Ion Beam Analysis including ToF-ERDA of complex composition layers

When developing new materials for example, for high-temperature nuclear reactors with the potential of hydrogen production, that are characterized by radiation, high temperature and corrosion resistance, it is indispensable the knowledge of their detailed elemental composition and its possible variation with depth from surface. Several analytical methods based on different physical principles are used to determine the depth distribution of elements in the surface layers of materials. For the quantitative determination of elemental depth profile to a depth of several micrometres are applicable established Ion Beam Analysis (IBA) methods such as RBS, EBS, NRA and ERDA. Their advantage is that they are considered to be absolute and to a certain extent non-destructive. Each of these methods is advantageously used to analyse a certain range of elements, sometimes depending on the combination of other elements present. The recently commissioned Time of Flight Elastic Recoil Analysis (ToF-ERDA) measuring system at the Slovak University of Technology MTF in Trnava significantly enhanced the Ion Beam Laboratory capability of a comprehensive elemental analysis of thin films to the depth of hundreds of nanometres. Using the primary analysing 50 MeV Au beam, the depth profiles of all elements from W to H can be obtained within a single measurement. Analysis of complex samples using traditional IBA methods and high-energy ToF-ERDA is discussed and compared. The first measurements on the new ToF-ERDA installation are also presented.


Introduction
One of the prospective ways of producing hydrogen as an energy carrier of the future is the use of a new generation of high-temperature nuclear reactors as the primary source of energy.Their development requires materials with enhanced specific properties in terms of resistance to radiation, high temperatures and corrosion.Also materials research, including hard coatings, new generation electrodes for electrochemical applications, semiconductor applications, requires knowledge of the detailed composition of elements.
Several standard analytical methods are used to determine the elemental composition in surface thin layers, based on different physical principles.They are based for example on electron microscopy (SEM, TEM, EDX, AES, …), X-ray techniques (XPS, XRF, XRD), optical methods (Ramman spectroscopy) or destructive methods like (SIMS, ICP-MS, …), etc. [1].Each of the methods has its specifics, advantages and disadvantages.
Considering the samples analysis from the surface to a depth of several micrometers also Ion Beam Analysis is an appropriate analytical tool.IBA is a set of established analytical methods based on energetic ion beams.Its advantage is non-destructive and absolute analysis, without the need of reference standards.The latter applies as long as we know all the relevant parameters entering into the evaluation of the measurement reliably, possibly traceable, such as primary beam energy, energy calibration of the detection system, measurement geometry, detector solid angle, number of incident particles, detector response, but also cross-sections, stopping powers, energy losses, etc.
Application of IBA using ion beams with energies of the order of 1 MeV/nucleon for surface thin film elemental depth profiling began in the 1970s [1].The development of the use of IBA methods for materials characterization led to the first meeting of the IBA Conference, which was held 50 years ago, in 1973 [2].At that time the main focus was on Rutherford Backscattering Spectrometry (RBS), Nuclear Reaction Analysis (NRA) and Proton Induced X-Rays Emission (PIXE).The depth distribution of light elements in heavy matrix using Elastic Recoil Detection Analysis (ERDA) was published by J. L 'Ecuyer et al. in 1976 [3].
Summarization of IBA methods can be found, for example, in the following publications.J.W. Mayer and E. Rimini edited the first summary of knowledge of IBA in Ion Beam Handbook for Material Analysis in 1977 [4].The development and application of IBA methods, including Elastic Backscattering Spectrometry (EBS) and ERDA, gradually took place.They were included in updated Handbook of Modern Ion Beam Materials Analysis by Y. Wang and M. Nastasi, published in 2010 [5].The last updated Ion Beam Analysis Fundamentals and Applications book by M. Nastasi, J. W. Mayer and Y. Wang was published in 2019 [6].
The above-mentioned IBA methods are being developed so far, as regards experimental equipment, detection systems and especially evaluation software.This particularly applies to the ERDA system, where a new user-friendly ToF-ERDA detection system evaluation has recently been developed [7].Also, the comprehensive evaluation ToF-ERDA program POTKU [8] is still under development.
Further, the analysis of the depth profile of the samples, which consists of a combination of heavy, medium and light elements, is considered.Our consideration includes the requirement for an analysis covering the studied layers to a depth of approximately 3 micrometers.

Ion Beam Analysis
Our goal is to compare the effectiveness and efficiency of standard IBA analysis and high-energy ToF-ERDA analysis (with a maximum energy of 50 MeV of the primary beam) in terms of sample composition data obtained.The first results from the new ToF-ERDA setup will be shown.

Standard IBA analysis
Hereafter we will refer to RBS, EBS, NRA, PIXE and ERDA with He or H primary analyzing beam as standard IBA analysis.
For the NRA analysis of some elements/isotopes, deuterium or 3He beams may also appear advantageous.But deuterium is in many laboratories excluded due to radiation protection risks.The use of 3 He is for routine analysis expensive or/and requires a special ion source with low gas consumption.
A 3 to 6 MeV He beam is required to cover by the RBS the entire thickness of the layer up to 3 micrometers, while containing heavy elements.Practically all laboratories with tandem accelerators can achieve most of this energy range, but this may be a problem for many single-ended accelerator laboratories.
According to the current International Atomic Energy Agency (IAEA) database [9], the majority of ion accelerators used for materials analysis have an accelerating voltage of up to 1.7 MeV or at most up to 3 MeV, some of which are single-ended but most are tandem accelerators.
Rutherford backscattering cross-section for light elements is mostly below 1% compared to that of heavy elements such as tungsten.Since the analysis of light ions in heavy matrix is usually a problem for RBS and standard PIXE can see elements with an atomic number higher than 11 [6], EBS and/or NRA are the main adepts for that task.Although X-ray detectors with a very thin entrance window or even windowless are already available today, they are not used for quantitative analysis due to the large self-absorption.PIXE cannot be considered as a depth profiling method for the type of samples considered below.

Time of Fligh -Elasic Recoil Analysis
ToF-ERDA analysis uses heavy energetic ions to recoil atoms from the surface layer of the sample and subsequently the mass and energy of individual recoiled atoms are measured and recorded.The atomic mass of the probing ion has to be higher than the mass of the heavier element in the sample to be detected.As for the lighter elements 35 Cl probing beam is common, for heavier ones 63 Cu or 127 I are used, for samples with heavy elements like tungsten, 197 Au beam can be applied.
The energy of primary ions should be preferably about 1 MeV/nucleon, like that for the mentioned RBS.But with the available 6 MV tandem machine it is not applicable for iodine or gold beams.Rather, it is feasible to reach an energy of up to 50 MeV.The latest generation of ToF-ERDA systems uses a gas ionizing chamber as an energy detector [7] of recoiled atoms.
The great advantage of the high energy ToF-ERDA analysis is that during the single measurement one can measure the depth concentration profile of all elements present in the sample.There is no need to know in advance the assumed composition of the sample, especially light impurities.

Experimental
The measurements presented below were performed at the IBA end station of the 6 MV tandetron accelerator [10,11] and on the newly installed ToF-ERDA setup.
Measurement parameters were always chosen so that the entire or maximum possible thickness of the analyzed layers was covered, and they were able to obtain as much information as possible with a minimum number of measurements.Specific values will be given in the presented measurements.
The basic RBS measurements were taken at 170°, while EBS and NRA were taken at the appropriate detection angle.The recoil angle during helium ERDA for hydrogen measurement was 30°.HPGe PIXE detector was at 45° to the beam axes.ToF-ERDA was measured at a 40.6° recoil angle and 29.5 mbar pressure of the GIC detector.
The Cu and Ti oxide/nitride thin films shown below were prepared by reactive magnetron sputtering.High entropy metal coatings TiNbZrHf with high thermal, mechanical and corrosion properties, were prepared by High Target Utilization Sputtering (HiTUS) technique from a TiNbZrHf target in 20 sccm Ar atmosphere, with 5 sccm C2H2 flows.More details can be found in Lofaj et al. [15].

Analysis Results
When comparing sample analysis using standard IBA methods (as specified above) and using ToF-ERDA, we start with a simple single-element thin film, layers of a heavier element on a light substrate.Then we will continue with two main elements of film containing light impurities on a light substrate.Finally, we will compare the complex layer containing eleven main and admixture elements.

Cu film on Si substrate
The 2 MeV He + RBS spectrum of the single Cu layer on Si substrate (sample Nr. 200-5) prepared by magnetron sputtering, brings the information, about 1894×10 15 at/cm 2 thickness of Cu layer, see Figure 1(a).No other elements in the Cu layer were indicated.1.
The resulting depth profile from 40 MeV I 7+ ToF-ERDA analysis gives the same (within uncertainty limits) 1885×10 15 at/cm 2 Cu layer thickness, which is 224 nm, see Figure 1(b).However, ToF-ERDA gave additional information about the presence of the surface and interface contamination, see Table 1.At a depth of 200 nm in the Si substrate, there is still 4 at.% of diffused copper.Table 2. Atomic elemental concentration of TiN layer between the two vertical lines, see Figure 3(c).2.

TiZrHfNb film on Si substrate
Finally, we compare standard IBA and ToF-ERDA analysis of a truly complex sample in which 11 elements were identified.The TiZrHfNb was prepared by direct current magnetron sputtering with 5 sccm C2H2 and 20 sccm Ar gas flow.A 5.65 MeV He + beam was used for IBA analysis to take advantage of the 12 C(α,α0) 12 C EBS signal enhancement and RBS of about 2 µm TiZrHfNb film, see Figure 4   Since the heaviest element in the sample was hafnium, a 45 MeV Au 8+ beam was used for ToF-ERDA analysis, see Figure 5(a).ToF-ERDA enabled the depth profile of the film to be monitored to a depth of 500 nm.ToF-ERDA established in TiZrHfNb film presence of O, Si, Ar and Fe, see Figure 5(b), which were not identified using the standard IBA.Although the PIXE measurement indicated the presence of Fe, there was no information about its localization, or distribution.Since ToF-ERDA does not distinguish the signal from neighbouring 90 Zr and 93 Nb, also PIXE data were used for their quantification.3. The estimated uncertainties of the presented results have not been evaluated in detail, but are in the order of units of percent.

Summary
If the analysis of thin layers with a complex composition is required, e.g. with a content ranging from heavy elements from Hf through Zr, Ti, etc., with a content of light elements ranging from Si through Mg, N, C, B to H, it becomes a rather challenging task.In such a case, it is necessary to use a combination of several, often even all the abovementioned standard IBA methods.Then the iterative evaluation of separate IBA data is necessary, which is a more demanding and time-consuming activity.However, the resulting depth profiles of thin films with a complex elemental composition with a thickness from few nanometers up to 3 micrometers can be reached.
As it was shown, standard IBA as well as high energy ToF-ERDA can provide quantitative elemental analysis within the given range of samples.But if the occurrence of some light elements is not expected in the sample, it may happen that they will be missed and will not be detected using the standard IBA analysis.
On the other hand, using high energy ToF-ERDA with a 45 MeV Au beam can provide complex analysis of all elements present in probing layer, which can exceed 0.5 µm depth.In the case of a complete analysis of the mentioned layers with a greater thickness of several µm, ToF-ERDA analysis must be supplemented by RBS, or other standard IBA methods, depending on the specific case.However, only a limited number of ion beam laboratories have these options available.

Figure 1 (
Figure 1(a). 2 MeV He + RBS spectrum of single Cu layer on Si.No light element contamination is evident.

Figure 1 (
Figure 1(b).Elemental depth profile in logarithmic scale derived from 40 MeV I 7+ ToF-ERDA of Cu layer on Si.Determined surface and interface light element contamination is given in Table1.

Figure 2 .
Figure 2. Depth profile of CuO layer on Si substrate, derived from 40 MeV I 7+ ToF-ERDA measurement.No contamination was detected.4.3.TiN film on Si plus 40 keV Ag implantation on Si substrate In the case of TiN/Si(Ag) sample only 40 MeV I 7+ ToF-ERDA measurement was done.At Figure 3(a) the ToF-ERDA histogram shows the 3D spectrum of Time of Flight (ToF) vs Energy of the detected particle, while the yield is expressed in color scale.Each banana-shaped part of the histogram represent a signal coming from an individual element.ToF-ERDA measurement also provides energy spectra of individual elements, see Figure 3(b).The total energy spectrum, created by summing the partials (not shown), would a similar character to that of the standard RBS spectrum.From the summed energy spectrum, it would not be possible to separate the individual signals of O and N located in the film, added to the Si-substrate signal, and determine the amount of O and N in the film.

Figure 3 (
Figure 3(a).40 MeV I 7+ ToF-ERDA histogram of TiN/Si(Ag) layer on Si substrate.Banana curves corresponding to individual elements are indicated.

6
The average concentration of elements in the TiN layer between the two vertical lines of the derived depth profile, is shown in Table2.The depth profile in Fig3(c) indicates surface oxygen contamination and subsurface 25.25×10 15 cm -2 implanted Ag profile.No interface contamination was indicated.

Figure 3 (
Figure 3(c).Depth profile derived from ToF-ERDA on Figure 3(a).The profile of implanted Ag and increased oxygen contamination can be seen on the surface of the sample.For elemental concentration of TiN film, see Table 2.
(a), and ERDA for hydrogen analysis, see Figure 4(b).Since it was not possible to separate the signal from Zr and Nb from the RBS spectra, the PIXE measurement was used to determine the ratio of Zr and Nb in the sample, and subsequently the concentrations of individual Zr and Nb were determined.The 14 N(α,p0) 17 O NRA for nitrogen and PIXE spectra are not shown.

Table 1 .
Surface and interface contamination of Cu film on Si substrate determined by ToF-ERDA measurement.

Table 3 .
Average atomic elemental concentrations of TiZrHfNb film derived from standards IBA and from ToF-ERDA measurement.