Chapter 5: Dark-Field Optical Microscopy Studies of Graphene Grown on Cu Foil: Direct
5.3. Results and discussion
5.3.3. Evaluation of graphene as an anti-oxidation barrier for Cu foil
122
123
seems to be oxidized even faster than bare Cu. After 4 days, the graphene grain appeared in the BF image (Figure 5.7e) and there was brightness increase of bright spots on G1 from DF image (Figure 5.7f). For the bare Cu, bright spots were mostly perceived at the locations near the graphene edges. Finally, after 28 days storage, the color of bare Cu turned to brown and graphene grain became very clear in the BF image from the color contrast (Figure 5.7g). DF image in Figure 5.7h shows the bare Cu like the DF image shown in Figure 5.1g, indicating the very deep oxidation. However, it is worthy to note that I found the density of bright spots at area G1 did not augment by time and the bright spot still could not be found on area G2.EBSD image revealed that the graphene grain actually crossed the Cu(211) and Cu(110) facet of Cu foil (the inset of Figure 5.7h). The above data indicates that the oxidation of graphene-coated Cu strongly relies on Cu crystal orientation rather than graphene. Graphene grown on Cu(110) accelerates the faster oxidation of the underlying Cu foil than bare Cu, instead of inhibition.
124
Figure 5.7 (a,c,e,g) BF optical images and (b,d,f,h) the corresponding DF optical images of graphene grown on oxygen-rich Cu foil after exposure at 40 ℃ with humidity of ~40 %. The insert in h shows the measured EBSD image overlapped with SEM image.All scar bars are 100 µm.
125
Further investigation on area G1 is shown in Figure 5.8. I firstly divided five regions according to the color and location from the BF image in Figure 5.8a. Compared with the corresponding DF image in Figure 5.8b, it can be seen that the dispensed bright spots appear orange color in BF images in Figure 5.8a (denoted as 2). AFM image and the height profile show that in these areas there is a ~10 nm thickness increase, indicating the formation of copper oxide layer. The oxygen concentrations of the areas marked in Figure 5.8a were measured by EDX (Figure 5.8e). Area 2 shows 15.2 % of oxygen, which is ~3.6 % higher than other graphene-coated Cu areas (denoted as 3 and 5 in Figure 5.8a) but very close to bare Cu (denoted as 1 in Figure 5.8a). Therefore, the data from EDX further proved that the observed bright spots are indeed the oxidized areas but coated by graphene. In addition, I noticed several bi-/multi-layer graphene grains in DF image (Figure 5.8b) and one of them was marked with blue dotted line (denoted as 4 in Figure 5.8a and the corresponding Raman spectrum shown in Figure 5.9a.) The oxygen concentration of the bilayers graphene grain is the lowest, only 10.7 %, which is the limitation detected by EDX, indicating the well oxidation protection provided by bi-/multi-layer graphene.10 Figure 5.9b shows another image showing that the survival of only one multilayers grain at the center after the oxidation in air.126
Figure 5.8 (a) BF optical image and (b) the corresponding DF optical image of the enlarged area as shown in Figure 5.7h. (c) AFM image scanned from the insert BF optical image. (e) Oxygen concentrations measured by EDX from the areas marked in (a).
127
Figure 5.9 (a) Raman spectra of graphene on Cu taken at the locations of 3 and 4 of Figure 5.8a. The stronger intensity of G peak than 2D at area 4 indicates the character of bi-/multi-layer graphene. (b) SEM image of another graphene grain on Cu foil after maintaining at 40 ℃ with humidity of ~40 % for 28 days.I further investigated the graphene quality after oxidation. Figure 5.10a shows one of the graphene grain after maintaining at 40 ℃ with ~40% humidity for 28 days. I divided the grain into three regions on the basis of Cu grain boundaries. Area 1: underlying Cu was well protected by graphene; area 2: Cu was partially oxidized, and area 3: the Cu surface was mostly oxidized. Then I transferred the graphene on SiO2/Si wafer. Figure 5.10b,c,d show the BF images of transferred graphene and the corresponding intensity of D-band to G-band (ID/IG), from area 1, 2 and 3 marked in Figure 5.10a. As can be observed, all the three areas show relatively uniform graphene film without any crack. The low value of ID/IG, which is commonly used to detect the structural defect of graphene,1 suggests the high quality of graphene. Figure 5.10h shows the Raman spectra of graphene measured at the locations circled in Figure 5.10b-d, and all showed the weak D band. Therefore, the above data demonstrate that the oxidation of underlying Cu may not be due to the low quality of graphene.
128
Figure 5.10 (a) BF optical image of graphene grown on Cu foil after staying at 40 ℃, 40% humidity. (b, c, d) BF optical images of graphene taken at the areas 1, 2 and 3 marked in (a), after transfer on SiO2/Si wafer. (e, f, g) are the corresponding Raman mappings ID/IG intensity ratio of the area shown in (b, c, d) respectively. (h) Raman spectra of graphene collected at the circled locations at (b-d).
Previous reports claimed that graphene-coated metal usually can be well protected even exposure to hash conditions from several minutes to hours compared with the bare metal.11-13 However, from Figure 5.7 and 5.8, graphene-coated Cu(110) was oxidized even faster than the bare Cu, although the size of the oxidized area will not enlarge with the exposure time. One of the previous reports claimed that the defects of graphene may accelerate the oxidation of underlying Cu by the electrochemistry response with Cu, oxygen, and water.10 However, from the Raman data provided in Figure 5.10, this may not the case in my system as the grown graphene shows high quality, regardless of the oxidation degree of the underlying Cu. Therefore, it is reluctant that oxygen atoms could oxidize Cu by penetrating graphene film here. It is highly noticeable that the Cu foils employed here is oxygen-rich Cu made by oxidation on a hot plate before CVD. As explained before, I employed oxygen-rich Cu in order to get large graphene grains with fewer GGBs. Thus the deterioration of oxidation resistance of graphene also cannot be ascribed to GGBs.
According to my former research in Chapter 4 and the relative literatures, the oxygen decomposed from oxide layer during heated up in inert gas, can dissolve into the Cu bulk at high temperature before the
129
graphene growth.21, 33 Therefore, I speculate that these dissolved oxygen prior to the graphene growth should be an important oxygen source of Cu oxidation. The dissolved oxygen atoms may diffuse out gently after leaving from CVD chamber. In this case, the top graphene layer may promote the oxidation of underlying Cu,20, 37 resulting in the faster oxidation (see the illustration in Figure 5.11a). As mentioned in 5.3.2, I observed the Cu oxidation even was stored in vacuum environment (Figure 5.2 d). The Cu oxidation here could be also ascribe to the dissolved oxygen in the bulk. The inhomogenous oxidation, as shown in Figures 5.7h and 5.10a, indicates that oxygen solubility or diffusion rate in Cu highly depends on the crystal orientation. Figure 5.12 shows the millimeter scale EBSD image overlapped with SEM of the graphene after oxidation for 28 days. The index of Cu crystals is only available for the graphene-coated areas because of the protection from oxidation. Since the bare Cu was covered with amorphous copper oxide layer, the crystal indices is unavailable and display black here. In addition, the oxidized graphene-coated areas also appeared darker (marked with dotted circle). It is noticeable that for the severe oxidation areas of graphene-coated Cu, the index of Cu crystals are main or near (110), which may suggest the relatively high oxygen solubility or diffusion rate of Cu(110).In order to further prove this idea, I grew graphene on an oxygen-free Cu foil with no pre-oxidation, then was stored at the same condition as in Figure 5.7. Figure 5.13 shows the BF and DF images taken randomly from the graphene grown on ~1 cm × 1 cm size of Cu foil after 11 days storage. It is noticeable that the graphene-coated Cu is effectively protected and I did not find any bright spots through the whole piece of Cu foil, which demonstrates that graphene coatings can inhibit the oxidation of oxygen-free Cu below, or at least in the storage environment I used here (Figure 5.11b). Therefore, my studies show that oxygen-free metals are highly desirable if people want to grow graphene film on the top as an oxidation barrier.
130
Figure 5.11 Illustrations of graphene grown on oxygen-rich (a) and oxygen-free Cu (b). For the dissolved oxygen atoms remaining in the oxygen-rich Cu after CVD, results in the Cu oxidization assisted by graphene on the top. On the other hand, graphene can protect the underlying Cu from being oxidized by the oxygen or water in air.
131
Figure 5.12 EBSD image overlapped with SEM image of graphene grown on oxygen-rich Cu foil then stored at 40 ℃, ~40 % humidity for 28 days. The graphene-coated areas that strongly oxidized were circled.132
Figure 5.13 BF optical images (top) and the corresponding DF optical images (bottom) of graphene grown on oxygen free Cu foil. CH4 15 ppm and the growth time 30 min. The as-grown graphene on Cu foil was then preserved at 40 ℃ with humidity of ~40 % for 11 days.