Applied Surface Science 45 (1990) 229-245 229

North-Holland

Thermal analysis of a cathode: temperature distribution

and warm-up time

A n j a n Sil *, N . K . S a m r i a ** a n d D.S. V e n k a t e s w a r l u *
Banaras Hindu University, Varanasi 221 005, India

Received 12 October 1989; accepted for publication 8 May 1990

This paper gives a theoretical estimate of the cathode temperature distribution at the surface and the warm-up time. Several cases
of practical interest have been studied from the thermodynamic considerations using the software specially developed for this
application. The software is in two parts, viz., the steady-state analysis which gives the cathode surface temperatures and the transient
analysis which yields the warm-up time. Temperature-dependent thermal conductivity, specific heat, density, and emissivity have
been used in the transient analysis which improves the accuracy. Scaling laws for the determination of cathode surface temperature
have been devised to cater for various sizes of cathodes used in industry. Not reported so far, our results show that a pancake-heater
configuration has a better thermal response but a non-uniform cathode surface temperature profile whereas the conventional
helical-heater configuration has a relatively poor thermal response but a more uniform temperature profile on the cathode. This
aspect gets amplified and is very revealing at larger cathode dimensions.

1. Introduction the authors' knowledge, n o theoretical studies have

b e e n reported o n the cathode surface t e m p e r a t u r e
Steady-state a n d t r a n s i e n t thermal analysis are estimation, T h e p u r p o s e of the present investiga-
i m p o r t a n t from the s t a n d p o i n t of cathode surface tions is to bridge the gap by evolving theoretical
t e m p e r a t u r e d i s t r i b u t i o n a n d w a r m - u p time, re- estimates of the cathode surface t e m p e r a t u r e s a n d
spectively. This s t u d y has a relevance to n o n - u n i - w a r m - u p times from the t h e r m o d y n a m i c consider-
form emission from the cathode which can take ations. T h u s this study would act as a first step to
place either due to different work f u n c t i o n values the designer a n d reduce the e x p e r i m e n t a l time
or different surface temperatures, or both, from considerably.
region to region across the entire surface area of The most c o m m o n l y used ceramic potted
the cathode. A surface t e m p e r a t u r e difference of h e a t e r - c a t h o d e assembly (fig. 1) has b e e n taken
5 0 ° C can give rise to a b o u t 100% t e m p e r a t u r e for our study, with purely radiative b o u n d a r i e s
limited emission increase a n d likewise a work (ignoring c o n v e c t i o n is fully justified in view of
f u n c t i o n increase of a b o u t 0.01 eV can give rise to the o p e r a t i o n in high v a c u u m of the order of 10 -6
a b o u t 10% r e d u c t i o n in emission (estimated for a T o r r or better). N o n l i n e a r i t y is thus i n t r o d u c e d
typical type B [1] cathode operating at 1 0 5 0 ° C i n t o the p r o b l e m , as a result of which the analyti-
with a work f u n c t i o n of 2.1 eV). Some studies cal solutions of the g o v e r n i n g e q u a t i o n s are dif-
have been d o n e to measure the p o i n t - t o - p o i n t ficult to o b t a i n . Therefore a n u m e r i c a l a p p r o a c h
work f u n c t i o n of the cathodes [2-7]. But in all using the finite difference t e c h n i q u e has b e e n ap-
these studies cathode surface t e m p e r a t u r e changes plied to o b t a i n the solution. T h e energy b a l a n c e
are n o t taken into account. However to the best of a n d the rate e q u a t i o n s are used to arrive at the
finite difference a p p r o x i m a t i o n s of the basic dif-
* Department of Electronics Engineering. ferential heat c o n d u c t i o n e q u a t i o n a n d the corre-
** Department of Mechanical Engineering. s p o n d i n g b o u n d a r y conditions. The difference

0169-4332/90/$03.50 © 1990 - Elsevier Science Publishers B.V. (North-Holland)

230 A. Sil et al. / Thermal analysis of a cathode

Z rately when all the interfaces in the thermal path

Cathode
are properly bonded. When the contact surfaces
/ are not bonded, but merely in pressure contact it
[ /" may be expected that large errors will arise de-
Potting l '
pending upon the accuracy with which the thermal
material~
transfer coefficient (surface contact coefficient) is
known. This factor is sensitive to the assembly
Sleeve Z3
geometry, materials, and surface finish, and also
Heater ~.~. depends to some extent on the contact pressure
[9]. The bulk porosity of the cathode and its
impregnants as well as the bulk porosity of the
potting have not been included in the present
analysis, as it would increase the complexity of the
(a) associated equations further. It is therefore as-
sumed that the cathode pellet and the potting are
Z non-porous.
The basic equations are presented for the case
T Cathode where the interfaces between the metal and the
ceramic are ideal which is justified because the
cathode pellet is assembled with the potted heater
Heater, - ~ . : housed in the supporting structure of m o l y b d e n u m
• ".'.',",." .L':,,.....,.: --F
by brazing or heli-arc or laser welding. In actual
material~

Sleeve ~--- ~.
~.~
iiiii!iii:i:f!i,! 1
2 fabrication it is assumed that the potting will be
made by plasma potter so as to achieve the near

L ::::::::::::::::::::::~::~..::.:-fi
ol-r,-t i
ii ideal bonding between the metal and the ceramic.
Even though the analysis and the software devel-
oped are general in nature, in view of the complex-
ities involved, the problem has been solved for
(b) typical geometries in both the steady and the
Fig. l. Potted heater-cathode assembly with (a) helical healer unsteady case to obtain the cathode surface tem-
and (b) pancake heater. peratures and the warm-up time, respectively. The
dimensions chosen for different components of
the assembly are given in table 1. The present
equations are then solved [8] by the iterative analysis also incorporates the steady-state temper-
method for the steady-state case and by the ex- ature distribution on concave cathodes, to take
plicit method in the unsteady-state case. into account the curvature effects of the cathode
The numerical scheme used in the heat transfer surface. The concave cathode parameters are taken
analysis, takes into account the heat losses by in this analysis as that of Vaughan's Pierce elec-
conduction through the cathode support structure tron gun [10]. These parameters and the dimen-
(sleeve) which is an essential part of the heater- sions of other components of the assembly are
cathode package. To account for the cathode sus- given in table 2. The results obtained for the
pension, one would need the precise knowledge of concave cathode case are compared to those of the
the complete assembly (i.e. not only the geometry flat cathode case which are of significance to the
of the heater-cathode assembly should be known, thermal design of the assembly.
but also the geometry of the total tube assembly In practical designs, one uses coaxial heat
should be known in detail). shields around the cylindrical part of the h e a t e r -
In general, it may be expected that the thermal cathode assembly so as to conserve some of the
behaviour of the assembly is described fairly accu- radiative heat flux and thereby improve the ther-
 A. Sil et al. / Thermal analysis of a cathode 231

Table 1 source of Spectra-Mat, Inc., California, are pre-

Dimensions of the components Value sented for the sake of c o m p l e t e n e s s [11].
(nun)
Cathode (tungsten) pellet
Thickness (z 3 - z 2 ) 2
Diameter (2r3) 12 2. Nomenclature

Supporting structure sleeve (molybdenum)

Outer diameter (OD) (2r3) 12 A Area, m 2
Inner diameter (ID) (2r2) 11.6 Ainsh Area of the i n n e r surface of the
Height (z2) 10 heat shield, m 2
Heater coil (tungsten) Asl Area of the sleeve surface, m 2
(a) Helical shape C Heat capacity, J k g - I K -
Diameter (2rI) 6 R a d i a t i o n c o n s t a n t ( = 14360
CI
Length (z~) 8
#K)
(b) Pancake shape
Diameter (2r1) 6 Ex Spectral emissivity c o r r e s p o n d i n g
to wavelength )t
The thickness of the coil is assumed to be negligible in the
cases of heater configuration. Fsl - sh View factor
hr(T) Radiative heat transfer coefficient,
W m -2 K -1
k T h e r m a l conductivity, W m ~ K - 1
/sh Thickness of the heat shield, m
mal efficiency considerably. Suitable b o u n d a r y 0 Rate of heat generation, W m - 2
c o n d i t i o n s for this case u n d e r steady state have 0g I n p u t power, W
b e e n f o r m u l a t e d a n d our software has been suita- r, Ar R a d i a l c o o r d i n a t e s a n d mesh
bly modified to i n c l u d e the same a n d was analysed length, m
too. O u r theoretical results together with the ap- ~r Small i n c r e m e n t of r at r = r, m
p r o x i m a t e t e m p e r a t u r e versus heater characteris- r~ R a d i u s of c u r v a t u r e of concave
tics reported in Technical Bulletin -#118, ion cathode, m
rk C a t h o d e disk radius, m
R c, Rd, R~j Radiative energy from cathode, di-
electric b o t t o m a n d sleeve surfaces,
W
Table 2 Total radiative energy, W
R t
Dimensions of the components Value S(r, z) Space f u n c t i o n
(ram) t, A t T i m e a n d size of time step, s
Concave cathode (tungsten) T ( r , z, t) T e m p e r a t u r e field in the body, K
Radius of curvature, rc 15.215 Tamb S u r r o u n d i n g temperature, K
Disk radius, r k 8.026 rB Brightness t e m p e r a t u r e of the
Thickness at the centre 1.883
cathode surface, K
Cathode pellet (tungsten) riosh T e m p e r a t u r e of the i n n e r surface
Radius 8.026 of the heat shield, K
Thickness (uniform) 2.991
rough T e m p e r a t u r e of the outer surface
Sleeve (molybdenum) of the heat shield, K
OD 16.0 T e m p e r a t u r e of the sleeve surface,
ID 14.0 G
Height 4.0 K
L True t e m p e r a t u r e of the c a t h o d e
Heater (tungsten) coil of pancake shape surface, K
Diameter 12.0
dv Volume, m 3
232 A. S i l et aL / T h e r m a l analysis o f a c a t h o d e

z, Az z coordinate and mesh length, m Assuming that the e n v i r o n m e n t is an isother-

8z Small increment of z at z = z, m mal heat sink at T~mb, the following b o u n d a r y
conditions are written as
Greek symbols"
_ k d ( T ) OT(r, O, t)
O~ Scaling factor Oz
y T h e r m a l diffusivity, m 2 s 1
= O £ d ( V ) [ T 4 ( r , 0, t) - r : 4 b ] , 0<r<r 2
£ Emissivity of the surface material
p Density, kg m - 3
(3a)
0 S t e f a n - B o l t z m a n n constant, W
m-ZK 4 for the dielectric b o t t o m surface,
6(r - q), - k ~ , ( T ) OT(r, O, t)
Dirac delta functions
~ ( Z - - Z1) Oz
~r,r I '
K r o n e c k e r delta = o % , ( T ) [ T 4 ( r , 0, t) - T~4nb], r2<r<r 3
~z,z 1
u( r - rl) , (3b)
Unit step functions
U ( Z - - Z1)
and
Subscripts"
_ k s l ( T ) O T ( ~ , z, t)
i, j At or associated with element Or
(i, j ) 4
= o % I ( T ) [ T 4 ( F 3 , Z, ' ) -- Tamb] , O<Z<Z 2
c, d, sl, sh Used for cathode, dielectric, sleeve
(3c)
and heat shield
for the sleeve surface,
Superscripts
0T( r, z 3, t )
p,p+l Used for temperature at time t -k~.(T) Oz
and t + A t
= o ¢c(T)[T4(r, z 3, t ) - T~4mb], 0 < r < < ~
(3d)
3. Analysis
and
The basic heat conduction equation for a rota-
tion-symmetrical system with a constant heat t)
-'k.(T) Or
source is, in cylindrical co-ordinates,
= o %(T) [T4(r3, z, I) - T~,4b], Z2<Z<Z 3
r Or k ( T ) r~7r + ~Tz ~z + QS(r, z)
(3e)
OT
= C ( T ) p ( T ) ~ 7 , for the unsteady case, (1) for the cathode surface.
= O, for the steady case, (2) However, in the presence of coaxial heat shields
surrounding the cylindrical part of the h e a t e r -
where
cathode package Tamb depends upon the radiative
r1
S(r, z ) = ~ - ~ r S ( r - rl) [1 - u ( z - z~)], exchange between the participating radiative
surfaces and thus Tam b becomes a variable in this
for the helical heater assembly, and case. Tamh remains the same for other radiative
r1 exchanges (i.e. cathode surface and dielectric
S(r, z ) = ~ - ~ z S ( Z - Z,) [ 1 - u ( r - rl) 1,
surface to environment). We can therefore set the
for the p a n c a k e heater assembly. following continuity equation to satisfy the heat
 A. Sil et al. / Thermalanalysisof a cathode 233

Table 3
Range of temperatures and values of the coefficients for the approximation of thermal conductivity of c o m p o n e n t materials with an
accuracy greater than 2%

Material k0 fll /32 f13 /34 Temperature

(W/[Link]) (oC-1×10-3) (oC-2)<10-6) (o C 3)<10 9) (oC-4×10-12) range(OC)

Alumina (99.5%) 39.4 - 3.159 5.322 - 4.531 1.526 0 - 800

Tungsten 182.0 - 0.962 0.907 - 0.315 0 0-1300
Molybdenum 139.0 - 0.275 0 0 0 0-1000

transfer process from the sleeve surface to the applied. Thus at three different interfaces we have
outer surface of the coaxial heat shields, as the following equations:
4- i4sh)
~r
z, ,) Jd =ks' (r)rt Or
z, ,)
]~,'
-1 -- %j
+ - - 1+ -1 - (sh
-

- - Asl
-

Esl Fsl- sh (sh Jinsh O<Z<Z2, (8a)

Vinsh- 7"ou.
= k~h l~h k d ( T ) [ ~T(r' z2' t) ] d kc(T) [LT(r, zZ2, t) 1,
= o ~sh(To4sh - T4mb). (4) 0<r<r 2, (88)
The unsteady-state temperature calculation and
considers any significant temperature variation of
all the thermal properties involved. Thus the poly- ksl(T)[ ~T(r' z2' t) ] =kc(T)[ OT(r' z2' t) ]
nomial k(T)
[12] and the radiative heat transfer ~Z sl ~Z c'

coefficients hr(T )
for all the materials have to be r 2 < r < r 3. (8c)
recomputed at each time step chosen before going
on to the next step and the average values of p, C
and ( recommended [13,14] have to be used for
4. S c a l i n g l a w s
corresponding small temperature ranges, is k(T)
given by:
This paper extends the analysis presented herein
k(T) =k0(a + ... (5) in forms which enable predictions to be made on
the effect of the cathode size parameters on the
where ill, f12..... /3, are the coefficients of the
steady temperature profiles of the assembly by
temperature and their values corresponding to the
range of the temperature are given in table 3. applying c o m m o n scaling laws of "constant input
heat power density scaling", although the mesh
However, the steady-state calculation has been
performed with constant thermal properties, using become coarse for larger scaling factors.
the recommended values [13,14] at - 1100 K. The The physical dimensions of the different com-
ponents in the assembly are scaled by a c o m m o n
initial condition is given by
factor a, and power fed to the heater is multiplied
T(r, z, 0) = Tamb. (6) by a 2. In all sizes of the assembly temperature
On the axis of the assembly, i.e. at r = 0, ther- profiles are presented with respect to distances.
mally insulated boundary conditions are assumed. The scaling starts from the dimensions of the
Thus we have, typical geometry of the assembly for which the
numerical formulations have been developed to
OT(O, z, t)/Or=O, 0<z<z 3. (7)
the extreme cases of small and large cathode di-
At the interface between the two materials the ameters, say 1 to 120 m m including proportionate
condition of the continuity in heat flux [15] is variation of the pellet thickness.
234 A. Sil et aL / Thermal analysis of a cathode

The assembly is scaled by the following linear

transformations of both the coordinates r, z and i Z(mm )
input power ( P ) as
12
r" = ra, (9a) 11

z" = za, (9b) 1O

9
P ' = Po~2. (9c)
IB

7
i,]÷l
6
5. S o f t w a r e d e v e l o p m e n t i-l,j [~ j i+l,j
5
iti-1
The problem is basically of three-dimensional
nature with cylindrical coordinates (r, @, z), but 3
the angular symmetry of the geometry of the as- 2
sembly has been exploited to reduce the problem I
into two-dimensional coordinates r and z in eq. (a)
m.
(1). Moreover, the axisymmetric nature of the 1 2 3 ~. 5 5'86 r ( m m ) ~"
cylinder makes it possible to analyse diametrically Fig. 2a. Vertical section of a potted heater cathode assembly
one half of the body, and interpret the results for with mesh generation (84 elements) for steady-state analysis.
the whole body by putting the constraint O T / O r =
Oat r=0.
The finite-difference formulations of the elements and using Gauss-Seidel iterative satis-
governing differential equation (1) and the faction of the boundary conditions. The flow di-
boundary conditions (3a) to (3e) have been arrived agrams of the software packages are given in figs.
at by subdividing the one half of the assembly into 3a and 3b.
84 elements for steady-state analysis and 24 ele-
ments for transient-state analysis. The proper step
sizes in space and time for transient analysis are . Z(mrn)
estimated by carrying out several test calculations t

satisfying the stringent stability and convergence

condition [16]. The time step At that has been 12
used in the transient analysis is 0.001 s.
Figs. 2a and 2b show tw0-dimensional grid 10
structures of the assembly for steady- and tran-
sient-state analyses. Here the general element
(i, j ) is surrounded by elements ( i - 1, j), (i +
1, j ) , (i, j - 1 ) and (i, j + 1). The explicit form i,]+1
6
of the temperatures 7:i,I p+I (subjected to the
boundary conditions (3a) to (3e)) and the output i-~,j i,'j i*l,j
radiative power during the heating-up process are 4

given in the appendix.

The finite-difference form of the steady-state i, j'-I
2
temperature distribution has been derived from
eq. (1) by putting ~ T / ~ t = O. (b)
The system of equations thus obtained have o 2 4 5.B 6 r(-~'-~'m
)
been solved by initiating the calculation with sui- Fig. 2b. Vertical section of a potted heater-cathode assembly
table guess temperature field T(r, z) values at the with mesh generation (24 elements) for transient-state analysis.
 A. Sil et aL / Thermal analysis of a cathode 235

6. Results and discussions

Data viz. Assembly geometry Dimensionsof the \ The present investigation has manifold objec-
"•Readcomponents~grid spacings, constant thermophysical
properties. \ tives, viz. surface temperature distribution of
cathode and other boundaries, isotherms inside
Subdivision of the assembly into the assembly, scaling of geometry, curvature ef-
number of elements.
fects, incorporation of heat shields and warm-up
time calculations. The following sections present
Initialisation of temperaturesat
all the elements. the results and discussions in each of these cases.

-1
Compute new temperature field 6.1. Surface temperature distribution of the cathode
from previous values using
Gauss-Seidet iteration. and assembly

Insert boundary conditions Since the uniformity of surface temperature of

to compute boundary point
temperatures. any cathode is prerequisite to the uniform electron
emission with a very narrow thermal spread of
electron velocity leading to suppression of thermal
_ No ~ yes noise in the vacuum device, the steady-state tem-
(a) perature distribution at the operating temperature
of a standard type B dispenser cathode has been
Fig. 3a. Flow diagram (steady-state analysis). computed. The results have been taken for various
input heater powers to judge the o p t i m u m heater
power required to obtain the desired cathode tem-
perature. Both the conventional helical and the
pancake heater shapes are considered in the analy-
sis to determine if there is any significant dif-
ference in the cathode surface temperature pro-
I,oitiaise
atorialproperties]
Input heater power.
files.
Fig. 4a shows the radial temperature profiles
for various input heater powers. It is seen that the
~ Read: AssemIDlygeometry, \
dimensions of the components,\ cathode assembled with pancake-heater configura-
Grid spacings, Initial \
temperatures at all the \ tion, attains higher surface temperature than that
nodat points. \
with helical heater. This feature perhaps may be
attributed to the fact that the heat flux towards
Estimate: (a) k(T)'s using polynomial
expression of T. [ the cathode is larger for the pancake heater. It is
(D) hr {T) ~s using the cubic [ interesting to notice t h a t the surface temperature
dependence of T. ]
and select the right values of P, C and E / distribution is uniform at all the power levels
for the correspondingtemperature ranges.J
considered for the helical heater assembly whereas
for the pancake the radial temperature gradient of
I atculatethe temperaturesat ~ . ~
everytime steP 6t using the surface increases with the input power.
Explicit method.
~ Print : T(i), Fig. 4b shows the longitudinal variation of the
Radiative steady-state temperature on the sleeve surface. It
energy
can be seen that the temperature distribution at-

(b) ~es No
tains its m a x i m u m value at the mid-point of the
heater coil height which is also intuitive for any
helical heater. However, for the pancake heater,
Fig. 3b. Flow diagram (transient-state analysis). with the increase of input power the sleeve surface
 236 A. Sil et aL / Thermal analysis of a cathode

-- FORPANCAKEHEATER lower than that of the sleeve. For the pancake

. . . . FORHELICALHEATER heater, the temperature of the dielectric b o t t o m
1660 ~" ,~P =4~
I 1650 P /
surface is increasing along the radius, which is also
1660 - I
evident from fig. 4a.
1640 1. i
1650 Fig. 5 shows the plot of the average values of
1630 1.
-
1640 the steady-state temperature (Ts) on the cathode
1674 - 16201--
1630 for various values of the input power ( P ) . The
1672 1610 I-
1620 . x
average cathode surface temperature versus heater
1670 -~ P=~Ow 16001- x~
- ~ 1610 power plot has a tendency to saturate at higher
1668 159q~ \
1600 power levels. This is due to radiative exchange to
1666 1562
_ ~ 1590 the environment at the boundaries which limits
1664 155BJ- P=4OW]
1580 the temperature.
1662 - 15541-Av /v
u 1570
1660
." p=40W 1560
1592 6.2. Isotherms
.~ 159o 1530 W /
1520 The spatial variation of steady isothermal field
154/~
1510 P=30
1542 i -.-.-, inside the assembly is very much sensitive to the
1500 heater shape. Figs. 6a and 6b show these maps
15~O
1490 / ~..~~W. plotted for helical- and pancake-heater packaging,
ae 1538- P:30W \
1480 respectively. The isotherms are drawn for a typical
uJ 1536,-..- P:30W
e,. 148~ 1470 input heater power of 30 W. The dashed lines in
uJ 1482 ~" 1460~ the figures represent the actual position of the
1380 1370"~- P=20W / heating source used in the computation. It is ob-
. ~ P : 20W
1378 136°r /
1376 135o~ _/s:2ow
1339:l_ _ _ 13.;-7--- ....
2000
,337 ; P:2°~-
1152 :::U , o
~= 19oo --PANCAKE HEATER
1158 LU
b- Q HELICAL H E A T ~
1132~p=10W 113ot;7; , -w- 1 O
T 1800
1130}--tw, , , i , 112o£-. . . . . . I
110~0 2 4 6 11oo~. ~ %' ; '~ '1; B tu 1700
RADIAL LONGITUDINAL RADIAL -r
DISTANCE (turn) DISTANCE(ram) DISTANCE (turn)
(a) (b) (c) 16oo

Fig. 4. (a) Radial temperature distribution on cathode; (b)

longitudinal temperature distribution on sleeve and (c) radial < 1500
temperature distribution on dielectric bottom surface for vari-
~E 1400
ous values of power. w

~= 13oo
average temperature gradient increases towards
the cathode. 1200
Fig. 4c shows the radial temperature distribu- uJ
tion on the bottom surface of the dielectric. For ~' 11oo
the helical heater case the temperature falls on # I I I I I
1000
both sides of the source. The falling is much faster 0 10 20 30 40 50 60
towards the interface between the dielectric and INPUT HEATER POWER, P(W)
the sleeve. This may be attributed to the fact that Fig. 5. Steady-state temperature of the cathode versus input
the thermal conductivity of the dielectric is much heater power.
 A. Sil et aL / Thermal analysis of a cathode 237

served from plots that the isotherms encircle near

the heating source and they get flattened near the
boundaries. Isotherms are shown for one half of
the assembly as the isotherms on the other half are
just the mirror image of the figure confirming that
the slopes of the contours are zero on the axis of
the body. The perpendicular meeting of the iso-
therms on the axis at r = 0 is characterised by eq.
(7) in the analysis and hence the assumption of an
insulated boundary at r = 0 is justified.

12t
11

1485 K
tn

RADIAL D I S T A N C E ( m r n ) -- "~
(b) ~ rl 1 ]j
r2 r3 -I

Fig. 6b. Steady temperature field in the potted heater-cathode

assembly with pancake heater; input power is 30 W.

6. 3. Effect of scaling of cathode size parameters on

steady-state temperature distribution

The industry uses cathodes of various sizes

ranging from 1 to 120 mm diameter. To take into
account these sizes, simple scaling laws have been
devised (see eqs. (9a) to (9c)) for constant heater
6
RADIAL DISTANCE (turn) power density. The temperature profiles along the
radial direction of the cathode pellet have been
i r, , ] computed and are shown in fig. 7a.
~2 r3
Fig. 6a. Steady temperature field in the potted heater-cathode It can be seen from fig. 7a that the mean value
assembly with helical heater; input power is 30 W. of the steady temperature of the cathode increases
238 A. Sil et al. / Thermal analysis of a cathode

FOR PANCAKE HEATER

.... FOR HELICAL HEATER
1595
r'=r x E
r'-- r X0E
1575 \\

160(

1500 i
¢y
223 i
F--
i', :-
1415 1400 ~ ""
ua 1395 w,o.~

1375
1300

1335 ~ i
1315 1200 7
/
1295 J ~ , fa

1275
1100; ] l J "f'°i 1100L~ ~ ' ~ J ~ 8
2 4 0 2
RADIAL LONGITUDINAL RADIAL
DISTANCE,r'( mm) DISTANCE, z' (ram) DISTANCE,r'( mm)
(a) (b) (c)
Fig. 7. Effect of scaling of dimensions on temperature distributions at constant input heater power density in the potted
heater-cathode assembly: (a) radial temperature distributions on cathode of various sizes, (b) longitudinal temperature distributions
on sleeve of various sizes, and (c) radial temperature distributions on dielectric bottom surface for various assemblies.

with increase in the assembly size, when the pan- temperature of the cathode surface decreases with
cake heater is used. This is because the thermal the increase in the assembly size. The fall of the
transmittance towards the cathode is higher than steady temperature value is also understood from
the thermal transmittance towards the bottom di- the concept of the thermal transmittance as long
electric surface, when the finite value of the con- as the heater location is fixed. It is also observed
tact resistance of the interfaces is neglected. For from the set of curves that the radial temperature
the helical heater-cathode assembly the steady distributions are almost uniform for the helical
 A. Sil et aL / Thermal analysis of a cathode 239

1600~ the helical heater assembly has a slow decrease of

~ I- PANCAKE
HEATER temperature with the size parameters. F r o m the
~
u,~ F .... HELICAL HEATER figure, it can now be easily concluded that the
pancake heater could be preferred for its efficient
oO I-INPUT HEATFLUX ~-0.7073553x106(W/m21 /
thermal response in situations where high thermal
response is a must, even though the steady-state
1500- temperature distribution on the cathode surface is
nonuniform. However, the helical heater has a
more uniform temperature distribution on the
cathode surface, though it does not give rise to
high thermal response.

i 1400 6. 4. The temperature profile on the concave cathode

surface

It is of interest to numerically appreciate the

effect of having a concave cathode surface on the
steady-state radial temperature distributions.
1300 t [ I I J I J I J I While making the study, however, we confine our
20 40 60 00 100 120
CATHODE PELLET DIAMETFR~2r3 from) results for practical values of cathode parameters
Fig. 8. Steady-state temperature versus cathode pellet size reported by Vaughan [10]. Fig. 9a illustrates the
parameters (diameter and thickness). steady temperature profile on a concave surface
and the result has been compared with that of a
pellet surface. The thickness of the pellet has been
heater as compared to that with the pancake heater
determined from the same thermal-mass consider-
where the radial temperature distribution is non-
ation. The temperature plot on the concave
uniform. The oscillations observed in the tempera-
cathode surface has been made by projecting the
ture distribution curves for the pancake heater at
temperature values of the curvilinear points on the
large scaling factors are mainly attributable to the
corresponding vertical points of the cathode disk
coarse mesh size. Fig. 7b shows the steady temper-
radius.
ature profiles in the sleeve. As can be seen, all the
Figs. 9b and 9c show the steady temperature
profiles of the scaled assemblies meet each other
profiles of the sleeve and dielectric bottom surfaces
near the mid-point of the sleeve height. This has
of the assembly made with the concave cathode in
been observed for both heater configurations. It
one situation and the pellet in the other.
can also be seen from the curves that the average
It is seen from fig. 9a that along the disk radius,
temperature gradient decreases with the increase
the temperature varies gradually but the variation
in assembly size. Fig. 7c shows the bottom dielec-
from point to point is higher compared to that of
tric surface steady temperature profiles along the
the flat pellet.
radial direction. As shown, increasing the size of
the assembly results in increasing its mean tem- 6. 5. Effect o f heat shields
perature when the helical heater is used. On the
other hand, for the pancake heater, it results in a To consider the effect of heat shields, we have
fall of the mean steady temperature with the size chosen the Spectra-Mat standard 600 ion source
of the assembly. design [11], as its typical experimental data is
The concise representation of surface tempera- readily available. Even though ion and electron
ture variation with cathode pellet dimensions is sources are widely different from an application
shown in fig. 8. It is interesting to note that as the point of view, their thermal designs are similar
pellet size increases the temperature (T~) increases and therefore we modified the software for the
rapidly for the pancake heater assembly whereas steady-state analysis suitably using variable
240 A. Sil et aL / Thermal analysis of a cathode

1316
a)
1315
8 o26 mm
1314 INPUT HEATER P0WER=20W

v
o...
1313 J'r c ==
z /--1.~K = a.o26 ~m
uJ
p,,
1312 A /J'r c = 15'215 rnm
; //-IRK= 8.026
1295 ~ rK =8"026 mm
129b, V " ~ 15"215 rnm 6 CATHODE
,,=,
1292j-- o r~?.~..-i~I
o 12 4\6
e r
-'POTTED HEATER
:2299~F , , , , , , , \ (PANCAKESHAPE)
0 1 2 3 4 5 6 7 8 ~-DIELECTRIC SURFACE
RADIAL DISTANCE OF CATHODEDISC
(mm)

1328
(c)
131~'(b ) 132Lo
13121 J ~ 1320
A 1311r/ rc=OO
o 1316
1310.~ 1312
129~" ,~= 13oe rc =
n- 1293 130b, 15.215turn

L,U
1292 13oo
I'--

1291 ~ 1296
1290 1292
12BC I J I 1288
0 1 2 3 4 0 1 2 3 4 5 6
LONGITUDINAL DISTANCE RADIAL DISTANCE ( r a m )
(ram)

Fig. 9. Effect of having a concave surface to the cathode: (a) cathode disk radial temperature distributions, (b) longitudinal
temperature distributions on the sleeve and (c) radial temperature distributions on the dielectric bottom surface of the assembly.

thermo-physical properties (thermal conductivity show that in the presence of the heat shields the
and emissivity) to include this design. Fig. 10a theoretical curve has the same trend as the experi-
shows the geometry of the ion source, and fig. 10b mental one but there is about 20% difference in
shows its discretized form for analysis. In fig. 10c, the temperature values for corresponding filament
the computed results are presented by the dotted powers. The discrepancy in the result is due to the
curve, whereas the reported typical experimental fact that the model analysed is a simplified one,
approximate result is shown by continuous line. and does not include the following factors:
The improvement of thermal efficiency with heat (1) the porosity factor in the emitter as well as
shields over the case when heat shields are not potting structure;
present (as dashed curve) is also depicted in fig. (2) condition of the interface formed across the
10c wherein heater power versus cathode surface intermetallic joints;
(brightness) temperature is shown. The results (3) the location of the heat shields;
 A. Sil et aL / Thermal analysis of a cathode 241

t--~15.2~, ~l---~ f-- emitterP°r°ustungstenmateriacontaining

l

I 1"~12 ~o-~rwp
~ -,E ~ 0"0508mm thick moly hea~

~ O
13 I// // O ~
f ~ 0--t-4~-- 0 5 0 8 mm ill moly bitilar heater
~ ~ ,~ I FUll wrapped with 0"25b,mm mow
[~J~J at exit leads
II 11-~ Re Support struts brazed

(a)
TU
t'- 1.016

0"7 ~'Z(mm )

! , I
! I r !.-.
1 / I , M"
• I I _

J , I I ...... r(mm)
10"9 -I
- - 12'7
15"2 - -
(b)
All dimensions ~n m m

160

1~0
Typical e x p e r i m e n t o l c u r v e available
-----Without heat shield (Theoretical)
~ 1 2 0 _. ......... W i t h heat shield ( T h e o r e t i c a l )
n
I / /
-----With h e a t shield including the losses /
J 100 t h r o u g h 1"016(~ s u p p o r t s . / / • ....

~. 80 J l / " " ""

~.....
t L.-
c 60 ./ . i . : ......
/ ~ ...-.-
E
---~ ~0 I- . .~...=.--

20
(c)
800 900 1000 1100 1200 1300 lt, O0 1500 1600
Emitter f a c e t e m p e r a t u r e ~ TB (°C] ( B r i g h t n e s s )

Fig. 10. (a) Spectra-Mat standard 600 ion source. (b) Vertical section of the ion source with mesh generation for steady-state analysis.
(c) Temperature versus heater characteristics of Spectra-Mat standard 600 ion source.

(4) actual emissivities of the sleeve surface and (5) heat losses by radiation and conduction
that of the surface of the heat shield (which sig- through rhenium support struts which are brazed
nificantly depend on the surface condition and to a m o l y b d e n u m body (sleeve) at three places at
contribute to the radiative exchange); 120 o apart;
242 A. Sil et aL / Thermal analysis of a cathode

1700
o f ~ = 1, Power = 40W
~15oo
f ¢ =1,Power =20W
I-,<
1300 I ¢--2
o.
uJ
I- 11 O0

14_
""
:D 900
u3

oN
,-I-
700

o 500

300 I I I I I I I I I I
0 200 400 600 800 1000 1200 lt*O0 1600 1800 2000 2200
TIME (sec.)

Fig. 11. Temperature-time plots of the cathode surface of the typical size heater-cathode assembly and few of its scaled sizes with
pancake heater configuration.

(6) the brightness temperatures of the emitting 10 ° C than that obtained by steady-state analysis.
side were calculated from true temperatures by This is perhaps due to the fact that in the transient
using the following formula [1] case the total energy balance condition is not
reached owing to the fact that less number of
1 1 2.303~ log( E x )
elements and only a finite number of time steps
TB T. C1
have been used in the calculation. The transient
which is not truly applicable to the emitter of an thermal profile has also been computed for higher
electron or ion source; and value of the input power for one geometry ( a =
(7) the emitter materials which are conventionally
fused into the porous plug of tungsten are not
taken into account.

6.6 Warm-up time 1331 I

1330
The warm-up time of the cathode has been 1329
determined from the time-temperature history of
the assembly. Fig. 11 shows the variation of the ~¢r
1326
1327
surface temperature of the cathode versus time. 1326
The different thermal profiles are for various val-
ues of the scaling factor a = 0.5, 1.0, 2.0, 5.0, and 132gI
10.0. One ( a = 1.0) of the representative plots of 1323|
temperature versus time has been compared with 1322 ~"
the result (fig. 12) specially computed from 1321 ~-
steady-state analysis using temperature-dependent 13201 I i J, I
0 1 2 3 4 5 6
thermal conductivity and emissivity. The calcula- RADIAL DISTANCE (mrn)
tions have been done for the pancake heater con- Fig. 12. Steady-state radial temperature distribution on the
figuration only. It is found that the steady temper- cathode pellet by using temperature-dependent thermal con-
ature attained from transient analysis is less by ductivity and emissivity.
 A. Sil et al. / Thermal analysis of a cathode 243

1.0) of the assembly. This is to illustrate the possi- made use of for electronic switching from high to
bility of reducing the warm-up time of the cathode low power values to reduce the warm-up time
by electronic power switching from higher to low considerably.
value. Parameters like heater location, helix diameter,
sleeve thickness can be optimized for the chosen
geometry of the assembly. This analysis is also
7. Conclusions
useful to estimate safe input power levels to avoid
The present analysis has been undertaken to excessive heating leading to the melting of materi-
simulate the potted heater-cathode assembly of als constituting the assembly. It would also throw
cylindrical geometry and determine the tempera- light on the processes and intermediate materials
ture distribution on the cathode surface and its required to mechanically and thermally bond the
warm-up time. Due to the nonlinearities involved cathode to its supporting sleeve.
in the basic equations and boundary conditions,
the solutions have been obtained for a typical
geometry having ideal interfaces. Acknowledgements
Dimensional scaling laws of constant input
power density have been devised to include practi- The authors are grateful for the financial assis-
cal cases of interest, i.e. 1 to 120 m m diameter tance of the Department of Electronics, Govern-
cathodes. Curvature effects of the cathode surface ment of India.
have been included and compared to that of a flat
cathode of the same thermal mass. The entire
analysis has been done for a single size of potted Appendix
heater-cathode assembly. The results have been
computed separately for two of the commonly All the interior element temperatures are given
used heater configurations to compare the dif- by
ference between using a helical heater as opposed ~p+l= 1- At
to a pancake heater. The comparative results show i,j
Pi,j(T) Ci4(T ) dv/,j
that though it suffers from the relative difficulty
of attaining efficient thermal response, the helical X( Mi-,,j(T) + Mi+,4(T)
heater would be used where a uniform cathode
temperature distribution is required. The fact that +M,,j_,(T) + M, j+,(T))/ ZP.
• I,J
cathode size parameters have a significant in-
fluence on the temperature profiles has been re- + At
vealed by dimensional scaling at constant input pi,j(T) Ci4(T ) dvi,j
heat power density.
Heat shields around the heater-cathode pack-
× [M,_,.j(T) 7:"
i+,,j

age have also been considered as a boundary +M,,j_,(T) T,Py_ 1 + M,,y+I(T ) TiPj+ 1
condition. The steady-state software has been sui-
tably modified and used to analyse the Spectra- +Qs P(r, z ) ] , (10)
Mat standard 600 ion source design and compared where
with the published typical experimental data avail-
Az
able. The theoretical results have been found to be P(r, z) = ~ r , q [ 1 - u(z -- z,)]
within 20% of the experimental (typical) values.
Within the limitations of the theoretical model for the helical heater, and
(listed at 6.5) the agreement seems to be reasona-
P(r, z)= ( r - X A r ) Ar ¢ , : , [ 1 - u ( r - r , ) ]
ble.
r2
The transient analysis results of temperature
versus time for different input powers can be for the pancake heater.
244 A. Sil et al. / Thermalanalysisof a cathode
The coefficients in eq. (10) are given by At
ki-l,j(T) Ai-1/2,j + M d ( r ) ) T,Py + P d ( r ) C d ( T ) dvi 4
Mi-l,J (T) = Ar '
x [Mi_,.j(T ) Tf_Lj+ Mi+,4(T ) T.P+I,j
ki+i4(T) Ai+1/24
M,+,,j(T) = Ar ' +M<j+1(T) TiPj+, + Md(r) ramb]
ki,i-l(T) Ai,j-1/2 (llc)
Mi,J -I(T) = AZ '
for the potting surface, where
ki,j+l(T) Ai,j+l/2
Mi,J +'(T) = Az Ma(T ) = [ h , ( T ) ] d Ai,j-l/z.
The temperatures at the boundary points of the In the above formulations h , ( T ) ' s are expressed
assembly are represented as as

Tp+l [ At T2(r, z3, t)

'4 = 1 - pc(T ) Q ( T ) dvi, j [h,(T)] =o%(T) T2mb 1 + 2
c Tamb

X (M,_,,j(T) + M,+,,j(T) + Mi,j_1(T ) T(r, z3, t ) ]

x 1+ Tamb ,

+Mc(r)) %+ o (r) Cc(r)d ,,j 3 [ T2(r3, z, t)

[hr(Z)]sl = a ¢sl(T) Tamb [1 + 2
X[M, ,,j(T) TiPl,j+ Mi+I,j(T) Tf+1,j Tamb

+M.j_,(T) %_, + Me(T) Lm ] [ T(r3' z' t)]

X 1 -I- Tamb ,
(Ua)
for the cathode surface, where 13 TZ(r, O, t) 1
[hr(T)]d-~O~d(T) Tamb 4- Ta2mb
Mc(r ) = [ h , ( T ) ] c A,.;+,/2;
[ T(r,O,t)]
Tp+l [ At X 1 "4- Zamb ,
,,j = 1 - ps,(T) C~,(T) dv,,j
defining the radiative heat transfer coefficients
X (Mi_I,y(T) + M,,j_,(T) + Mi,j+,(T) [16]. The output radiative energies from cathode,

]
+Ms,(r)) r,5+ Os,(r) Cs,(r)d~,,j
sleeve and dielectric surfaces are given by
R¢ = Y'~Ac[h,(T)] c [r(r, z 3, t ) - T~mb],
X[Mi-I,j(T) TiP-I,j+ Mi,j-I(T) TiS-1 Rs, = 2 A ~ [ h , ( T ) ] s L [T(r3, z, t) - Tamb],
+Mi,j+1(Y) Tp,j+l 4- msl(T ) Tamb] Rd = 2 A d [ h r ( T ) ] d [ T ( r , O, t) - Tamb] •
(llb)
for the sleeve surface, where
References
Msl(r ) = [hr(T)]sl Ai+l/2,j;
and [1] J.L Cronin, IEE Proc., Vol. 128, Pt. I, No. 1 (1981) pp.
19-32.
Te+, [ At [2] C.G.J. Jansen, A. Venema and Th.H. Weekers, Philips
,4 = 1 - pa(T ) Cd(T ) dvi, j Tech. Rev. 24 (1962-63) 402.
[3] J.C. Tonnerre, D. Brion, P. Palluel and A.M. Shroff, Appl.
X (Mi_,,j(T) + M i + , , j ( T ) + M,,j+,(T) Surf. Sci. 16 (1983) 238.
[4] G. Eng and H.K.A. Kan, Appl. Surf. Sci. 8 (1981) 81.
 A. Sil et al. / Thermal analysis of a cathode 245

[5] G.A. Haas and R.E. Thomas, J. Appl. Phys. 38 (1967) [12] R. Crivello and R.W. Grow, IEEE Trans. Electron De-
3969. vices ED-35 (1988) 1701.
[6] R.E. Thomas and J.W. Gibson, Appl. Surf. Sci. 29 (1987) [13] W.M. Rohsenow, J.P. Hartnett and E.N. Ganic, Eds.,
49. Handbook of Heat Transfer Fundamentals (McGraw-Hill,
[7] N.R. Roy, S.K. Pardhasaradhi and S.N. Sengupta, IVS New York, 1985).
Bull. 18 (1987) 20. [14] Y.S. Touloukian an C.Y. Ho, Eds., Thermophysical Prop-
[8] S. Kakac and Y. Yener, Eds., Heat conduction, 2nd ed. erties of Matter, Vol. 1 (Plenum, New York, 1972).
(Hemisphere, New York, 1985) pp. 318-323. [15] B. Sunden, Int. Commun. Heat Mass Transfer 13 (1986)
[9] J.A. Lucken, IEEE Trans. Electron Devices ED-16 (1969) 515.
813. [16] M.N. Ozisik, Ed., Heat Conduction (Wiley, New York,
[10] J.R.M. Vaughan, IEEE Trans. Electron Devices, ED-28 1980) p. 487.
(1981) 37.
[11] Spectra-Mat Technical Bulletin #118, Ion Sources, TB
118 (November 1980).