of 5
IEEE
TRANSACTIONS
ON
NUCLEAR
SCIENCE,
VOL.
31,
NO.
6,
DECEMBER
1990
2214
Advances
in
the
Development
of
Encapsulants
for
Mercuric Iodide
X-Ray
Detectors
Abstract-
Advances
in the
development
of
protective impermeable
encapsulants
with
high
transparency
to
ultra-low-energy X-rays
for
use
on
Hgl,
X-ray
detectors
are
reported. Various
X-ray
fluorescence spec-
tra
from coated
detectors
are presented.
The
X-ray
absorption
in the
encapsulants has
been
analyzed using
characteristic
radiation
from
vari-
ous
elements. Results suggest that
low-energy
cutoffs
for
the
detectors
are
not
determined
solely
by the
encapsulating
coatings
presently
em-
ployed
but are
also
influenced by
the front electrode and surface effects,
which
can
affect the
local
electric field
or
the
surface recombination
velocity.
An
energy resolution
of
182
eV
(FWHM)
has
been achieved
for
Ni
L
lines
at
850
eV.
Improved detector sensitivity
to
X-ray
energies
under
700
eV
is demonstrated.
I.
INTRODUCTION
ERCURIC iodide
has
received considerable interest as
M
a material useful
for solid-state
X-ray
detector spec-
trometer applications
[
11
-
[3].
Bare HgI, detectors
are
capa-
ble
of
operation to quite
low photon
energies, while retaining
excellent resolution; however, proper HgI, surface passiva-
tion and
device encapsulation
are
critical
for insuring their
long
term
reliability.
For the
last
several years
we
have
made
a major effort
to develop
a protective impermeable
X-ray
transparent coating that
will
not
degrade the detector perfor-
mance,
while
still
assuring good protection
in
various
ad-
verse environmental conditions. Several excellent protective
surface treatments have been
identified and
previously re-
ported, including polymethylmethacrylate (PMMA)
and
Parylene@
[4]
-
[6].
The encapsulation
of
detectors introduces an additional
X-ray attenuation
by
the protective material
itself.
This
is
particularly
critical
for low-energy X-rays.
In
this paper
we
report
on
the results
of
recent efforts to improve the low-en-
ergy response
of
these detectors through the use
of
higher
transparency coatings,
while still
maintaining,
or
even
im-
proving, long-term device
reliability.
Results
using
Parylene-C
as
an
encapsulant have
shown
that this
material quite adequately meets the protection
needs
Manuscript received January 17,
1990;
revised
June
26,
1990.
This
work
was sponsored
by
the National Aeronautics
and Space
Administration and
supported
in
part
by
NIH
grant
#5R01
GM
37161.
J.
S.
Iwanczyk
and
Y.
J.
Wang
are
with
Xsirius, Inc. (formerly
the
Institute of
Physics,
University of
Southern California), 4640
Admiralty
Way,
Suite
214,
Marina
del
Ray, CA
90292.
J.
G.
Bradley is with the Jet Propulsion
Laboratory, California
Institute of
Technology,
4800
Oak
Grove Drive, Pasadena,
CA
91 125.
A.
L.
Albee is with the
California
Institute of
Technology,
1201 East
California
Blvd.,
Pasadena,
CA
91 125.
W.
F.
Schnepple is with Xsirius Scientific,
Inc.,
4640
Admiralty
Way,
Suite
214,
Marina
del
Ray,
CA
90292.
IEEE
Log
Number
9038740.
@
Registered
service
mark of Union
Carbide Corporation.
BRADLEY, A.
L.
ALBEE,
AND
W.
F.
SCHNEPPLE
for
HgI,
X-ray
detectors
[2].
In
the interest
of
reducing the
attenuation
of
the coating for very-low-energy X-rays, how-
ever,
it appeared
that
some improvement could
be
gained
by
changing to Parylene-N. Unlike Parylene-C the Parylene-N
composition contains
no
chlorine
and
should therefore
be
more transparent
to
low-energy X-rays.
To
investigate this
alternative coating,
we
initiated
a
program to adapt our
Parylene-C processes
and
procedures
to
permit the deposition
of
thin films
of
Parylene-N
and
to
perform testing
on
coated
X-ray detectors.
While the attenuation
of
the encapsulant has,
until
re-
cently,
been
the
most
critical
parameter
in
determining the
low-energy
cutoff
of
the response
of
the
X-ray
detectors,
it is
not
the
only
factor.
As
a result
of
this present work, other
important influences were found to
be
the attenuation caused
by
the detector’s entrance electrode
and
also the condition of
the detector surface,
which
can
affect
the electric
field and
the surface recombination velocity. These topics
will be
discussed
in
the paper.
11.
EXPERIMENTAL
A.
Parylene Encapsulation
We have
found
it impossible
in
the
past
to achieve good
low-energy
cutoffs with
commercially supplied relatively thick
Parylene-C coatings.
The best
of
these coatings did
not
transmit X-rays
below
approximately
lo00
eV
or
sometimes
even higher energies.
As
a consequence,
we
have devoted
considerable program effort to developing and improving our
own
in-house Parylene-C deposition capability at the Univer-
sity
of
Southern California (USC)
[2],
[3].
The improved
low-energy cutoffs obtained from our Parylene-C depositions
have
been
accomplished primarily through
a carefully con-
trolled reduction
of
the thickness
of
the Parylene-C detector
encapsulating material. This reduction
in
thickness was
made
possible because
of
the development
of
improved in-house
deposition techniques, including carefully controlling the
amount
of
Parylene material loaded into the source
boat
and
its
heat-up temperature ramp
rate.
As
a result
of
the
im-
proved processing,
the
films which
were deposited were
highly
transparent and completely free
of
the quantity
of
gases
that
are
often present
in
commercially produced films,
and which
give
them
a cloudy
or
milky
appearance.
“Gas
bubbles,”
if they
occur
in
the films, increase the probability
of
the occurrence
of
pinholes
in
the film.
As
a consequence,
“gas bubble” type
films
require
a greater deposition thick-
ness
to
insure
a seal
for
the detector. The
new
clear
films
can
still
provide
vacuum
integrity for the detector
but
need
only
be in
much
thinner deposits.
0018-9499/90/1200-2214$01
.OO
O
1990 IEEE
T
IWANCZYK
et
al.:
ENCAPSULANTS
FOR
X-RAY
DETECTORS
2215
To
enhance
our
control
over
the encapsulation
process,
we
have developed
a dynamic measurement technique that
con-
tinuously monitors
the
thickness of the Parylene
film
during
its
deposition. Using an optical method frequently employed
for the measurement of thin films,
the
technique detects
the
successive interference maxima and minima that
are
succes-
sively created
as
the Parylene thickness slowly builds up on
a
reflective witness slide placed near the
detectors. A
small
He-Ne
laser provides
a
monochromatic light
source for
illuminating the witness slide.
The
intensity
of
the reflected
beam
is
detected
by
a silicon photodiode whose amplified
output
is
recorded
on a strip chart recorder.
Counting
the
number
of
successive maxima and minima recorded
yields,
through
a
simple calculation,
an indication of the film's
thickness.
The
method yields an accuracy
of
one-quarter
wavelength
or
better
for
these
coatings.
In
film
thickness, this
is equivalent
to
an
uncertainty
of
less than
0.1 pm.
Although Parylene-C provides good protection,
it
does
contain
chlorine, a
relatively
high-Z element. The chlorine
provides
for
improved cross-linking of the polymer; how-
ever,
it also
produces undesirable
strong
attenuation
for
very
soft
X-rays.
We
therefore
have made
some experiments
to
deposit thin films of
chlorine-free Parylene-N,
which should
have less attenuation than the Parylene-C films of equal
thickness.
However, compared
to
Parylene-C deposition
runs,
these have proved
to
be much
more
difficult.
Initially, the resulting Parylene-N films
were
not
clear
but
instead
were
cloudy in
their appearance.
In
depositing
Pary-
lene, the evaporated
dimer
is first
passed through
a high-tem-
perature region of the
system, where
it
is
cracked
to the
monomer.
The
Parylene-N monomer then passes
to a
plasma
region,
where cross-linking
is enhanced
and the
polymer then
deposits
onto
the detector
surfaces
[7].
The
plasma was
specially introduced
into the
Parylene-N processing
proce-
dures to create a
film with improved physical
properties.
To
reduce the cloudiness
of
the
deposits,
it was found necessary
to
raise the temperature of
the
hot
zone
considerably
above
that
used
for
Parylene-C
(i.e.,
775°C
versus
650°C).
In
addition,
there
was
a frequent tendency
to produce
powdery
deposits
on
the
detector. These appear to come from stream-
ing of the uncracked
dimer
and unpolymerized monomer
depositing directly
onto the detector.
Such deposits
are
of
extremely poor quality in that the
powder
will subsequently
come
off,
leaving
a large
number of pinholes in the
film
that
remains.
Further
improvements in repeatability
for
this pro-
cessing
are
needed;
however,
initial
data
from
Parylene-N
coated detectors
are
encouraging.
An
alternative deposition
geometry,
in
which the
detector surface
to
be coated
is
shielded
by
a baffle
from the directly streaming particles, has
already resulted in
a much-improved
deposit, free
of pin-
holes. As described
in
the next section on detector results,
the low-energy
cutoff
is improved with Parylene-N compared
to
that
for
Parylene-C.
B.
X-Ray
Detector
Testing and
Results
Windowless Detector Probe System:
Low-energy X-ray
testing
of
the detectors was
done
using both the scanning
electron microscope and particle analyzer
(SEMPA)
instru-
ment
at
the Jet Propulsion Laboratory
(JPL)
and
a window-
less detector probe test system assembled at USC. The
SEMPA
is a multifunction prototype instrument designed
for
SEM and
XRF
applications
on
NASA space missions.
In
this
application, the
HgI, device
is
used
as
an
XRF
radiation
detector
to
provide elemental analysis
of
the sample under
test and must be stored and operated
for
extended periods
of
time
in
the
high vacuum environmental conditions, which
will exist on long duration
space
missions. Because
HgI,
has
a measurable vapor
pressure
even at room
temperature, a
gas-tight encapsulation coating
is
thus mandatory
to
prevent
gradual
sublimation loss
of
the
material;
however,
at
the
same time,
the coating should provide good transmission
for
X-rays down
to
the
characteristic
energy of sodium (ca
1
keV)
.
The
important influences
of
both the Parylene coating and
the near-surface electric
field
persuaded
us
of the need
to
have available
at
USC
a quick
and
convenient method
of
evaluating the effects of changes in detector processing.
Consequently, an
XRF
test system was built at USC using
a
modified
Tracor
analysis instrument.
In this
case
X-rays
rather than the
electrons
employed in
the SEMPA were
used
as
the primary radiation
to excite
various fluorescence
tar-
gets. The
system consisted
of
an
X-ray
generator
tube (with
rhodium anode and beryllium exit
window), a
rotating
carousel containing multiple
targets, and a
probe head con-
taining the HgI, detector
and
its
charge-sensitive preampli-
fier.
The
target
carrier
could be rotated without breaking
vacuum
to
position any
one
of several
targets
into the pri-
mary X-ray beam
from
the
tube.
Fluorescence X-rays com-
ing
from the
target
were
then detected
by
the
probe
head.
Because
there
was
no
window
in
front
of
the
detector
in
the USC
system,
the detector
probe,
X-ray tube and sample
holder were kept under continuous vacuum in
order
to pre-
vent any moisture condensation
during
detector and
FET
cooling and
to assure
minimum absorption
for
low-energy
X-rays.
The
total volume evacuated was much
larger
than the
volume inside
the
detector
probe.
The
USC system was used
to
excite radiation
from a
number of
different targets,
including
Al,
Cu,
Ni,
a teflon
specimen
(for
generating fluorine
K
X-rays,
677 eV). The
X-ray tube had
a thin beryllium
exit
window and was
oper-
ated
at
voltages up
to
13
kV. This
produced
an
output of
not
only the usual Bremsstrahlung
spectrum
but
also
fairly strong
L
line radiations
from the
tube's rhodium
anode.
The
USC system had somewhat more electronic
noise, due
to
certain construction specifics that
were
different
on
the
USC
probe, as compared to
the noise obtained on the
JPL
SEMPA
(175 eV
versus
152 eV
FWHM
pulser widths,
respectively).
This
noise,
however,
was easily low enough to
allow
for
evaluations
and comparisons to
be made between
different detectors and coatings.
Fig.
1
presents
a spectrum taken
from
the HgI, detector
probe (so-called
Mark
I1
version) showing
a copper
fluores-
cence spectrum
as
obtained
on
the
JPL
SEMPA
instrument.
The Cu
L
lines, at an
energy
of
around
940
eV,
are
barely
visible
in
this figure. The
HgI,
detector
used in the
Mark
I1
probe
to
take this spectrum was built using
a Parylene-C
11
_I
I
II
I
2216
.
I
IEEE
TRANSACTIONS
ON
NUCLEAR
SCIENCE,
VOL.
37,
NO.
6,
DECEMBER
1990
I
1
I
I
I I I
I
a
1
I
I
I
I
I
I
IUJJ
I I I
1
I
~SUI~~I~LLLLLLLI
HgI,
Spectrometer (Mark
111
Probe)
Cu
Target
9e1cc/or
Tgrf
RerirllS
t
.uL-
IILU.LLl
L1.l
UULddLdd
HgI,
Spectroiwter
(Mark
I1
Probe)
Cu Target
cu-ka
Energy
(keV)
Fig.
1.
X-ray
fluorescence
spectrum
for copper.
Cu-ka
I:
II
8
;
111
Energy
(lcev)
Fig.
2.
X-ray
fluorescence
spectrum
for copper.
coating
supplied
by
an
outside vendor. The encapsulation
film
thickness
was
estimated
at
7-8
pm,
and its effect
in
reducing the sensitivity
to
the
L
line
is clearly apparent.
Fig.
2
shows another spectrum taken
on
the SEMPA
using
the same
Cu
fluorescence source; however, this time
it
is
taken
with
a newer Mark
I11
probe,
which
incorporated
a
detector encapsulated
at
USC.
In
contrast to Fig.
1,
this
spectrum clearly shows
a higher-intensity
well-defined
(190
eV
FWHM)
peak
for
the combined
Cu
L
lines
(L,,
L,
at
928,
948
eV,
respectively). The reduced attenuation from the
much
thinner (ca.
3.5
pm)
USC
Parylene-C coating
is
thus
clearly evident. Also shown
are
the
K,
and
K,
lines
at
8.0
Hg12
Spectrometer
Ni
Target
....-
1114Sb1
Energy
(keV)
I
9
IO
Fig.
3.
X-ray
fluorescence
spectrum
for nickel.
and
8.9
keV
(with
resolution
of
218
eV
for the
8.0
keV
peak).
Fig.
3
shows Mark
I11
probe SEMPA instrument
data
taken
with
the same HgI, detector as was
used
for
Fig.
2.
The
figure
shows the fluorescence spectrum from
a Ni
target
excited
with
electrons.
In
addition to the
Ni
K,
and
K,
lines
at
7.5
and
8.3
keV, the combined
Ni
L
lines
are
clearly
visible
at
849
eV
with an
energy resolution
of
182
eV
(FWHM).
Fig.
4
shows four different spectra taken
using
the win-
dowless
USC
system and
a Parylene-C coated detector. The
low-energy
cutoff
permits the transmission
and
detection at
fluorine K-line energies,
as
shown
by
the
well-defined peak
at
677
eV.
In
addition to the fluorine
K,
spectrum, three
other fluorescent spectra peaks
are
shown
in
the figure:
Ni
L
(0.86
keV),
Cu
L
(0.94
keV), and
A1
K
(1.486
keV).
In
all
cases, the
peaks
are
clearly resolved above the backgrounds.
A
major objective
of
this testing was
to
study both
the
attenuation
of
the
Parylene coatings
and
also changes
in
the
detector front
window
transmission due
to
processing varia-
tions
(e.g.,
in
the
Pd
metallization). The approach employed
was
to measure the
X-ray
responses
of
different detectors to
both
the
K
and
L
X-ray
lines from elements
such
as
Cu
or
Ni. The concept was
that
the ratio
of
the
K/L
responses for
one
detector
could be
compared to
that
for another detector,
perhaps
with
a different coating, to assess low-energy sensi-
tivity.
KIL
ratios were
used
since under
fixed
excitation
conditions, the ratios should
be
relatively independent
of
detector-to-detector geometry
or
configuration variations.
Based upon
cross-section data
for
the
thin
Parylene coat-
ings
under
consideration, photons at
any
of
these
K
line
energies
(ca.
7-8
keV)
were
all highly
penetrating,
i.e.,
the
coatings were essentially transparent. The
L
line photons
on
the other hand, were
in
the energy range
(<
1
keV)
where
attenuation
in
the coating becomes very significant. Hence,
any
differences
noted
in
the ratios for different detector/coat-
IWANCZYK
er
al.:
ENCAPSULANTS
FOR
X-RAY
DETECTORS
I
.
:*
2217
.
FK
0
677
KeV
.
..
y,;\,l
..
'.
+"
.$
(I
1
2
.
FK
0.677
KeV
t
I
1
1
(I
1
2
:.
j:
(3
...
I
t
1
I
MI.
0.OG
KeV
..
.
..
..
-
i
i
0
1
2
,
E11ergy
(Icev)
,
(I
I
2
5
Eiieigy
(Kev)
Eiirigy
(KeQ
Fig.
4.
X-ray
fluorescence spectrum
for
F,
Ni,
Cu,
and
AI.
TABLE
I
Calculated
Parylene-N
Parylene-C
Ratio
1/10,
%
1/10?
%
LP-NILP-C
Parylene-N
=
3.6
pn
Parylene-C
=
Experimental
Target
L
Line
Energy
3.6
3.3
3.5
3.1
3.3
3.5
3.1
P-N
P-C
eV
pm
Pm
Irm
Ctm
pm
pm
Ctm
Km/L
Km/L
LP
NILP
C
cu
925
31.0
31.3 29.5
21.5
1.18
1.25
1.35
64
SI
I
26
Ni
849
28.6
23.5 21.5
20.0
1.22
1.33
1.43
I
40
209
292
Note:
K,
/L
is
the ratio
of
the transmitted
K,
line
and
L
line X-rays intensities measured.
Lp
/LP.c
is the
ratio
of
the
transmitted
L
line X-ray
intensities
through the
Parylene-N
and
Parylene-C coatings.
ing
combinations could
be
assigned, almost entirely, to dif-
ferences
in
L
line responses.
Measuring the same detector at two different targets
would
also provide two
KIL
ratios. Again, since the transparency
of
any of
the coatings
was
very high for
K
line energies,
changes
in
the ratio responses could
be
attributed to
different
L
line
responses
.
It was
thought
that
these
K /L
ratios for
a
unit could then
be
examined to see
if
they
changed
in
agreement
with
an expected calculated change due to the
known
thickness
of
the Parylene coating. Determining abso-
lute
theoretical
K/L
ratios existing
at
the front entrance
window of
the detector proved extremely
difficult
due
to such
things as geometry factors, the presence
of
a complicated
excitation spectrum from the X-ray tube, self-absorption
in
the fluorescing target,
etc.
[8].
The ratio technique therefore
was used only
as a
relative
method
to compare the low-en-
ergy sensitivity
of
one detector to another.
Table
I
shows calculated and measured data for two HgI,
detectors, one coated
with
Parylene-N
and
the other
with
Parylene-C. The theoretical
X-ray
percentage transmissions
were calculated
using
linear attenuation
coefficients
calcu-
lated
from cross sections appropriate to the two different
compositions
of
the materials. The measured data were
based
upon
a summation
of
the counts
in
a
K
or
L
peak
in
the
output spectrum,
which
were above background. The thick-
ness
of
3.6
pm cited for the Parylene-N coated detector
was
a value measured during deposition using the laser interfer-
ence technique. The thickness
of
the Parylene-C coating
on
the other detector was
not
measured during
its
deposition
run. Rather,
a standard amount
of
dimer
was
processed,
T
II
I
2218
IEEE
TRANSACTlONS
ON
NUCLEAR
SCIENCE.
VOL.
31.
NO.
6,
DECEMBER
1990
TABLE
I1
Calculated
Pd
HgI
2
IlIo>
%
I/Io.
%
L
Line Energy
100
150
200
100
200
300
1000
4000
loo00
Target
eV
AAA
AAAAAA
94.6
89.4
84.5
57.1
10.6
0.37
94.1
88.5
83.3
54.3
8.7
0.22
cu
925
92.5
86.7
82.0
Ni
849
89.5
84.5
80.0
which
ordinarily gives
a final
film thickness
on
the detector
of
about
3.5
pn
f
0.2
pm.
Calculated percentage transmis-
sion values
for Parylene-C thickness
of
3.3,
3.5,
and
3.7
pm
are
shown
in the table.
Table
I1
shows similar transmission values as calculated
for various thickness
of
the
Pd
electrode
and
for
a (possible)
HgI,
dead
layer.
A
Pd
thickness
of
150
A
is
the
nominal
value
assumed
for
most of
our metallization deposition runs.
The thickness
of an
HgI, dead layer
is unknown; however,
it
is unlikely
to
be
more
than
a hundred angstroms
or
so,
since
otherwise the low-energy sensitivities already achieved
would
not
otherwise
be
possible.
111.
ANALYSIS
OF
RESULTS
Reviewing the data
in
these tables,
it
appears
that
the
experimental
Lp.,,,/Lp-c
data
in
the
last
column of Table
I
matches
closest the middle column data
for
the calculated
Lp.N/Lp-c
ratio.
This corresponds to thickness
of
3.5
pm
for the Parylene-C (calculated),
and
3.6
pm
for
the Parylene-
N
(measured). While the agreement
between
the
Cu
ratio
data
is
good
(1.25
calculated versus
1.26
experimental), the
lower-energy
Ni
data
are
less
so.
In
this case, the calculated
value
of
1.33
is
5%
lower than the experimental value
of
1.40.
While the true cause
of
the discrepancy
is unknown
at this
time, several possibilities exist
to
explain the difference.
First, something
on
the order
of
a 2%
difference can
be
attributed
simply
to the increased attenuation for the lower
energy
Ni
L
line
in
the
Pd
front electrode, provided
both
the
Parylene-C
and
Parylene-N detectors
had
the same
Pd
thick-
ness.
If
there
was
some difference
in
the
Pd
thicknesses for
the two detectors,
as
there
might well have
been, even more
variation
is
possible
(e.g.,
there
is
about
5%
difference
in
transmission
between
100
and
150
A
films. Moreover, either
a diminished
electric
field
and/or
a high surface recombina-
tion velocity
near the front surface
will
reduce the charge
signal
collected
in
the detector
and
could also
be invoked to
help explain
the discrepancy
between
the calculated
and
experimental
ratios.
These
effects
are
conveniently described
by
an effective
dead layer
of
HgI, near the surface. The
calculated attenuation for various assumed dead layer thick-
nesses
are
shown
in
Table
11.
The data
are
such
that the
existence of
a dead layer cannot
be
precluded.
In
therefore appears,
on
the basis
of
the data obtained,
that
any one
or
more
of
several different
phenomena
could
be
determining the low-energy sensitivity of these detectors.
Further
tests
will be needed
to clarify
and
resolve the relative
importance
of
the separate elements
of
attenuation
in
the
encapsulation coating, attenuation
in
the
Pd
electrode,
and
attenuation
or
electric field/surface recombination
effects
in
the HgI,
itself.
IV.
CONCLUSION
Based upon
the results obtained
to
date,
we
are
optimistic
about continued future progress
in
the areas
of
energy resolu-
tion, more optimal detector encapsulation,
and
enhanced
sensitivity
to
low-energy X-rays. Examination
of
the
K
/
L
line ratios appears to
be
a good technique for the relative
ranking
of
different detectors
based on
their low-energy
response. The data obtained
so
far,
however,
are
insufficient
to provide
a clear-cut answer to the question
of which
factor,
i.e.,
the coating, the near-surface electric
field,
or
the recom-
bination velocity,
is presently limiting the low-energy cutoff.
Future efforts
will be
directed towards gaining
an
improved
understanding
of
these factors.
Even
at the present state
of
development, however, HgI,
devices have already
met
the low-energy sensitivity
and
resolution
needs
for
the
JPL
SEMPA instrument. These HgI,
detectors under development
offer
substantial advantages
in
terms
of
decreased power requirements, the absence
of
the
need
for cryogenic cooling,
and
fewer thermal/mechanical
design
difficulties.
At
the same time, their performance ap-
proaches that
of
cryogenically cooled Si(Li) units.
V.
ACKNOWLEDGMENT
The valuable technical assistance
of
Mrs.
F.
Riquelme,
Messrs.
N.
Dorri
and
B.
Dancy
of
USC,
and Mr.
V.
Taylor
at JPL are
greatly appreciated.
REFERENCES
S.
Swierkowki,
G.
Armantrout,
and
R.
Wichner, “Recent advances
with
HgI,
X-ray
detectors,”
ZEEE
Trans.
Nucl. Sci.,
vol. NS-21,
M.
Slapa,
G.
Huth,
W.
Seibt,
M.
Schieber,
and
P.
Randtke, “Capa-
bilities
of
mercuric iodide
as
a room temperature X-ray
detector,”
ZEEE
Trans.
Nucl. Sci.,
vol. NS-23, no.
1,
pp.
102-111,
Feb.
1976.
A.
Dabrowski
and
G.
Huth, “Toward the energy resolution limit
of
mercuric iodide in room temperature
low
energy X-ray spectrometry,”
IEEE
Trans.
Nucl.
Sei.,
vol.
NS-25,
no.
1,
pp.
205-211,
Feb.
1978.
J.
S.
Iwanczyk,
“Advances
in
mercuric iodide X-ray detectors
and
low noise
preamplification
systems,”
Nucl.
Instrumen. Merhods,
vol.
A283,
no.
2,
pp.
208-214,
Nov.
1989.
J.
S.
Iwanczyk,
Y.
J.
Wang,
J.
G.
Bradley,
J.
M.
Conley,
A.
L.
Albee,
and
T.
E.
Economou,
“Performance
and
durability
of
HgI,
X-ray detectors for space
missions,”
IEEE
Trans.
Nucl. Sci.,
vol.
NS-36,
no.
I,
pp.
841-845,
Feb.
1989.
J.
G.
Bradley,
J.
M.
Conley,
A.
L.
Albee,
J.
S.
Iwanczyk,
A.
J.
Dabrowski,
and
W.
K.
Warburton, “Practical applications
of
HgI,
detectors
to
a space-flight scanning electron
microscope,”
Nucl.
Instrumen.
Methods,
vol.
A283, no.
2,
pp.
348-351, Nov.
1989.
1.
M.
Williams
and
J.
T.
Rowan,
“Diagnosis
and therapy
for a
recalcitrant Parylene
coater,”
J.
Vac. Sci.
Techno/.,
vol. AS, no.
4,
Jul./Aug.
1987.
Eugene
P.
Berth,
Principles and Practice
of
X-ray
Spectrometric
Andysis.
pp.
302-304,
1974.
New
York: Plenum,
1975,
pp.
113-
120.