Journal of Microscopy, Vol. 203, Pt 2, February 2001, pp. 188±194.

Received 25 August 2000; accepted 11 December 2000

Microscopy and computational modelling to elucidate the

enhancement factor for field electron emitters

C. J. EDGCOMBE* & U. VALDREÁ ²

*PCS/MP Group, Cavendish Laboratory, Madingley Road, Cambridge CB3 0HE, U.K. and Granta
Electronics Ltd, 25 St Peter's Rd, Coton, Cambridge CB3 7PR, U.K.
²INFM and Department of Physics, University of Bologna, via Irnerio 46, I-40126 Bologna, Italy

Key words. Carbon contamination, carbon nanotips, field emitters, field

enhancement factor, field factor, finite element computation, Fowler±Nordheim
plots, reduction factor, scanning electron microscopy, threshold field, threshold
voltage.

Summary (i) Theoretical and computational papers; (ii) Experimental

work; (iii) Combined computational and experimental
We report on the computation of the electric field at the
approach.
surface of single-tip field emitters for a variety of geometries
It is usually difficult, if not impossible, to compare the
and wide range of geometrical parameters. In conjunction
results of (i) with those of (ii). For instance, the field quoted
with experimental work, this has allowed the determination
in experimental work is sometimes a macroscopic field Fm
of quantities useful for characterizing and comparing the
(also incorrectly called applied field), when this can be
performance of field emitters. The ratio of the field at the tip
calculated easily (see section 3.1). In planar geometries, Fm
surface to field at a tip supporting base (enhancement
is obtained by dividing the voltage difference V between tip
factor) has been calculated for hemispherical tips with
and extraction electrode (the applied voltage) by their
parallel or conical shanks, for ratios of tip length to tip
separation d:
radius from 1 to 3000. Enhancement factors greater than
1000 are achievable with suitable tip geometry. The Fm ˆ V/d 1†
threshold voltage dependence on the tip±anode separation
This applies in particular to geometries used for flat panel
for cylindrical tips facing a flat anode has also been
displays. In materials such as alumina or coated or etched
calculated and reported.
silicon, the height of asperities (potential field emitting
sites) from the base may be a few nanometres. This is of
the same order as the radius of curvature of the emitting
1. Introduction
regions which are separated from the anode by a distance
Cold emitters (field electron emitters, FEE) and related of the order of 100 mm. Then the field at these regions
devices have become the subject of intense fundamental and may be of the same order as the macroscopic field specified
technological investigations not only for microscopy but by Eq. (1).
also because of their great potential for industrial applica- In theoretical work on single tips, the field of interest, Ft,
tions, in particular as flat panel displays and electron guns. is that at the tip surface (or very near to it, say 1 nm away)
The vast amount of literature recently published on FEEs along the tip axis. In the literature, this field strength is
(for example at recent conferences: Proceedings of the 45th related to other parameters in at least two further ways.
Int. Field Emission Symp., Irbid, Jordan, 1998, Ultramicro- One is via the relation
scopy, 79 (1999), 1±293; Technical Digest 12th Int. Vacuum
Microelectr. Conference, Darmstadt, Germany 1999; Abstracts Ft1 ˆ V/ka 2†
of 1st European Field Emission Workshop, Toledo, Spain where a is the radius of curvature of the tip and k is a
1999; Abstracts of 2nd European Field Emission Workshop, dimensionless constant (named here the reduction factor).
Segovia, Spain, 2000) can be divided into three categories: For long tips supported on a distant conducting base and
Correspondence: Dr C. J. Edgcombe, PCS/MP Group, Cavendish Laboratory, placed near the anode, k has a value of about 5 (Gomer,
Madingley Road, Cambridge CB3 0HE, U.K. Tel.: 144 (0)1223 337335; fax: 144 1961), but for short tips with length equal to 100 times the
(0)1223 337333; e-mail: cje1@[Link] end radius, k can be of the order of 100 (Edgcombe & ValdreÁ,

188 q 2001 The Royal Microscopical Society

 E N H A N C E M EN T FAC TO R F O R F I E L D E L E C T RO N EM I T T E R S 189

1998). The second relation used (Dyke & Dolan, 1956; p. ordinary SEM (Antognozzi et al., 1997). We need to grow
97) is tips perpendicular to a substrate and to measure their
length and apical radius before and after use in the
Ft2 ˆ bV 3†
apparatus for field emission experiments. For this purpose
where b, a geometrical field factor, is a function of the shape we use a special but simple holder (Fig. 1), which also
and size of the tip support and the tip itself. Here b is a allows observation at the eucentric height. The tips are
quantity with dimension of length21. Values for b can grown at the top of a tungsten ball with a diameter of about
range from 104 to above 107 m21. In the following, Ft1 and 150 mm. The tungsten balls are obtained by fusing ordinary
Ft2 will be indicated by Ft. electron microscope filaments; they present a very smooth
The ratio Ft/Fm is called the field enhancement factor and surface, not likely to give rise to field emission under our
is dimensionless. It has also been represented by b in the experimental conditions (Edgcombe & ValdreÁ, 1998).
literature, and the two possible uses of b are liable to lead to Growth and observation of the tip are performed in a
confusion. Here, for consistency with notation used Philips SEM 501. The tungsten ball (not discernible in
recently, the field enhancement factor is denoted by g: Fig. 1) is at the end of the shank W, in turn supported by a
glass socket G. The tip must be grown at the top of the ball
g ˆ Ft /Fm 4† when its stem is placed vertically inside the microscope. The
It can be found for a particular geometry by computation. tilting stage of the PSEM 501 allows only 108 tilt clockwise
Other definitions frequently used, although completely (CW) and 608 tilt anti-clockwise (ACW) from the horizontal
arbitrary and not formally accepted, are those of threshold position of the traverse stage. In order to be able to observe
voltage Vth and threshold field Fth. They specify the voltage the tip through 908, it is necessary to keep the glass socket
and field required to obtain a stated emission current, inclined by 308 with respect to the traverse stage. For this
typically 1 nA. purpose a special SEM stub is used (Fig. 1). By combining
Reported values for g range from a few tens to ten the 608 ACW maximum tilt angle with the rotation around
thousand or more (Xu, 1995; Davydov et al., 1999). The the stub axis, the stem W supporting the ball is brought
knowledge of g enables the field Ft at the surface of the parallel to the microscope axis. The area selected for the tip
emitting tip to be found from the macroscopic field, hence b growth is then placed at the eucentric height. The progress
from Eq. (3) and k from Eq. (2) if a is known (and, vice versa, of the growth is checked by intermittently measuring the tip
a can be found if k is known). As Ft is one of the quantities length. For this operation the beam scanning is made
present in the Fowler±Nordheim (F±N) equation, knowl- operative to form an image at low magnification, the tilt is
edge of its value enables evaluation of the other quantities. brought to zero (i.e. traverse stage horizontal), the specimen
This paper describes the calculation of Ft and g for a
range of configurations and of Vth for Fth ˆ 109 V/m as a
function of the tip±anode separation s, and reports on some
applications.
It should be stressed that it is only by a simultaneous use of
computational work, electron microscopy observations and
the production of F±N plots that it is possible to determine
important but sometimes unknown quantities such as the
work function of the tip f, the tip length L or radius a and
the emitting surface area. For instance, if f is known, as is the
case for tungsten, Ft and g can be found experimentally from
the F±N plots. Alternatively, if f is not known, as with some
forms of carbon, it can be found by relying on computation of
Ft and g and by use of F±N plots (Edgcombe & ValdreÁ, 2000).
The same approach can be used to test the validity of certain Fig. 1. The simple jig H attached to an ordinary SEM stub S is used
assumptions (see Section 4). The values of g, calculated for for the growth of carbon contamination tips; it allows the
measurement of tip diameter and tip length without breaking the
specific dimensions with an accuracy of about 0.5%, have
vacuum to align the tip axis perpendicular to the electron beam. It
been fitted by approximate expressions which can be used to
is also used to control the tip length (see text). The carbon tip is
find g for other dimensions within a few percent. grown on a tungsten ball (not visible) formed at the end of the
tungsten wire W of 125 mm diameter. G is the glass socket of a
2. Experiments Philips microscope filament. Wire W is spot-welded to one of the
two electrical leads L carried by G. C is a graphite cup, which has a
Our experimental work refers to the characterization of double function of enhancing the contamination for the tip growth
carbon tips we produce by carbon contamination in an and of preventing the charging of glass socket G.

q 2001 The Royal Microscopical Society, Journal of Microscopy, 203, 188±194

190 C . J. E D G C O M B E A N D U. VA L D R EÁ

is rotated by 1808, then the stage is tilted 308 ACW in order

to observe the tip perpendicularly to its axis. This procedure
is repeated until a tip of the specified length and shape is
obtained. Because of the limited resolution of an ordinary
SEM, the measured value of the tip apex diameter is affected
by a large uncertainty estimated as 20±30%.
The socket G with the tip has until now been transferred
to the UHV apparatus for the field emission experiments.
When necessary, the socket is re-mounted in the special
stub, which allows easy finding of the tip for observation in
the SEM, in particular to check for possible changes in the
tip geometry as a result of operation or discharges.
Fig. 2. Computed field enhancement factor g ˆ Ft/Fm as function
3. Computation of tip length/radius for a tip of semi-angle 208, on a ball of radius
63 mm. Between the tip and the front face of the hollow cylindrical
The axial field near the surface of various tip configurations anode of bore 1 mm, the spacing s was maintained at 0.1 mm. The
with cylindrical symmetry has been calculated by the use of `total' tip length L is the sum of the conical length of the shaft and
the program CIELAS2 (Granta Electronics, Cambridge, the tip radius a. The geometry of tip and anode is shown
U.K.), based on the finite element method. The geometries schematically in the inset.
have been chosen to model those found most often in the surface of the tip and the front face of the anode,
experimental work. They have included cylindrical and whereas for the planar anode and cathode (Fig. 5) the same
frusto-conical shanks ending with a hemisphere, for which spacing was maintained between the planar surfaces. The
our measured quantities can be directly and accurately tip radius was kept constant at 10 nm in most cases, but
related to computed ones. Calculations have been made was varied in some early runs to determine that the
with a hollow anode of bore 1 mm and with a planar anode. enhancement factor depended on the ratio (total tip length
L/radius a). The relations between g and (L/a) were of
3.1. Determination of Ft and Fm similar form for the cylindrical shank between parallel
planes (as for nanotubes) and for the cylindrical or conical
For accuracy in calculating Ft and the current density, the shank mounted on a ball near a pierced anode. These
mesh-point spacing of finite elements near the tip should be relations can be fitted by expressions of the form
small compared with tip dimensions. At the same time it is
desirable to represent, in the same geometry, the anode of g , b ‰ L/a† 1 hŠC 5†
much larger dimensions. This can be achieved with a
in the range 1 # L/a # 3000. The behaviour of g for (L/
suitably graded mesh. The field on axis at the hemispherical
a) @ 1 shows that the exponent C is in the range 0.7±0.9
surface of a tip whose radius of curvature is a can be found by
for all the configurations tested; it is smaller for the conical
first determining the voltage V1 (relative to the tip) at a
tips than for cylindrical ones. The parameter h can be
distance dz from the tip. Then the field at the surface is given
approximately by (V1/dz) (1 1 dz/a). This expression is exact adjusted to fit the curvature of the graph at small values of
for a completely spherical system. Comparison of the field
calculated in this way for a tip with that for a pair of
concentric spheres with similar geometry and meshing
showed agreement within 0.5%. Values of Ft have been
computed as a function of tip length L, tip radius a and tip
taper angle a for single tips standing on a ball (Figs 2±4)
(Edgcombe, 1999; Edgcombe & ValdreÁ, 2000) and on a flat
plane (Fig. 5). For non-planar geometry, the magnitude of Fm
is obtained similarly by finding the voltage near the base, after
the tip structure has been removed from the computation.

3.2. Determination of g
Fig. 3. Computed field enhancement factor as function of tip
In our calculations we used tips of length ranging from length/radius for a tip of semi-angle 108, on a ball of radius 63 mm,
10 nm to 30 mm. For the geometries with hollow anode near a hollow cylindrical anode as for Fig. 2. The geometry of tip
(Figs 2±4) a spacing of 0.1 mm was maintained between and anode is shown schematically in the inset.

q 2001 The Royal Microscopical Society, Journal of Microscopy, 203, 188±194

 E N H A N C E M EN T FAC TO R F O R F I E L D E L E C T RO N EM I T T E R S 191

Computations have been carried out for both plane and

pierced anodes, at differing spacings from the cathode. Our
calculations showed that, with the tip-to-anode distance s
equal to 100 mm or more, the effect of varying this distance
on the ratio Ft/Fm was small. With the values given in
Figs 2±5 for g and an anode voltage of 1300 V (our
threshold voltage), the magnitude of electric field near the
tip surface ranges from 0.039  109 V m21 to
24  109 V m21.
These results, together with field emission (I±V) data,
have been used to obtain a consistent set of values for L, a
and the work function f, when combined with an improved
version of the F±N equation. This includes a linear
Fig. 4. Computed field enhancement factor g ˆ Ft/Fm as function
approximation for the term v(y) introduced by Nordheim
of tip length/tip radius for a cylindrical tip on a ball of radius
(Dyke & Dolan, 1956; p. 93) to account for the change of
63 mm, near a hollow cylindrical anode as for Fig. 2. The geometry
of tip and anode is shown schematically in the inset.
potential distribution caused by electronic charges near the
surface (Edgcombe, 1999; Edgcombe & ValdreÁ, 2000). The
outcome is a work function of 3.8 eV for our carbon tips
(L/a). For the cylindrical shank between our parallel planes, and an effective emitting area of 5.0  10216 m2.
and in the range 4 # L/a # 3000, g is represented within
^ 3% by
3.3. Determination of Vth
g < 1:2‰L/a 1 2:15Š0:90 6†
For a given geometry it is often desirable to know the anode
As a rule of thumb for the same geometry, the enhancement
voltage that must be applied to produce a given value of Ft,
factor is given within 25% by
typical of that needed for the onset of field emission. In the
g , 0:7 L/a† 7† literature, results are often given for the threshold field Fth and
the related threshold voltage Vth needed to obtain a specified
for 30 # (L/a) # about 2000. Within this range lie
current such as 1 nA (see Section 1). The way in which Vth
nanotube emitters as well as our configurations. Another
varies with the nominal anode-cathode spacing s (measured
rule of thumb,
along the axis of the system) depends on the geometry. When
g , 2 1 L/a the electron beam is to be transmitted through the anode, as for
as given by Xu (1995), provides a closer fit for smaller microscopy, typically a tubular anode of internal diameter
values of L/a. Relations similar to (6) for other geometries about 1 mm is used, with s set to 0.1 mm (Figs 2±4). In this
are shown in Figs 2±4; for L/a , 1000 they fit the geometry, reduction of s to zero makes little change in the
computed values within 7% for a 108 cone or a cylindrical shortest distance from tip surface to anode, and so Vth is
tip on a ball, and within 12% for a cone of 208 semi-angle. expected to vary little with s. This has been verified by
computation. In contrast, experiments on field emitters for
displays commonly use a flat anode surface (Figs 5 and 6), for
which s is identical with the shortest distance from tip to anode.
In this case, Vth can be expected to vary nearly linearly with s
when s is much greater than the tip length L.
Calculations have been made for this latter geometry, to
find the variation of Vth when s is comparable with, or
smaller than, the tip length L, and a result is shown in
Fig. 6(a,b) for Fth ˆ 109 V m21. As s decreases to become
comparable with the tip length, the slope of the graph of log
Vth vs. log s decreases from 1 to about 0.5. Then as s
decreases further, towards the tip radius, the slope increases
again.
Fig. 5. Computed field enhancement factor as function of tip
length/radius for a cylindrical tip between planar electrodes. The 4. Discussion
distance d between the planes was kept at 0.1 mm and the tip
radius at 10 nm. The geometry of tip and anode is shown In work on field emission, the experimental data usually
schematically in the inset. available are: current and voltage readings, spacing d

q 2001 The Royal Microscopical Society, Journal of Microscopy, 203, 188±194

192 C . J. E D G C O M B E A N D U. VA L D R EÁ

by using the analytical result for concentric spheres: Fm < V

D/b (D 2 b) ˆ 18110 V (V m21). The value of g for this
geometry is 101, so the calculated b is 18110
 101 ˆ 1.82  106 m21. The voltage needed to obtain
measurable emission was about 1300 V, and thus the
calculated threshold field is about 2.4  109 V m21, in
good agreement with the known range of field for emission
into vacuum (Gomer, 1961, p53).
Planar geometry is mostly used for studies of large-area
emission. Consequently, detailed measurements of both
geometry and I±V relation are not often reported for an
isolated emitter of the shape considered here, on a plane
base. However, in an array of emitters, the individual tips
may be spaced far enough apart to avoid screening each
other, and then the results for single emitters can be
applied.
De Heer et al. (1995) found that when the I±V
characteristics of an array of carbon nanotubes were fitted
to the F±N equation and expected work function, values of
g in the range 500±1300 were needed for consistency,
rather than the range 5±10 estimated for gated molybde-
num tips. This can be understood by considering the shapes
of the nanotubes. The lengths were of the order of 1 mm,
while the radii of some tips were less than 1 nm. Figure 5
predicts that when L/a is in the range 1000±2000, the
Fig. 6. (a) Computed threshold voltage Vth (in volts) for a threshold
values of g will be in the range 600±1000, which is in
field of 109 V m21 as a function of tip-anode spacing s (in mm) for agreement with the values deduced by de Heer et al. The
cylindrical tips in front of a flat anode with L/a ˆ 100 (L ˆ 1 mm, density of nanotubes over their emitting area was about
a ˆ 10 nm), shown with logarithmic scales; (b) As for (a) but with 108 cm22, suggesting that the average spacing was about
linear scales to show Vth for low values of s, comparable to the 1 mm, similar to the tip length. Preliminary calculations
apical radius a (ˆ 10 nm) of the tip. suggest that at this tip spacing, the tip field is not greatly
reduced by screening.
between cathode and anode, electrode geometry and some The emitting arrays described by Collins & Zettl (1997)
details of the emitter. Computation can provide in addition showed F±N plots with two regions of different slopes. It
values of Ft for a given voltage V, then b ˆ Ft/V and finally appears from the description that the emitters protruded
g ˆ Ft/Fm, once Fm has been calculated (as in Section 3.1) from the surface of the epoxy support by about 1 mm and
for the particular configuration. For planar geometry (see some tip radii were as small as 1 nm. With the
Fig. 5) assumption that the surface of the epoxy base was
approximately an equipotential, the field enhancement
g ˆ Ft d/V ˆ Ft L 1 s†/V: 8†
factor corresponding to these dimensions would be about
This is equivalent to saying that for planar geometry, b ˆ g/d, 600, and with a cathode-to-grid spacing of 50 mm, b
with the value of g given by Eq. (7) or (8) or Fig. 5. Thus, if would be about 1.2  107 m21. Then from the observed
the geometry of the emitter and its support corresponds to value of slope in the low-field region, the consistent value
one of Figs 2±5, values of Ft can be estimated from values of of work function can be calculated as about 2.8 eV.
V, supplying one quantity in the F±N relation. Although this value is probably too small, it shows that a
The following examples show how the computed value of lower value can be obtained than the 12 eV, which was
g can be used to help comparison of quantities and to check found by assuming a reduction factor of 10. Collins and
consistency. Zettl also found that the emission characteristics were not
A conical tip with L/a ˆ 200 was grown as described very sensitive to variation in cathode-to-grid spacing d.
above on a ball of radius b ˆ 63 mm. The surface of the ball This may perhaps be an example of the hollow-anode
was placed 0.1 mm from the face of a hollow anode of bore effect described above in section 3.3. When d is less than
1 mm (as in the inset of Fig. 3). The smallest distance the spacing between grid wires, the distance D from a
between anode and cathode surface, D, was thus about point on the cathode to the nearest grid wire varies more
510 mm. For this geometry, an estimate of Fm can be made slowly than d, and so the field at the emitting point also

q 2001 The Royal Microscopical Society, Journal of Microscopy, 203, 188±194

 E N H A N C E M EN T FAC TO R F O R F I E L D E L E C T RO N EM I T T E R S 193

varies more slowly than if it were inversely proportional threshold voltage has been found for a given material in one
to d. of the geometries for which g is known, the threshold
voltage for the same material in another geometry of known
g can also be estimated.
5. Conclusions
We envisage the modification of a scanning electron
The following points and results have been presented and microscope in Bologna in order to perform in-situ emission
commented upon. experiments on tips, arrays of tips and films. This implies the
(i) We have proposed unambiguous definitions of incorporation of a micromanipulator to provide additional
quantities encountered in experimental and theoretical degrees of freedom for the setting of the anode±emitter
work to ease the comparison between the results obtained relative spacing and to improve the SEM vacuum, if found
by different researchers. necessary (Bishop et al., 2000; Wilshaw et al., 2000).
(ii) We have emphasized the necessity of combined Investigations in a SEM will allow both the growth of the
experimental, theoretical and computational investigations tips, the characterization of the emitters, the control of the
in order to derive several important unknown quantities anode±tip distance and the direct assessment of the tip
characterizing the field emission properties of new materi- shape and size before and after emission. Our SEM has an
als. The experimental work is performed, in our case, both ultimate spot size of 5 nm, which roughly corresponds to
inside electron microscopes, mainly for growing carbon tips the minimum obtainable radius a; in addition the above spot
and/or measuring geometrical parameters (e.g. tip length size limits the accuracy of the measurements of the radius
and apical radius), and in dedicated equipment for itself and of the tip length L. It is therefore advisable to
collecting field emission data (e.g. Fowler±Nordheim plots) perform experiments in SEMs equipped with a LaB6 electron
under ultra high vacuum (5.1029 Torr). source or, even better, with a field emission gun; this will
(iii) The field Ft at the surface of a tip placed in front of also bring the bonus of a better vacuum in the specimen
a hollow anode is practically the same as that for the chamber.
same tip and base placed in front of a flat anode, as long
as the tip±anode distance s is much greater than the tip
Acknowledgement
length L.
We have calculated the magnitude of the electric field Ft We thank Dr A. Alessandrini for assistance in the growth of
at the surface of a tip, on its axis, for tips of varied sizes and the carbon tips.
geometry (cylinder and frusto-cone), supported on a plane or
on a sphere with a long shaft. The computed values of Ft
References
have then been used to calculate the enhancement factor g.
The results show a rather simple relation between g and the Antognozzi, M., Sentimenti, A. & ValdreÁ, U. (1997) Fabrication of
ratio L/a (total tip length L divided by the tip apical radius a). nanotips by carbon contamination in a scanning electron
(iv) For a given value of L/a, the configuration which microscope for use in scanning probe microscopy and field
emission. J. Microsc. Microanal. Microstruct. 8, 355±368.
achieves the highest value of g, especially for L/a greater
Bishop, H.E., Burden, A.P. & Tuck, R.A. (2000) Characterisation of
than 1000, is the cylindrical shank (ending with a
field emitting structures by in-situ scanning electron microscopy.
hemisphere) placed between planar electrodes (Fig. 5). If
Abstracts of the Meeting on Microscopy for the Development of
the shank is conical, its angle has a substantial effect on g, Displays, Rutherford Appleton Laboratory, Chilton, UK, 13 June
for L/a greater than 100. 2000. Royal Microscopical Society, Oxford.
(v) The plots of log g vs. log (L/a) are practically linear for Collins, P.G. & Zettl, A. (1997) Unique characteristics of cold
L/a greater than about 30 for cylindrical tips, whereas for a cathode carbon-nanotube-matrix field emitters. Phys. Rev. B, 55,
conical tip on a ball the slope of the log/log plot decreases 9391±9399.
above L/a , 1000. Davydov, D.N., Sattari, P.A., AlMawlawi, D., Osida, A., Haslett, T.L.
(vi) By computation we have found the dependence of the & Moskovits, M. (1999) Field emitters based on porous
threshold voltage Vth for Ft ˆ 109 V m21 from the tip-to- aluminium oxide templates. J. Appl. Phys. 86, 3983±3987.
anode distance s in the case of a cylindrical tip of length L De Heer, W.A., ChaÃtelain, A. & Ugarte, D. (1995) A carbon
nanotube field-emission source. Science, 270, 1179±1180.
ending with a hemisphere of radius a, supported by a plane
Dyke, W.P. & Dolan, W.W. (1956) Field emission. Adv. Electronics
and placed at a distance d from a flat anode. The plot of log
Electron Phys. 8, 90±187.
Vth vs. log s shows roughly two main regions; for s @ L, the
Edgcombe, C.J. (1999) The electron optics of cold field emitters.
slope is about 1, while for s , L the slope is about 0.5. In Proceedings EMAG 1999 (ed. by C. J. Kiely), pp. 347±350.
the case of hollow anodes of bore radius much greater than Institute of Physics Conference Series 161, Bristol.
a, Vth is insensitive to s. Edgcombe, C.J. & ValdreÁ, U. (1998) Determination of F-N
In general, the relation between the threshold current parameters for carbon contamination grown nano-tip field
and Fth depends on the emitting material. Where the emitters: a combined experimental and computational approach.

q 2001 The Royal Microscopical Society, Journal of Microscopy, 203, 188±194

194 C . J. E D G C O M B E A N D U. VA L D R EÁ

Proc. Int. Centennial Symp. on the Electron, Cambridge, September, Use of an SEM to characterise the electron emission
1997 (ed. by A. Kirkland and P. D. Brown), pp. 318±325. IOM properties of field emission arrays. Abstracts of the Meeting
Communications Ltd, London. on Microscopy for the Development of Displays, Rutherford
Edgcombe, C.J. & ValdreÁ, U. (2000) Field emission and electron Appleton Laboratory, Chilton, UK, 13 June 2000. Royal
microscopy. Microsc. Microanal. 6, 380±387. Microscopical Society, Oxford.
Gomer, R. (1961) Field Emission and Field Ionization. Harvard Xu, N.S. (1995) The physical origin of pre-breakdown electron
University Press, Cambridge, Mass. `pin-holes'. High Voltage Vacuum Insulation (ed. by R. V. Latham),
Wilshaw, P.R., Boswell, E., Huq, E., Holland, E. & Li, Y. (2000) pp. 115±164. Academic Press, London.

q 2001 The Royal Microscopical Society, Journal of Microscopy, 203, 188±194