Acta Polytechnica CTU Proceedings https://doi.org/10.14311/APP.2024.50.0088 Acta Polytechnica CTU Proceedings 50:88–93, 2024 © 2024 The Author(s). Licensed under a CC-BY 4.0 licence Published by the Czech Technical University in Prague OPTIMIZATION OF FIB MILLING PROCEDURE FOR MICROMECHANICAL TESTING OF CEMENT PASTES Jan Procházkaa,∗, Jiří Němeček 1,b, Patricie Halodováa, Jiří Němeček 2,b a Research centre Řež, Hlavní 130, Řež, 250 68 Husinec, Czech Republic b Czech Technical University in Prague, Faculty of Civil Engineering, Thákurova 7, 166 29 Prague 6 – Dejvice, Czech Republic ∗ corresponding author: jan.prochazka@cvrez.cz Abstract. This study focuses on optimizing the Focused Ion Beam (FIB) preparation process for micro-beams made from cement paste prepared from Portland cement CEM I-42.5R, which includes four phases: inner and outer products, Portlandite, and residual clinker. The effects of the primary electron beam, focused ion beam, image drift, imaging issues, and sample moisture cycling were investigated. Recommendations on specific fabrication procedures and FIB milling parameters, such as accelerating voltage, probe current, and time, were provided. Keywords: Cement paste, focused ion beam, nanoindentation, micro-beam, C–S–H gel, scanning electron microscopy. 1. Introduction Ordinary cement paste is a heterogeneous material, primarily composed of Calcium-Silica-Hydrate (C–S– H) gel, Portlandite (CH), residual clinker (C), porosity, and other minor phases. The C–S–H gel exists in two variants with different densities. Low-density C–S–H gel, along with other minor phases, forms the outer product (OP), while high-density C–S–H gel, also with other minor phases, forms the inner product (IP) [1, 2]. These phases exhibit variable chemical compositions and different mechanical properties. Nanoindenta- tion can be used to assess the mechanical properties, such as the Young’s modulus or hardness of these micrometer-sized phases [3–5]. However, nanoindentation cannot directly measure the compressive and tensile strengths or fracture en- ergy of these phases. These properties can be deter- mined by fabricating micro-pillars or micro-beams us- ing focused ion beam (FIB) techniques. Subsequently, compressing the micro-pillars [6, 7] or bending the micro-beams [8, 9] with a nanoindenter can be used to derive the elastic and fracture properties of the indi- vidual phases. Despite the indisputable advantages of the FIB milling technique, several potential undesir- able effects associated with ion beam nanofabrication, such as temperature-induced phase transformations, gallium implantation, and unintended material rede- position, can significantly influence the measurement results [9]. Therefore, when using an ion beam for micro-beam fabrication, it is essential to select appropriate sputter- 1Postdoctoral researcher at Czech Technical University in Prague, ORCID: 0000-0002-5635-695X. 2Professor at Czech Technical University in Prague, ORCID: 0000-0002-3565-8182. ing parameters to minimize the impact of nanofabrica- tion on the mechanical testing results. This study elu- cidates the crucial steps involved in preparing micro- beams using FIB techniques and discusses the utiliza- tion of optimized parameters to ensure the efficacy and accuracy of both the manufacturing process and subsequent nanomechanical testing. 2. Materials and methods 2.1. Material The cement paste samples were mixed from Portland cement CEM I-42.R (Českomoravský cement, Czech Republic) of the chemical composition shown in Ta- ble 1 without chemical admixtures and the water-to- cement ratio of 0.4. A fresh paste was cast in cylin- drical molds (27mm in diameter, 70mm in height) and vibrated to avoid excessive air voids in the struc- ture. Then, the samples were stored in lime water for 20 months. Detailed procedure and sample labeling are described in [10]. 2.2. Sample preparation After the curing period in lime water, the samples were dried at 50 °C for 3 days. Then, the samples were ground for 1 min on SiC paper grit 2 000 and for 2.5 min on SiC paper grit 4 000 with no lubrica- tion. The debris was continuously swept away from the papers throughout the grinding procedure. Final polishing took place on a soft cloth with alcohol-based diamond suspensions for 15 seconds with 0.25µm grains. The samples were ultrasonically cleaned in an alcohol bath for 1 min after each polishing step [2]. After the polishing, three coarse base-defining marks were engraved in the corners of the analyzed fragment using a diamond tip. 88 https://doi.org/10.14311/APP.2024.50.0088 https://creativecommons.org/licenses/by/4.0/ https://www.cvut.cz/en vol. 50/2024 FIB procedure optimization for cement pastes Component CaO SiO2 Al2O3 Fe2O3 MgO Na2O K2O SO3 MnO CEM I (wt. %) 63.77 20.51 4.74 3.30 1.05 0.15 0.95 3.07 0.09 Table 1. Chemical composition of the used Portland cement (XRF data supplied by the manufacturer). Process step Label deposition Trench milling Rough undercut Fine polishing Shortening Pattern Lines Text Stairs rectangles Polishing rectangle Polishing rectangles Polishing rectangle Dimensions (µm) 175 × 45 × 0.1 25 × 15 × 15 24 × 15 × 15 20 × 3 × 10 3 × 2 × 10 8 × 5 × 10 IB (pA) ∼ 200 ∼ 1 000 ∼ 1 000 ∼ 600 ∼ 600 ts (min) ∼ 240 2× ∼ 120 2× ∼ 90 ∼ 10 θe (°) 55 55 23 23 55 / 23 θB (°) 0 0 32 32 0 / 32 Table 2. Recommended FIB parameters and estimated process times for different steps in micro-beam preparation routine. 2.3. Imagining and ion machining Polished samples were glued to the SEM stub using colloidal silver lacquer (Agar Scientific), and the same lacquer was applied to the sides of the samples to reduce charging. The location of the four main hydration phases (OP, IP, CH, and C) with large enough dimensions suitable for micro-beam fabrication were identified using a Scanning electron microscope (SEM) Phenom XL Desktop SEM and back-scatter electron (BSE) detector. Also, the coordinates of locations and base- defining marks were recorded for further aid. The micro-beams and labels for nanoindentation tests were prepared using a dual-beam FIB-SEM Tes- can Lyra3 GMU, equipped with Ga+ COBRA FIB column tilted 55° to the electron column and gas in- jection system with platinum, tungsten, and carbon precursors for ion-beam assisted deposition (IBAD). The microscope is equipped with external and in-beam detectors for both BSE and secondary electrons (SE). 5-degree-of-freedom stage allows sample position- ing with adjustable tilt around axis perpendicullar to both electon and ion beams and (actual) rotation around axis perpendicullar to the sample surface. The controlling system allows software adjustments of im- age shift and (apparent) rotation. The image is not deformed by apparent rotation only if the sample is perpedicullar to the beam used. At first, the coarse base-defining marks were made with two platinum alignment tags: one located inside the scratch mark, labeled as coarse, and one on the polished surface, labeled as fine. This set of three base-defining marks allows easier beam positioning using microscopes ranging from optical microscopes to SEMs. (a) SE (b) BSE Figure 1. SEM images of the label, alignment tags, and outline for the beam in residual clinker using (a) SE and (b) BSE detectors. 2.4. Micro-beams After the marking, the regions of interest were iden- tified and labeled using platinum IBAD. Additional beam outlines and alignment tags were added for eas- ier and more precise ion milling routines, see Figure 1. The positioning and orientation of a cantilever must allow enough room for nanoindentation measurement of the selected phase in the proximity of the can- tilever root. The root shall be as far as possible from phase interfaces too. The desired dimensions of the final equilateral triangular cantilever for testing are 20 × 8µm. The process of cantilever preparation is divided in four steps: high-current trench milling, high-current rough cantilever undercut, low-current cantilever side- wall and cantilever base polishing, and low-current cantilever shortening. Recommended parameters and estimated process times are listed in Table 2. 89 J. Procházka, J. Němeček, P. Halodová, J. Němeček Acta Polytechnica CTU Proceedings (a) SE, perpendicular to the electron beam. (b) BSE, perpendicular to the electron beam. (c) In-Beam BSE, perpendicullar to the electron beam. (d) SE, perpendicular to the ion beam. (e) In-Beam BSE, perpendicular to the ion beam. Figure 2. Comparison of SEM images of residual clinker grain with the sample surface (a–c) perpendicular to electon beam and tilted to the ion beam, and (d–e) tilted to the electron beam and perpendicular to the ion beam, using different detectors (a) SE, (b) In-beam BSE. 3. Results and Discussion 3.1. Imaging issues As the cement paste is poorly conductive in a vac- uum chamber, surface charging is an issue for SEM imaging using SE because the charged surface has a higher secondary electron emission rate as shown in Figures 2(a) and 2(d) due to the negative charge build-up on the scanned surface. BSE imaging is less affected by immediate surface charge, see Figure 2(b), because the detector collects the signal from electrons interacting elastically with the sample surface. Due to their higher energy, BSEs are less susceptible to deflec- tion by the electric fields caused by surface charging. In-chamber BSE detectors have better signal than their in-beam (IB) counterparts that are located in- side the SEM column, see Figures 2(b) and 2(c). On the other hand, In-chamber detectors are prone to mechanical damage and cannot be used with tilted specimens during ion fabrication. In cases where the sample was tilted for FIB pro- cesses or cross-section documentation, a combination of the SE detector, primarily for focusing, and the in-beam BSE detector, providing phase and chemical contrast, was used. Otherwise, the in-chamber BSE detector was used. 3.2. Effect of primary electron beam Continual SEM scanning loads the entire surface of the sample in the view field. In the case of a cemen- titious sample, the energy deposited by the incident electron beam on the surface cannot dissipate through the material, leading to energy absorption and subse- quent thermal effects, such as localized heat damage of the surface. The amount of heat generated depends on the current of the electron beam, and the actual BI dE,R (nm) dE,D (nm) IE (pA) 1 7.8 13 ∼ 9 2 7.4 12 ∼ 11 3 6.6 11 ∼ 20 4 6.3 11 ∼ 24 5 5.9 10 ∼ 35 6 5.4 9.3 ∼ 50 7 5.0 8.5 ∼ 67 8 4.6 7.8 ∼ 95 9 4.5 7.2 ∼ 130 10 4.6 6.6 ∼ 190 11 5.3 6.1 ∼ 270 12 6.9 5.7 ∼ 330 13 10 5.5 ∼ 450 14 15 5.6 ∼ 550 15 24 6.1 ∼ 750 16 37 7.4 ∼ 880 17 62 11 ∼ 1 000 Table 3. Electron beam current IB and spot sizes dB for resolution-optimised (-R) and focal-depth- optimised (-D) scanning modes, respectively, with acceleration voltage of 15 keV. temperature rise depends on several factors, including exposure time, thermal conductivity, heat capacity, and the geometry of the sample [11]. The microscope used in the study offers two beam- focusing setups optimized for improved resolution or enhanced depth of focus. Differences between esti- mated spot sizes dE of both modes and beam current IE for beam intensity (BI ) indices are listed in Table 3. Figure 3(a) shows the micro-beam prepared under continuous electron imaging with electron beam inten- sity BI 17 overheating the sample surface. Electron 90 vol. 50/2024 FIB procedure optimization for cement pastes (a) Damaged micro-beam prepared in IP under con- tinuous electron imaging. (b) Micro-beam prepared in IP with electron imaging deactivated. Figure 3. Comparison of micro-beams damaged and undamaged by extensive electron beam heating. imaging using BI 7 and deactivating the continuous ac- quisition during the FIB process minimizes the surface degradation, see Figure 3(b). BI 7 with pixel dwell time of 3.2µs and resolution of 512 × 512 px offers an acceptable compromise between low sample heat loading and reasonable preview image quality. To suppress the surface heat loading even further, the continuous image acquisition should be turned off during FIB processes. 3.3. Effect of ion beam The FIB, as well as the electron beam or laser beam, is assumed to be elliptical in the plane perpendicular to the beam with a Gaussian-type power profile. As the beam current is defined by masking the defocused beam in the column, the different current settings lead to different standard deviations of the power profile. A higher standard deviation results in blunter edges for higher current ratings. Figure 4 shows four identical ion beam etching routines of rectangular object, 1µm wide, 8µm long and deep, in Portlandite using four different FIB presets defined in Table 4. It is clearly seen that with higher beam currents, the edges of the etched objects are less defined and blunter compared to those prepared using lower beam currents. As described in [12], higher current ratings lead to higher temperature rises and broader heated volumes. Figure 4. Comparison of FIB etching routines per- formed with different ion beam presets. Preset dA (µm) dB (nm) IB (pA) ts (min) 1 500 130.1 ∼ 7 000 0:20 2 200 109.6 ∼ 1 000 2:13 3 150 118.2 ∼ 600 4:09 4 90 121.9 ∼ 200 13:33 Table 4. FIB preset parameters for ion acceleration voltage of 30 keV and time needed for etching 1µm wide, 8µm long and deep rectangular trench in Port- landite. In the case of micro-beams prepared in cement paste phases, both effects contribute to artifacts created on the sides of the micro-beam and the base. The Gaus- sian beam broadening primarily causes blunt edges, while beam-induced heating can lead to geometrical changes in the micro-beam due to possible phase trans- formations, decomposition, or the release of adsorbed gel water in C-S-H products, resulting in shrinkage and microcracking [9]. Figure 5(a) shows preparation of a micro-beam in Portlandite interrupted after the second stage of rough undercut using the IB ∼ 7 nA because of excessive heat loading induced by the FIB. Reducing the beam current to IB ∼ 1 nA eliminated the excessive heat loading, see Figure 5(b). 3.4. Effect of image drift Because of the low conductivity of cement paste in the vacuum chamber, the accumulation of positive charge due to Ga+ ion implantation and secondary electron emission during the Ga+ ion beam fabrica- tion can lead to significant charging effects, causing apparent image movement called image drift. This effect can be mitigated by drift correction, where the milling/polishing routine is automatically paused and the reference area is scanned to estimate corrections. The FIB/SEM used in this study allows for either finer pattern correction, where the FIB schedule pat- tern coordinates are altered for the compensation, or coarser but wider-range image shift correction, where the FIB image shift is adjusted for compensation. In the case of cement pastes, the period between correction scans had to be reduced to 30–60 seconds. Even though the wider-range correction method of 91 J. Procházka, J. Němeček, P. Halodová, J. Němeček Acta Polytechnica CTU Proceedings (a) Micro-beam undercut performed using IB ∼ 7 nA (b) Micro-beam undercut performed using IB ∼ 1 nA Figure 5. Comparison of micro-beams undercut pre- pared in Portlandite using ion beam current of (a) 7 nA and (b) 1 nA. image shift was used, it could not always compen- sate for the drift. In such cases, additional polishing steps were added to achieve the desired micro-beam geometry. See Figure 6(a). 3.5. Effect of moisture cycling Hydration and dehydration processes in C–S–H gel are associated with volumetric and mechanical changes in the gel [13]. The changes of water partial pressure in the sample between the storage in the given atmo- sphere and micro-beam preparation in the SEM evac- uated chamber induce high artificial stress in the sam- ples, see Figure 7. It was found that week-long storage of the samples in the vacuum of SEM (p ∼ 10−6 Pa) causes significantly fewer cracks in the samples than at- mospheric storage overnight between FIB procedures. Attention must be given to this effect when trans- porting ready-made micro-beams for nanoindentation testing to another laboratory. We recommend trans- porting and storing samples in a vacuum to mitigate potential issues. 4. Conclusions The preparation of micro-beams in cement paste is a time-consuming process with a significant failure rate. During the preparation of beams for nanoinden- tation testing using FIB techniques, several critical parameters require careful assessment to ensure the (a) Micro-beam in OP phase recovered by additional polishing. (b) Micro-beam prepared in IP phase. Figure 6. Comparison of micro-beams (a) affected and (b) not affected by excessive sample drift. Figure 7. Sample surface damaged by moisture cy- cling. procedure’s success. The general recommendations are given as: • Parameters BI 7 of the electron beam equal to electron current Ie ∼ 70 pA combined with higher- gain SE and in-beam BSE detectors are suitable to use to avoid surface overheating and to ensure reasonable image quality. Additionally, continuous image acquisition should be turned off during the FIB milling for the same reason. • High currents used during FIB milling can lead to excessive heat loading of the sample, resulting in phase transformation, visible bubbles or cracks on the micro-beam, or even geometry changes. Safe current values for different milling steps of Port- landite are provided in Table 2. • To avoid image drift caused by charging due to the accumulation of Ga+ ions during FIB fabrication, 92 vol. 50/2024 FIB procedure optimization for cement pastes it is recommended to perform drift correction by rescanning the measured area at intervals of 30 to 60 seconds. • Cyclic water resaturation of the cement paste can lead to the creation of a significant number of micro- cracks, potentially damaging already fabricated micro-beams. Therefore, it is advisable to maintain the sample under vacuum conditions throughout the preparation and testing phases to effectively mitigate these effects. List of symbols BI Beam Intensity index BSE Back Scattered Electron (detector) dB Ion Beam estimated spot size [nm] dE Electron Beam estimated spot size [nm] FIB Focused Ion Beam IB Ion Beam current [A] IE Electron Beam current [A] IBAD Ion Beam Assisted chemical vapour Deposition SE Secondary Electron (detector) SEM Scanning Electron Microscope ts estimated process time for given beam preparation step [min] θe Sample tilt angle respective to the electron beam [°] θB Sample tilt angle respective to the ion beam [°] C Residual Clinker C–S–H Calcium – Silicate – Hydrates CH Portlandite IP Inner Product OP Outer Product Acknowledgements This work was financially supported by the project of the Czech Science Foundation grant number 23-05435S. The presented results were obtained using the CICRR infrastructure, which is financially supported by the Min- istry of Education and Culture – project LM2023041. References [1] K. L. Scrivener. Backscattered electron imaging of cementitious microstructures: understanding and quantification. Cement and Concrete Composites 26(8):935–945, 2004. https: //doi.org/10.1016/j.cemconcomp.2004.02.029 [2] J. Němeček, J. Lukeš, J. Němeček. High-speed mechanical mapping of blended cement pastes and its comparison with standard modes of nanoindentation. Materials Today Communications 23:100806, 2020. https://doi.org/10.1016/j.mtcomm.2019.100806 [3] W. C. Oliver, G. M. Pharr. An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments. Journal of Materials Research 7(6):1564–1583, 1992. https://doi.org/10.1557/JMR.1992.1564 [4] C. Hu, Z. Li. A review on the mechanical properties of cement-based materials measured by nanoindentation. Construction and Building Materials 90:80–90, 2015. https: //doi.org/10.1016/j.conbuildmat.2015.05.008 [5] Z. Luo, W. Li, K. Wang, S. P. Shah. Research progress in advanced nanomechanical characterization of cement-based materials. Cement and Concrete Composites 94:277–295, 2018. https: //doi.org/10.1016/j.cemconcomp.2018.09.016 [6] R. Shahrin, C. P. Bobko. Micropillar compression investigation of size effect on microscale strength and failure mechanism of Calcium-Silicate-Hydrates (C-S-H) in cement paste. Cement and Concrete Research 125:105863, 2019. https://doi.org/10.1016/j.cemconres.2019.105863 [7] C. Puttbach, G. S. Prinz, C. D. Murray. Strength and stiffness characterization of ultra-high performance concrete (UHPC) cement paste phases through in-situ micro-mechanical testing. Cement and Concrete Composites 149:105520, 2024. https: //doi.org/10.1016/j.cemconcomp.2024.105520 [8] J. Němeček, V. Králík, V. Šmilauer, et al. Tensile strength of hydrated cement paste phases assessed by micro-bending tests and nanoindentation. Cement and Concrete Composites 73:164–173, 2016. https: //doi.org/10.1016/j.cemconcomp.2016.07.010 [9] J. Němeček, J. Maňák, J. Němeček, T. Krejčí. Effect of vacuum and Focused Ion Beam generated heat on fracture properties of hydrated cement paste. Cement and Concrete Composites 100:139–149, 2019. https: //doi.org/10.1016/j.cemconcomp.2019.03.027 [10] J. Němeček, P. Trávníček, M. Keppert, et al. Nanomechanical analysis of Gamma-irradiated cement paste exposed to different humidities. Construction and Building Materials 393:131969, 2023. https: //doi.org/10.1016/j.conbuildmat.2023.131969 [11] J. Goldstein. Scanning Electron Microscopy and X-Ray Microanalysis: A Text for Biologists, Materials Scientists, and Geologists. Springer US, 1981. [12] T. Ishitani, H. Kaga. Calculation of local temperature rise in focused-ion-beam sample preparation. Journal of Electron Microscopy 44(5):331–336, 1995. https: //doi.org/10.1093/oxfordjournals.jmicro.a051185 [13] M. B. Pinson, E. Masoero, P. A. Bonnaud, et al. Hysteresis from multiscale porosity: modeling water sorption and shrinkage in cement paste. Physical Review Applied 3(6):064009, 2015. https://doi.org/10.1103/PhysRevApplied.3.064009 93 https://doi.org/10.1016/j.cemconcomp.2004.02.029 https://doi.org/10.1016/j.cemconcomp.2004.02.029 https://doi.org/10.1016/j.mtcomm.2019.100806 https://doi.org/10.1557/JMR.1992.1564 https://doi.org/10.1016/j.conbuildmat.2015.05.008 https://doi.org/10.1016/j.conbuildmat.2015.05.008 https://doi.org/10.1016/j.cemconcomp.2018.09.016 https://doi.org/10.1016/j.cemconcomp.2018.09.016 https://doi.org/10.1016/j.cemconres.2019.105863 https://doi.org/10.1016/j.cemconcomp.2024.105520 https://doi.org/10.1016/j.cemconcomp.2024.105520 https://doi.org/10.1016/j.cemconcomp.2016.07.010 https://doi.org/10.1016/j.cemconcomp.2016.07.010 https://doi.org/10.1016/j.cemconcomp.2019.03.027 https://doi.org/10.1016/j.cemconcomp.2019.03.027 https://doi.org/10.1016/j.conbuildmat.2023.131969 https://doi.org/10.1016/j.conbuildmat.2023.131969 https://doi.org/10.1093/oxfordjournals.jmicro.a051185 https://doi.org/10.1093/oxfordjournals.jmicro.a051185 https://doi.org/10.1103/PhysRevApplied.3.064009 Acta Polytechnica CTU Proceedings 50:88–93, 2024 1 Introduction 2 Materials and methods 2.1 Material 2.2 Sample preparation 2.3 Imagining and ion machining 2.4 Micro-beams 3 Results and Discussion 3.1 Imaging issues 3.2 Effect of primary electron beam 3.3 Effect of ion beam 3.4 Effect of image drift 3.5 Effect of moisture cycling 4 Conclusions List of symbols Acknowledgements References