at much higher energies, as well as the possibility of simultaneous FIB milling and sample analysis. The novel method taught herein favorably extends the availability of the PIXE technique from conventional, relatively inaccessible particle accelerator facilities to FIB microscopes that can be placed in standard laboratory facilities.

[0144] The novel VLE-PIXE technique described herein is made possible by using a multi-gas inlet inductively coupled plasma ion source to implement proton beams and proton / heavy ion mixed beams on a FIB microscope. Beam composition characterization demonstrates the generation of high proton currents by the plasma FIB system, a result that is important for the performance of the PIXE technique and is unique to plasma FIBs compared to previous hydrogen FIB manifestations such as GFIS.

[0145] The roles of various spurious signal sources, such as stray electrons, backscattered ions, and neutralized particles, were investigated, and methods were developed to eliminate such signals, including manipulating the SEM immersion magnetic field clamp and FIB column lens bias. Furthermore, reducing the FIB acceleration voltage to 24 keV can limit the influence of backscattered ions on the X-ray spectrum. Ensuring sufficient vacuum is established within the FIB column and microscope chamber before the experiment was found to favorably reduce secondary particle collisions resulting from ion-gas collisions within the FIB column and chamber.

[0146] Trace elemental analysis of standard materials using a hydrogen beam doped with low concentrations of Xe allowed us to validate the performance of this technique using metrics such as signal-to-noise ratio and detection limit. Direct comparisons between similar techniques, SEM-EDS, established that the sensitivity of doped VLE-PIXE is superior to EDS for almost all elements analyzed due to the absence of the broad bremsstrahlung background typical of electron X-ray spectra. The absence of the bremsstrahlung background also demonstrated a significant increase in sensitivity, enabling the identification of several peaks using VLE-PIXE that could not be identified using SEM-EDS. These observations open up the possibility of producing enhanced elemental composition analysis of samples using dual-beam instruments by employing SEM-EDS for quantifying some elements and VLE-PIXE for others, where the quantification technique for each element uses a method that provides superior results for that element. The compositional mapping capability of the VLE-PIXE technique was investigated and determined to be useful in several situations. The performance of the doped beam VLE-PIXE technique has been established to be comparable to PIXE performed at much higher energies, as evidenced by the significant increase in the X-ray generation cross-section compared to proton-only beams.

[0147] Investigations into beam doping methods demonstrated that the signal produced when both protons and heavy ions are incident at the same location on the sample is greater than the sum of the individual components. Xe was determined to be the optimal dopant species due to its relatively high performance and the absence of additional X-ray peaks generated by the Xe emitter. The proportion of dopant species was shown to significantly increase the emission of X-rays from the sample, reaching a maximum with approximately 80% Ar species, with the remainder being hydrogen. This creates the possibility of both heavy-seed doped hydrogen beams and hydrogen-doped heavy-seed beams, each with its own potential applications. Theoretical models of the mechanism behind the signal-enhancing effect of heavy ions will be investigated in future studies.

[0148] The description contained herein is intended to serve as a basic explanation. The present invention is not intended to be limited in scope by the specific embodiments described herein, which are intended as single examples of individual aspects of the invention. Functionally equivalent methods and components are within the scope of the invention. In addition to those shown and described herein, various other variations of the invention will be apparent to those skilled in the art from the foregoing description and the accompanying drawings. The specific numerical values ​​of the apparatus operating parameters presented herein are provided as typical operating parameters to be determined for a particular apparatus and are not intended to be limiting in any way.

[0149] List of References [1] Young, Richard J., and Mary V. Moore. Dual-beam (FIB-SEM) systems. Introduction to focused ion beams. Springer, Boston, MA, 2005.247-268. [2]McLean P Echlin, Marcus Straw, Steven Randolph, Jorge Filevich, and Tresa M Pollock.The TriBeam system:Femtosecond laser ablation in situ SEM.Materials Characterization,100:1-12,2015. [3]Yongqi Fu and Kok Ann Bryan Ngoi.Focused ion beam direct fabrication of micro-optical elements:features compared with laser beam and electron beam direct writing,2004. [4]Lucille A Giannuzzi et al.Introduction to focused ion beams:instrumentation, theory, techniques and practice.Springer Science&Business Media,2004. [5] Lucille A Giannuzzi and Frederick A Stevie. A review of focused ion beam milling techniques for TEM specimen preparation. Micron, 30(3): 197 - 204, 1999. [6] Richard Young, C Rue, S Randolph, C Chandler, G Franz, R Schampers, A Klumpp, and L Kwakman. A comparison of xenon plasma FIB technology with conventional gallium LMIS FIB: imaging, milling, and gas - assisted applications. Microscopy and Microanalysis, 17(S2): 652 - 653, 2011. [7] TL Burnett, R Kelley, B Winiarski, L Contreras, M Daly, A Gholinia, MG Burke, and PJ Withers. Large volume serial section tomography by Xe Plasma FIB dual beam microscopy. Ultramicroscopy, 161: 119 - 129, 2016. [8] Naohiro Shimizu and SR Hart. Applications of the ion microprobe to geochemistry and cosmochemistry. Annual Review of Earth and Planetary Sciences, 10(1): 483 - 526, 1982. [9] Sarah Witman. Ten things you might not know about pArticle accelerators. SymmetrymAgazine (Department of Energy), 2014.

[10] Waldemar Scharf. Particle accelerators and their uses. CRC Press, 1986.

[11] R Curtis Bird and James Stanislaus Williams. Ion beams form Aterials analysis. Elsevier, 1990.

[12] Sven AE Johansson and Thomas B Johansson. Analytical application of pArticle induced X-ray emission. Nuclear Instruments and Methods, 137(3):473-516, 1976.

[13] CG Ryan. Quantitative trace element imaging using PIXE and the nuclear microprobe. International Journal of Imaging Systems and Technology, 11(4):219-230, 2000.

[14] Dale Newbury, Nicholas Ritchie, Michael Mengason, and Keana Scott. SEM / EDS Trace Analysis: Limits Imposed by Fluorescence of the Detector. Microscopy and Microanalysis, 23(S1):1026-1027, 2017.

[15] Rainer Siegele, David D Cohen, and Nick Dytlewski. The ANSTO high energy heavy ion microprobe. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with mAterials and Atoms, 158(14):31-38, 1999.

[16] A Denker, J Opitz Coutureau, M Griesser, R Denk, and H Winter. Nondestructive analysis of coins using high-energy PIXE. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 226(1-2):163-171, 2004.

[17] K Ishii and S Morita. Theoretical estimation of PIXE detection limits. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 34(2):209-216, 1988.

[18] Keizo Ishii and Susumu Morita. Continuous backgrounds in PIXE. International Journal of PIXE, 1(01):1-29, 1990.

[19] Javier Miranda. Low energy PIXE: advantages, drawbacks, and applications. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 118(1-4):346-351, 1996.

[20] J Miranda. Low energy PIXE revisited: is it still alive? Revista Mexicana de Fisica, 56(1):82-84, 2010.

[21] Yoshiko Moriya, Yasuro Ato, and Sohji Miyagawa. Sensitivity in light element analysis by 2 MeV and 150 keV proton and photon induced X-rays. Nuclear Instruments and Methods, 150(3):523-528, 1978.

[22] SA Correa, E Pitthan, MV Moro, and D Primetzhofer. A multipurpose set-up using keV ions for nuclear reaction analysis, high-resolution backscattering spectrometry, low-energy PIXE and in-situ irradiation experiments. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 478:104-110, 2020.

[23] Wataru Kada, Atsuya Kishi, Masato Sueyasu, Fuminobu Sato, Yushi Kato, and Toshiyuki Iida. Development of a low-energy PIXE analysis system based on an ion implanter. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 318:51-54, 2014.

[24] Milo Boirot. Implementation of low-energy PIXE at the new scattering chamber of the 350 kV implanter, 2019.

[25] M Soueidan, M Roumie, and B Nsouli. PIXE detection limit for aluminium thin film deposited on Si-based matrix. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 406:71-74, 2017.

[26] SJC do Carmo, FIGM Borges, AMF Trindade, and CAN Conde. K X-ray production cross sections in aluminium for 15, 20 and 25 keV protons. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 293:16-20, 2012.

[27] Gregory Lapicki. The status of theoretical K-shell ionization cross sections by protons. X-Ray Spectrometry: An International Journal, 34(4):269-278, 2005.

[28] Lucille A Giannuzzi. Particle-induced x-ray analysis using focused ion beams. Scanning, 27(4):165-169, 2005.

[29] David C Joy, Harry M Meyer, Mehdi Bolorizadeh, Yinghong Lin, and Dale Newbury. On the Production of X-rays by Low Energy Ion Beams. Scanning: The Journal of Scanning Microscopies, 29(1):1-4, 2007.

[30] A Pillay and M Peisach. Production of charge-induced X-rays during PIXE studies using light and heavy ion-beams. Journal of radioanalytical and nuclear chemistry, 200(1):53-74, 1995.

[31] Jurgen H Gross. Mass spectrometry: a textbook. Springer Science&Business Media, 2006.

[32] PD Prewett and DK Jefferies. Characteristics of a gallium liquid metal field emission ion source. Journal of Physics D: Applied Physics, 13(9):1747, 1980.

[33] RG Forbest. Understanding how the liquid-metal ion source works. Vacuum, 48(1):85-97, 1997.

[34] John Morgan, John Notte, Raymond Hill, and Bill Ward. An introduction to the helium ion microscope. Microscopy today, 14(4):24-31, 2006.

[35] Gregor Hlawacek, Vasilisa Veligura, Raoul van Gastel, and Bene Poelsema. Helium ion microscopy. Journal of Vacuum Science&Technology B, Nanotechnology and Microelectronics: Materials, Processing, Measurement, and Phenomena, 32(2):020801, 2014.

[36] Hironori Moritani, Radovan Urban, Mark Salomons, Robert Wolkow, and Jason Pitters. Hydrogen Ion Beams fromnAnostructured Gas Field Ion Sources. Microscopy and Microanalysis, 22(S3):612 - 613, 2016.

[37] Hironori Moritani, Radovan Urban, Kyle Nova, Mark Salomons, Robert Wolkow, and Jason Pitters. Selective production of hydrogen ion species at atomically designednAnotips. Ultramicroscopy, 186:42 - 48, 2018.

[38] ShinichimAtsubara, Hiroyasu Shichi, Yoshimi Kawanami, and Tomihiro Hashizume. Novel Scanning Ion Microscope with H3+ Gas Field Ionization Source. Microscopy and Microanalysis, 22(S3):614 - 615, 2016.

[39] John A Notte, Lou Farkas, Raymond Hill, and Bill Ward. An Introduction to Helium Ion Microscopy and itsnAnotechnology Applications. NanoScience and Technology Institute, 2006.

[40] R Hill and FHM Faridur Rahman. Advances in helium ion microscopy. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 645(1): 96 - 101, 2011.

[41] Jacques Gierak, Paulm Azarov, Lars Bruchhaus, Ralf Jede, and Lothar Bischoff. Review of electrohydrodynamical ion sources and their applications to focused ion beam technology. Journal of Vacuum Science & Technology B, Nanotechnology and Microelectronics: Materials, Processing, Measurement, and Phenomena, 36(6): 06J101, 2018.

[42] Hiroyasu Shichi, Shinichim Atsubara, and Tomihiro Hashizume. Comparison of Characteristics of Neon, Argon, and Krypton Ion Emissions from Gas Field Ionization Source with Single Atom Tip. Microscopy and Microanalysis, 23(S1): 274 - 275, 2017.

[43] Anthony Graupera and Charles Otis. Inductively - coupled plasma ion source for use with a focused ion beam column with selectable ions, September 2 2014. US Patent 8,822,913.

[44] Jacques Gierak. Focused ion beam technology and ultimate applications. Semiconductor science and technology, 24(4):043001, 2009.

[45] BW Ward, John A Notte, and NP Economou. Helium ion microscope: A new tool for Nanoscale microscopy and metrology. Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures Processing, Measurement, and Phenomena, 24(6):2871 - 2874, 2006.

[46] NS Smith, WP Skoczylas, SM Kellogg, DE Kinion, PP Tesch, O Sutherland, Ane Aanesland, and RW Boswell. High brightness inductively coupled plasma source for high current focused ion beam applications. Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures Processing, Measurement, and Phenomena, 24(6):2902 - 2906, 2006.

[47] Gorelick Sergey, Korneev Denis, Handley Ava, Gervinskas Gediminas, Oorschot Viola, Pocock Roger, et al. Oxygen plasma focused ion beam scanning electron microscopy for biological samples. BioRxiv, page 457820, 2018.

[48] KH Bai, SJ You, Hong-Young Chang, and HS Uhm. Plasma Parameters Analysis of Various Mixed Gas Inductively Coupled Plasmas. Physics of Plasmas, 9(6): 2831 - 2838, 2002.

[49] Osamu Fukumasa, R Italani, and S Saeki. Numerical Simulation of Hydrogen Ion Species in the Steady-State Plasma of a Low-Pressure Ion Source. Journal of Physics D: Applied Physics, 18(12): 2433, 1985.

[50] Katsushige Tsuno, Nobuo Handa, and Sunaom Atsumoto. Immersion Lenses for Low-Voltage SEM and LEEM. In Electron-Beam Sources and Charged-Particle Optics, volume 2522, pages 243 - 252. International Society for Optics and Photonics, 1995.

[51] BL Thiel, M Toth, RPM Schroemges, JJ Scholtz, G Van Veen, and WR Knowles. Two-stage Gas Amplifier for Ultrahigh Resolution Low Vacuum Scanning Electron Microscopy. Review of Scientific Instruments, 77(3): 033705, 2006.

[52] Thierry Czerwiec, F Greer, and DB Graves. Nitrogen dissociation in a low pressure cylindrical ICP discharge studied by actinometry and mass spectrometry. Journal of Physics D: Applied Physics, 38(24):4278, 2005.

[53] Mark W Kiehlbauch and David B Graves. Inductively coupled plasmas in oxygen: Modeling and experiment. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films, 21(3):660 - 670, 2003.

[54] Denis Pick, Matthias Leiterer, and Jürgen W Einax. Reduction of polyatomic interferences in biological material using dynamic reaction cell ICP - MS. Microchemical Journal, 95(2):315 - 319, 2010.

[55] HM Sullivan. Vacuum pumping equipment and systems. Review of Scientific Instruments, 19(1):1 - 15, 1948.

[56] Tom Wirtz, Olivier De Castro, Jean - Nicolas Audinot, and Patrick Philipp. Imaging and analytics on the helium ion microscope. Annual Review of Analytical Chemistry, 12:523 - 543, 2019.

[57] Werner Brandt and Roman Laubert. Pauli excitation of atoms in collision. Physical Review Letters, 24(19):1037, 1970.

[58] FW Saris, WF Van der Weg, H Tawara, and Roman Laubert. Radiative transitions between quasimolecular levels during energetic atom-atom collisions. Physical Review Letters, 28(12):717, 1972.

[59] James Rm Acdonald and Matt D Brown. Noncharacteristic X-Ray Bands Produced in Targets of C, Al, Si, and Solid Ar by Argon Ions at keV Energies. Physical Review Letters, 29(1):4, 1972.

[60] Michael H Loretto. Electron beam analysis of mAterials. Springer Science&Business Media, 2012.

[61] Xavier Llovet, Cedric J Powell, Francesc Salvat, and Aleksander Jablonski. Cross sections for inner-shell ionization by electron impact. Journal of Physical and Chemical Reference Data, 43(1):013102, 2014.

[62] David D Cohen. K and L shell X-ray cross sections for use in PIXE analysis systems. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 49(1-4): 1-9, 1990.

[63] Usama M El Ghawi, Beni M Bahal, and Salem K Al Arbi. Possibility of using a low-energy proton beam for Particle-induced X-ray emission microanalysis. Radiation research, 131(3): 243-248, 1992.

[64] Jean-Louis Pouchou. Standardless x-ray analysis of bulk specimens. Microchimica Acta, 114(1): 33-52, 1994.

[65] David A Armbruster and Terry Pry. Limit of blank, limit of detection and limit of quantitation. The clinical biochemist reviews, 29(Suppl 1): S49, 2008.

[66] WJ Teesdale, J Am Maxwell, A Perujo, JL Campbell, L Van der Zwan, and TE Jackman. Limits of detection and quantitation in PIXE analysis of thick targets. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 35(1): 57-66, 1988.

[67] M Khaliquzzaman, ST Lam, DM Sheppard, and LG Stephens Newsham. Dependence of X-ray yields on different parameters for light element matrices in thick target PIXE and use of standards for calibration in such analysis. Nuclear Instruments and Methods in Physics Research, 216(3):481-488, 1983.

[68] JL Campbell, J AmAxwell, WJ Teesdale, J-X Wang, and LJ Cabri. MicroPIXE as a complement to electron probe microanalysis in mineralogy. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with materials and Atoms, 44(3):347-356, 1990.

[69] CG Ryan, DR Cousens, SH Sie, WL Griffin, GF Suter, and E Clayton. Quantitative pixe microanalysis of geological material using the CSIRO proton microprobe. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with materials and Atoms, 47(1):55-71, 1990.

[70] Alessandro Borghi, Roberto Cossio, and CmAzzoli. A mineralogical application of micro-PIXE technique: Yttrium zoning in garnet from metamorphic rocks and its petrologic meaning. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions withmAterials and Atoms, 189(1-4):412-417, 2002.

[71] B Nsouli, M Roumie, K Zahraman, JP Thomas, and MnAsreddine. On the phosphorus characterization in thin SiO2(P,B)CVD layer deposited onto a silicon substrate by PIXE. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions withmAterials and Atoms, 192(3):311-317, 2002.

[72] K Zahraman, B Nsouli, M Roumi’e, JP Thomas, and S Danel. On the optimization of the PIXE technique for thickness uniformity control of ultra-thin chromium layers deposited onto large surface quartz substrate. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions withmAterials and Atoms, 249(1-2):447-450, 2006.

[73] D Zudhistira, V Viswanathan, Vn Arang, JM Chin, S Sharang, K Novotny, and J Vincen Obona. Precision Xe plasma FIB delayering for physical failure analysis of sub-20nm microprocessor devices. In ISTFA 2017, pages 574-579. ASM International, 2017.

[74] Roger Alvis, Trevan Landin, Chad Rue, Peter Carleson, Oleg Sidorov, Andrew Erickson, Sean Zumwalt, Sinjin Dixon Warren, Wan-Yi Liu, ShihHsin Chang, et al. Plasma FIB DualBeam Delayering for Atomic Forcen AnoProbing of 14nm FinFET Devices in an SRAM Array. In ISTFA 2015, pages 388-400. ASM International, 2015.

[75] David Halliday, Robert Resnick, and Jearl Walker. Fundamentals of physics. John Wiley & Sons, 2013.

[76] CE Sosolik, AC Lavery, EB Dahl, and BH Cooper. A technique for accurate measurements of ion beam current density using a Faraday cup. Review of Scientific Instruments, 71(9):3326-3330, 2000.

Claims

1. An analytical method, The process involves introducing an ion mixture containing protons and non-hydrogen ions onto a sample, wherein the kinetic energy of the ions in the mixture is 50 kiloelectron volts (keV) or less. An analytical method comprising detecting and measuring X-rays emitted from a sample in response to collisions of the protons and non-hydrogen ions onto the sample.

2. The analytical method according to claim 1, wherein guiding the ions includes guiding separate proton and non-hydrogen ion beams onto the sample, and the separate proton and non-hydrogen ion beams overlap.

3. The analytical method according to claim 1, wherein the introduction of the ions is performed by a focused ion beam column of an independent focused ion beam microscope.

4. The analytical method according to claim 1, wherein the mixture of protons and non-hydrogen ions is generated by passing a mixture of hydrogen gas and a second gas through a plasma ion source.

5. The analytical method according to claim 4, wherein the plasma ion source is an inductively coupled plasma ion source.

6. The analytical method according to claim 5, wherein the hydrogen gas is supplied from a source of purified hydrogen gas, the second gas is supplied from a source of purified second gas, and the hydrogen gas and the second gas are mixed in a manifold fluidly coupled to the inductively coupled plasma ion source.

7. The analytical method according to claim 1, further comprising measuring the proportion of ion species constituting the ion.

8. Measuring the proportion of the aforementioned ion species The separation of the ion beam into a plurality of individual beamlets, thereby separating the ion species according to their respective mass-to-charge ratios, The analytical method according to claim 7, comprising measuring the current of each beamlet.

9. Measuring the current of each of the beamlets includes introducing each beamlet into an isolation Faraday cup, and the isolation Faraday cup is A grounded first electrode having an aperture that receives the beamlets one at a time through it during operation, A second electrode, separated from the first electrode and having a charge collection cavity, One or more electrical insulators between the first electrode and the second electrode, The analytical method according to claim 8, further comprising a potentiometer electrically coupled to the second electrode.

10. The analysis method according to claim 9, wherein measuring the current of each of the beamlets includes scanning each beamlet across the aperture of the grounded first electrode.

11. The analytical method according to claim 9, wherein measuring the current of each of the beamlets involves moving the isolation Faraday cup so that each beamlet then passes through the aperture of the grounded first electrode and reaches the second electrode.

12. The analytical method according to claim 1, wherein the introduction of the ions is performed by a focused ion beam column of a dual-beam apparatus, which also includes a scanning electron microscope column (SEM column).

13. The analytical method according to claim 12, further comprising generating an image of a sample region surrounding the position on the sample where the proton and the non-hydrogen ion collide with the sample, wherein the image is generated by performing a raster scan of an electron beam across the sample region while detecting either secondary electrons emitted by the sample or electrons backscattered by the sample.

14. To detect and measure X-rays emitted from a sample in response to electron collisions from the SEM column onto the sample region, The analytical method according to claim 12, further comprising: combining information derived from detecting and measuring X-rays emitted from the sample in response to collisions of protons and non-hydrogen ions onto the sample with information derived from detecting and measuring X-rays emitted from the sample in response to collisions of electrons from the SEM column onto the sample region, thereby producing an enhanced compositional analysis of the sample region.

15. The method involves measuring the proportion of ion species that make up the ion beam, The ion beam is passed through an immersion magnetic field generated by the magnetic immersion lens of the SEM column, thereby separating the ion beam into a plurality of individual beamlets, and the ion species are separated into the beamlets according to their respective mass-to-charge ratios. The analytical method according to claim 12, further comprising measuring the current of each beamlet, thereby determining the proportion of the ion species constituting the ion beam.

16. Measuring the current of each of the beamlets includes introducing each beamlet into an isolation Faraday cup, and the isolation Faraday cup is A grounded first electrode having an aperture that receives the beamlets one at a time through it during operation, A second electrode, separated from the first electrode and having a charge collection cavity, One or more electrical insulators between the first electrode and the second electrode, The analytical method according to claim 15, further comprising a potentiometer electrically coupled to the second electrode.

17. The analytical method according to claim 1, wherein the kinetic energy of the ions in the mixture is 30 keV or less.

18. The analytical method according to claim 1, wherein the kinetic energy of the ions in the mixture is 24 keV or less.

19. The analytical method according to claim 1, wherein introducing the ions, which include the mixture of protons and non-hydrogen ions, onto the sample comprises focusing a beam containing the mixture of protons and non-hydrogen ions onto the sample.

20. The analytical method according to claim 1, wherein the non-hydrogen ion includes Ar+ ions, Xe+ ions, or Kr+ ions.

21. A method for milling the surface area of ​​a sample (sample surface), The method involves guiding an ion containing a mixture of protons and non-hydrogen ions onto the region, wherein the kinetic energy of the ions in the mixture is 50 kiloelectron volts (keV) or less, the ion beam is raster-scanned across the region, and collisions of the non-hydrogen ions cause sputtering on the sample surface within the region. A method for milling a region of a sample surface, comprising detecting and measuring X-rays emitted from the sample in response to collisions of the protons and non-hydrogen ions onto the sample.

22. A method for milling a region of a sample surface according to claim 21, wherein guiding the ions onto the region includes guiding separate proton and non-hydrogen ion beams onto the region, wherein the separate proton and non-hydrogen ion beams overlap.

23. A method for milling a region of a sample surface according to claim 21, wherein guiding the ions onto the region includes focusing the ions onto the region.

24. A method for milling a region of a sample surface according to claim 21, further comprising detecting a change in the emitted X-rays and stopping the guidance of the ions onto the region.

25. The method for milling a region of a sample surface according to claim 21, wherein the non-hydrogen ions include any of Ar+ ions, Xe+ ions, or Kr+ ions.

26. The method for milling a region of a sample surface according to claim 21, wherein the introduction of the ions is performed by a focused ion beam column of a dual-beam apparatus, including a scanning electron microscope (SEM) column.

27. A method for milling a region of a sample surface according to claim 21, further comprising generating an image of the region, wherein the image is generated by performing a raster scan of an electron beam over the region while detecting either secondary electrons emitted by the sample or electrons backscattered by the sample.

28. A method for milling a region of a sample surface according to claim 21, wherein the kinetic energy of the ions in the mixture is 30 keV or less.

Images