Non-destructive electron microscopy imaging and analysis of biological samples with graphene coating

In electron microscopy (EM), charging of non-conductive biological samples by focused electron beams hinders their high-resolution imaging. Gold or platinum coatings have been commonly used to prevent such sample charging, but it disables further quantitative and qualitative chemical analyses such as energy dispersive spectroscopy (EDS). Here we report that graphene-coating on biological samples enables non-destructive high-resolution imaging by EM as well as chemical analysis by EDS, utilizing graphene’s transparency to electron beams, high conductivity, outstanding mechanical strength and flexibility. We believe that the graphene-coated imaging and analysis would provide us a new opportunity to explore various biological phenomena unseen before due to the limitation in sample preparation and image resolution, which will broaden our understanding on the life mechanism of various living organisms.

Comprehensive understanding of biological objectstheir chemical, physiochemical and biological characteristics-can be effectively achieved through electron microscopy (EM) analysis [1][2][3][4], preferably without any fixation or auxiliary surface treatment. Scanning electron microscope (SEM) and transmission electron microscope (TEM) are increasingly more employed as they provide direct imaging of specimen's morphological structures with high-resolution [5,6]. Recently, EM analysis has extended its use to construct three-dimensional structure of the biological specimen with combination of serial block-face sectioning or focused ion beam [7,8]. In addition, the unique interaction between electron beams and specimen enables various physical and chemical analyses such as energy dispersive spectroscopy (EDS), electron probe micro analysis, and electron energy loss spectroscopy [9,10]. Nevertheless, charge accumulation by electron beams and shrinkage of samples by dehydration in vacuum have always hindered EM-mediated biological studies as it distorts the morphological and chemical characteristics of the specimens [11][12][13][14][15][16]. For these reasons, various coating and sample preparation methods for EM analysis have been developed to enhance the image contrast from non-conducting biological specimen [1,5,[17][18][19][20][21][22][23][24][25][26][27][28]. In particular, a metal-and carbon-coating methods have been widely employed to dissipate the accumulated charges on non-conducting surface [23][24][25][26][27].
However, the relatively thick coating layer hampers from studying the fine structures of the specimen at nanometer scale because of the large size of metal or Original content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence.
Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. carbon grains. In addition, x-ray fluorescence signals required for EDS analysis are screened by metal layers [29]. Furthermore, it is usually difficult to use the metal-coated samples for further analyses such as TEM that requires electron-transparency. Here we report that graphene-coating on biological samples enables non-destructive high-resolution imaging by SEM and TEM as well as chemical analysis by EDS, utilizing graphene's transparency to electron beams, high conductivity, outstanding mechanical strength and flexibility [30][31][32].
Recent progresses in large scale synthesis of high quality graphene films using chemical vapor deposition (CVD) methods have widened its potential in practical device applications as well as unique interests in basic scientific researches [33][34][35][36]. The feasibility of the large scale fabrication of continuous graphene films as well as easy transfer onto diverse biological objects opens up a unique opportunity to create new hetero-interfaces or interfaces with non-conducing biological samples. As demonstrated in a recent work [37], the in situ high-resolution EM imaging of nanocrystal growth has been achieved by using graphene liquid cells to encapsulate nanoscale materials as well as their environment (i.e. liquid) and separate them from the vacuum environment. In this regard, graphene mediated coating on biological samples can provide high-resolution EM imaging and chemical analysis due to the excellent electron and heat flow thorough the graphene and electron-transparency [44]. Here, with taking all these advantages of graphene films, we have employed continuous graphene films as coating for biological samples and exploited them for non-destructive high-resolution EM imaging and chemical analysis.
As schematically displayed in figure 1, the unique feasibility and availability of continuous graphene films at large scale enables the conformal coating of biological objects including leaves, ants, spiders, neuron cells, Escherichia coli (E. coli), proteins, and polypeptides whose sizes range from several centimeters down to few nanometers. Atomically-thin and electrically-conducting graphene membranes were prepared on non-conducting biological surface by isolating graphene films from copper (Cu) foils after CVD growth, followed by conformal coating onto biological samples as illustrated in figure 1(c). Compared to other conventional sample preparation methods including fixation and metal sputter coating (figures 1(a) and (b)), the present method based on graphene coating is relatively simple, bio-friendly and non-destructive, which is particularly advantageous for preserving the chemical information of samples for further experiments.
Monolayer graphene film was synthesized on high-purity Cu foil using CVD method (please see supplementary materials). Continuous graphene films coated with a poly(methyl methacrylate) (PMMA) layer can be isolated from Cu foils and transferred to a target surface after wet chemical etching [35]. The PMMA was removed by using acetone before the Cu etching. The number of graphene layers was controlled by repeating this transfer process. We found that triple-layered (3-layer) graphene films provide optimum electron transparency and mechanical stability for SEM analysis (figures S1-S3). The biological specimens were cleaned and positioned onto a metallic sample stage for SEM imaging. The 3-layer graphene sheets were then transferred on top of the biological specimen by scooping from bottom side, followed by drying in a desiccator.
To demonstrate the advantages of using graphene membrane for SEM imaging, we have selected several representative biological specimens (ants, bee's wings, water fleas, E. coli and ferritin) that are different in terms of size, surface hardness, and morphology. The 3-layer graphene mostly covered these millimeter to nanometer sized samples, and only shows minor fractures around needle-like structures (figure 2(c)). In contrast, the use of graphene oxide (GO) and reduced GO resulted in incomplete coating due to their poor mechanical strength and difference in hydrophobicity ( figure S4). An SEM imaging on a carbon coated samples also showed large carbon grains that distort the sample's morphology (figure S5). The high-magnification FE-SEM images of a graphene-coated ant clearly show not only unique micro-patterns but also nanopores as small as 40 nm (figure 2(b)) that are invisible in platinum (Pt)-coated or carbon-coated samples (figures 2(i) and S5). Such fine and clear observation of the surface structures implies that the adhesion between graphene and the sample (mostly by van der Waals interaction) is strong enough to maintain its morphology [38] and stable up to acceleration voltage of 20 kV (figure S6). The needle-like structures on bee's wings result in punctures on graphene, but the surface still shows conformal graphene coating that enables stable SEM imaging (figure 2(c)). We also performed SEM imaging on a 1.5 mm long water flea (Daphnia pulex) covered with 3-layer graphene films. High-magnification SEM images of the water flea (area P1 in figure 2(d)) clearly display the unique features of a water flea on its dorsal carapace (figure 2(e)). Interestingly, the graphene film mostly covers the needle-like surface on its antenna without much tearing (figure 2(f)).
The advantages of graphene coating compared with a conventional metal coating method were demonstrated under identical conditions (figures 2(g)-(i)). We observed that the bare gaster surface of an ant is strongly charged even at low acceleration voltages (<2 kV) (figure 2(g)), and the bare eye surface is immediately burning at 5 kV, while the graphene-coated area does not show any damage even with high acceleration voltages up to 20 kV (figures 2(g), (h), and S6).
Unlike the above mentioned hard-surfaced insects, soft biological objects such as tissues, cells and bacteria need an additional treatment for EM analysis, including aldehyde fixation, osmium tetroxide staining, and critical point drying. In this process, the use of organic solvents often distorts the samples' original contents and disables further qualitative or quantitative chemical analyses. In this regard, the simple graphene-coating method can be advantageous because biological samples close to their native structures can be imaged and preserved for further analyses. If combined with conventional fixation methods, it would be more effective for the high-resolution EM imaging of biological samples (figure S7). We also demonstrate that common bacteria, E. coli, cultured in a liquid medium can be imaged after monolayer graphenecoating that protects E. coli from sudden vacuum drying as well as e-beam damage (figures 2(j)-(l)) in SEM. Another graphene layer on bottom side was used to seal the liquid environment by π−π interaction with top graphene layer. Recently, an environmental SEM (ESEM) has been utilized to observe the native structures of biological samples without conductive coating, but its resolution is still limited due to the charging problem associated with low conductivity of biological surfaces (figures S7 and S8) [39][40][41]. We also observed that untreated E. coli shrank by dehydration in a vacuum chamber during ESEM imaging ( figure  S9). On the other hand, the graphene coating not only provides higher resolution than ESEM but also stabilizes the liquid-containing samples that can be easily damaged by intense electron beams.
We also performed EM imaging of ferritin, an intracellular protein that stores and releases iron to control the iron concentration in living organisms [42]. The ferritin particles in water are encapsulated between two monolayer graphene films. The individual ferritin particles are clearly observed in SEM, in which the iron cores look brighter ( figure 2(m)). In a low-magnification TEM image, spherical protein shell as well as hydrous ferric oxide cores were identified from their different contrast, and at high-magnification, the lattice fringe of the iron core was clearly resolved with atomic resolution in an aqueous medium ( figure 2(o)). Likewise, we also demonstrate that the hydrated structure of plasmid deoxyribonucleic acids of E. coli sandwiched between graphene layers can be successfully imaged by SEM and TEM (figure S10).
We compared the performances of graphenecoating and Pt-coating methods in chemical analysis by EDS. All the experimental conditions and parameters including spot sizes and signal collection time were identical. The results showed that EDS signals from graphene-coated samples ( figure 3(a)) are 2-3 times stronger than Pt-coated samples ( figure 3(c)), which facilitates the qualitative and quantitative chemical analyses on nitrogen-containing chitin (from ants) and oxygen-rich cellulose (from leaves). The non-destructive analysis enabled by the graphene coating is particularly efficient for element-specific EDS mapping. The water flea sample was fed with 25 nm cerium oxide nanoparticles (CeO 2 NPs) to stain its digestive pathway (please see supplementary materials for experimental details). The CeO 2 NPs are clearly visualized in the Ce-selective EDS mapping of and (e)). The other EDS analyses also indicate that the graphene-coated method is superior to Pt-coating in terms of signal intensity (figure S11). We attribute the signal reduction in the Pt-coated samples to the absorption and scattering of incident electrons and x-ray fluorescence radiation by thick Pt layers, which will be further discussed in figure 4.
The outstanding performance of atomically thin graphene membrane as protective coating for EM analysis was theoretically confirmed by Monte Carlo simulations (please see supplementary materials for detailed methods). The 1 nm graphene-coated chitin (Gr/chitin) was compared with 10 nm Pt coated chitin (Pt/chitin). As seen in the electron trajectory images ( figure 4(a)), incident electrons can easily pass through the thin graphene and penetrate deep into the chitin layer, while the Pt layer blocks electron penetration because of the large nucleus radius (i.e. large electron scattering cross-section) of Pt (Z=78) (see supplementary figure S12 for cross-section calculations). As the acceleration voltage increases from 2 to 10 kV, the maximum penetration depth of electrons increased from 140 to 2250 nm for Gr/chitin. This indicates that graphene membrane is more transparent to lower accelerating voltages. On the contrary, the electrons irradiated to Pt/chitin show less penetration depths with larger scattering angles. The amount of penetrated electrons is directly related to the x-ray signals  (figure 4(c)), resulting in large difference in x-ray absorption intensity between Gr/chitin and Pt/chitin. Figure 4(b) shows the cross-section profiles of energy loss along the simulated electron trajectories [43], which is related to the intensity of EDS signals. From 2 and 5 kV, most of the energy loss happens inside the 10 nm Pt layer. Major energy loss still occurs near the Pt layer at 10 kV. On the other hand, energy loss in Gr/ chitin mostly takes place inside the chitin layer even at 2 kV, indicating that the graphene is almost free from electron energy loss and background EDS signals. Figure 3(e) shows the depth profiles of Rho-Z x-ray intensity of carbon for Gr/chitin (red) and Pt/chitin (gray). As the accelerating voltage increases from 2 to 10 kV, the total x-ray intensities for Pt/chitin and Gr/ chitin increased from 6 to 259 and from 87 to 470, respectively (see also supplementary figure S14 for the simulations for nitrogen and oxygen analysis). This indicates that graphene is superior to Pt for protective coating for EDS analysis. The experimental EDS spectra with varying accelerating voltages coincide with the simulation results (figure S15).
In conclusion, we demonstrated that graphene's outstanding mechanical strength, conductivity, flexibility, and transparency to electron beams enable the simple and non-destructive imaging and analysis of various biological samples with high resolution that can be hardly achieved in bare or metal-coated samples. The graphene coating effectively prevent charge accumulation by spreading e-beam induced charges and heats over large surface area, and the high mechanical strength and flexibility of graphene allows conformal coating by excellent adhesion with various biological interfaces. We believe that the graphene-coated imaging and analysis would provide us a new opportunity to explore various biological phenomena unseen before due to the limitation in sample preparation and image resolution, which will broaden our understanding on the life mechanism of various living organisms.