Development of TEM and SEM high brightness electron guns using cold-field emission from a carbon nanotip

Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Ultramicroscopy journal homepage: www.elsevier.com/locate/ultramic

Development of TEM and SEM high brightness electron guns using cold-field emission from a carbon nanotip F. Houdellier a,n, L. de Knoop a, C. Gatel a, A. Masseboeuf a, S. Mamishin b, Y. Taniguchi b, M. Delmas a, M. Monthioux a, M.J. Hÿtch a, E. Snoeck a a b

CEMES-CNRS, 29 Rue Jeanne Marvig, 31055 Toulouse, France Hitachi High-Technologies Corporation, 882, Ichige, Hitachinaka, Ibaraki 312-8504, Japan

art ic l e i nf o

a b s t r a c t

Article history: Received 6 August 2014 Received in revised form 19 November 2014 Accepted 20 November 2014

A newly developed carbon cone nanotip (CCnT) has been used as field emission cathode both in low voltage SEM (30 kV) electron source and high voltage TEM (200 kV) electron source. The results clearly show, for both technologies, an unprecedented stability of the emission and the probe current with almost no decay during 1 h, as well as a very small noise (rms less than 0.5%) compared to standard sources which use tungsten tips as emitting cathode. In addition, quantitative electric field mapping around the FE tip have been performed using in situ electron holography experiments during the emission of the new tip. These results show the advantage of the very high aspect ratio of the new CCnT which induces a strong enhancement of the electric field at the apex of the tip, leading to very small extraction voltage (some hundred of volts) for which the field emission will start. The combination of these experiments with emission current measurements has also allowed to extract an exit work function value of 4.8 eV. & 2014 Elsevier B.V. All rights reserved.

Keywords: Field emission Electrons sources Electron holography Carbon tips

1. State of the art in cold field emission source The room temperature field emission (FE) source, named coldfield emission (CFE) source to distinguish it from the high-temperature thermal-field emission (TFE) and Schottky emission (SE) ones, is currently used in transmission electron microscope (TEM), scanning electron microscope (SEM) and scanning transmission electron microscope (STEM) instruments. The fundamental physics of field emission has been covered by many reviews [1–3]. Crewe reported the first successful development of a SEM equipped with a CFE gun (CFEG) in 1964 [4]. Crewe and his group were convinced of the performances of CFE and benefited from the progress of UHV technology. They were pioneers in the development of an advanced STEM equipped with a CFEG in the early 1960s [5]. Field emission electrons are obtained by applying an extraction voltage of several thousand volts at the apex of a metal tip of a radius of less than 100 nm. FE electrons are emitted by tunnel effect through the potential barrier near the Fermi level because the barrier is thinned by the application of negative electrostatic field strength of 107 V/cm at the metal surface. As the emitter works at room temperature, the energy spread of the emitted electrons is narrow (0.2–0.3 eV) [31]. n

Corresponding author.

Due to the very small apex of the FE tip, the CFEG is the brightest electron source available leading to a highly coherent electron beam with values of some 109 A/cm2 sr [31]. Standard FE metallic tips used in CFE source are made of tungsten oriented in [310] zone axis for which the exit work function is the smallest even if other crystal orientation have been also used like the [111] orientation. The exit work functions are respectively 4.34 eV for [310] orientation and 4.47 eV for [111] orientation [6]. However, tungsten tips remain an unrivaled cathode material in the CFEG technology since more than 40 years. The main challenge of the CFE electron source with tungsten FE tip is to keep a beam stability over a sufficient duration [7]. It is indeed very problematic to achieve a stable emission together with a high intensity of the beam current at the same time. The FE current fluctuates, and instability of the emission current increases with the probe current. This instability is one of the main problems that must be overcome in FE technology for practical and routine use. These instabilities have multiple origins but are mainly due to the FE tip contamination, which can be related to the vacuum environment, anode degasing and so on [7]. If the lifetime of the FE tip is very important when the gun is used as a simple diode (like in the Butler system used in a SEM with low acceleration voltage [8]), it may be much shorter for a FE gun used in electron microscopes working at high voltage (100 kV

http://dx.doi.org/10.1016/j.ultramic.2014.11.021 0304-3991/& 2014 Elsevier B.V. All rights reserved.

Please cite this article as: F. Houdellier, et al., Development of TEM and SEM high brightness electron guns using cold-field emission from a carbon nanotip, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.11.021i

F. Houdellier et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

2

or more) [9]. In this case, the instabilities of the current may cause the destruction of the tip by disruptive breakdowns. To prevent this unwanted event, the tip is regularly flashed by heating it up during some microseconds in order to evaporate the adsorbates deposited on it. The need for UHV (10
9 Pa) and regular flashing of the tip makes the technology complicated and expensive. In 1998 [10] and 2000 [11], new field emission sources, using carbon nanotubes (CNTs) as FE tips were demonstrated. Additional studies have showed disordered matrix arrays [12], multiwalled CNT yarns [13] or thin films of CNTs [14] to perform and study field emission (an exhaustive review of CFE with CNTs has been published by de Jonge and Bonard [15]). The CNTs exhibit remarkable properties that should make them excellent candidates as field emitters. First, their high aspect ratio and the small tip radius enhance the local electric field, which thus ensures a low extraction voltage to be used for starting the emission process. Second, they are chemically less reactive than standard tungsten emitters therefore leading to a more stable electron beam. Third, they also present high mechanical strength and are capable of emitting a large amount of electrons of uniform properties. In 2002, de Jonge et al. demonstrated for the first time the cold field emission from a single CNT source [16], and many other studies have been performed as well [17,18]. Despite being ideal candidates for FE sources of high brightness, CNTs suffer from a couple of important problems. Their huge aspect ratio makes them likely to vibrate, which is unwanted for a CFEG [19]. In addition, their nanometer size makes them difficult to handle, which is a critical criteria for possible industrial application in commercial microscopes. In order to combine the advantage of tungsten emitter and the good properties of CNTs we used carbon cone nanotips (CCnT) as field emitters [20]. As we will show in the following, they exhibit a micrometer-large carbon fiber segment as a base from which a carbon cone with nanometer-sized apex protrudes. The conical shape prevents vibration and the microfiber support enables easy handling.

2. CCnT as field emitter: preparation The synthesis of CCnTs was first reported by Jacobsen and Monthioux [20] in 1997. The CCnTs were produced using a catalytic method to create multiwall carbon nanotubes with a diameter of around 5 nm. This was followed by the deposition of pyrolytic carbon onto them, using chemical vapor deposition from a methane carbon source at a temperature of  1300 °C [for more details see reference [21]]. The resulting CCnTs exhibit a unique shape with a “large” microfiber segment of few micrometer in diameter consisting of graphene stacks coarsely concentrically displayed, ended by two smooth carbon cones on each side of the microfiber segment. The cone angle is of few degrees only and the cone apex is of some nm only (between 1–10 nm in general), i.e. equal to the CNT diameter onto which both the fiber segment and the cones have grown and which is shown to emerge from the cone apex in Fig. 1. Those morphological features make the cone tip very sharp.

3. Mounting of the carbon nanocone tip as field emission cathode In order to mount the CCnT onto a FE tip which can be used as a cathode for SEM and TEM sources, we used a FIB-SEM procedure to select and manipulate them, as illustrated in Fig. 2 and described in detail by Houdellier et al. in [22]. A standard tungsten FE

Fig. 1. Two examples of the CCnT-bearing objects as described in the text, with enlargement showing the large base consisting of a carbon microfiber segment with rough surface (designated as “disordered” carbon in the figure) and a portion of a smooth carbon cone from which the primary CNT are seen to emerge.

tip and a preparation of several CCnT-bearing objects are loaded in the FIB. The tungsten FE tip in Fig. 2 was electro-chemically etched under DC current using standard preparation method [23]. A suitable CCnT-bearing object is then selected (Fig. 2a). A micro tweezer, is then used to grab the CCnT-bearing object (Fig. 2b) which is cut with the ion beam (Fig. 2(c) and (d)). The tungsten tip is also cut with the ion beam to create a flat surface (Fig. 2e) to easily weld the half CCnT-bearing object on it. The gasinjection system (GIS) is then used to seal the CCnT onto the end of the tungsten tip (Fig. 2f) resulting in the structure shown in Fig. 2h where the CCnT is attached at the apex of the standard FE W tip.

4. Field emission properties of the new carbon cone nanotip In order to determine the main physical properties of the carbon cone involved in FE (exit work function, field enhancement factor, etc.), in situ field emission TEM experiments have been performed using a dedicated Nanofactory holder (double tilt, with field emission, STM and nanoindentation capabilities). In situ field emission experiments were carried out in an image Cs-corrected Schottky-FEG FEI Tecnai F20, operated at 100 kV. A TEM overview micrograph is reported in Fig. 3. The CCnT mounted as described in the previous paragraph is set and electrically connected on the piezoelectric actuator, which allows moving it. The CCnT is facing to a gold electrode. The distance “d” between the tip and the anode was set to less than 1 μm. The voltage between the tip and the anode is increased until the onset voltage is reached and the cold-field emission process starts. The emitted electrons current is collected by the positive anode and recorded. Holograms were acquired around the CCnT apex every 10 V to analyze the local changes of the electrostatic field before and during the field emission process.

Please cite this article as: F. Houdellier, et al., Development of TEM and SEM high brightness electron guns using cold-field emission from a carbon nanotip, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.11.021i

F. Houdellier et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

3

Fig. 2. CCnT emitter preparation in the FIB-SEM. (a) Overview of the carbon tape with CCnT-bearing objects deposited. (b)–(d) Cutting one CCnT-bearing object in two parts and securing it with the micro tweezers. (e)–(h) The CCnT being transported and welded to the etched then truncated W wire.

4.1. Fowler–Nordheim curve Fig. 4c shows the field emission ie(V) curve that was obtained during the in situ field emission process [24]. The voltage V on the

anode was increased smoothly by 1 V/s. Holograms were acquired after each voltage step. As can be seen in Fig. 4c, the data is quite noisy due to the low vacuum around the CCnT in the objective lens of the TEM (some 10
5 Pa). However this noise, due to collision of

Please cite this article as: F. Houdellier, et al., Development of TEM and SEM high brightness electron guns using cold-field emission from a carbon nanotip, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.11.021i

F. Houdellier et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

4

4.2. In situ electron holography Electron holography is an interferometric TEM technique, which allows measuring the phase shift of an electron wave experienced when interacting with an electrostatic and a magnetic field [27]. When interacting with the three dimensional electrostatic potential V(x,y,z) only the phase shift is expressed as:

φ (x , y ) =

∫beampath CEV (x, y, z)dz

(5)

where CE is an interaction constant depending of the electron beam energy [27]. The phase shift measurement therefore allows mapping the local two dimensional projected electrostatic potential V(x,y) as it is illustrated in Fig. 4a. The electric field around the tip for each bias can then be mapped as show in Fig. 4b and compared to simulations. Electron holography experiments were performed for various tipanode distances and electric bias. For the experiment reported here, the tip-anode distance was d¼680 nm. The voltage V between the CCnT and the anode was increased by 1 V steps, from 0 V to 95 V, with one step per second. Holograms (2048  2048 pixels) were recorded with an exposure time of 5 s, every 10 V or less. 4.3. Exit work function of CCnT tips Fig. 3. TEM micrograph showing the overview of the CCnT facing the anode. The SEM inset image shows the micro tweezers releasing the CCnT after having been attached to a cut W wire.

various molecules with the tip during emission, will not affect the holography results and the corresponding measurements of the electric field strength. Fig. 4d reported the variation ln(ie/V2) versus 1/V (Fowler–Nordheim plot). The local electric field Eloc in the following Fowler–Nordheim (F–N) ie(Eloc) equation is the local field around the emitting tip (see [24]).

ie = A

⎛ 6.44 × 109ϕ 3/2 ⎞ ⎛ 10.4 ⎞ 1.54 × 10−6 2 ⎟⎟ ⎟exp⎜⎜ − Eloc exp⎜ ϕ Eloc ϕ ⎝ ⎠ ⎝ ⎠

4.4. Charge measurements

(1)

where A is the emitting area and ϕ the exit work function of the tip (in eV). Due to the important aspect ratio of the tip, the local field is enhanced around the apex. If we compare this local field with the reference electric field from two parallel plates, E|| ¼V/d, (V being the applied voltage and d the separation distance between the two plates), the field is enhanced by a factor γ known as “field enhancement” factor [25]. The F–N equation ie(V) can then be rewritten as follows:

Eloc = γE|| = γ

ie = A

V d

2 ⎛ 6.44 × 109ϕ 3/2d ⎞ ⎛ 10.4 ⎞ 1.54 × 10−6 ⎛ V ⎞ 2 ⎟⎟ ⎜ ⎟ γ exp⎜ ⎟exp⎜⎜ − ⎝d ⎠ ϕ Vγ ⎝ ϕ ⎠ ⎝ ⎠

(2)

(3)

The slope of the logarithmic F–N plot gives the work function ϕ of the CCnT:

−

6.44 × 109ϕ 3/2d γ

Combining the F–N results of Fig. 4(c) and (d) with electric field measurement obtained by electron holography and quantitatively fitted by finite element simulation (carried out using Comsol Multiphysics software), an exit work function ϕ of 4.83 70.3 eV, which is close to that of standard carbon nanotubes, has been found [28]. This value remains a little bit higher than the 4.34 eV of a [310] oriented tungsten tip, but the high aspect ratio of the CCnT counterbalance this effect regarding the onset of field emission. The details of the simulations and the analysis of the hologram phase can be found in the paper published by de Knoop et al. [26].

(4)

As “V” and “d” are known (the latter can be measured directly by TEM), only γ needs to be found in order to get ϕ. The local electric field Eloc can however be used to calculate γ. To do so, de Knoop et al. have simulated the local electrostatic field obtained by in situ electron holography with finite element modeling [26].

Gatel et al. [29] proposed a smart way to measure the number of electric charges from the analysis of the phase shift maps obtained in electron holography experiments. Their approach combines the Gauss's Law together with the Aharonov–Bohm equation. They demonstrated that the amount of charge Qtot contained within a volume can be determined performing an integration around the projected contour of the volume in the electron phase map. This can be summarized using the formula:

Q tot = −

εo CE

∮C ∇RφQ (R(l)). N(R(l))dl

(6)

where l is a line parameter along which the integration is performed, N is the outward normal and R¼(x, y) is the vectorial component in two dimensions (see [29]). Applying this equation onto the phase maps in Fig. 5a–c the number of electrons within the dotted integration contour can then be measured. The integration path is indicated by the arrows around the contours. Several integrations were performed on contours with increasing size, allowing mapping the distribution of charges within the CCnT (along its length toward its apex) and in the vacuum. The red line in Fig. 5a–c corresponds to the starting position (x ¼0): a first integration has been done (Fig. 5a) then a second (Fig. 5b) etc. until a nth one (n E1800) with a total length of xE 320 nm. This allows retrieving the electric charge distribution along the CCnT as reported in Fig. 5d for various extraction voltages. For example, the total number of electrons for 10 V is 154 74 electrons, 302 74 electrons for V¼20 V. As observed in Fig. 5d, the number of electrons increases with the applied voltage

Please cite this article as: F. Houdellier, et al., Development of TEM and SEM high brightness electron guns using cold-field emission from a carbon nanotip, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.11.021i

F. Houdellier et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

5

Fig. 4. (a) The unwrapped phase map extracted from the 80 V bias hologram. (b) The gradient of the phase, which qualitatively represents the electric field around the CCnT tips. (c) The ie(V)-curve obtained from the onset of field emission at 80 V. (d) Fowler–Nordheim plot with the blue line being a linear fit. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

and once the field emission started at 80 V, the number of electrons increases slower and for 80 and 95 V extraction voltages, values of 1193 75 electrons and 123375 electrons, respectively, were obtained. Using these results, we can easily calculate the electrons surface density (see [32]). New experiments and simulations of these results are under progress using various analytical models and finite element modeling and will be the subject of a dedicated paper.

5. Application of CCnT tips to a field emission gun in SEM CCnT tips have been tested as FE cathodes using a standard Butler field emission gun in a commercial FEG Scanning Electron Microscope (Hitachi SU8200) working at 30 kV and equipped with NEG (Non-Evaporate Getter) pumping system with an extracting anode with a voltage V1 and the accelerating anode with voltage V0 [8]. The tip has been mounted using the FIB procedure described before and reported in [22]. The final picture of one tip

mounted on a standard Hitachi tungsten tip is reported in Fig. 6a. The new tip has then been installed as the cathode of the FE gun of the SU8200, with a 10 mm tip-anode distance. Even for a relatively small emission current of 40 nA, the tip needs to be cleaned to remove contamination and avoid quick and erratic changes in the current. The results presented in the Fig. 6(b) and (c) shows the evolution of the emission current before and after a flash cleaning respectively. After flash cleaning, the emission current appears to be very stable and the noise very small as shown in Fig. 6c. No decay has been observed during one hour with a root mean square (rms) less than 0.5% which are good values compare to state of the art CFEG using [310] oriented tips where the decay can be around 15% for one hour with a rms of 1%. The emission was then increased up to 5 μA applying an extracting voltage V1 of 408 V. Even if it is, as expected, noisier, the emission current remains extremely stable as reported in Fig. 6d and almost no decay was observed. Due to the very small apex of the tip, the field enhancement factor is much higher than for standard tungsten FE tips, as demonstrated by the in situ electron holography experiment [26]. To

Please cite this article as: F. Houdellier, et al., Development of TEM and SEM high brightness electron guns using cold-field emission from a carbon nanotip, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.11.021i

F. Houdellier et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

6

Fig. 5. (a)–(c) Unwrapped 10 V phase map. Counting of electrons as a function of position. The dotted lines with arrows in (a)–(c) show the 1st, 2nd and nth integration contour paths. (d) The number of electrons inside the nth integration path for 0 to 95 V. The black vertical line indicates the CCnT/vacuum interface. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)

obtain an optimized probe current for the SEM electron beam, the ratio of the Butler optics V0/V1 or the distance between the tip and extracting anode need to be adjusted (see Fig. 7). On the SU8200 SEM, none of these two options are easy to tune, and we choose to continue our investigations in a high voltage FE source, namely HF-2000 TEM and HD-2700 STEM, where a third anode (called the “gun lens”) located below the extracting anode allows to easily adjust the focal point of the source relatively to the first condenser lens (C1) of the microscope by adjusting a voltage V2 on it (see Fig. 7). In this case the V2/V1 ratio is tuned to achieve this operation, keeping V0 constant. Work is under progress to determine the best configuration to maximize the probe current by adjusting the height of the carbon tip relatively to the extracting anode using the Butler configuration (see Fig. 7). These first results obtained in a state-of-the-art SEM (namely the SU8200 CFEG of Hitachi) clearly showed the advanced performances of the CCnT as FE cathode regarding the beam stability, which is essential in SEM. As opposed to standard tungsten tips, almost no decay of the emitted current is observed during more than one hour. Furthermore, after a first flash-cleaning procedure, the beam noise remains very low (rms less than 0.5%), even if, as for standard FE tips, it increases with the emission current.

6. High voltage TEM CCnT electron source A 200 kV FE source of a 20-years old Hitachi HF-2000 has been modified in order to test the CCnT tips capabilities and adjust the

electron optics conditions to maximize the ratio between the probe current on the TEM sample and the emission current of the source [22]. The vacuum set-up has been modified using state of the art vacuum technologies adding a NEG pumping system close to the tip and additional differential pumping below the gun valve of the source. This was done to guarantee the best vacuum possible (around 2  10
9 Pa) and really compare the properties of the CCnT tip itself. An image of the modified microscope is reported in Fig. 8a. As for the 30 kV SU8200 source, the tip-anode V1 distance was adjusted around 10 mm. The standard evolution of the emitted current is shown in Fig. 8b. The evolution of the emitted current is reproducible even for different technologies (i.e. high voltage 200 kV HF2000 and low voltage 30 kV SU8200), especially regarding the extraction voltage for which the field emission starts. In order to maximize the probe current, the ratio V2/V1 between the gun lens voltage and the extracting anode has been adjusted with the help of ray tracing simulations done using the electrostatic simulation software Simion [30]. The simulation reported in Fig. 8c shows the position of the cross over inside the accelerating stage when setting the standard ratio at 7, as used with regular W FE tips, and an extraction voltage of 380 V. This extraction voltage leads to an emission current of 1 μA with the CCnT cathode (see Fig. 8.b). The best illumination condition was found tuning the ratio down to 4.5 for which the electron beam was almost parallel at the exit of the gun (see Fig. 8d). Using this condition the first cross-over inside the column, is formed close the back focal plane of the first condensor lens (C1), which maximizes the probe current and minimize the effect of C1 aberrations and therefore optimizes the probe brightness.

Please cite this article as: F. Houdellier, et al., Development of TEM and SEM high brightness electron guns using cold-field emission from a carbon nanotip, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.11.021i

F. Houdellier et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

7

Fig. 6. (a) The CCnT tip used as cathode in a SU8200 SEM, (b) 40 nA emission current without flashing, (c) 40 nA emission current after a flash cleaning procedure and (d) 5 μA emission current where small jumps of some 10 nA can be identified.

Fig. 7. Schematic comparison between low voltage SEM FEG optics (named Butler electron optics) and high voltage TEM FEG where the gun lens V2 need to used before the accelerator to adjust the electrons beam at the exit of the gun.

The experimental results presented in Fig. 9b confirm the better stability of the emitted current from the CCnT compared to a standard tungsten tip using the same emission current of 30 μA, even under a high voltage environment of 200 kV. In this case the calculated ratio of 4.5 has been also used, as this value is not dependent to the emitted current value. Indeed, the gun lens voltage is automatically adjusted when the extraction voltage change in order to keep the same illumination conditions. The two Fresnel images in Fig. 9a correspond to the defocused edge of the condensor 1 aperture both taken under the same conditions (emission current, lenses, cross-over position, etc.) one with the

standard tungsten FE tip and the other with the new CCnT tip. Fresnel fringes can be detected at a much farer distance from the edge of the aperture using CCnT tip than with the standard W tip. The two line profiles presented in Fig. 10 clearly show this effect. This indicates that the spatial coherence of the beam is higher in the case of a CCnT FE source due to the much sharper size of its apex. This property, essential for electron interferometry experiments, is directly related to the higher brightness of the beam. Measurements of the probe brightness, which depends on the convergence angle of the probe, its shape and the probe current, were carried out. A Faraday cup was used to measure the probe

Please cite this article as: F. Houdellier, et al., Development of TEM and SEM high brightness electron guns using cold-field emission from a carbon nanotip, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.11.021i

8

F. Houdellier et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Fig. 8. (a) Modified HF-2000 FE-TEM. (b) Emission current evolution as a function of the extraction voltage V1. (c) Simion ray tracing simulations for a 1 μA emission current with a ratio V2/V1 of 7. (d) Simion ray tracing simulations for a 1 μA emission current with a ratio V2/V1 of 4.5.

Fig. 10. Intensity profile obtained from the line scan of the Fresnel fringes observed in Fig. 9a.

7. Discussion and perspectives Fig. 9. (a) Fresnel fringes observed at 200 kV on the edge of a defocused condensor 1 defocused aperture. (b) Current stability comparison at 200 kV under the emission current conditions. Inset shows the stability of CCnT cathode where small jumps of some 10 nA can be identified.

current, the FWHM of the probe shape was then been determined using a TEM image, and the convergence angle calculated using the CBED disk in diffraction mode of the microscope. The effect of the illumination system aberrations have not been considered, which should lead to a wrong value of the absolute brightness, even when minimizing the convergence angle. However, this brightness measurement was determined and compared to a standard tungsten tip under the exact same conditions. The measurements reported by Houdellier et al. [22], demonstrate that the CCnT source is 3 times brighter than a standard tungsten FE source.

We aim at developing CCnT FE tips as new sources for TEM and SEM applications. Studies and tests have been carried out and are still under progress in a modern dedicated Hitachi HD-2700 STEM. They are extremely promising showing an unprecedented stability with almost no decay over 9 h at acceleration voltage of 200 kV. As the optics of the HD-2700 source is similar to that of the HF-2000 one, optimal optical conditions were easily been found and tests showed that the CCnT source can be used in STEM imaging. Regarding TEM applications, the performances of the CCnT source need to be investigated more deeply for electron holography and high-resolution purposes to better understand the effect of the tip regarding the spatial coherence of the beam. Current experiments are under progress to determine the coherent current Ic as defined by Krivanek et al. [31] and a new measurement of the absolute brightness is also a work in progress.

Please cite this article as: F. Houdellier, et al., Development of TEM and SEM high brightness electron guns using cold-field emission from a carbon nanotip, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.11.021i

F. Houdellier et al. / Ultramicroscopy ∎ (∎∎∎∎) ∎∎∎–∎∎∎

For SEM applications, the good stability of the CCnT tip with no decay of current during one hour can be a very good advantage for practical applications. The main concern will be to find the correct optical conditions using the standard and simple Butler geometry, which have to be used to maximize the beam brightness. This work is also under progress. Regarding quantitative electric field studies of the field emission process, new in situ electron holography experiments have been obtained to better analyze the FE mechanisms just after the field emission starts. Using in situ electron diffraction we are also currently investigating the structural changes of the tip under emission. We hope, that all these studies will explain the origin of small nA jumps in the emission current, which can be seen in Fig. 6d and Fig. 9b when the tip is used both as FE-SEM and FETEM cathodes.

Acknowledgements The authors acknowledge the European Integrated Infrastructure Initiative Reference 312483-ESTEEM2, and the French “Investissement d’Avenir” Program Reference no. ANR-10-EQPX38-01, and the French ANR–Emergence Program ANR-12-EMMA0044-01 for partial funding.

References [1] W.P. Dyke, W.W. Dolan, Field emission, Adv. Electron. Electron Phys. 8 (1956) 89. [2] E.W. Müller, Das Feldionenmikroskop, Z. Physik 131 (1951) 136. [3] L.W. Swanson, A.E. Bell, Recent advances in field electron microscopy of metals, Adv. Electron. Electron Phys. 32 (1973) 193. [4] A.V. Crewe, Scanning techniques for high voltage microscope, in: Proceedings of the AMU-ANL High Voltage Electron Microscope Mtg. (Argonne National Laboratory), 1964, 68. [5] A.V. Crewe, J. Wall, J. Langmore, The visibility of single atoms, Science 168 (1970) 1338. [6] L.W. Swanson, Current fluctuations from various crystal faces of a clean tungsten field emitter, Surf. Sci. 70 (65, 1978) 1. [7] H. Todokoro, N. Saito, S. Yamamoto, Role of ion bombardment in field emission current instability, Jpn. J. Appl. Phys. 21 (1982) 1513. [8] H. Shimoyama, M. Ichihashi, N. Tamura, Emitter and electron gun, J. Electron Microsc. 38 (1989) S1. [9] A. Tonomura, Application of electron holography using a field-emission electron microscope, J. Electron Microsc. 33 (1984) 101.

9

[10] Q.H. Wang, A.A. Setlur, J.M. Lauerhaas, J.Y. Dai, E.W. Seelig, R. Chang, A nanotube-based field-emission flat panel display, Appl. Phys. Lett. 72 (1998) 2912. [11] Y. Saito, S. Uemura, Field emission from carbon nanotubes and its application to electron sources, Carbon 38 (69, 2000) 1. [12] P. Collins, A. Zettl, Unique characteristics of cold cathode carbon nanotube matrix field emitters, Phys. Rev. B 55 (91, 1997) 93. [13] G. Chen, D.H. Shin, S. Roth, C.J. Lee, Field emission characteristics of point emitters fabricated by a multiwalled carbon nanotube yarn, Nanotechnology 20 (2009) 315201. [14] L. Nilsson, O. Groening, P. Groening, O. Kuettel, L. Schlapbach, Characterization of thin film electron emitters by scanning anode field emission microscopy, J. Appl. Phys. 90 (2001) 768. [15] N. de Jonge, J.M. Bonard, Carbon nanotube electron sources and applications, Philos. Trans. R. Soc. A: Math. Phys. Eng. Sci. (2004). [16] N. de Jonge, Y. Lamy, K. Schoots, T.H. Oosterkamp, High brightness electron beam from a multi-walled carbon nanotube, Nature 420 (93, 2002) 3. [17] J.-M. Bonard, H. Kind, T. Stö ckli, L.-O. Nilsson, Field emission from carbon nanotubes: the first five years, Solid-State Electron. 45 (2001) 893. [18] P. Collins, A. Zettl, A simple and robust electron beam source from carbon nanotubes, Appl. Phys. Lett. 69 (1996) 1969. [19] R. Gao, Z. Pan, Z.L. Wang, Work function at the tips of multiwalled carbon nanotubes, Appl. Phys. Lett. 78 (2001) 1757. [20] R.L. Jacobsen, M. Monthioux, Carbon beads with protruding cones, Nature 385 (1997) 211. [21] H. Allouche, M. Monthioux, R.L. Jacobsen, Chemical vapor deposition of pyrolytic carbon on carbon nanotubes. Part 1. Synthesis and morphology, Carbon 41 (2003) 2897. [22] F. Houdellier, A. Masseboeuf, M. Monthioux, M.J. Hÿ tch, New carbon cone nanotip for use in a highly coherent cold field emission electron microscope, Carbon 50 (2012) 2037. [23] O. Albrektsen, H.W.M. Salemink, K.A. Morch, A.R. Tholen, Reliable tip preparation for high resolution scanning tunneling microscopy, J. Vac. Sci. Technol. B 12 (1994) 3187. [24] F.R.S. R. Fowler, L. Nordheim, Electron emission in intense electric fields, Proc. R. Soc. Lond. 119 (73, 1928) 1. [25] W.P. Dyke, W.W. Dolan, Field emission, Adv. Electron. Electron Phys. III (1956) 129. [26] L. de Knoop, F. Houdellier, C. Gatel, A. Masseboeuf, M. Monthioux, M.J. Hÿtch, Determining the work function of a carbon-cone cold-field emitter by in situ electron holography, Micron 63 (2014) 2. [27] A. Tonomura, Electron-holographic interference microscopy, Adv. Phys. 41 (1992) 59. [28] P. Liu, Q. Sun, F. Zhu, K. Liu, K. Jiang, L. Liu, Q. Li, S. Fan, Measuring the work function of carbon nanotubes with thermionic method, Nano Lett. 8 (2008) 647. [29] C. Gatel, A. Lubk, G. Pozzi, E. Snoeck, M. Hÿ tch, Counting elementary charges on nanoparticles by electron holography, Phys. Rev. Lett. 111 (2013) 025501. [30] 〈http://simion.com〉. [31] O.L. Krivanek, T.C. Lovejoy, N. Dellby, R.W. Carpenter, Monochromated STEM with a 30 meV-wide, atom-sized electron probe, Microscopy 62 (3, 2013). [32] L. de Knoop, Université Paul Sabatier Toulouse III, PhD Thesis, February 2014.

Please cite this article as: F. Houdellier, et al., Development of TEM and SEM high brightness electron guns using cold-field emission from a carbon nanotip, Ultramicroscopy (2014), http://dx.doi.org/10.1016/j.ultramic.2014.11.021i