Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 Contents lists available at ScienceDirect 2 3 4 Ultramicroscopy 5 6 journal homepage: www.elsevier.com/locate/ultramic 7 8 9 10 11 12 Specimen preparation for correlating transmission electron microscopy 13 and atom probe tomography of mesoscale features 14 a b,c b,c a,n Q115 Matthew I. Hartshorne , Dieter Isheim , David N. Seidman , Mitra L. Taheri 16 a 17 Department of Materials Science and Engineering, Drexel University, 3141 Chestnut Street, Philadelphia, PA 19104, USA b 18 Department of Materials Science and Engineering, Northwestern University, 2220 Campus Drive, Evanston, IL 60208-3108, USA c Northwestern University Center for Atom-Probe Tomography, 2220 Campus Drive, Evanston, IL 60208-3108, USA 19 20 21 article info abstract 22 23 Article history: Atom-probe tomography (APT) provides atomic-scale spatial and compositional resolution that is ideally 24 Received 20 November 2013 suited for the analysis of grain boundaries. The small sample volume analyzed in APT presents, however, 25 Received in revised form a challenge for capturing mesoscale features, such as grain boundaries. A new site-specific method 26 8 May 2014 utilizing transmission electron microscopy (TEM) for the precise selection and isolation of mesoscale Accepted 24 May 2014 27 microstructural features in a focused-ion-beam (FIB) microscope lift-out sample, from below the original 28 surface of the bulk sample, for targeted preparation of an APT microtip by FIB–SEM microscopy is 29 Keywords: presented. This methodology is demonstrated for the targeted extraction of a prior austenite grain 30 Atom probe tomography boundary in a martensitic steel alloy; it can, however, be easily applied to other mesoscale features, such Focused ion beam 31 as heterophase interfaces, precipitates, and the tips of cracks. Transmission electron microscopy & 32 2014 Elsevier B.V. All rights reserved. 33 34 67 35 68 36 69 37 1. Introduction features of interest within the microtip in materials that do not 70 38 pulse electropolish reliably, such as cemented carbides [19,26]. 71 39 Atom-probe tomography (APT) and atom-probe field-ion These methodologies may involve many steps depending on the 72 40 microcopy (APFIM) and have been used to analyze point, line feature being investigated, and relies heavily on the desired 73 41 and planar imperfections in materials since the beginning of their feature occurring within a reasonable distance from the apex of 74 – 42 development in the late 1960s [1 11]. The use of APT to study the electropolished microtip. With the development of focused 75 43 grain boundaries (GBs) in metallic alloys is of particular interest, as ion-beam (FIB) microscopes, it became possible to produce APT 76 44 their structure and composition can determine many properties of specimens with their microtips containing GBs and other features 77 45 the bulk material, such as ductility, toughness and corrosion of interest, with both predictability and accuracy [27–31]. Various 78 – – 46 resistance [12 18]. FIB SEM-microscope based methodologies have been developed 79 47 In the early years of the FIM and the APFIM techniques, to lift-out sections of GBs and place them on single microtips or on 80 48 specimens were prepared by electropolishing to form a microtip silicon microrods residing on microtip arrays for analyses [27,28]. 81 49 geometry, which would permit both field ionization and field Alternatively, techniques involving FIB–SEM microscopy ion-beam 82 50 evaporation of specimens [6,19]. While successful, electropolishing milling and analyzing specimens in situ have also been successful 83 51 alone does not readily permit site specific targeting of imperfec- for the study of different features, such as GBs and heterophase 84 52 tions or mesoscale features. To address this problem, methods interfaces [29,32], which are central to the control of materials 85 53 have been developed using transmission electron microscopy properties. In steels, for example, it has been demonstrated that 86 54 (TEM) to identify features of interest and then electropolishing segregation of elements such as copper, tin, antimony, phosphorus 87 55 repeatedly to position the feature of interest in the microtip, which and sulfur to austenite GBs during hot working leads to a loss of 88 – – 56 is subsequently analyzed [19 25]. In addition to the pulse electro- ductility [12 14]. 89 57 polishing approach, broad ion beams have been used to position Despite their importance in determining material properties, 90 58 microstructural features with a spacing in the 5–40 μmrange 91 (corresponding to ASTM grain size 6–12 in metals [33])havenot 92 59 n Corresponding author. Tel.: þ1 215 895 6618. been analyzed readily with APT, due to difficulties in preparation of 93 60 E-mail addresses: [email protected] (M.I. Hartshorne), microtips using either electropolishing or existing FIB–SEM micro- 94 61 [email protected] (D. Isheim), 62 [email protected] (D.N. Seidman), scopy lift-out methods. An important example of these mesoscale 95 63 [email protected] (M.L. Taheri). features is prior austenite GBs, which serve as nucleation sites for 96 64 97 http://dx.doi.org/10.1016/j.ultramic.2014.05.005 98 65 0304-3991/& 2014 Elsevier B.V. All rights reserved. 66 99

Please cite this article as: M.I. Hartshorne, et al., Specimen preparation for correlating transmission electron microscopy and atom probe tomography of mesoscale features, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.05.005i 2 M.I. Hartshorne et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 carbides and intermetallic precipitates, further reducing ductility, This methodology as demonstrated in this paper, can be used to 67 2 toughness and corrosion resistance [13,15–18]; therefore, character- select desired mesoscale features from among large numbers of 68 3 ization of compositional changes at these features is critical for microscale features at depths of several microns below the surface 69 4 understanding the evolution of microstructures in steels. During an of the sample, which has applications beyond grain boundaries. 70 5 electropolishing procedure, a large number of specimens and Irradiated materials [40,41], such as nuclear reactor steels, can be 71 6 repeated electropolishing steps are required to locate an interface. analyzed at a specific depth with the specimen’s z-axis parallel to 72 7 For example, in the methods presented by Henjered and Norden the surface of the sample to ensure that all features within the 73 8 [19] and Krakauer and Seidman [20,22–25], an approximately dataset received the same dose or to analyze a section of a GB at a 74 9 1–1.5 μm long section of the specimen is rendered electron trans- specific depth from the irradiated surface; current methods only 75 10 parent after each electropolishing step [19,22]. As such, if 500 nm of allow for microtips to be prepared parallel to the bulk sample 76 11 the viewed volume is removed with each polishing step, 20 surface if they are extracted from a region very close to the bulk 77 12 electropolishing steps will be required to view 10 μm of the speci- sample’s surface. Thin-films and semiconductor device structures 78 13 men. In addition to this large number of steps, materials with a [42,43], such as electrode-gate-substrate triple points, can also be 79 14 mixture of microscale and mesoscale features, such as the lath prepared with the device-substrate interface parallel to the ana- 80 15 boundaries and prior austenite GBs in martensitic steels, require lysis direction in the APT to minimize the probability of the 81 16 more intensive analyses employing selected area diffraction (SAD), specimen rupturing during analysis. 82 17 using a TEM, of each boundary encountered to determine whether 83 18 or not it is of interest. While this approach is certainly possible it 84 19 requires a significant amount of TEM analysis. 2. Samples and experimental apparatus 85 20 More recently, specimen preparation has been performed using 86 21 FIB–SEM microscope lift-out procedures, which are more site- A martensitic ultrahigh strength steel, currently marketed as 87 22 specific than conventional electropolishing techniques; standard PremoMet™ by Carpenter Technology Corporation, was used to 88 23 lift-out techniques require, however, a straight and several micron validate the methodology presented in this article. This steel was 89 24 long section of a GB to extract a specimen [27]. Cairney et al., Felfer austenitized at 1191 K for 1.5 h, quenched, cryogenically treated at 90 25 et al. and Tytko et al. have demonstrated methods that have 200 K for 8 h and tempered at 533 K for 2 h. The microstructure 91 26 successfully analyzed grains in the 5–40 μm size range, which contains a prior austenite grain diameter of approximately 10 μm 92 27 involves lifting out a rod of material with a GB located near its end, and a sub-micron martensitic lath diameter. As such, the features 93 28 then shaping the end of the rod into an APT specimen [34–38]. of interest, the prior austenite GBs, must be precisely selected 94 29 These techniques provide an excellent methodology for the fast, and distinguished carefully from among the more numerous lath 95 30 reliable and relatively damage-free preparation of samples con- boundaries. 96 31 taining GBs, but they have two major limitations when considering FIB–SEM microscopy ion-beam milling was performed using a FEI 97 32 preparation methods for materials containing mesoscale features. Strata DB-235 dual-beam FIB microscope equipped with an Omnip- 98 33 First, it does not permit TEM observation of the broader micro- robe Model 100.7 micromanipulator. The ion-beam column in this 99 34 structure in the vicinity of the tip to be prepared prior to shaping instrument permits Ga þ ion beam milling at either 5 or 30 kV. 100 35 the APT microtip. This step can be critical for targeting the correct Ar þ ion milling was also performed in a Fischione model 1010 ion 101 36 features in some materials, such as a martensitic steel with a small mill to further thin the lift-out specimen. TEM imaging was performed 102

37 lath size, which is discussed in this article and maraging-TRIP in a JEOL JEM-2100 with a LaB6 electron source, a standard JEOL 103 38 steels [39]. Second, as with most FIB–SEM microscope techniques, double-tilt specimen holder, and a Fischione Model 2050 on-axis 104 39 the GB must be in a region that is fairly close to the surface of the rotation tomography holder designed to hold wire specimens and 105 40 bulk sample. While this typically is not an issue, samples where permit 3601 rotation of a specimen. The APT analyses were conducted 106 41 the GBs have been etched or otherwise corroded require that the utilizing a LEAP 4000X-Si (Cameca, Madison, WI) at a base tempera- 107 42 APT microtip be extracted from below the surface of the bulk ture of 80 K, a pulse energy of 20–50 pJ, a pulse repetition rate 108 43 sample to avoid the region of the GB that has been attacked by the of 500 kHz, and an ion detection rate of 0.20 to 0.5 ion pulse1. 109 44 etchant. This instrument utilizes a local-electrode and laser pulsing with a 110 45 In this article, we present a methodology for analyzing mesoscale picosecond 355 nm wavelength ultraviolet laser, which enables the 111 46 features, such as GBs or heterophase interfaces, at depths of several analysis of insulating materials and minimizes specimen fracture [44]. 112 47 microns below a sample surface. The methodology involves targeting 113 48 the feature utilizing the ion channeling contrast generated by the 114 49 interaction of the ion beam with different crystal orientations, 3. Specimen preparation 115 50 followed by an iterative FIB–SEM microscope ion-beam milling and 116 51 TEM imaging process [34–36]. The FIB lift-out specimen produced is The specimen preparation technique can be separated into three 117 52 significantly larger than used in other approaches so that the atom distinct phases: (A) feature-specific lift-out and observation by 118 53 probe specimen can be formed from beneath the layer initially TEM; (B) transfer to an electropolished wire; and (C) a final 119 54 damaged by the ion channeling contrast imaging, as the Gaþ damage specimen sharpening procedure based on the TEM images and 120 55 is dependent only on the ion acceleration voltage and sample examination of the microtip in TEM. Specifically, the methodology 121 56 crystallography [34–36]. Our new approach permits either GBs or involves: (1) locating a boundary of interest using ion-channeling 122 57 heterophase interfaces, in this case prior austenite GBs spaced contrast; (2) extracting the specimen with the micromanipulator 123 58 approximately 10 mm apart, amidst the lath boundaries, to be located and attaching it to an Omniprobe grid; (3) thinning the sample until 124 59 precisely in the lift-out specimen prior to the formation of an APT it is electron transparent using a FIB–SEM microscope and Arþ ion 125 60 specimen. In this manner, the probability is increased of capturing milling; (4) imaging utilizing TEM to locate the mesoscale feature of 126 61 the desired mesoscale feature in the microtip of an APT specimen interest, here a prior austenite grain boundary; (5) transferring the 127 62 and readily identifying the correct boundary employing a single-tilt specimen onto an electropolished wire suitable for loading into the 128 63 TEM specimen holder. Samples prepared utilizing this methodology APT; (6) cutting the specimen, using ion-beam milling, based on the 129 64 display a high success rate in locating the desired GB in the analyzed TEM images to target the feature of interest; (7) annular milling to 130 65 region for analysis by TEM and subsequently studying it during an form an APT microtip of the desired location; and (8) observation of 131 66 APT analysis. the final APT microtip with TEM by capturing a TEM tilt series for 132

Please cite this article as: M.I. Hartshorne, et al., Specimen preparation for correlating transmission electron microscopy and atom probe tomography of mesoscale features, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.05.005i M.I. Hartshorne et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 3

1 direct correlation with the APT analyses, which involves character- time, 22 12 6 mm3 are then ion milled on either side of the Pt 67 2 izing the GB and its location in the APT microtip. The sample is cap with the bulk sample tilted normal to the Gaþ ion beam. 68 3 imaged in step 4 using a standard TEM sample holder, and in step The edges of the lift-out specimen are trimmed by milling a 69 4 7 using the Fischione on-axis holder. After this preparation protocol, 24 3 4 mm3 rectangular pattern on both sides of the specimen, 70 5 the microtip is transferred to the APT for analysis. tilting the sample 71.51 as needed to expose the area to be 71 6 trimmed, to produce parallel sides on the specimen. 72 7 3.1. Feature-specific lift-out The specimen undercuts are made at a 71 stage tilt angle, after 73 8 which the micromanipulator is inserted at 01 tilt and attached to 74 þ 9 The first step of preparation is to locate the region of interest the lift-out specimen before it is Ga ion milled free from the bulk 75 10 (ROI) using secondary electron imaging. After the area has been sample. This specimen is then transferred to an Omniprobe grid 76 11 located, the GB may be targeted by operating the ion beam and Pt-welded on each side. Thinning is accomplished with a 77 12 at a low brightness/high contrast setting to enhance the ion- pattern which mills each raster line of the pattern to full depth 78 13 channeling contrast. This method works well with martensitic before moving to the next raster line, located closer to the center 79 þ 14 steels, where it permits prior austenite GBs to be identified by of the specimen, using a 300 pA Ga ion-beam current. The 80 15 following the morphology of the prior austenite grains, Fig. 1a. Ion specimen is first tilted to 1.51 to mill the side not visible to the 81 16 imaging may not be required in materials that form a passivation SEM beam, then the final milling is completed for þ1.51, so that 82 17 layer because the grain structure may still be resolved by electron progress can be monitored using the SEM beam. The final thick- 83 18 channeling contrast in spite of, for example, the existence of a thin ness of the specimen is approximately 250–300 nm; below this 84 19 oxide layer. Gallium ion implantation for this method is less of thickness the martensitic steel samples begin to bend due to 85 20 concern than in other APT preparation techniques as the microtip residual stress and ion implantation [49,50], making later thinning 86 21 is formed from material that is several microns below the exposed and specimen manipulation more difficult. 87 22 surface of the sample. Transport of ions in matter (TRIM) [45] Following thinning the specimen is removed from the FIB–SEM 88 23 calculations for 30 kV Gaþ ions in pure Fe and the alloy discussed microscope and transferred to the TEM. Even though the specimen 89 24 this article yield a stopping range of approximately 11 nm, which is relatively thick for TEM, the martensitic lath structure is still 90 25 is significantly less than the 27 nm reported for Si, although ion resolvable using a large condenser aperture and a slightly con- 91 26 channeling and radiation-damage enhanced diffusion can increase densing the electron beam, Fig. 1c. The micrographs captured at 92 27 the implantation depth by up to 20 times [46–48]. Therefore, if ion this stage are saved for later use in the specimen sharpening step 93 28 channeling contrast is used gallium implantation into the microtip to target the prior austenite GB. The GB of interest is identified in 94 29 can be avoided by forming the microtip from a feature 300 nm or the TEM micrograph by tracing a dividing line between lath 95 30 more below the surface that was imaged. orientations, and is located in the FIB–SEM microscope using the 96 31 Once the ROI is located, a standard TEM sample is prepared edge contours of the TEM sample as a reference, Fig. 1d. 97 þ 32 perpendicular to the GB of interest. A nominally 3 20 1 mm3 Additionally, some of the specimens were subjected to Ar ion 98 33 thick Pt capping layer is deposited to protect the sample from Ga þ milling at this stage to thin the specimen further. The specimen 99 34 ion implantation, Fig. 1b. Stepped trenches, used to reduce milling shown in Fig. 4 underwent argon milling for a total of 10 min at 100 35 101 36 102 37 103 38 104 39 105 40 106 41 107 42 108 43 109 44 110 45 111 46 112 47 113 48 114 49 115 50 116 51 117 52 118 53 119 54 120 55 121 56 122 57 123 58 124 59 125 60 126 61 127 62 128 63 129 Fig. 1. (a) FIB–SEM microscope ion-beam image showing the microsctructure of the martensitic steel with selected prior austenite grain boundaries indicated by arrows. 64 130 (b) Pt capping layers for TEM samples placed perpendicular to grain boundaries. (c) TEM image of a thick lift-out, with a visible lath structure visible. (d) The same image, 65 contrast enhanced, with a section of the prior austenite grain boundary marked with a dotted line and the desired location for the APT sample marked with a white arrow. 131 66 White lines connect the region of interest with the edge topography. 132

Please cite this article as: M.I. Hartshorne, et al., Specimen preparation for correlating transmission electron microscopy and atom probe tomography of mesoscale features, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.05.005i 4 M.I. Hartshorne et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 5 kV, 10 min at 3 kV, and 8 min at 2 kV; all steps utilized 5 mA. prevent accidental exposure to Gaþ while shifting the Gaþ ion 67 2 We have found, however, that argon ion milling of the specimen beam back to the region to be milled. The first two cuts are made 68 3 should be avoided when possible while the specimen is in the TEM in this position, after which the Ga þ ion beam is shifted below the 69 4 lift-out specimen geometry due to potential damage to the plane of the specimen, then shifted to the edge of the post to finish 70 5 specimen. detaching the section containing the sample. 71 6 After the section of the grid containing the specimen of interest 72 7 is lifted out, the chamber is vented, and the Omniprobe lift-out 73 3.2. Transfer to electropolished wire 8 stage is removed and replaced with a custom fabricated stage that 74 9 securely holds a 1.02 mm diam. wire in either vertical or horizon- 75 After TEM imaging, the specimen is transferred back to the FIB– 10 tal positions. An approximately 20 mm long tungsten wire seg- 76 SEM microscope, and the Omniprobe is attached to the post of the 11 ment, sharpened by electropolishing in potassium hydroxide 77 Omniprobe grid, using electron-beam deposited Pt, on which the 12 [51–53], is placed in the stage in the horizontal position, and the 78 specimen is located. The section of the grid post attached to both 13 chamber is re-evacuated. The tip of the wire, now aligned 79 the specimen and Omniprobe is then cut free using the a 30 kV þ 14 þ perpendicular to the Ga ion beam, is milled flat in a region of 80 20 nA Ga ion-beam current, the largest aperture available on the 15 approximately the same diameter as the section of the grid post 81 DB-235 FIB–SEM instrument, to minimize the time needed to ion 16 that has been removed. The grid section containing the specimen 82 mill through the thickness of the grid post, Fig. 2a. 17 is placed flush with the wire tip and attached with Pt on both 83 During the steps described, the specimen is not imaged, as the 18 þ sides; first on one side, Fig. 2b, then after releasing the micro- 84 high Ga ion currents used to cut the grid can render the sample 19 manipulator, the stage is rotated 1801 and the Pt is deposited on 85 unusable due to Ga ion implantation into the sample along its 20 þ the opposite side of the wire/grid-section connection. The cham- 86 thinnest dimension. To prevent Ga ion implantation, the oppo- 21 ber is then vented, the wire moved from the horizontal to the 87 site side of the grid post from the specimen is initially used to 22 þ þ vertical position, and the chamber re-evacuated for the final 88 focus the Ga ion beam. The Ga ion beam is then shifted until 23 specimen sharpening stage, described below. 89 the Pt weld holding the specimen to the post is observable. The 24 90 micromanipulator is then moved to the grid post near the speci- 25 91 men weld using only the electron beam and attached with 3.3. Final specimen sharpening 26 þ 92 e -beam deposited Pt. The 20 nA Ga ion beam current aperture 27 þ 93 is selected and the Ga ion beam is again focused onto the far side The electron beam is focused on the specimen (the highest 28 94 of the post. The micromanipulator is used as a fiducial marker to point on the sample stage) for the final sharpening procedure, 29 95 30 96 31 97 32 98 33 99 34 100 35 101 36 102 37 103 38 104 39 105 40 106 41 107 42 108 43 109 44 110 45 111 46 112 47 Fig. 2. (a) FIB–SEM microscope image showing micromanipulator lifting out grid post section containing the specimen. (b) SEM image showing grid section on wire stub. 113 48 114 49 115 50 116 51 117 52 118 53 119 54 120 55 121 56 122 57 123 58 124 59 125 60 126 61 127 62 128 63 129 64 130 65 131 66 Fig. 3. (a) FIB image of TEM sample. (b) Rough-cut sample, with region of interest at the apex. 132

Please cite this article as: M.I. Hartshorne, et al., Specimen preparation for correlating transmission electron microscopy and atom probe tomography of mesoscale features, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.05.005i M.I. Hartshorne et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 5

1 after which the stage is rotated and tilted to accommodate any The TEM micrographs provide the requisite information for 67 2 small mis-alignments, such that the specimen sits in a precisely calibrating the APT reconstruction dimensionally, which can be 68 3 vertical position, with its broad side parallel to the electron-beam performed in the absence of crystallographic poles in the field- 69 4 and facing the Ga þ ion-beam. The 50 pA Gaþ ion beam is then desorption images. Reconstructions of the dataset with different 70 5 focused on the grid section, turned off, translated over the speci- reconstruction parameters were overlaid onto the TEM micro- 71 6 men and a single high-resolution image recorded, Fig. 3a. The graph using Adobe Photoshop; the difference in magnifications 72 7 edges of the specimen are then used as fiducial markers with between the TEM micrograph and the APT 3-D reconstruction was 73 8 which the TEM images are aligned with the FIB images, and the then calculated using ImageJ [63] (Appendix B). The dataset 74 9 specimen is Ga þ ion beam milled into a wedge shape with the ROI displayed, Fig. 4, contains both the GB and a vanadium carbide 75 10 just below the apex of the wedge, Fig. 3b. precipitate, which serve as fiducial markers for calibration pur- 76 11 Pt is electron-beam deposited on the tip of the wedge in poses. To minimize errors due to a possible relative rotation of 77 12 preparation for the final thinning step, after which the specimen the APT 3-D reconstruction about its z-axis relative to the TEM 78 13 is tilted so that the Gaþ ion beam is parallel with the specimen’s micrograph, only the z-axis component of distances between 79 14 long axis. An annular milling pattern is utilized to form the wedge points are used to evaluate the fit and serve as the best possible 80 15 into an APT specimen using a 0.50 μm inner radius pattern at reference points between the TEM image and APT reconstruction, 81 16 300 pA beam current followed by a 0.15 μm inner radius pattern at as the z-axis aligns closely with the axis of rotation of the TEM 82 17 50 pA beam current. A final step is performed at 5 kV, the lowest holder. 83 18 Ga þ ion acceleration voltage available in the DB235, using 3 μm Based on this calculation, an evaporation field of 29.5 V nm1 84 19 circular pattern and a slightly defocused ion beam. This step is at an estimated field factor of k¼3.3 and an image compression 85 20 performed to remove as much of the radiation region damaged by factor of 1.65 provides the best fit along the z-axis, while main- 86 21 the 30 kV Gaþ ions as possible [27], to remove the remainder of the taining an appropriate angle of the GB with respect to the 87 22 Pt cover to prevent specimen fracture [54], and if required to erode principal axes of the 3-D reconstruction. The evaporation field, 88 23 slightly the microtip to position precisely the GB. The defocused 29.5 V nm1, is 16% smaller than the empirical value for iron of 89 24 Ga þ ion-beam with a circular pattern is utilized because it has been 35 V nm1, [9–11,64]; the recorded base temperature of 80 K is, 90 25 found to maintain the shape of the microtip, similar to in broad however, lower than the actual specimen temperatures during UV 91 26 beam ion milling of surfaces [55], while a focused beam has been laser-assisted field evaporation due to heating by the laser beam 92 27 found to introduce asymmetries while milling at 5 kV in the DB-235 (Appendix C). This lower evaporation field based on elevated 93 28 FIB–SEM microscope. A bright-field TEM tomographic tilt series is temperature is consistent with the results of Wada and Arslan 94 29 recorded after the final ion-milling step, two TEM images are et al. [65,66]. Finally, the calibration methodology herein is only an 95 30 displayed in Fig. 4e and f, using the Fischione on-axis rotation estimate, while other methods based on field-desorption images 96 31 tomography specimen holder to load the electropolished wires on provide greater precision [66–70]. 97 32 which the APT specimens sit. The Gaþ implantation into the analyzed volume was minimal in 98 33 this procedure, Fig. 4c, d and h, i, with no Ga concentration along the 99 34 GB as occurred with Ar, Fig. 4b. This difference was likely due to the 100 35 4. Results and discussion fact that only three images were recorded at 50 pA with the Gaþ ion 101 36 beam at a significant angle to the thinnest part of the sample, whereas 102 37 4.1. Analysis of PremoMet steel thesampleunderwent28minofArþ ion milling, even though the 103 38 projected stopping range for the kV Ar þ is only 4.0 nm as previously 104 39 Atotaloffive APT specimens of PremoMet steel have been described. The majority of the Ga implantation is confined to a small 105 40 prepared using this method, from one of which 49 million ions were region on one side of the apex of the microtip, Fig. 4candd.The 106 41 collected by APT, without the specimen fracturing (Appendix A). This dotted lines in Fig. 4d correspond to the direction of the concentration 107 42 specimen is presented in Fig. 4, with reconstruction containing profiles for Ga, Fig. 4g and h, which indicate a maximum Ga 108 43 carbon shown in (a), argon in (b), gallium in (c) and (d), TEM images concentration of 3.68 at%; this is greater than the maximum of 2 at% 109 44 at two tilts shown in (e) and (f), gallium concentration profiles in (f) found by Thompson et al. [27] using their site-specific technique; it 110 45 and (g) and concentration of carbon, argon and gallium across the decreases, however, to o2at%at3nmand o1at%at8nmfromthe 111 46 prior austenite grain boundary in (i). The carbon and argon segrega- edge of the dataset. The total Ga concentration in the dataset is 112 47 tion displayed in Fig. 4a clearly highlights the prior austenite GB 0.036 at%, and the concentration in the vertical cylinder in the center 113 48 displayed in Fig. 4e and f. Additionally, Fig. 4b indicates that the Ar þ of the dataset, Fig. 4d, is 0.027 at%. TRIM calculations yield a stopping 114 49 ions from the explorative ion-milling penetrated deep within the range in iron of 10.9 nm (skewness¼0.68, straggle¼5.3 nm, 115 50 approximately 300 nm thick sample at the 2–5 kV accelerating kurtosis¼2.59) for 30 kV Gaþ . The target is, however, amorphous in 116 51 voltages, even though the TRIM calculations yield a stopping range the TRIM calculations; ion channeling in a crystalline materials results 117 52 in iron of 4.0 nm (skewness¼0.63, straggle¼2.2 nm, kurtosis¼3.22) in much greater penetrations. Materials with a smaller atomic number, 118 53 for 5 kV Arþ .Thisislikelyduetotheincreasedacceptanceanglefor such as silicon, also have a larger stopping range, which increases the 119 54 ion channeling at low energies followed by diffusion to the grain probability that Ga þ ions will penetrate to the region or feature of 120 55 boundary [56–59] although other explanations are possible, such as interest. Therefore, great care must be taken to minimize Ga þ ion 121 56 species specific enhanced grain boundary diffusion similar to Ga in exposure, even though the region calculated to be most heavily 122 57 the aluminum alloy system [60,61].Thebright-field TEM micro- damaged is removed during the annular milling step, as the sample 123 58 graphs in Fig. 4eandfarefromtheTEMtiltseries,displayingtwo is about 300 nm thick when the image during the final sharpening 124 59 tilts with the grains on either side of the prior austenite GB, which procedure is recorded. 125 60 highlights the curved nature of the prior austenite GB between it and 126 61 the adjoining grain. Fig. 4g is an APT 1-D concentration profile 127 62 obtained using an analysis cylinder placed perpendicular to this GB, 4.2. Proposed combinations with other analysis methods 128 63 which demonstrates that the carbon is enriched to 10.75 at%, while 129 64 the argon concentration along the GB from the ion-milling step is As demonstrated in the preparation section, our procedure can 130 65 0.56 at%. Further information on the composition of the prior be used successfully in conjunction with other characterization 131 66 austenite boundary show will be provided in a future article [62]. techniques, such as TEM. Since the specimen is kept relatively 132

Please cite this article as: M.I. Hartshorne, et al., Specimen preparation for correlating transmission electron microscopy and atom probe tomography of mesoscale features, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.05.005i 6 M.I. Hartshorne et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 67 2 68 3 69 4 70 5 71 6 72 7 73 8 74 9 75 10 76 11 77 12 78 13 79 14 80 15 81 16 82 17 83 18 84 19 85 20 86 21 87 22 88 23 89 24 90 25 91 26 92 27 93 28 94 29 95 30 96 Fig. 4. (a) Carbon segregation to prior austenite grain boundary and at a vanadium carbide precipitate. (b) Argon channeling along grain boundary from ion milling. ((c) and 97 31 (d)) Gallium implantation into sample from dual-beam FIB microscope, dotted lines denoting regions for concentration profiles ((h) and (i)). ((e) and (f)) TEM images of APT 32 specimen before evaporation, clearly showing the location of the prior austenite grain boundary. (g) C (black), Ar (pink) and Ga (orange) concentration profiles across the 98 33 grain boundary (cylindrical volume delineated in red near the top of the reconstruction in (a)). (h) Ga concentration profile horizontally across the top of tip as shown in (d). 99 34 (i) Ga concentration profile vertically down the axis of the tip as shown in (d). (For interpretation of the references to color in this figure legend, the reader is referred to the 100 35 web version of this article.) 101 36 102 37 thick to prevent bending, while on the Omniprobe grid, electron 4.3. Comparison to current techniques 103 38 backscattering diffraction (EBSD) may also be employed to analyze 104 39 the orientation relationships within the specimen, especially if the Our results demonstrate that this specimen preparation tech- 105 40 FIB–SEM microscope utilized is equipped with an EBSD system. nique provides a precise method for targeting mesoscale features 106 41 Orientation imaging microscopy (OIM) in the TEM could also be in the 5–40 mm size range located several microns below the 107 42 used in conjunction with this method, but will be significantly sample surface. The original pulsed electropolishing techniques 108 43 more difficult due to the specimen remaining thicker than is provided a methodology to locate features of interest within an 109 44 optimal for this technique. These techniques can be used to APT specimen volume; they function, however, as an iterative 110 45 characterize the orientation relationship at the grain boundary selection tool, where the sample is polished until a feature of 111 46 selected for APT analysis [71]. interest is located and positioned in the volume of the microtip to 112 47 EBSD may be performed on the bulk sample surface before the be analyzed [19,20,22]. As discussed, these methods require multi- 113 48 lift-out process described herein, to select a grain boundary or ple repetitive polishing steps, each of which is accompanied by 114 49 heterophase interface, as an alternative to using ion channeling extensive TEM diffraction analyses to locate the mesoscale feature 115 50 contrast [71]. This variation of the outlined methodology could be of interest, especially so in the case of martensitic steels due to the 116 51 used when ion channeling is of great concern, such as during the large number of lath boundaries contained within each grain. 117 52 lift-out of GBs in a corroded sample; we have shown, however, While these methods have proven to be successful, they do not 118 53 using APT analyses that Ga implantation during the initial step of provide the most convenient route for preparing site-specific 119 54 this technique is generally not an issue due to the depth from samples by current standards. 120 55 which the sample can be taken, even though TRIM calculations are FIB–SEM microscope based techniques, Kelly et al. [72] and 121 56 known to understate the penetration of gallium due to ion Thompson et al. [27], provide methods to lift-out areas containing 122 57 channeling and radiation assisted diffusion. specific GBs, and the newer method described by Felfer et al. [36] 123 58 The specimen preparation on a sharpened wire suitable for a integrates TEM into the FIB–SEM microscope preparation proce- 124 59 TEM on-axis tomography holder is important since it allows a dure. These methods, however, do not provide information on 125 60 microtip to be easily manipulated, while minimizing the possibi- where the feature is located below the sample’ssurface.Byfirst 126 61 lity of mechanical damage during handling. This specimen geo- lifting out a relatively large TEM sample, the method demonstrated 127 62 metry also facilitates examination of a specimen after preparation in this article permits a mesoscale feature to be precisely selected 128 63 by recording a TEM tilt series to confirm the presence and location from its surrounding microscale features at a depth of several 129 64 of the mesoscale feature of interest, and in principle permits the microns below the surface. This new capability is useful for several 130 65 creation of an electron tomographic reconstruction if a suitable problems. First, it permits the use of ion-channeling contrast, as the 131 66 contrast mechanism is available [66]. sample can be extracted at depths greater than the maximum range 132

Please cite this article as: M.I. Hartshorne, et al., Specimen preparation for correlating transmission electron microscopy and atom probe tomography of mesoscale features, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.05.005i M.I. Hartshorne et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 7

1 of an implanted Ga atom. Second, specimens can be prepared from References 67 2 features such as crack tips without the crack or pre-crack oxide 68 3 running through the rest of the sample as in other methods [73,74]. [1] E.W. Müller, The atom-probe field ion microscope, Naturwissenschaften 57 69 4 Moreover, these techniques suggest preparation from a plan view of (1970) 222–320. 70 ̈ 5 the crack, which requires a location that is only partially corroded or [2] E.W. Muller, J.A. Panitz, S.B. McLane, The Atom-Probe Field Ion Microscope, 71 Rev. Sci. Instrum. 39 (1968) 83. 6 cracked to ensure that an APT specimen can be fabricated from near [3] S.S. Brenner, J.T. McKinney, Construction and performance of an FIM-atom 72 7 the crack’s tip. By preparing the TEM specimen in a cross-sectional probe, Surf. Sci. 23 (1970) 88–111. 73 8 geometry across the crack, the method demonstrated herein [4] A. Cerezo, T.J. Godfrey, G.D.W. Smith, Application of a position-sensitive 74 detector to atom probe microanalysis, Rev. Sci. Instrum. 59 (1988) 862–866. 9 provides a route for targeting the region immediately below the [5] D. Blavette, B. Deconihout, A. Bostel, J.M. Sarrau, M. Bouet, A. Menand, The 75 10 crack tip instead of a section of a GB or interface that, while next to tomographic atom probe: a quantitative three-dimensional nanoanalytical 76 11 a section that was corroded, was not itself attacked. instrument on an atomic scale, Rev. Sci. Instrum. 64 (1993) 2911–2919. 77 [6] T.F. Kelly, M.K. Miller, Invited review article: atom probe tomography, Rev. Sci. 12 Instrum. 78 (2007) 031101. 78 13 [7] D.N. Seidman, Three-dimensional atom-probe tomography: advances and 79 – 14 5. Conclusions applications, Annu. Rev. Mater. Res. 37 (2007) 127 158. 80 [8] D.N. Seidman, K. Stiller, An atom-probe tomography primer, MRS Bull 34 15 (2009) 717–724. 81 16 A new method for the preparation of atom-probe tomograph [9] M.K. Miller, A. Cerezo, M.G. Hetherington, G.D.W. Smith, Atom Probe Field Ion 82 17 Microscopy, Clarendon Press, Oxford, UK, 1996. 83 microtips containing mesoscale features for correlative investi- — 18 [10] M.K. Miller, Atom Probe Tomography Analysis at the Atomic Level, Springer 84 gations employing atom-probe tomography (APT), transmission ScienceþBusiness Media, LLC, New York, NY, 2000. 19 electron microscopy (TEM) and additional auxiliary methods is [11] B. Gault, M.P. Moody, J.M. Cairney, S.P. Ringer, Atom Probe Microscopy, 85 20 demonstrated. Springer ScienceþBusiness Media, LLC, New York, NY, 2012. 86 21 [12] W.T. Nachtrab, Y.T. Chou, Grain boundary segregation of copper, tin and 87 This preparation technique is applied to the study of prior- antimony in C–Mn steels at 900 1C, J. Mater. Sci. 19 (1984) 2136–2144. 22 austenite grain boundaries (GB) in PremoMet, a commercial [13] N.S. Lim, C.W. Bang, S. Das, H.W. Jin, R. Ayer, C.G. Park, Influence of tempering 88 23 martensitic ultrahigh strength steel alloy, and the specimens temperature on both the microstructural evolution and elemental distribution 89 24 fi in AISI 4340 steels, Met. Mater. Int. 18 (2012) 87–94. 90 created with this technique are shown to eld-evaporate [14] D. Isheim, R.P. Kolli, M.E. Fine, D.N. Seidman, An atom-probe tomographic 25 smoothly and produce useful APT and TEM datasets. study of the temporal evolution of the nanostructure of Fe–Cu based high- 91 26 Lift-out specimens are prepared perpendicular to the grain strength low-carbon steels, Scr. Mater. 55 (2006) 35–40. 92 27 [15] P. Kritzer, Corrosion in high-temperature and supercritical water and aqueous 93 boundary, analyzed in TEM, and a microtip containing the GB is – 28 solutions: a review, J. Supercrit. Fluids 29 (2004) 1 29. 94 prepared based on the TEM analysis. [16] M.W.A. Rashid, M. Gakim, Z.M. Rosli, M.A. Azam, Formation of Cr23C6 during 29 The APT microtips produced by this method are well suited for the sensitization of AISI 304 stainless steel and its effect to pitting corrosion, 95 – 30 analysis using an on-axis tomography holder in a TEM. Calibra- Int. J. Electrochem. Sci. 7 (2012) 9465 9477. 96 31 [17] N. Parvathavarthini, S. Mulki, R.K. Dayal, I. Samajdar, K.V. Mani, B. Raj, 97 tion of the APT dataset using the TEM images is demonstrated. Sensitization control in AISI 316L(N) austenitic stainless steel: defining the þ 32 The need to avoid extended Ar ion milling, often used for role of the nature of grain boundary, Corros. Sci. 51 (2009) 2144–2150. 98 33 cleaning TEM foils, is also discussed, as the 3-D reconstructed [18] H.-C. Lee, S.-H. Mun, D. McKenzie, Electron microscopy study on the grain- 99 34 boundary precipitation during aging of Fe–10Ni–5Mn steel, Metall. Mater. 100 APT dataset clearly shows argon concentrated along the grain Trans. A 34 (2003) 2421–2428. 35 boundary. [19] A. Henjered, H. Norden, A controlled specimen preparation technique for 101 36 This technique permits the preparation of specimens contain- interface studies with atom-probe field-ion microscopy, J. Phys. E: Sci. 102 – 37 ing mesoscale features, such as GBs, from materials in which Instrum. 16 (1983) 617 619. 103 38 [20] B.W. Krakauer, J.G. Hu, S.M. Kuo, R.L. Mallick, A. Seki, D.N. Seidman, et al., A 104 their spacing and the proximity of other features renders both system for systematically preparing atom-probe field-ion-microscope speci- 39 electropolishing and traditional lift-out techniques difficult. mens for the study of internal interfaces, Rev. Sci. Instrum. 61 (1990) 105 – 40 Potential future applications include the study of specific grain 3390 3398. 106 41 [21] L. Karlsson, H. Nordén, Overview no. 63 Non-equilibrium grain boundary 107 boundaries that have experienced corrosive attack, hydrogen segregation of boron in austenitic stainless steel-II. Fine scale segregation 42 embrittlement or irradiation, as well as semiconductor device behaviour, Acta Metall. 36 (1988) 13–24. 108 43 structures and interfaces. [22] B.W. Krakauer, D.N. Seidman, Systematic procedures for atom‐probe field‐ion 109 44 microscopy studies of grain boundary segregation, Rev. Sci. Instrum. 63 (1992) 110 4071–4079. 45 [23] B.W. Krakauer, D.N. Seidman, Subnanometer scale study of segregation at 111 46 grain boundaries in an Fe(Si) alloy, Acta Mater. 46 (1998) 6145–6161. 112 47 [24] B.W. Krakauer, D.N. Seidman, Distribution of grain boundaries in an Fe(Si) 113 Acknowledgements alloy, Interface Sci. 40 (2000) 27–40. 48 [25] B.W. Krakauer, D.N. Seidman, Absolute atomic-scale measurements of the 114 49 Gibbsian interfacial excess of solute at internal interfaces, Phys. Rev. B: 115 50 MLT and MIH acknowledge funding support from the Carpenter Condens. Matter 48 (1993) 6724–6728. 116 fi 51 Q2 Technology Corporation and the National Science Foundation’s [26] K. Alexander, P. Angelini, M. Miller, Precision ion milling of eld-ion speci- 117 Faculty Early Career Development Program (grant 1150807). APT mens, J. Phys. Colloques 50 (1989) 549. 52 [27] K. Thompson, D. Lawrence, D.J. Larson, J.D. Olson, T.F. Kelly, B. Gorman, in situ 118 53 experiments conducted at NUCAPT, a shared research facility at site-specific specimen preparation for atom probe tomography, Ultramicro- 119 54 the Materials Research Center of Northwestern University, were scopy 107 (2007) 131–139. 120 partially supported by the NSF-MRSEC program (DMR-0520513). [28] M.K. Miller, K.F. Russell, Atom probe specimen preparation with a dual beam 55 SEM/FIB miller, Ultramicroscopy 107 (2007) 761–766. 121 56 The LEAP tomograph at NUCAPT was purchased and upgraded [29] M.L. Taheri, J.T. Sebastian, B.W. Reed, D.N. Seidman, A.D. Rollett, Site-specific 122 57 with funding from NSF-MRI (DMR-0420532) and ONR-DURIP atomic scale analysis of solute segregation to a coincidence site lattice grain 123 (N00014-0400798, N00014-0610539, N00014-0910781) grants. boundary, Ultramicroscopy 110 (2010) 278–284. 58 [30] B. Gault, E.A. Marquis, D.W. Saxey, G.M. Hughes, D. Mangelinck, E.S. Toberer, 124 59 Additional instrumentation at NUCAPT was supported by the et al., High-resolution nanostructural investigation of Zn4Sb3 alloys, Scr. 125 60 Initiative for Sustainability and Energy at Northwestern (ISEN). Mater. 63 (2010) 784–787. 126 61 [31] E.A. Marquis, R. Hu, T. Rousseau, A systematic approach for the study of 127 radiation-induced segregation/depletion at grain boundaries in steels, J. Nucl. 62 Mater. 413 (2011) 1–4. 128 63 Appendix A. Supplementary information [32] H.O. Colijn, T.F. Kelly, R.M. Ulfig, R.G. Buchheit,Site-specific FIB preparation of 129 64 atom probe samples, in: Microscopy and Microanalysis, Microscopy Society of 130 America, 2004, pp. 1150–1152. 65 Supplementary data associated with this article can be found in [33] ASTM Standard E112-10: Standard Test Methods for Determining Average 131 66 theonlineversionathttp://dx.doi.org/10.1016/j.ultramic.2014.05.005. Grain Size, ASTM International, West Conshohocken, PA, 1996. 132

Please cite this article as: M.I. Hartshorne, et al., Specimen preparation for correlating transmission electron microscopy and atom probe tomography of mesoscale features, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.05.005i 8 M.I. Hartshorne et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 [34] J.M. Cairney, D.W. Saxey, D. McGrouther, S.P. Ringer, Site-specific specimen [54] S. Kölling, W. Vandervorst, Failure mechanisms of silicon-based atom-probe 44 2 preparation for atom probe tomography of grain boundaries, Physica B 394 tips, Ultramicroscopy 109 (2009) 486–491. 45 (2007) 267–269. [55] P. Sigmund, A mechanism of surface micro-roughening by ion bombardment, 3 [35] P. Felfer, S.P. Ringer, J.M. Cairney, Shaping the lens of the atom probe: J. Mater. Sci. 8 (1973) 1545–1553. 46 4 fabrication of site specific, oriented specimens and application to grain [56] J. Lindhard, Influence of crystal lattice on motion of energetic charged 47 5 boundary analysis, Ultramicroscopy 111 (2011) 435–439. particles, Munksgaard (1965). 48 [57] G. Ayrault, R.S. Averback, D.N. Seidman, A new approach for the study of 6 [36] P.J. Felfer, T. Alam, S.P. Ringer, J.M. Cairney, A reproducible method for 49 damage-free site-specific preparation of atom probe tips from interfaces, transmission sputtering, Scr. Metall. 12 (1978) 119–123. 7 Microsc. Res. Tech. 75 (2012) 484–491. [58] G. Ayrault, D.N. Seidman, Transmission sputtering of gold thin films by low- 50 8 [37] T. Alam, P.J. Felfer, M. Chaturvedi, L.T. Stephenson, M.R. Kilburn, J.M. Cairney, energy (o1 keV) xenon ions. I. The system and the measurement, J. Appl. 51 Segregation of B, P, and C in the Ni-based superalloy, inconel 718, Metall. Phys. 53 (1982) 6968–6978. 9 — 52 Mater. Trans. A 43 (2012) 2183–2191. [59] H. Trinkaus, B.N. Singh, Helium accumulation in metals during irradiation – 10 [38] D. Tytko, P.-P. Choi, J. Klöwer, A. Kostka, G. Inden, D. Raabe, Microstructural where do we stand? J. Nucl. Mater. 323 (2003) 229 242. 53 11 evolution of a Ni-based superalloy (617B) at 700 1C studied by electron [60] W. Ludwig, D. Bellet, Penetration of liquid gallium into the grain boundaries of 54 12 microscopy and atom probe tomography, Acta Mater. 60 (2012) 1731–1740. aluminium: a synchrotron radiation microtomographic investigation, Mater. 55 Sci. Eng., A 281 (2000) 198–203. 13 [39] O. Dmitrieva, D. Ponge, G. Inden, J. Millán, P. Choi, J. Sietsma, et al., Chemical 56 gradients across phase boundaries between martensite and austenite in steel [61] R. Hugo, R. Hoagland, The kinetics of gallium penetration into aluminum grain — 14 studied by atom probe tomography and simulation, Acta Mater. 59 (2011) boundaries in situ TEM observations and atomistic models, Acta Mater. 48 57 – 15 364–374. (2000) 1949 1957. 58 [40] Z. Jiao, G.S. Was, Segregation behavior in proton- and heavy-ion-irradiated [62] M.I. Hartshorne, C. McCormick, M. Schmidt, P. Novotny, D. Isheim, D.N. 16 Seidman, et al., Analysis of a new high toughness ultra high strength 59 ferritic–martensitic alloys, Acta Mater. 59 (2011) 4467–4481. martensitic steel by transmission electron microscopy and atom probe 17 [41] M.K. Miller, Y. Zhang, Fabrication and characterization of APT specimens from 60 tomography, Metall. Mater. Trans. A To be submitted. Q3 18 high dose heavy ion irradiated materials, Ultramicroscopy 111 (2011) 672–675. 61 [63] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, NIH Image to ImageJ: 25 years of [42] K. Inoue, F. Yano, A. Nishida, H. Takamizawa, T. Tsunomura, Y. Nagai, et al., 19 image analysis, Nat. Methods 9 (2012) 671–675. 62 Dopant distributions in n-MOSFET structure observed by atom probe tomo- 20 [64] T.T. Tsong, Atom-Probe Field Ion Microscopy, Cambridge University Press, 63 graphy, Ultramicroscopy 109 (2009) 1479–1484. 2005. 21 [43] T.F. Kelly, D.J. Larson, K. Thompson, R.L. Alvis, J.H. Bunton, J.D. Olson, et al., 64 [65] M. Wada, On the thermally activated field evaporation of surface atoms, J. Atom probe tomography of electronic materials, Annu. Rev. Mater. Res. 37 65 22 Phys. 45 (1984) 89–94. (2007) 681–727. 23 [66] I. Arslan, E.A. Marquis, M. Homer, M.A. Hekmaty, N.C. Bartelt, Towards better 66 [44] J.H. Bunton, J.D. Olson, D.R. Lenz, T.F. Kelly, Advances in pulsed-laser atom 24 3-D reconstructions by combining electron tomography and atom-probe 67 probe: instrument and specimen design for optimum performance, Microsc. tomography, Ultramicroscopy 108 (2008) 1579–1585. 25 – 68 Microanal. 13 (2007) 418 427. [67] L. Yao, M.P. Moody, J.M. Cairney, D. Haley, A.V. Ceguerra, C. Zhu, et al., 26 [45] J.F. Ziegler, J. Biersack, M. Ziegler, Stopping Range of Ions in Matter (2013). Crystallographic structural analysis in atom probe microscopy via 3D hough 69 27 [46] F. Laruelle, A. Bagchi, M. Tsuchiya, J. Merz, P.M. Petroff, Focused ion beam transformation, Ultramicroscopy 111 (2011) 458–463. 70 channeling effects and ultimate sizes of GaAlAs/GaAs nanostructures, Appl. [68] V.J. Araullo-Peters, B. Gault, S.L. Shrestha, L. Yao, M.P. Moody, S.P. Ringer, et al., 28 – 71 Phys. Lett. 56 (1990) 1561 1563. Atom probe crystallography: atomic-scale 3-D orientation mapping, Scr. 29 [47] N.G. Stoffel, S.A. Schwarz, M.A.A. Pudensi, K. Kash, L.T. Florez, J.P. Harbison, Mater. 66 (2012) 907–910. 72 30 et al., Evidence for heavy ion channeling in AlGaAs at low energies, Appl. Phys. [69] B. Gault, F. de Geuser, L.T. Stephenson, M.P. Moody, B.C. Muddle, S.P. Ringer, 73 Lett. 60 (1992) 1603–1605. 31 Estimation of the Reconstruction Parameters for Atom Probe Tomography, 74 [48] S.M. Myers, Ion-beam-induced migration and its effect on concentration Microsc. Microanal. 14 (2008) 296–305. 32 profiles, Nucl. Instrum. Methods 168 (1980) 265–274. [70] B. Gault, D. Haley, F. de Geuser, M.P. Moody, E.A. Marquis, D.J. Larson, et al., 75 33 [49] L. Frey, C. Lehrer, H. Ryssel, Nanoscale effects in focused ion beam processing, Advances in the reconstruction of atom probe tomography data, Ultramicro- 76 – 34 Appl. Phys. A 76 (2003) 1017 1023. scopy 111 (2011) 448–457. 77 [50] C. Borschel, S. Spindler, D. Lerose, A. Bochmann, S.H. Christiansen, S. Nietzsche, [71] S.I. Baik, M.J. Olszta, S.M. Bruemmer, D.N. Seidman, Grain-boundary structure 35 et al., Permanent bending and alignment of ZnO nanowires, Nanotechnology and segregation behavior in a nickel-base stainless alloy, Scr. Mater. 66 (2012) 78 36 22 (2011) 185307. 809–812. 79 ̃ 37 [51] A.I. Oliva, A. Romero G., J.L. Pena, E. Anguiano, M. Aguilar, Electrochemical [72] T. Kelly, R. Martens, S. Goodman, Methods of Sampling Specimens for 80 preparation of tungsten tips for a scanning tunneling microscope, Rev. Sci. Microanalysis, US Pat. 6,700,121., 2004. 38 Instrum. 67 (1996) 1917. [73] A. Cerezo, P.H. Clifton, M.J. Galtrey, C.J. Humphreys, T.F. Kelly, D.J. Larson, et al., 81 39 [52] D. Hubel, Tungsten microelectrode for recording from single units, Science 125 Atom probe tomography today, Mater. Today 10 (2007) 36–42. 82 40 (1957) 549–550 (80-. ). [74] S. Lozano-Perez, A guide on FIB preparation of samples containing stress 83 41 [53] J. Fasth, B. Loberg, H. Norden, Preparation of contamination-free tungsten corrosion crack tips for TEM and atom-probe analysis, Micron 39 (2008) 84 specimens for the field-ion microscope, J. Sci. Instrum. 44 (1967) 1044–1045. 320–328. 42 85 43

Please cite this article as: M.I. Hartshorne, et al., Specimen preparation for correlating transmission electron microscopy and atom probe tomography of mesoscale features, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.05.005i