represents AFM images for un-irradiated (central image) and irradiated FLG samples at different primary Ar+
ion energies ranging from 0.25 to 5 keV. At first glance, roughness and well-known ripple structures of varying sizes and wavelengths [15
] are visible after ion beam exposure. Careful analysis of the roughness parameters, mean roughness (Ra
) and RMS roughness (Rq
) reveals that the roughness is not monotonically increasing as a function of irradiation energy; instead, a reduction in the roughness is found at 1.5 keV as indicated in Figure
- a sign of sputtering completion of FLG first layer at this irradiation energy.
Sequence of FLG AFM height images as a function Ar+irradiation energy 0.25 to 5 keV. The image in the middle was obtained from the un-irradiated FLG sample.
RaandRqfor FLG as a function of Ar+ion irradiation energy. Ranging from 0.25 to 5 keV.
The evolution of the large scale morphology was probed by two-dimensional power spectral density (PSD) analysis [16
]. PSD function provides a representation of the amplitude of a surface's roughness as a function of the spatial frequency. As shown in Figure
, all PSD functions show a linear dependence at high spatial frequency (f
) with ripple wavelength (λ
) of =24 nm. After irradiation at 0.25 keV, the PSD spectrum superimposes to that of the un-irradiated sample, indicating negligible damage of the FLG surface. Interestingly, a decrease in the graphene PSD intensity is observed at energies higher than 0.25 keV for the spatial frequencies between 60 and 120 μm−1
(i.e., the area enclosed by a rectangle and labeled as first layer).
Figure 3 Two-dimensional PSD spectra calculated from 1×1 μm AFM images (energies: 0.25 to 5 keV). Obtained after irradiation of the samples at different energies from 0.25 to 5 keV. The spectrum for the un-irradiated sample is (more ...)
This suggests strong suppression, smoothing and lateral coarsening of ripple structures (i.e., decrease in PSD intensity, hence roughness amplitude) at these energies. At low frequency regime, the spectra show a switching behavior where the PSD intensities are seen to be higher than that of un-irradiated and increase by a factor of 10 and 100 for the 1.5 and 5 keV irradiations, respectively. This increase is a sign of large amplitude ripple formation with long periodicity of λ
72 nm that resulted from sputtering of both the first and the second graphene layers. It should be noted that the irradiation by 1.5 and 5 keV will only affect the first and the second FLG layers; hence, the assignment of ‘second layer’ and ‘first layer’ is used in Figure
. This is supported by a comparison of depth data obtained from 1
1 μm AFM images after irradiation of the samples at 1.5 and 5 keV. The depth is based on the accumulation of data within specified area of the image, application of a Gaussian low-pass filter to the data to remove noise, and then obtaining depth comparisons between dominant features in a consistent, statistical manner [18
]. The results of this method are depicted in the inset of Figure
. Considering 0.34 nm as the thickness of individual atomic planes in graphite [19
], the depth distribution seen in the histograms confirms the sputtering of one and two layers after 1.5 and 5 keV irradiations, respectively.
Figure 4 Two-dimensional PSD spectra calculated from 1×1 μm AFM images (energies: 1.5 and 5 keV). Obtained after irradiation of the samples at 1.5 and 5 keV. The spectrum for the un-irradiated sample is also plotted for comparison. (more ...)
In order to further shed light on the morphology of the two layers after the 5-keV ion irradiation, the roughness exponent α has been determined from the 5-keV PSD spectrum in the high frequency region. It is evident that the spectrum shows bimodal intensity trend (i.e., two slopes as shown in Figure
) which obeys the inverse power law K f−β
are the spectral index and spectral strength, respectively [20
]. The two slopes of the PSD in this region were found to be β1=
6.0 and β2=
6.5. Since β
is related to α by the equation α
], where the line scan dimension d
is 2, then α1
2 and α2
2.25. Here, α1
attributes to the roughness exponent associated to the first FLG layer and α2
could be assigned to roughness exponent of the second FLG layer.
shows the XPS survey spectra indicating the C (1s
), Ar (2p
), and Ar (2s
) core level excitations recorded from FLG samples after Ar+
ion irradiation at different energies. It is clear from the XPS spectra and from the inset of Figure
that, as the irradiation energy increases, Ar+
ions are implanted into the FLG and reached the saturation limit at energies
XPS spectra after Ar+ion irradiation at different energies from 0.25 to 5 keV. The spectrum for the un-irradiated sample is also plotted as a reference. Inset shows the Ar+ ion concentration as a function of irradiation energy.
Also, variation in the effect of ion irradiation can be observed from the newly developed shoulder (shoulder labeled A) in the XPS spectra along with the increasing of the irradiation energy. The origin of shoulder A is unclear at the moment; however, complex defects previously theoretically predicted [7
] to be associated with Ar+
irradiation might be the reason. In addition, the commonly observed π
π* transition in graphene at binding energy of 290 eV [22
] seems to gradually disappear as the irradiation energy is increased, indicating the FLG electronic damage. This is supported by the following further XPS investigations carried out to determine the transition from the sp2
-bonding to sp3
-hybridized state as a result of Ar+
It can be concluded that the resilience (i.e., only two layers being sputtered with Ar+
at 5 keV for 30 min) of FLG to the Ar+
ion irradiation stems from the combined effect of the Ar+
implantation as evident from Figure
and back scattering or reflection of the incoming ions due to the sample tilting. As the implanted Ar+
ions build up, they electrostatically repel subsequent incoming incident ions, therefore reducing number of ions involved in collisions. This charge repulsion can be understood from the ion concentration saturation observed in the inset of Figure
at energies ≥3 keV; a caution should be taken here as re-sputtering of the implanted Ar+
ions can also take place and lead to the saturation observed. Auxiliary experiments at constant irradiation energy and at different tilt sample angles were carried out to confirm the effect of the reflection of the incoming ions on the resilience of FLG. Indeed, the results (to be reported elsewhere) obtained from the ion irradiation of FLG samples at constant energy of 1 keV, and different sample tilt angles (Φ
) from 0° to 80° show that the amount of implanted Ar+
ions seems to decrease as the tilt angle is increased.
shows the fitted C (1s
) XPS spectra of FLG samples for the un-irradiated and irradiated at energies 0.5 and 4 keV. The fitting of the spectra were done by Gaussian-Lorentzian functions with a Shirley background subtraction [23
]. The fitting yields two peaks positioned at 284.8 and 285.5 eV corresponding to sp2
-hybridized states, respectively [24
]. The sp2
peak appears to be shifted to higher binding energies at high Ar+
ion irradiation energies.
XPS C1sspectra. (a) Un-irradiated, (b) after 0.5 keV Ar+ ion irradiation, (c) after 4 keV Ar+ ion irradiation energy.
Quantification of the sp2
ratio along with the full width at half maximum (FWHM) of the sp2
peaks as a function of irradiation energy is shown in Figure
and its inset, respectively. For the un-irradiated sample, sp2
hybridization dominates (i.e., sp2
3.6, where sp2
78% and sp3
22%) - a FLG characteristic of a two-dimensional sheet of sp2
bonded carbon atoms in a honeycomb lattice [25
]. However, the presence of 22% of sp3
in the un-irradiated sample is attributed either to the edge of individual graphene layers or to the intrinsic sp3
defects reported to exist in the FLG samples [28
]. The well-known effect of transformation of sp2
hybridization due to oxygen exposure of FLG [29
] is ruled out in our case for the un-irradiated sample. This is justified by the absence of O 1s
peak in the XPS spectra.
sp2/sp3ratio as a function of Ar+ion irradiation energy. The inset shows FWHM of sp2 and sp3 components as a function of Ar+ ion irradiation energy.
On the other hand, the initial irradiation of the FLG causes fast transformation of sp2
hybridization as seen in region I of Figure
. The irradiation energies in this region induce irradiation damage dominated by large number of small pits (see AFM images in Figure
) which give less sp2
character due to the enhanced possibility of the damage of the edge of the graphene structure, consequently the formation of dangling bonds. An amorphous stable mixture containing equal sp2
hybridizations (i.e., sp2
1) is observed (region II) when employing irradiation energy in the range from 1.5 to 4 keV. At high irradiation energies above 4 keV, the number of transformed sp2
bonds seems to be slightly less as depicted in region III of Figure
compared to those of region II. This observation can be understood in two different ways by considering the morphology formed after the irradiation in the vertical and the lateral directions. First, in the vertical direction, the observation is supported with the conclusion of the depth distribution analysis where the third graphene layer was just found to be exposed at these high irradiation energies. Therefore, XPS is expected to directly detect the sp2
contribution from this layer and the layers beneath up to approximately 2 nm of the expected XPS sampling depth in the geometry used. In the lateral direction and as seen from the AFM image at 5 keV in Figure
, few large pits are formed due to the irradiation damage; hence, large untouched islands on the surface of the samples with better sp2
character are expected to exist. The reduced number and the large size of these pits are reflected from the long periodicity of λ
72 nm and two orders of magnitude increase in PSD intensity found in Figure
To further account for the effect of the ion irradiation, we adapt the commonly used first derivative of C KVV spectra obtained from XPS data [30
]. The selection of C KVV Auger transition reflects a self-convolution of the occupied valence band and can be used to distinguish carbon atoms at different hybridization states. Figure
shows the binding energy width (D
) between the most positive maximum and most negative minimum obtained from the first derivative Auger transition C KVV spectra recorded from FLG samples, as a function of Ar+
ion irradiation energy.
The first derivative Auger transition C KVV spectra. Recorded from FLG samples, as a function of Ar+ ion irradiation energy, indicating the values of D parameter.
values are within the reported values of 21 to 23 eV for HOPG (multi-layer graphene) [31
], but not sensitive to the irradiation energy. However, comparison between the intensities of shoulder A and peak B reflects the sp2
transition trend similar to that observed in Figure
. Here, the intensity difference between A and B reaches the minimum and stabilizes after ion irradiation in energy range between 1.5 and 4 keV. Slight increase in the intensity difference is found at energies above 4 keV. Although, the shape of derivative of Auger transition C KVV spectra gives the indication of the ion irradiation effect, the D
parameter values obtained from these spectra fail to account for sp2
hybridization transition, and reliable quantification of sp2
ratio is only obtained from the deconvolution of the XPS C (1s
) envelope as explained before and shown in Figure
. Our preliminary results to exploit the roughness created by ion irradiation (surfaces shown in Figure
) for nanobubble formation and, consequently, creating graphene surfaces with less friction for different applications are promising. Supplement data shown in Figure S1 in Additional file 1
support this claim where observed contact angle, size and density of nanobubbles deposited on the sputtered samples are observed to increase with surface roughness.
From all the above findings, it is clear that more than one factor can contribute to the sp2 to sp3 transformation mechanism. For example, irradiation energy will cause lattice displacement in the FLG which affects the FLG layer cross-linking, therefore contributing in the formation of new sp3 bonds. In addition, bonds can be broken to form sp3 due to the collision impact. This is evident from the XPS reduction of π electron (π bond) intensity as the irradiation energy is increased. Furthermore, bending (to be discussed later) and breaking of the graphene planes are likely to contribute to the sp2 to sp3 transformation observed.
We turn our discussion to the effect of Ar+
ion irradiation on the FLG edges. In contrary to the damage found in the flat regions, the edges of FLG are found to be much rougher; with four distinct features as shown in Figure
, (1) the upper edge regions are not smooth and show irradiation energy dependent wiggles, and (2) the shadowing effects similar to those found in oblique angle growth of thin films are observed [32
2-D and 3-D images showing the effect of ion irradiation on the FLG edges. (a) after Ar+ ion irradiation of 0.5 keV, (b) after Ar+ ion irradiation of 1 keV, (c) profile taken cross the length indicated by solid line in (b) showing shadowing effect.
In this case, the edges are seen to be lifted up and shadow the regions behind them; shadowing effects are seen from the cross section profile taken at a position shown by a solid line in Figure
b and illustrated in Figure
c. (3) The lower edges act as weak points where more carbon sputtering takes place, and (4) new features (shown by arrow in Figure
b) at an angle of 53° from the edges are formed as the irradiation energy is increased. All irradiation features and their roughness amplitudes after 3 keV ion irradiation can be seen in the supplement data shown in Figure S2a,b in Additional file 2
. The presence of these features is detected by the fast Fourier transform (FFT) analysis shown in the inset of Figure S2b in Additional file 2
During the ion irradiation, unusual defects and edge reconstructions were also observed; defects appear only after ion irradiation and were not found in the un-irradiated samples. Furthermore, these defects and edge reconstructions are similar to those reported by Zhang and Feng [33
] during the growth of graphene and by Mathew et al. [34
] after MeV proton beam irradiation of graphene, respectively. Figure
a shows the AFM image of the obtained FLG after Ar+
ion irradiation of 1.5 keV. Few holes (enclosed by circles) can be clearly seen, suggesting the possibility of creating nanoporous graphene layers by ion irradiation for filtration applications. Increasing the ion irradiation to 4.5 keV, the peeling of graphene layers in the direction normal to the edge takes place which is associated with triangular-shaped FLG folding (enclosed by a circle) as shown in Figure
b. Furthermore, FLG folding along the edge (indicated by white arrow) is seen to take place.
Figure 10 2-D AFM images showing localized hole defects and shape of FLG folding. (a) The localized hole defects after Ar+ ion irradiation of 1.5 keV. (b) Triangle shape of FLG folding (enclosed by a circle), multi-layer folding (indicated by white arrow), wrinkles (more ...)
This type of folding has been observed on the surface of FLG resembling single wall nanotubes or double wall nanotubes due to the single- or double-graphene layer folding back to itself [35
]. The folding observed might be due to the electron beam damage associated with scanning electron microscope and transmission electron microscope used while characterizing the samples in these studies. In our case, the folding is rather attributed to the irradiation process-induced damage which includes multi-layers of graphene as inferred from the thickness of folded layers ranging from 2 to 4 nm. Intense irradiation of the multi-folded layers can induce isolated particles (enclosed by a rectangle) and initiate wrinkles (indicated by black arrow) as seen in Figure
b. Similar edge reconstructions have been reported by Huang et al. [36
], which are due to the graphene layer-by-layer peeling as a result of void formations and their migration towards the graphene edges. This process was associated with atomic sublimation caused by Joule heating and facilitated by atomic displacement caused by high-energy electron irradiation. Despite the serious electronic and morphological damages at high energies, roughness, PSD, XPS, π
π* transition, sp2
FWHM values and Auger transition C KVV spectra data analysis reveal that, using low energy of 0.25 keV Ar+
ion irradiation at large sample tilt angle at room temperature, is an alternative way of peeling graphene layers without the need of Joule heating.