Fast, wafer-scale growth of a nanometer-thick graphite film on Ni foil and its structural analysis

A 25 µm Ni foil (99.5% purity, Goodfellow) was loaded into a commercial CVD reactor (4" BMPro Aixtron, figure SI1(a) (available online at stacks.iop.org/NANO/31/485605/mmedia)). The parameters profile is schematized in figure SI1(b). The system was purged with Ar and evacuated to 10^–3 mbar. The Ni foil was then heated to 900 °C with a ramp rate of 75 °C min^–1 in Ar/H₂ (500/1000 sccm) while a chamber pressure of 10 mbar was maintained [20]. After the pre-annealing step of the Ni foil (5–30 min), the NGF was deposited by exposure of the foil to a CH₄/H₂ flow (100 sccm each) [20] at 10–500 mbar in the temperature range of 700 °C–900 °C and, for 1–45 min. Then, the sample was cooled down at a rate of 40 °C min^–1, via an Ar flow (4000 sccm) [20]. The as-grown NGFs were transferred on SiO₂/Si substrate and Cu grid with lacey carbon film by the use of an improved wet-chemical process, as described elsewhere [30, 34, 35].

Scanning electron microscopy (SEM) was performed by a Nano FEI Nova and a Zeiss Merlin (both at 1–5 kV, 50 pA). The sample surface roughness was measured via atomic force microscopy (AFM, Dimension Icon SPM, Bruker). For transmission electron microscopy (TEM) investigation, the cross-section lamella was prepared with a FEI Helios G4 focused ion beam (FIB) SEM microscope, equipped with a field emission gun and a Leica cryo-holder stage. To protect the region of interest, we coated the sample with a 5 nm layer of Ir and 1.5 µm of Pt. Then, the milling procedure was carried out, with the cryo-stage set at −145 °C and by the use of an ion beam of 16 kV (to reduce sample damage and artifacts formation) with a current range of 10 nA to 12 pA. Once completed, the thin lamella was slowly warmed up to room temperature. The cross-section TEM analysis was performed via a FEI Titan 80–300 Cubed microscope equipped with a high-brilliance field emission gun (300 kV), a Wien-type FEI monochromator and a CEOS spherical aberration corrector (for the objective lens) to allow a final spatial resolution of 0.09 nm. High-resolution TEM (HRTEM) images were acquired in low-dose conditions (10–20 electrons/Ų/second) with a Gatan K2 IS direct detection camera. A setting of a low-dose beam condition was necessary to preserve the fine structure of the sample. For the graphene layers to be imaged, the Ni crystallites of the foil were oriented along a proper zone axis to have the Ni (1 1 1)//NGF (0 0 0 2) interface parallel to the electron beam. Plan-view TEM and selective area electron diffraction (SAED) measurements were performed on NGF transferred on the Cu grid with lacey carbon film.

Crystal orientation analysis of an annealed Ni foil surface was performed by electron backscattered spectroscopy (EBSD) using a FEI Quanta 600 SEM (at 20 kV). The Ni foils (as-received, annealed, and with NGF) were also analyzed with x-ray diffraction (XRD). The XRD patterns were obtained by a powder diffractometer (Brucker, D2 phaser with LYNXEYE detector) with a Cu radiation source.

Raman spectroscopy (Horiba LabRam-Amicus) was performed by a 473 nm laser with low excitation power (25%) to avoid heat-induced effects and 100x objective (lateral resolution 0.5 µm). The Raman mapping was performed with a 0.5 µm step (as the laser spot size is 0.5 µm) over an area of 8 × 37.5 µm².

Producing NGF on polycrystalline Ni requires crucial control of several process parameters. After several preliminary experiments and also considering the process parameters for graphene growth on Cu/Ni and Ni film substrates, from our previous work [20, 30], we grew the NGFs by keeping constant the following parameters: (i) reactor pressure of 10 mbar during the pre-annealing step, (ii) sample temperature of 700 ºC –900 ºC during the pre-annealing and growth steps, (iii) Ar/H₂ gas flow of 1000/1000 sccm during the pre-annealing, and (iv) Ar/H₂/CH₄ flow of 1000/100/100 sccm during the NGF growth step. The effect of other key parameters was studied in more detail (table 1): (i) Ni foil pre-annealing time (A_t, 5–30 min), (ii) chamber pressure during growth (P_g, 10–500 mbar), and (iii) growth time (G_t, 5–45 min). The morphological features of four representative NGF samples (listed in table 1) are shown in figure 1. The low-magnification SEM image (figure 1(a)) of NiAG1 sample shows the presence of bright and dark regions (a photo of it is presented in the inset of figure 1(a)). According to the literature, the change in shades on the SEM image corresponds to variations in the graphene film thickness [33]. As we discuss later, indeed, the bright regions consist of a few graphene layers (FLG), whereas the dark dendric-shaped areas are composed of multilayers (confirmed by Raman spectroscopy point measurements). Similar kind of dendric-shaped graphene growth, where samples were treated at lower chamber pressures, was observed previously [18, 20, 23, 30]. In the corresponding higher magnification image of NiAG1 (figure 1(b), inset), 1–10 µm grains on the Ni foil were observed in the bright regions. These grains were larger (2–20 µm) for the 15 min pre-annealed Ni foil (NiAG2 sample), as shown in figure 1(d). Similar to NiAG1, NiAG2 samples also resulted in growing non-uniform NGF (figure 1(c)). In the NiAG3 sample (G_t = 45 min), more surface coverage of NGF (dark areas) was observed (figure 1(e)) than in the former samples. This result indicates that longer annealing or longer growth times will not produce uniform NGF at 10 mbar chamber pressure during growth. Additional non-uniform NGF growth, resulting from process parameters in the given range (refer section 2.1) are shown in figure SI2. For the NiAG4 sample (annealing conditions same as the NiAG1 sample), the higher reactor pressure of 500 mbar during the growth led to a much more uniform NGF (figures 1(g) and (h)) than the previous samples. When P_g < 500 mbar or A_t > 5 min, no uniform growth (as that of NiAG4) was observed by SEM inspection (table 1 and figure SI2). Similarly, non-uniform growth was observed at a lower temperature (700 ºC) even though Pg = 500 mbar, A_t = 5 min (same as that of NiAG4), when G_t = 5 min (not shown) or G_t = 45 min (see results for NiAG7). Significant C deposition in the hot zone of the growth chamber was observed after the NiAG7 experiment, so we excluded longer time experiments at high pressure. We note that unlike NiAG1 and NiAG2, the underlying Ni grains were not seen in NiAG3 and NiAG4 samples, as the surface coverage of dark regions (NGF) was reasonably high. Also, it should be noted that in figures 1(b) and (d), the light lines are the Ni grain boundaries, as confirmed from high-magnification imaging (e.g., figure 1(b) inset), while the lines seen in figures 1(f) and (h) are folds (discussed later) in the sample NiAG3 and NiAG4.

Zoom In Zoom Out Reset image size

We further evaluated the extent of growth uniformity in the NiAG4 sample in a typical photo of the NGF/Ni sample, as shown in figure 2(a). Unlike the as-received Ni foil, the color of the treated sample changed from shiny metallic gray to dull gray, all over the 50 cm² sample (unlike for the NiAG1 sample, figure 1(a), inset). This visual change indicates a wafer-scale uniform growth over large-area. After its reproducibility was confirmed, the process parameters for the NiAG4 sample were identified as the optimal conditions for NGF growth.

Zoom In Zoom Out Reset image size

We cut the NiAG4 sample into pieces and inspected further to obtain detailed information on the quality of the as-produced NGFs. The high-magnification images were not recorded from the same regions as that of the corresponding low-magnification to avoid analyzing areas which were contaminated with the aggregation of secondary electrons from the electron beam (figure SI3). The bright areas coverage estimated from SEM images (an example shown in figures SI4(a) and (b)) was 0.1%–3% in 100 µm². Although the relative percentage of bright areas is significantly lower, hereafter, for further characterizations, we have chosen areas consisting of both bright and dark areas, irrespective of their actual relative area coverage. The light and dark contrast observed here is a simple and qualitative measure to recognize the presence of FLG and MLG areas in the NGF film and not a quantitative assessment for how many numbers of graphene layers are present. For quantitative analysis, regarding NGF thickness, we used additional techniques such as Raman, TEM and, AFM, as discussed below. The stepped Ni texture shown in the AFM image (figures SI5(b) and (c), marked with red arrow) can be associated with bright areas of SEM images (figure SI5(a)). The root mean square (RMS) roughness of 40 nm was found from the AFM image in figure SI5(b) (10 x 10 µm²). The fold defect density on the NGF surface was 20–70 folds/100 µm², as estimated from the SEM and AFM images (figures SI4(c) and (d)).

By using optical microscopy (NiAG4 sample), we observed the dominant presence of dark regions (figure 2(b)). A Raman map was performed to include both dark and bright regions marked with an overlapping grid in figure 2(b). A typical spectrum consists of the characteristic graphite peaks, such as G (E_2g mode, at 1580 cm⁻¹), D (defect-activated breathing mode at 1350 cm⁻¹), and 2D (overtone at 2750 cm⁻¹) [36]. The presence of G and 2D peaks pointed to a highly crystalline graphitic material. D peaks with small intensity were observed rarely and only in the few-layer graphene regions. Figure 2(c) presents the intensity ratio map of the 2D and G peaks (I_2D/I_G) [36]. The I_2D/I_G values correlate well with the number of graphene layers [18, 37]. Normalized Raman spectra (with respect to G peak intensity), taken at different positions and representative of the colors in the map, are shown in figure 2(d). Clearly, the I_2D/I_G values vary depending upon the probe position. The red area in figure 2(c) (corresponding to the bright region in figure2(b)) relates to bilayer- and trilayer-graphene (I_2D/I_G ratio of ∼1.1 and 0.9, respectively) [18, 37]. On the other hand, the blue area in figure 2(c) (corresponding to the dark regions, figure 2(b)) shows I_2D/I_G ratios of ∼0.3. These ratios are similar to graphite; thus, the NGF in these regions must be composed of four or more graphene layers. In addition, the commonly observed broadening and upshift of the 2D peak (related to the increasing number of graphene layers) are shown in figure 2(d). Finally, the G peak position is almost constant in all spectra, indicating the absence of chemical doping in the graphene layers.

The crystalline quality of the Ni foils before and after NGF growth was analyzed by XRD. The XRD patterns for the NiAR and NiA samples (table 1) are displayed in figure 3(a). In the NiAR, peaks at 2θ = 44.4º and 51.7º were observed. The 2θ values are in good agreement with the standard values of the (1 1 1) and (2 0 0) Ni crystal planes [38]. In the NiA, the (1 1 1) peak intensity increased by 1.5 times in relation to the (2 0 0). This higher intensity indicates a preferential (1 1 1) orientation of the grains in the annealed Ni foil. Also, as shown in figure SI6 (SEM image and EBSD mapping of the sample surface), Ni surface grains with mainly (1 1 1) orientation resulted from the recrystallization process during annealing [30, 34, 39]. In the NGF-growth samples (NiAG2 and NiAG4), an additional peak in the XRD pattern was observed at 2θ = 25.6º (figure 2(b)). This 2θ value corresponds to the 0.335 nm interplanar distance of (0 0 0 2) graphite planes [28, 40]. A relatively lower (0 0 0 2) peak intensity was observed for the 15 min pre-annealed Ni sample than the 5 min one, which correlates well with the formation of the thinner and non-uniform NGF on the NiAG2 sample (figure 1). Longer annealing offers large Ni (111) grain formation, which is favorable for FLG growth [30, 41]. However, not all grains have (111) surface orientation. Smaller grains with different orientation and grain boundaries will contribute to MLG growth, which leads to non-uniform NGF growth at low P_g during growth [18, 30].

Zoom In Zoom Out Reset image size

The as-grown film was transferred on a SiO₂/Si substrate (digital photo of a typical sample shown in figure 4(a)) by an improved wet chemical transfer process [30, 34]. A uniform film with dark grey color can be seen with the naked eye. Typical SEM images (figures 4(b) and (c)) for the same sample confirm continuous large-area NGF growth (major dark areas). Further studies on the as-grown and transferred film properties such as conductivity and transparency is reported elsewhere [42].

Zoom In Zoom Out Reset image size

Researchers have attempted to control the graphene/ultra-thin graphite films by tuning process parameters such as growth temperature, hydrocarbon exposure (limiting the amount of C by maintaining ultra-high vacuum conditions), substrate thickness (precipitation control), and post-cooling rates (fast cooling, slow precipitation) [41, 43]. The solubility and diffusivity of C in Ni at 900 °C are 0.8 at. % and 8.5 × 10⁻⁸ cm²s^–2, respectively [27]. At higher temperatures, these values are higher [27, 28]. After the growth step, when the Ni catalyst is cooled below 900 °C, the C precipitates at the surface of the Ni as a graphite thin film [43]. However, the rates of C precipitation differ for differently oriented grains and at the grain boundaries, so achieving uniformity in thickness over large-area polycrystalline Ni substrate has remained challenging [41, 44]. We obtained extended NGFs by tuning the A_t and P_g. Extended A_t favors larger Ni (1 1 1) grain formation (figures 1(b) and (d)) whereas short A_t results in a fewer Ni (1 1 1) grains with smaller sizes (figure SI6, 1–10 µm). Ultrathin graphite film formation is favored on Ni (1 1 1) [44, 45]. In our case, short A_t time might be responsible for dominant NGF growth with low FLG coverage (0.1%–3% areas in 100 µm²). Another critical process parameter is P_g. When P_g was higher (500 mbar), more C was dissolved in the Ni foil, and it resulted in relatively more uniform NGF upon cooling than those grown at 10 mbar. Thus, we were able to produce the NGFs with uniformity at mm²-scale when the two key process parameters A_t and P_g were controlled.

As discussed above, we attribute the contrast in SEM images to variations in the thickness of NGF. Figure 5(a) shows a thicker sector of the NGF that was characterized by a black contrast in the SEM secondary-electron imaging (figures SI7(a) and (b)). The TEM observation revealed a highly oriented, defect-free state of the graphene layers, also confirmed by the fast Fourier transform (FFT) pattern (figure 5(a), inset). The graphene layers have a typical (0 0 0 2) lattice-spacing of 0.33 ± 0.01 nm in this section, and the NGF thickness is about 100 nm, with a flat surface. And moving laterally along the NGF, we found two characteristic structural faults. In the first case, due to tangential stress along the graphite film, the NGF forms rippled textures associated with local delamination of the graphitic layers (figure 5(b)). This phenomenon, named as ripplocation, is well known and documented in those layered crystalline materials where the combination of the tangential stress and the strain perpendicular to the layer can generate folded structures without brittle damage [46–48]. In the second case, in those areas of the Ni substrate characterized by a gentle zig-zag topography of a few tens of nanometers, the graphene layers arrange by following the concave and convex interface (figure SI8(c)). As a result, they form symmetrical chevron-like nano-folds with dislocation fields along the fold axis (figure 5(c)). Differently from these, when the Ni substrate shows extended zig-zag surfaces of hundreds of nanometers (steps with small slope, figure SI8(d)), the NGF cannot arrange by a perfect folding, but it gives rise to the formation of kinked structures characterized by brittle damage and very sharp dislocation fields separating the NGF blocks with different orientation. In particular, the graphitic structure in the core region of these kink bands shows large delamination and the gaps filled by amorphous carbon (figures 5(d) and (e)). The HRTEM image of the gap region at the interface NGF/Ni is shown in figure 5(e). The graphene layers exhibit a lattice spacing slightly expanded, 0.35 ± 0.01 nm (instead of 0.33 ± 0.01 nm), compatible with the formation of a mixed NiC_x phase [41]. The FFT pattern of the Ni substrate (figure 5(f)) is consistent with the Ni [−1 1 2] zone axis projection, displaying how the Ni (1 1 1) surfaces are exposed to the NGF. We observed that the secondary phases of NiC_x were not formed continuously along the Ni substrate. Where present, they displayed an arrangement of the basal lattice planes parallel to those of the Ni (1 1 1), resulting in an epitaxial symmetry relationship with the substrate. Figures 5(g) and (h) show a different region of the host Ni substrate where the NiC_x phase interposes between the NGF and Ni substrate, preserving the same orientation of the Ni (1 1 1) lattice planes. The restructuring of NGF into ripplocation, nanofolds and kinks can occur during the cooling process as there is a significant difference in thermal expansion coefficients between Ni and graphite [49, 50].

Zoom In Zoom Out Reset image size

In addition to the dark contrast regions, a thinner sector of the NGF (characterized by a bright contrast in SEM secondary-electron imaging, refer to figures SI7(a) and (c)) was also analyzed. Figure 6 displays the corresponding cross-section TEM analysis. The continuous NGF at the nanometer scale is shown in figures 6(a) and (b): a gentle folding texture in the NGF that links areas with few hundreds of graphene layers and few tens of graphene layers. Such links were observed in the areas where the Ni substrate has steps of large slopes (figure SI8(e)). For example, as shown in figures 6(c)–(g), FLG is in contact with Ni grains (∼1–10 µm, figures SI8(a) and (b)) with nanoscale surface steps. These grains could be correlated to the large (111) grains observed in NiA sample (figure SI6), which further undergo size reduction and recrystallization during the growth step; similar grain transformation were observed previously [20]. HRTEM imaging of the bi-layer graphene (BLG) regions (figures 6(d) and (e)) reveals a perfect and continuous crystal structure showing the typical graphite (0 0 0 2) lattice spacing of 0.33 ± 0.01 nm. In certain regions, at the interface of FLG/Ni (figures 6(d) and (e)), two layers characterized by a lattice spacing of 0.35 ± 0.01 nm were observed. We attribute this extended spacing to the presence of a mixed NiC_x phase, as in the case of a thicker NGF sector (figure 5(e)). Here, for proper alignment of the FLG/Ni interface for HRTEM imaging, the Ni crystallite of this sector was oriented along the [–1 1 0] zone axis with the Ni (1 1 1) surface parallel to the electron beam. The HRTEM that measured a graphitic (0 0 0 2) lattice spacing of 0.33 ± 0.01 nm for the BLG is shown in figures 6(f) and (g). Here the Ni crystal was aligned along the [2–2–1] zone axis to expose the Ni (1 1 1) surface correctly for the imaging. Overall and thanks to a cryo-FIB preparation associated with a gentler ion beam setting, not only could we able to preserve the structure of the FLG sectors, including areas with BLG, but we could also image these layers successfully.

Zoom In Zoom Out Reset image size

Interestingly, in some of the FLG sectors, we found that an atomic interface mediates the contact between the FLG layers and the Ni substrate via an epitaxial relationship, as shown in figures 7(a) and (b). In particular, this sharp interface layer of C atoms follows the flat Ni substrate facet (1–1 1) perfectly. In this HRTEM image, we can see its basal and perpendicular lattice planes with 0.17 ± 0.01 nm and 0.21 ± 0.01 nm, respectively. The distance between the interface and the topmost Ni layer was found to be 0.19 ± 0.01 nm. This distance matches with the distance between graphene and the topmost layer of Ni(111) measured by low energy electron diffraction pattern analysis (0.21 ± 0.007 nm) [51] and theoretical calculations (0.205 nm) [52]. These values indicate that the interface layer is nothing but a tightly bonded graphene with a Ni substrate (in this region).

Zoom In Zoom Out Reset image size

The FLG layers growing on such interface are homogeneous, with a slightly wavy profile ascribable to the tangential stress. After the growth step, as the temperature decreases lower than 900 ºC, the C solubility in Ni decreases, the carbon atoms are expelled to the surface and form a separated phase [43]. The occurrence of the interface layer suggests that an atomic C layer grows on the flat Ni (111) substrate by C precipitation. In parallel, spontaneously expelled C atoms from the Ni grain boundaries form the additional FLG layers [45, 53]. These TEM observations (of the presence of NGF and FLG regions) agree with the SEM and Raman results (figures 1 and 2).

During the growth process, the C atoms are produced by pyrolytic reactions between H₂ and CH₄ . These C atoms dissolve into Ni substrate at elevated temperatures (the C diffusivity/solubility in Ni is temperature-dependent). Studies report that the NGF growth can occur by preferential carbon precipitation in C-enriched Ni films [18, 41] and nucleation of graphite/graphene layers on the metal substrate surface upon sample cooling steps. Studies also reported an epitaxial growth of structurally homogeneous graphene (0 0 0 2) with excellent physical properties on the lattices of Ni (1 1 1) [43]. The lattice mismatch ($\eta )$ can be defined as the absolute difference between two lattice spacings (${d_1},{\text{ }}{d_2}$) along a certain crystal direction, relative to the average of the lattice spacings [54]:

$$\begin{equation}\eta = \frac{{\left| {{d_1} - {d_2}} \right|}}{{\frac{1}{2}\left( {{d_1} + {d_2}} \right)}} \times 100\end{equation}$$

For example, in the case of FLG (figures 6(d), (e), and 7), we can identify the following epitaxial relationship:

• (a)  
  Ni (1 1 1)//FLG (0 0 0 2)
• (b)  
  Ni [–1 1 0]//FLG [1 0 –1 0]
• (c)  
  Ni [1 1 –2]//FLG [0 1 –1 0],

where the first term represents the planar alignment and the other two vector alignments, respectively. The corresponding calculated $\eta$ values are 49%, 19%, and 39%, respectively. These results are consistent with a semi-coherent NGF/Ni interface along which: (a) three Ni [11 1] lattice units fit two FLG [0 0 0 2] lattice units, (b) five Ni [–1 1 0] lattice units fit six FLG [1 0 –1 0] lattice units, and (c) three Ni [1 1 –2] lattice units fit four FLG [0 1 –1 0] lattice units. This epitaxial relationship is consistent with the literature data that indicates the Ni (1 1 1) surface as a preferable nucleation and growth site for graphene, with the value graphene/Ni adhesion energy being 72.70 Jm⁻² [55, 56].

The plan-view TEM imaging for the transferred NiAG4 sample on Cu grid (with lacey C film), presented in figure SI9, shows the presence of the FLG, MLG sections. Unlike SLG [18], we were successful in achieving cracks- and contaminant-free transfer of NGFs which contained FLG areas (few microns wide) partially suspended TEM grid with lacey carbon (figure SI9(a)). These FLG sections were found to be intact and well connected to the NGF. SAED patterns were acquired at 19 different spots to assess the structural features of the NGFs. The grain size of the NGFs varied in the range of a few μm² to tens of μm². In the FLG section, the SAED pattern was consistent with two sets of FLG {0 0 0 1} sheets characterized by a rotational mismatch of 20°, and thus giving rise to (1 0 –1 0) and (1 1 –2 0) spots with a very weak intensity (figure SI9(c)). In this sector, we could discriminate thinner FLG generating weak diffraction spots arranged circularly. But in the NGF section (figure SI9(f)), the SAED pattern was consistent with polycrystalline diffraction spots arranged circularly and showed different intensities. This presence of several spots is the typical effect of random rotational mismatched layers generating a kind of polycrystalline diffraction pattern (figures SI9(c)–(f)).

Generally, graphite films grown on Ni substrate have folds on the surface [23, 26]. These folds arise because the thermal expansion coefficients of Ni (4.7 to 7.7 × 10⁴/°C [49]) and graphite (0.4 to 1.3 × 10^–6/°C [50]) are different (for the temperature range 300 °C–1000 °C). Researchers have assumed that the compressive stress is developed between the Ni substrate and graphite film during the cooling step, and that is what gives rise to the formation of folds on graphene/graphite films [26, 32]. It has been also speculated that the surface steps could be responsible for fold formation [57]. Our results, obtained from SEM, AFM, Raman and TEM, reported in the results section confirm that a close connection exists among the number of graphene layers in NGFs, surface defects, and morphology of the substrate topography of Ni. Our cross-section TEM images evidence that Ni surface steps have a significant role in the fold formation (figures 5 and 6), which is demonstrated in a model shown in figure 8. More details regarding the various NGF/Ni samples (SEM and TEM) are provided in figure SI8. The model shows a continuous NGF and not a single continuous graphene layer on the top. Some graphene layers with exposed edges (especially in FLG areas) might exist on the NGF surface, as observed in figure 7. In the model, we show the compressed strain induced kink formation close to the large Ni grains with small slope. Also illustrated, nanofolds occurred at zig-zag steps of Ni surfaces while the ripples are not associated with surface steps.

Zoom In Zoom Out Reset image size

We believe that our method (with a 5 min growth step) could be applied to obtain continuous NGFs in higher throughput and at lower temperatures (900 ºC) than those of the recently proposed graphite film growth methods [15, 23, 58]. These reported other methods require complicated steps, longer process times and temperatures in the range 1000 ºC –3200 ºC. Finally our TEM observations revealed the following: (i) epitaxial growth of NGF on Ni (111) grains and (ii) the crucial role the polycrystalline nature and topography of the Ni foil in the formation of the NGF texture and the number of graphene layers. Our findings would be a guiding step to obtain the NGFs free of folds by precisely tuning the Ni grain-size and performing surface topography engineering.