вход по аккаунту



код для вставкиСкачать
Transmission Electron Microscopy Studies of (111) Twinned
Silver Halide Microcrystals
EMAT, University of Antwerp, Groenenborgerlaan 171, B-2020 Antwerp, Belgium
tabular; stacking fault; dislocation; short-range-order; surface modulation;
The present paper covers the results of different transmission electron microscopy
studies on silver halide microcrystals. Pure AgBr as well as core-shell AgBr-AgBrI crystals are
investigated. In the former parallel and non-parallel twinning modes yielding tabular and needle- or
tetrahedral-shaped microcrystals, respectively, are discussed. Also the short-range-order of Ag1
interstitials around the twin planes as determined from diffuse intensity in reciprocal space is
described. The latter yields a technique to determine the variant in which dislocations are located in
certain core-shell microcrystals. The introduction of iodine also results in the presence of inclined
stacking faults in the shell and of long-range iodine ordering on the crystal surface. Microsc. Res.
Tech. 42:85–99, 1998. r 1998 Wiley-Liss, Inc.
In preparing high-speed silver halide emulsions for
photographic applications, many variables have to be
considered at every stage of the design and manufacture. Almost every type of crystal is chemically and/or
spectrally sensitized to enhance its intrinsic and/or
spectral sensitivity. Although the largest increase in
sensitivity is obtained as a direct consequence of these
treatments, the optimization of the basic material for
the light sensitive system, i.e., the silver halide crystal
itself, still is of extreme importance. This is performed
in two ways, namely by increasing the probability of
light absorption and by using more efficiently the
photoelectrons created by the absorbed light.
By properly adjusting the precipitation conditions,
tabular crystals can be grown in which the volume can
be changed independently of the surface and vice versa,
resulting in the ability to deliver on demand a crystal
type having a wanted surface-to-volume ratio. In this
way, the increase in spectral sensitivity can be made
more or less equal for the different regions of the visible
spectrum. Tabular crystals have several other distinct
advantages that make them widely used in many
photographic applications. By using less silver, the
same surface can be covered compared to the use of
globular crystals, resulting in an equal speed/granularity ratio at reduced silver coverages. Moreover, the light
scattering is less in a layer containing tabular crystals
than it is in a globular crystal-containing layer. Since
background irradiation sensitivity increases with increasing crystal volume, an equal speed/quality ratio is
obtained at reduced background radiation. In the present studies the focus is on AgBr-based crystals.
The more effective use of photoelectrons generated by
absorbed light can be performed in several ways. The
general aim is to spatially separate photoelectrons from
photoholes in the same crystal, thus reducing the
probability of recombination. This is frequently done by
using core-shell type crystals, the core having a different halide composition than the shell. Moreover, if
tabular crystals are used with a core-shell type halide
distribution, the advantages of both techniques are
combined. An excellent review on the precipitation of
different kinds of core-shell type crystals is given by
Mitchell (1993). We will limit ourselves to a discussion
on AgBr-AgBrI core-shell tabular crystals.
The incorporation of iodine into an AgBr crystal has
several effects. If iodine is homogeneously mixed in
AgBr, an AgBrI phase results, which preferentially is
located in the same crystal next to a pure AgBr phase.
Since AgBrI introduces energy levels slightly above the
valence band of AgBr, it acts as a hole trap (Hamilton,
1988). Photoholes will be directed towards the AgBrI
region, while photoelectrons are mainly concentrated in
the AgBr region. Moreover, the introduction of iodine in
AgBr increases the concentration of interstitial Ag1
ions (Burt, 1977; Si-Yong, 1986), resulting in a diffusion
of silver ions from the AgBrI region to the AgBr region.
This again causes holes to drift from the AgBr region to
the AgBrI region (Mitchell, 1993). This effect is superimposed on, and enhances the effect arising from the
matching of electronic band structures (Granzer, 1989;
Pischel and Granzer, 1991). Iodine ions at low concentrations (below approximately 4 at.%) act as hole traps,
while at higher concentrations, the recombination probability is increased (Kahan, 1977; Si-Yong, 1986). The
same authors found in uniform AgBrI crystals an
increase in efficiency of latent image formation with
increasing iodine content up to 5 at.%, above which the
efficiency was seen to decrease. Possibly, lattice defects
are introduced at high levels of iodine that can trap
electrons and enhance the recombination probability.
AgBrI crystals respond better to chemical sensitization than pure AgBr crystals (Galvin, 1970). Also some
*Correspondence to: D. Schryvers, EMAT, University of Antwerp, Groenenborgerlaan 171, B-2020 Antwerp, Belgium.E-mail:
Contract grant sponsor: Flemish Institute for the Encouragement of Scientific
and Technological Research in the Industry (IWT).
Received 20 November 1998 ; accepted in revised form 13 January 1998
changes are observed in the spectral sensitization
process, since dyes on the surface of AgBrI crystals tend
to aggregate, while they do not on AgBr surfaces
(Steiger et al., 1991). An increased sensitivity after
chemical sensitization does not rule out the existence of
other processes that may cause a decrease in efficiency,
such as recombination for example; it merely indicates
that the former process dominates the latter one.
In the following, the combined results of several
transmission electron microscopy (TEM) and electron
diffraction (ED) studies on the effects of different twinning features on the shape of AgBr crystals will be
described. Also, different phenomena related to the incorporation of iodine ions using conversion and coprecipitation are discussed. Most results are presented only
briefly with the emphasis on specific features observed
with TEM, and the reader is referred to the more
extensive literature published earlier on these matters.
As the present review covers a large set of different
samples, a general description of the crystal preparation is irrelevant. Details on all cases described below
can be found in the respective publications given in
For the preparation of suitable TEM samples, 100 ml
of the emulsion was diluted in 25 ml distilled water,
after which it is assumed that additional growth of the
crystals stops. A drop hereof is then placed on the TEM
Cu grids, covered with a carbon film. After evaporation
of the water, a well-oriented monolayer of AgX tabular
crystals is left on the carbon film. All these preparations
are performed under red light, as is the transport and
insertion into the microscope. The X-ray diffractometry
is performed with a Siemens (Germany) diffractometer
D500/501 instrument operating with the Cu Ka line
(l 5 0.1540 nm). The TEM results were obtained in a
side entry JEOL (Tokyo, Japan) 100C, operating at 100
keV, and a Phillips (Eindhoven, The Netherlands) CM20
microscope, operating at 200 keV. Both microscopes are
equipped with a liquid nitrogen cooling holder. Cooling
the specimen in the microscope is necessary to avoid
reaction with the electron beam (Goessens et al., 1991).
Precipitation of AgBr crystals at high supersaturation results in twinning along {111} type planes (Jagannathan and Gokhale, 1991; Jagannathan, 1991). The
twin planes are formed during the crystal growth
process, and thus are not deformation twins that arise
by lattice deformation.
Much has been written about the effect of twinning
on the enhanced growth rate of crystals. Twins in grown
diamond films were observed to enhance the growth in
the direction lying in the twin plane (Shechtman et
al.,1993a,b). Flat, hexagonal-shaped diamond platelets
were observed during the initial stage of microwave
plasma assisted deposition of diamond (Angus et al.,
1992). A reentrant groove was seen to be present in the
small side faces of the platelets, at which preferential
adsorption of growth material occurs. Also in gold and
silver particles, prepared by the solution reduction of
either chloroauric acid or silvernitrate with citric acid,
plate-like morphologies were observed (Kirkland et al.,
1989). High resolution TEM (HRTEM) studies on these
crystals showed the presence of multiple (111) twin
planes parallel to the major (111) surface. The tabular
morphology of twinned C60 crystals is explained by the
presence of the same reentrant groove on the small
sides of the crystals (Li et al., 1994b). A similar mechanism of growth at the reentrant groove is believed to
operate in the transformation of silicon from the amorphous to crystalline form (Brokman et al., 1986; Drosd
and Washburn, 1982). Hamilton and Seidensticker
(1960) proposed a model for the enhanced lateral
growth of twinned platelets. A key feature for the rapid
growth was argued to be the reentrant groove between
two adjacent {111} type planes at the twin plane (Hamilton and Seidensticker, 1960).
This model (Hamilton and Seidensticker,1960) is also
used to explain the growth characteristics of twinned
silver halide crystals. Moreover, it can explain all
morphologies by assuming preferential growth at the
reentrant corners that emerge where the twin plane
reaches the surface. Jagannathan et al. (1993) calculated for silver bromide crystals a probability of adsorption of a growth species at the reentrant groove which is
about 50 times higher with respect to adsorption at a
surface site. An isolated atom at a surface site on a (111)
plane has three nearest neighbours, while it has four at
the reentrant groove, resulting in a stronger bonding,
thus explaining the difference in growth rate. Ming and
colleagues (Ming and Sunagawa, 1988; Ming and Li,
1991) came to the same conclusion using Monte Carlo
simulations. They showed the twin lamella to act as
self-perpetuating sites of enhanced nucleation. They
also compared growth by the screw dislocation mechanism and the growth at steps on the surface created by
stacking faults with the growth at reentrant corners,
and concluded that growth by nucleation at reentrant
grooves is the most efficient at high supersaturation
(Jin et al.,1989; Jin and Ming, 1989; Li et al., 1994a;
Ming et al., 1988).
Parallel Twinning and Stacking Faults
The most common crystal resulting from the twinning process is the tabular crystal, which can have a
hexagonal or triangular morphology and has a typical
diameter of the order of 1 µm, while the thickness is
100–200 nm. From ultramicrotomies of these crystals it
could be concluded that they have in most cases two or
three twin planes parallel to the major (111) surface
(Black et al., 1983; Hamilton and Brady, 1964; Klein et
al., 1964a). A typical electron microscopic image of AgBr
tabular crystals is shown in Figure 1.
Next to some smaller crystals, three hexagonal and
one triangular crystal can be seen. Under normal
conditions pure AgBr tabular crystals are defect free,
apart from the twin planes that are invisible in this
A schematic drawing of the top and side views of a
hexagonal tabular crystal is presented in Figure 2a and
b, respectively, together with the relevant lattice directions and planes. The unit cell axes a, b, and c are
chosen for the top and bottom variants (labelled 1), the
central twin variant (2) having a mirrored set. Since the
twin planes are perpendicular to the [111] direction, the
lattice directions in the plane of Figure 2a belong to
both twin variants. Approximately 5 % of the crystals in
Fig. 1. Electron microscopical BF image
of AgBr tabular crystals. No contrast is observed from internal structural defects. The
darker regions are bend contours.
Fig. 2. Schematic representation of the (a) top and
(b) side view of a tabular crystal. The relevant directions and planes are indicated. The full arrows in a
indicate the a, b, and c axes of the fcc unit cell. In b1 is
the top and bottom variant; 2 the central one. The angle
a in b is 70°32’.
the emulsions are triangular in shape (i.e., the length of
the short edges is less than 10 % that of the long edges).
Macroscopically, the tabular crystals grow primarily in
the 71128 directions. On a microscopic scale, it is believed that the lateral growth of tabular crystals actually occurs along the (111), (11) and (111) planes (this
incomplete set is further on denoted as 51116) inclined
over an angle a of 70° 32’ with respect to the (111)
surfaces, and more specifically at the reentrant groove
between two 51116 planes belonging to different variants
(Berriman and Herz, 1957; Hamilton and Brady, 1958).
This mechanism indeed explains the shape difference
between crystals with an even (hexagonal) and an odd
(triangular) number of twin planes. In a hexagonal
crystal each side has one reentrant groove along which
preferential adsorption takes place. In triangular crys-
tals three sides have two reentrant grooves (these are
the faster growing sides), while the other three
sides have one reentrant groove (these are the slower
growing sides). The faster growing sides grow themselves out of existence, thus resulting in a triangular
It should be noted that the construction with parallel
twin variants implies the same 61116 edge plane on
opposite sides A and B of a given variant, as schematically represented in Figure 2b. Also the sequence from
top to bottom of the crystal of the acute and obtuse
corners changes between the A and B type edges.
Although the actual growth thus probably occurs along
the 51116 planes and the edges of the crystal consist of
planes in different orientations (see Fig. 2b), we will
frequently use the averaged growth direction 71128 or
Fig. 3. ED patterns of a AgBr tabular grain. a: [111] zone. b: [221] zone. The small white arrows
indicate the presence of small spots at commensurate positions.
the averaged 51126 projection when the description in
terms of the edge orientation is appropriate.
The side faces of tabular crystals have always been
thought of as being built up by 51116 planes, the
enhanced lateral growth being completely the result of
the presence of reentrant grooves. The argument behind this viewpoint is the fact that the tabular crystals
are precipitated at conditions under which (111) type
planes are the slow growing (and thus the stable) ones.
However, no firm experimental evidence has yet been
presented showing the side face geometry. Thus, it can
not be excluded that other non-(111) type planes take
part in the growth process. Different models have been
proposed, consisting of a cooperation of (100) surface
growth and reentrant corner effect (Ming and Sunagawa, 1988), or of exclusive (100) surface growth
(Jagannathan et al., 1993). In the latter model, the
surface at the side of the tabular crystal in between the
two twin planes would be composed of (100) planes,
being atomically rougher than (111) type planes, and
thus being the faster growing planes. The (100) growth
planes are self-perpetuating due to the presence of the
two parallel twin planes (Jagannathan et al., 1993). In
the light of the lack of experimental evidence for either
model, the experiments in this work will be interpreted
primarily based on the influence of the reentrant groove
on crystal growth.
When a well-defined AgBr tabular crystal is investigated as such in a TEM instrument, an electron diffraction tilt sequence around a [220] direction will confirm
the existence of the two twin variants from the sequence of obtained diffraction zones. Besides the expected Bragg reflections, sharp but weak spots appear
at commensurate positions in between the Bragg reflections (arrows in the [111] zone of Fig. 3a). When tilting
the specimen around, e.g., a [220] direction, these spots
do not disappear (see Fig. 3b) and their distance to the
spot in the perfect [111] orientation varies linearly with
the cosine of the tilt angle. This indicates that they are
due to the intersection of the Ewald sphere with a thin
line of diffracted intensity parallel to [111]. In fact,
when the entire reciprocal space is scanned, it is seen
that a streak runs through every Bragg reflection for
which h 1 k 1 l C 3n, n being an integer. The spot at 1/3
[224] in Figure 3a, e.g., is the intersection of the streak
through the (111) reflection above this [111] section.
Based on the length and sharpness of the streaks in
reciprocal space it is concluded that they originate from
a large number of non-periodic planar defects parallel
to the (111) surface of the tabular crystals. Two or three
twin planes with an average distance of 30–50 nm can
indeed not account for this diffraction effect (Amelinckx
et al., 1978; Herremans, 1983; Herremans et al., 1980,
1985; Van Dyck et al., 1984). Based on the fact that no
streaking is observed for h 1 k 1 l 5 3n, and thus g.R 5
n, n being an integer, it is concluded that the causing
defects are stacking faults (SF’s) with a displacement
vector of 1/3[111]. As extrinsic faults only alter the
Ag-Ag or Br-Br nearest neighbour connections, they are
probably preferred over intrinsic ones that affect both.
Non-Parallel Twinning
In emulsions precipitated at high pAg values, most
crystals have two or three parallel (111) twin planes,
resulting in thin tabular sheet crystals with a triangular or hexagonal form. However, less than 1% of the
total population of crystals shows different morphologies, resulting from non-parallel twin planes. Two
different kinds of morphologies will be discussed here,
i.e., needle-shaped crystals and crystals that have a
tetrahedral form. In both cases, it will be shown that
these particular crystal shapes result from the presence
of non-parallel twin planes.
Needle Crystals. The diameter of needles (about
100–150 nm) was found to be of the same order of
magnitude as the lateral size of the tabular microcrystals after the physical ripening, after which the latter
grow to become tabular crystals. This indicates that the
growth of the needles only occurs along the long edge of
the crystal and that growth in all other directions is
Fig. 4. a: ED pattern showing the superposition of three [012]
patterns. The 200 type spots belonging to the three variants are
indicated. The subindex denotes the variant. b: Schematic representation of the pattern in a. The squares, triangles, and circles denote
reflections from variant 1, 2, and 3, respectively. Open circles denote
reflections originating from double diffraction between variant 1 and 3
and between variant 2 and 3.
inhibited. The length is about ten times the diameter of
the tabular crystals after the precipitation, indicating a
unilateral growth that is about ten times faster than
the growth of tabular crystals.
If a needle-shaped crystal is examined in an electron
microscope, a diffraction pattern as the one shown in
Figure 4a can be obtained upon proper tilting of the
needle. The diffraction pattern can be indexed by
assuming a superposition of three [012] zones, i.e., by
assuming the existence of three variants. A schematic
representation of the diffraction pattern in Figure 4a is
shown in Figure 4b, in which the three 200 diffraction
spots are indicated. The subscript indicates the variant
to which the spot belongs. The open circles indicate
diffraction spots that arise from double diffraction
between variants 1 and 3 or between variants 2 and 3.
It should be noted that no double diffraction spots are
present from variants 1 and 2, indicating that there
exists no spatial overlap between them.
The corresponding bright field (BF) transmission
image of the needle in the same orientation is shown in
Figure 5a, in the proper orientation relationship with
respect to the diffraction pattern of Figure 4a. Over the
entire length of the needle, a fine black contrast line is
visible, suggesting the existence of a planar defect
parallel to the viewing direction. When the needle is
examined in the dark field (DF) mode using the 2001
diffraction spot, only one half of the needle, i.e., on the
lower side of the contrast line, lights up, which is shown
in Figure 5b. The other side is visible when using the
2002 diffraction spot to image the crystal. When using
the 2003 spot to image the crystal, parts of the needle on
both sides of the contrast line light up, as is shown in
Figure 5c, confirming the indexing and double diffraction of the diffraction pattern in Figure 4a and b.
However, it can also be seen in Figure 5c that the entire
needle does not contribute to the image, but that on
both sides near the surface of the needle, dark regions
remain. This indicates that, under this projection,
variant 3 does not extend over the entire width of the
needle but does extend up to the tip of the needle. It
should also be noted that variant 1 (Fig. 5b) extends
from the contrast line up to the edge of the crystal (as
does variant 2), but that these variants do not extend
up to the tip of the needle, as is seen in Figure 5b.
Combining the information from the DF transmission images with the existence of double diffraction
spots between variants 1 and 3 and between variants 2
and 3, a schematic view perpendicular to the viewing
direction in Figure 5 can be deduced, as is shown in
Figure 6.
The edge morphology under this projection cannot be
deduced from the current experiments. It is, therefore,
represented as a circular edge. The numbers in the
different parts of the needle denote the variant.
The twin plane between variants 1 and 2 can be
deduced from the diffraction pattern in Figure 4a and b.
Indeed, it is seen that the 442 diffraction spot for
variant 1 and the 442 diffraction spot for variant 2 are
common, the [221] direction (respectively, the [221]
direction) being perpendicular to the interface (and to
the long axis of the needle), indicating that this planar
defect is a (221) type twin plane, which is viewed edge
on. The viewing directions for the three variants in
Figure 6, i.e., the [021] direction for variant 3, the [542]
direction for variant 1, and the [542] direction for
variant 2, do not necessarily correspond to the long axis
(and thus the growth direction) of the needle. In Figure
5a, the needle may very well be rotated along a
direction perpendicular to the long axis, so that the
[012] direction (viewing direction in Fig. 5a) is not
perpendicular to the long edge. Therefore, the 1–3 and
2–3 interfaces are for now still represented by a dashed
line since the exact orientation of these interfaces can
not be deduced from the diffraction pattern and images
in Figures 4 and 5.
Fig. 6. Schematic representation of a side view with respect to the
viewing direction of Figure 5, as indicated.The numbers in the needle
indicate the variant. The interface between variant 1 and 2 is a (221)
type twin plane. The interfaces between variant 1 and 3 and between
variant 2 and 3 are represented by dashed lines.
Fig. 5. a: Bright field electron transmission image of part of a
needle crystal. A black contrast line is visible over the entire length of
the needle. b: DF image of the same needle using the 200 spot of
variant 1. c: DF image of the same needle using the 200 spot of variant
3. The needle is correctly oriented with respect to the diffraction
pattern of Figure 4.
By tilting the needle around the [242]1 (5[242]3)
direction other zones are obtained on which the same
exercise using multiple patterns can be performed.
From this it is concluded that the remaining interfaces
(dashed in Fig. 6) are parallel with (111) type planes
and that the long edge of the needle, and thus the
average growth direction, is proven to be a [110] type
direction for all three variants. Moreover, by combining
all information from reciprocal and real space it could
be concluded that the surfaces of these needles actually
consist of {111} planes as proposed by Klein et al. in
1964 (1964b), and shown in Figure 7, resulting in three
reentrant groves at the tip. The one-dimensional growth
of the needle-shaped crystals is thus due to the presence of three non-parallel twin planes in much the same
way as parallel twins give rise to tabular crystals.
Tetrahedral Crystals. A second kind of morphology related to the occurrence of non-parallel twinning of
{111} type planes, is shown in Figure 8a. The image was
Fig. 7. Schematic representation of the end of the needle showing a
possible edge morphology. Growth of the needle occurs at the three
reentrant grooves (dashed lines) originating from the intersection of
(111) type growth planes. The interfaces between the different variants are indicated by the three shaded regions.
recorded under two beam conditions using a 220 type
diffraction spot. The crystals show in transmission a
triangular morphology, and their size is typically 400
nm or more, i.e., larger than the width of the needle
crystals discussed previously, to such an extent that in
most crystals transmission of the electron beam is
greatly hampered due to absorption. However, the
thickness is not uniform, and some parts of the crystals
are thin enough to result in observable diffraction
Fig. 8. a: DF electron microscopic image of a tetrahedral crystal
using a 220 type spot. Two regions of darker contrast indicate the
presence of two interfaces inclined over an angle with respect to the
electron beam. Thickness fringes are visible near all three edges of the
crystal, revealing the tetrahedral shape. b: Schematic representation
of the diffraction pattern of the crystal in a. The diffraction pattern
consists of the overlap between a [111] type zone, and two [115] type
zones, each zone originating from a different variant. All spots of the
different variants coincide in this orientation.
spots. It can be seen in Figure 8a that near the edge, the
crystal is transparent to the electron beam, while
absorption of the electron beam is dominant in the
central part of the crystal. This is consistent with the
contrast fringes observed near (and parallel to) all
edges of the crystal, which indicates an increasing
thickness when going from the edge of the crystal to the
central part.
Apart from the thickness fringes, a different phenomenon is observed in the upper part of the crystal in
Figure 8a. At two of the three edges of the crystal,
triangular darker regions are observed, starting at the
edge and going to the centre of the crystal. This
difference in contrast is due to the particular choice of
diffraction spot used to image the crystal, i.e., three
different variants are present, and the diffraction spot
used to image the crystal only belongs to one variant.
Thus the two other variants do not contribute to the
image formation. The two regions of darker contrast
can be readily interpreted as being the result of the
presence of two twin related variants on top of the
central variant. The two twin related variants, which
will be called variants 2 and 3, are inclined over an
angle with respect to the electron beam. The central
larger variant, exhibiting a more uniform contrast, will
be called variant 1. From tilting experiments and
examining the thickness fringes in variant 1, it can be
seen that variant 1 has a tetrahedral morphology, while
variants 2 and 3 are thin twinned slabs with a typical
thickness of a few tens of nm. This results in a
morphology of the whole crystal, which is almost tetrahedral, as is confirmed by examining carbon replicas of
these crystals shadowed with a gold film (Van Roost,
personal communication).
The corresponding diffraction pattern at zero tilt
angle can be indexed as a [111] zone, without any
observable extra spots caused by the other two variants. However, by a combination of tilting and dark
field imaging, it could be concluded that the diffraction
pattern consists of a superposition of a [111] zone and
two [115] type zones, each one belonging to different
variants, the [111] zone belonging to variant 1. A
schematic representation of the diffraction pattern is
shown in Figure 8b, in which the relevant diffraction
spots are indexed, the subscript indicating the variant
to which the spot belongs. All spots of the different
variants coincide in this orientation.
The presence of the three different variants is revealed in the diffraction pattern when tilting the crystal
slightly away from the exact [111] orientation (with
respect to variant 1) around the [202] direction, as is
shown in Figure 9a. This direction coincides with a
[114] type direction for variants 2 and 3, as can be seen
in Figure 8b. Figure 9b gives a schematic representation of this diffraction pattern. The black circles, squares,
and triangles denote diffraction spots from variant 1,
variant 2, and variant 3, respectively. Open squares
and triangles indicate double diffraction spots between
variant 1 and 2 and 1 and 3, respectively. The diffraction pattern in Figure 9a shows distinct features that
can be attributed to variants 2 and 3. It can be seen that
the row of diffraction spots on the tilt axis are unsplit,
while the rows parallel to the latter show spot splitting.
The diffuse spots on these latter rows indicate the
positions of 220 type spots, which belong to variant 1
and are positioned in reciprocal space slightly above or
below the Ewald sphere. It can be seen that the 220
type spots of variants 2 and 3, which are indicated by
white arrows in Figure 9a, are displaced from these
positions in the [552] direction, i.e., they do not longer
coincide. Moreover, they are displaced in the opposite
direction when tilting in the opposite sense, the amount
of displacement in reciprocal space being proportional
to the tilt angle. From these observations it could be
concluded that the spots of variants 2 and 3 are in fact
the intersections of the Ewald sphere with a streak
running in the [111] direction, both for variants 2 and 3.
This is in agreement with the fact that the [111]
direction projected onto the [115] zone in reciprocal
space coincides with the [552] direction, which explains
Fig. 9. a: ED pattern obtained when tilting the crystal in Figure 8
slightly away from the exact [111] orientation (with respect to variant
1) around the [220] direction. This direction coincides with a [114] type
direction for variants 2 and 3. The diffuse spots on these latter rows
indicate the positions of [220] type spots of variant 1 slightly above or
below the Ewald sphere. b: Schematic representation of the diffraction
pattern in a including indexation.
Fig. 10. Perspective drawing of a tetrahedral crystal. The larger, central variant 1 is in
contact with two twin related, thin slabs
(variants 2 and 3), enclosing an angle of 70.5°
with each other. An extra (441) type mirror
plane is introduced due to the (111) type
twinning. A view along the direction indicated in a by an arrow is shown in b. This
direction corresponds to the viewing direction in Figure 8.
the observed direction of movement of the 220 type
spots of variants 2 and 3 in the latter direction.
When it is assumed that all ending planes of the
crystal are (111) type planes (which is not unreasonable
given the precipitation conditions), and taking the
double non-parallel twinning into account, a crystal
like the one shown in Figure 10 is obtained. When the
crystal is viewed along the direction indicated in Figure
10a, the situation as illustrated in Figure 10b is obtained, which corresponds to the image shown in
Figure 8a.
Based on the TEM observations on tabular, needleand tetrahedral-shaped crystals, a general growth model
explaining these different morphologies was proposed
(Goessens et al.,1997b).
Location of Ag1 Interstitials
When using long exposure times (30–45 seconds) and
a defocused electron beam, weak diffuse intensities
become visible on electron diffraction patterns of pure
AgBr and AgBr-AgBrI core-shell tabular crystals. No
increase in intensity of the diffuse contours was observed after irradiation of the crystals for 30 minutes,
indicating that the observed effect is not beam induced.
This diffuse intensity is especially visible when the
crystal is tilted away from an exact zone. Moreover, a
different geometry is observed between hexagonal and
triangular tabular crystals. In Figure 11a and b, diffraction patterns of a hexagonal and triangular tabular
crystal, respectively, are shown. Here, it can be seen
that the diffuse lines at the 224 spots have a symmetry
close to threefold and sixfold, respectively. In fact, the
complete geometry of the diffuse intensity found in
triangular crystals can be considered as a superposition
of the diffuse intensity observed in hexagonal crystals
plus the mirror image hereof with respect to the (111)
reciprocal plane corresponding with the normal to the
Fig. 11. Electron micrographs showing diffuse intensity contours
in hexagonal (a) and triangular (b) tabular AgBr crystals. The
micrographs were obtained by tilting slightly away from the [111]
crystal surface. On the other hand, the diffuse intensities observed in ED patterns of cubic as well as octahedral crystals show the same symmetry as in the hexagonal tabular crystals. So it seems that the diffuse
intensity observed in hexagonal crystals reflects the
inherent ordering in AgBr material. The different behaviour in triangular crystals can be explained by the
different variant morphology of triangular crystals
with respect to hexagonal ones.
The appearance of diffuse intensity contours in ED
patterns can be explained by the existence of shortrange-order (SRO) in the crystal (De Ridder et al., 1976,
1977; Van Dyck et al., 1977). In the present case, SRO of
interstitial Ag1 ions located close to the surface or to
twin planes is considered. It can be shown that the
occurrence of the diffuse intensity can be attributed to
the presence of clusters of Ag1 ions of definite size on
the simple cubic lattice of interstitial positions. On this
lattice, the clusters consist of small portions of (001)
planes as shown in Figure 12.
The observed ordering is independent of the crystal
morphology, and seems to be inherent to the AgBr
material. Arguments are given in favour of the idea
that interstitial Ag1 ions are mainly concentrated in a
sub-surface layer (Goessens et al., 1994). The different
geometry of the diffuse intensity locus observed between triangular and hexagonal tabular crystals can be
explained in this way. Critical investigation of the
symmetry of the diffuse intensity indicates that the
common definition of hexagonal vs. triangular crystals
and the correlated implication of an even or odd num-
Fig. 12. Schematic representation of the cluster of Ag1 ions as
derived from the cluster theory. The atom positions (dark circles) are
situated on a (010) plane of the simple cubic interstitial lattice. The
position vectors rk are indicated. The distance between two adjacent
atom positions is half the length of the AgBr unit cell.
ber of twin planes, respectively, is not always justified
(Goessens et al.,1997a).
In this section several effects of the incorporation of
iodine in the shell of core/shell AgBr/AgBrI tabular
crystals will be discussed. When an X-ray spectrum is
made of an emulsion of AgBr/AgBrI tabular crystals,
the diffraction angle of the extra peak due to the AgBrI
regions decreases when the percentage iodine increases
from which an increase of the {222} planar distance
from 0.166 nm to 0.169 nm can be calculated. This
increase implies that some of the Br- ions are substituted by the larger I- ions without changing the NaCl
structure. Both core and shell thus exhibit the NaCl fcc
structure, the former having a smaller lattice parameter than the latter. These results indicate that most, if
not all, of the I- ions are uniformly distributed in the
shell of the mixed crystals and that, at least for the 5
at.% and the 10 at.% cases, no other phases are to be
Inclined Stacking Faults
In Figure 13 two DF images of a typical tabular
crystal with a uniformly mixed shell (iodine coprecipitated at 10 at.% of the anion concentration) are presented. In the shell region numerous contrast lines,
parallel with all three 71128 edges are visible. These
Fig. 13. Two two-beam situations in which extinction of the line contrast occurs. The operating
diffraction vectors are indicated. Notice the extinction of the contrast lines in one segment (a and b).
indicate the existence of crystallographic defects in the
shell region and parallel to the edges. Under the
present two-beam conditions, the contrast lines are
found to disappear in some segments of the shell. From
these extinctions, the nature of the crystal defect can be
determined by considering the product g.R 5 0 conditions (Brown, 1971). Careful examination of the sign of
the contrast lines reveals that the black-white contrast
reverses between 1g and 2g DF images. Combining all
different results from images as the ones shown in
Figure 13 indicates that these contrast lines are due to
51116 fcc stacking faults (SF) with R 5 1/3[111].
Although one does not a priori know in which variant
the stacking faults occur, the actual choice of the
present reflections does not yield conflicting results for
both variants. Indeed, the 220 reflection belongs to both
variants and the extinction rule thus applies to the
(111) planes for both variants. The 204 reflection,
however, only belongs to one of the two twin variants.
As a result, the other variant will simply not contribute
to the diffraction and the extinction condition reveals
the present information about only one variant. Tilting
the sample so as to obtain the same two-beam conditions but for the other variant results in the same but
reverse situation, again giving extinction.
In Figure 13a a strong black or white contrast is
observed at the interface between the AgBr core and the
AgBrI shell in those segments without extinction. In
Figure 13b this contrast is less prominent, but it is now
also observed at the interfaces of the segments showing
extinction for the stacking faults. This can be explained
by assuming atom shifts at the interfaces perpendicular to the respective 51116 habit planes and continuously
decreasing when moving into the shell. The extinction
thus only occurs for g.R 5 0, as in Figure 13a, and not
for g.R integer but different from 0, as in Figure 13b.
The different nature (black vs. white) on both sides of
the crystal is evident from the increase of the lattice
parameter, which implies a displacement field pointing
away from the centre on all interfaces. As a result, the
sign of g.R changes and thus also the nature of the
contrast. This strong contrast thus results from the
distortion field due to the interface misfit between the
AgBr and AgBrI phases.
All tabular crystals with uniformly mixed shells
grown at high pAg exhibit the stacking faults presented
in the previous section. Since most crystals do not
reveal any dislocations and since in those that do, the
dislocation direction or place is not correlated with the
stacking faults, these stacking faults must end at the
surface or at an interface such as the (111) twin planes.
From the X-ray diffractometry it follows that the crystal structure of the shell is still of the NaCl type but
with a slightly expanded lattice parameter, the change
being linearly related to the amount of I2 ions that
randomly replace some of the Br2 ions. The lack of
added diffraction intensity in the ED pattern for the
mixed shell with respect to the pure core or crystal
confirms this conclusion. Thus, except for the interface
region between core and shell, the structure of the shell
should be of a uniform fcc type.
At first sight, no lattice defects are expected away
from the misfit region at the interface. However, when
it is considered that the lateral growth of the crystals
occurs at 51116 faces joining at reentrant (and protruding) angles and that the lattice parameter in the shell is
slightly larger than in the core, it follows that the (111)
planes joining on both sides of the twin boundary can be
forced closer together than in a free growing situation.
As a result, an energetically unfavourable configuration of, e.g., nearest neighbour Br2 anions would occur,
as indicated by the double arrow in the centre of
Figure 14.
Here, a [110] section at a (111) twin plane is presented schematically, with the AgBr core on the left and
the AgBrI shell on the right, the interface being clearly
indicated; in this picture the crystal thus grows from
left to right. Possibly this anion-anion configuration
causes too much strain in the lattice, which then could
be accommodated by the introduction of a SF in one of
both twin variants. Alternatively, when at a given
moment during the growth a larger iodine anion is
attached at one side of the twin plane, the extra stress
can now introduce an SF in the adjacent variant, again
Fig. 14. [110] projection showing the possible atomic configurations yielding SF’s in the shell region as
a result of the increase in lattice parameter and local substitution of Br2 by I2 anions.
increasing the distance between the anions. The displacement vector indicated in the drawing is R 5
1/6[112] being equivalent with R 5 1/3[111]. When all
(111) lattice planes are considered, the present SF
configuration is that of an extrinsic one. The stacking
can indeed be described as ...c8ab8ca8bc8ba8bc8... showing two non-penetrating hexagonal blocks (unprimed 5
anions; primed 5 cations) (Sprackling, 1984). It should
be noted that this definition of extrinsic is not consistent with the one used by Fontaine (1967) and Tasker et
al. (1981), who only use the stacking of one of the
In Figure 14 it is seen that this type of SF does not
introduce nearest neighbour anion configurations, as
would occur for an intrinsic SF, and is thus more likely
to occur. In fact, when this configuration is compared to
that around the twin plane, it can also be interpreted as
a small (111) twin slab. This type of configuration was
not considered by Fontaine (1967) and Tasker et al.
(1981) who calculated the energies of planar defects in
ionic crystals. Sprackling (1984) already pointed out
that this twin configuration is expected to have a lower
energy than the stacking faults considered in the
atomistic calculations. From Figure 14 it can also be
seen that, after the growth of an SF, the atom configurations at the (111) twin plane is changed, which could
lead to a shift of the twin plane to an adjacent (111)
plane. It is clear that these variations plus the continuing mixing with iodine will yield stacking faults in the
entire shell between the interface and the edge. It can,
moreover, be imagined that the varying atom configura-
tions around this twin plane in the shell could locally
provide extra open spaces, which could improve the
cation mobility on this twin boundary.
Recently the existence of these stacking faults was
confirmed by Chen et al. using TEM and HRTEM
images of cross-sectioned AgBrI tabular crystals (Chen
et al., 1997).
Long-Range Iodine Ordering in Surface Layer
The crystals studied above were precipitated at a pAg
of 8.8 at 70°C. It cannot be excluded that some thickness growth occurs at this pAg value. As a result of a
small thickness growth, long period fringes caused by
the long-range ordering of iodine ions in the surface
layer are observed both in the core region as in the shell
region. Due to the relatively low magnification of the
crystals in the previous sections, these fringes are not
When the core region of a AgBr-AgBrI core-shell
tabular crystal is investigated by conventional twobeam bright and dark field TEM, an effect of this
iodine-enriched thickness layer can be observed. In
Figure 15 a DF image of such a tabular crystal is
shown. The local iodine concentration in the shell is 10
at.%, i.e., with respect to the anions. The image was
recorded under two beam conditions, using the 220
diffraction spot. It can be seen that fine parallel fringes
extend over the entire area of the crystal, both in the
core and in the shell. An enlarged part of the image in
the core region is shown as an inset. Two sets of fringes
can be observed, running parallel to the edges of the
Fig. 15. Electron microscopic dark field image of a tabular crystal, using the 220 diffraction beam,
showing fringe patterns in two ,110. type directions. The insets show a Fourier spectrum (top) and an
enlarged part of the core region (bottom).
crystal in the [101] direction and the [011] direction.
The Fourier transform of this image (see inset) shows
spots in directions perpendicular to the above-mentioned directions, enclosing an angle of 60°. When
examining the same crystal under other two-beam
conditions, e.g., using the 202 diffraction spot, the
fringes running in the [101] direction become extinct,
while fringes are now observed in the [110] and the
[011] directions. As expected, the third combination is
obtained using a 022 reflection. These observations can
be summarized by stating that three sets of fringes are
present, but due to the particular choice of the diffraction vector used to image the microcrystal, the fringes
parallel to the direction of the operating diffraction
vector are extinct. When combining the different extinctions for the 220 type diffraction spots, and using the
g.R 5 0 extinction criterion, i.e., 2h - 2k 5 0, it can be
concluded that the displacement vector has to be of the
form: R 5 [hhl]. The value of l can not unambiguously
be determined from the observed extinctions.
It has to be emphasized that the observed effect
arises due to the addition of iodine during the growth of
the crystal, and that these fringes were never observed
in pure AgBr tabular crystals. When the concentration
of iodine in the shell of the crystal is varied, the spacing
between the fringes changes from about 6.5 nm for the 5
at.% case to about 5.6 nm for the 10 at.% case.
The present observations can be explained by assuming a long-range ordering of the iodine ions in the
surface layer of mixed AgBr-AgBrI core-shell type
tabular crystals (Goessens et al., 1995). This ordering is
due to a preferential adsorption on the reconstructed
AgBr surface as shown schematically in Figure 16.
This preferential adsorption is argued to be a consequence of the existence of a long period charge density
modulation on the surface of the crystals.
Variant-Selective Dislocation-Imaging
In Figure 17a and b, two-beam dark field images are
shown of a AgBr crystal in which the iodine is build in
by a conversion process. The diffraction vectors used to
image the crystal are indicated. The incorporation of
iodine by conversion results in the presence of a fine
contrast band between the core and the shell, parallel
to the edges of the crystal. The contrast band becomes
extinct when imaging the crystal with a 220 type
diffraction spot parallel to the contrast band, as can be
seen in Figure 17b. This indicates that the contrast
band results from local strain in a direction perpendicular to the band, as discussed before. This strain is
produced by the lattice distortion created by the incorporation of a high amount of iodine. Due to the local
high concentration of iodine, a large number of dislocations are also created, starting at the strain region and
extending to the edge of the crystal. Since the concentration of iodine along the interface is probably not constant, the dislocations are not equidistant, but show a
more or less random occurrence. As can be seen by
comparing Figure 17a and b, these dislocations become
extinct for 220 type diffraction vectors. As a reference,
the arrows in Figure 17 are placed at the same place
with respect to the crystal. The behaviour of the
dislocation contrast under different two-beam situations is in agreement with a Burgers vector of the type
a/2[110], as would be expected for NaCl type crystals
(Fontaine, 1967; Sprackling, 1984). There seems to be
Fig. 16. Schematic representation of the
charge modulation as a function of distance
on the (111) surface of a tabular crystal. The
viewing direction is along one set of fringes.
Conversion of bromine by iodine will preferably occur in the regions of higher charge,
resulting in a composition modulation. The
subindex c stands for complex.
Fig. 17. Enlarged part of a crystal in which iodine ions are
incorporated in the AgBr matrix by conversion. The BF images are
recorded under two-beam conditions using 220 type of diffraction
vectors (indicated in the figure). A black contrast band is visible
parallel to the edge of the crystal in a, and is extinct in b. Dislocations
are observed running from this band to the edge of the crystal,
indicated by the arrowheads. Some dislocations are seen to be extinct
under the current imaging conditions. A second group of dislocations
start at the contrast band and extend towards the core of the crystal
(small black arrows).
no correlation between the type of Burgers vector, i.e.,
of the type a/2[110] and a/2[110], and the localisation of
the dislocation in the crystal.
A general property of these dislocations is that they
run from the strained region more or less perpendicular
to the interface towards the edges of the crystal through
the shell region, i.e., that they lie more or less in [112]
type directions. However, in many crystals a second
group of dislocations can be observed. The latter also
originate at the strained region but extend to the
interior of the crystal, i.e., towards the core region, as
can be seen in Figure 17b. The same group of dislocations is present in Figure17a, but their contrast is
partly obscured by the interface contrast. Since this
dislocation contrast does not extend all over the crystal,
but is seen to stop inside the core, it has to be concluded
that they end at the top or bottom surface. These
dislocations have the same Burgers vector of the type
a/2[110]. Since they end at the surface of the crystal,
they might serve as preferred nucleation sites in the
chemical sensitization process and for the latent image
(Mitchell, 1994).
The observation of diffuse intensity contours in tabular crystals combined with two-beam dark field techniques, allows us to selectively image the crystal defects present in the different twin variants in a tabular
crystal (Goessens et al.,1997a). The reciprocal lattice of
a twinned tabular crystal consists of Bragg spots origi-
nating from the AgBr fcc structure and from the same
structure mirrored over the (111) plane. In principle,
one does not a priori know which zone belongs to which
variant. However, since it is known that the diffuse
intensity in the diffraction patterns originates from
ordering in the sub-surface layer, and since in doubly
twinned tabular crystals the top and bottom subsurface layer are situated in the same variant, a
differentiation can be made by examining the geometrical features of the diffuse intensity contours with
respect to the Bragg spots.
The correlation of specific diffuse intensity features
with Bragg reflections of a given variant enables one to
perform two-beam dark field imaging using diffraction
spots belonging only to variant 1 or variant 2. Recording an image under such two-beam conditions, an
image of the top and bottom variant can be obtained,
since the other variant simply does not contribute to the
electron diffraction. The same is valid for different
kinds of defect contrast, such as dislocation contrast,
i.e., dislocations present in variant 1 will not be visible
when viewing the crystal under two-beam dark field
conditions using a diffraction spot of variant 2 and vice
versa. The latter contrast results from local changes in
diffraction conditions induced by the presence of the
dislocation. If extinction of the contrast is accounted for,
this allows us to determine whether a dislocation is
present in variant 1 or variant 2. In Figure 18a–d, part
Fig. 18. Two-beam dark field images showing the defect structure
of a AgBr crystal in which iodine ions are incorporated by conversion.
Part of the same crystal is shown under different diffraction conditions. a was recorded using the 202 spot, which is common to both
variants. Under these imaging conditions, defect contrasts originating
from the two variants will be superimposed on each other. b shows the
same crystal imaged with the 111 beam of variant 1. c and d were
obtained using the 111 spot and the 024 spot of variant 2, respectively.
All arrowheads are placed at the same position.
of the same crystal is shown under different diffraction
conditions. Figure 18a was recorded using the 202 spot,
which is common to both variants. Under these imaging
conditions, defect contrasts originating from the two
variants will be superimposed. Figure 18b shows the
same crystal imaged with the 111 beam of variant 1.
Figure 18c and d were obtained using the 111 spot and
the 024 spot of variant 2, respectively.
A first remarkable observation is the fact that the set
of dislocations starting at the misfit region and extending downwards towards the core of the crystal (small
black arrowheads in Fig. 18a and b) are visible in
Figure 18b while they are not visible in Figure 18c and
d. This implies that they are present exclusively in the
top or bottom variant, starting inside the crystal at,
e.g., the high iodine region at the twin plane and
reaching the surface in the core region. The dislocation
in the core of the crystal in the left part of the crystal is
present in variant 1 for the same reason.
Still, difficulties in image interpretation arise for the
dislocations extending from the misfit region to the
edge of the crystal due to their high density in a region
with prominent strain contrast. Moreover, it is possible
that a dislocation in one variant induces lattice distortion in the other variant. In some cases, however, the
method is still applicable, as for the dislocations indicated by white arrowheads in Figure 18.
The dislocation indicated by the first white arrowhead from the left of the figure (which is extinct in Fig.
18a) is situated in variant 2. No contrast is observed in
Figure 18c due to extinction. However, the dislocation is
clearly visible in Figure 18d. A Burgers vector a/2[101]
satisfies all the extinction and visibility conditions. The
next indicated dislocation, which seems to be dissociated into two dislocations, is present in variant 1 and
does not show extinction under the given imaging
conditions. The next dislocation again is situated in
variant 2 and shows the same image characteristics as
the first one. Finally, the last indicated dislocation is
situated in variant 2, and does not show extinction
under the present imaging conditions.
The present review describes most of the results of
our transmission electron microscopy and diffraction
studies on defects in tabular and other twinned silver
halide microcrystals. These defects range from zerodimensional ones like the short-range-ordering of Ag1
ion interstitials over one-dimensional dislocations to
two-dimensional stacking faults and twin planes. Some
are intrinsic to the basic AgBr material while others
result from the introduction of iodine.
The authors like to thank R. De Keyzer and C. Van
Roost for their continuous practical support and scientific input and the Flemish Institute for the Encouragement of Scientific and Technological Research in the
Industry (IWT) for financial support.
Amelinckx, S., Gevers, R., and Van Landuyt, J. (1978) In: Diffraction
and Imaging Techniques in Material Science, Vol. 1 and 2. NorthHolland, Amsterdam.
Angus, J.C., Sunkara, M., Sahaida, S.R., and Glass, J.T. (1992)
Twinning and faceting in early stages of diamond growth by
chemical vapor deposition. J. Mater. Res., 7:3001–3009
Berriman, R.W., and Herz, R.H. (1957) Twinning and the tabular
growth of silver bromide crystals. Nature, 180:293–294.
Black, D.L., Timmons, J.A., and Bottger, G.L. (1983) Direct observation of twin planes in silver halide grains. Proc. East-West Symp. 2,
Hawaii C71–C76.
Brokman, A., Gat, R., and Alpern, Y. (1986) Morphological effects
during low pressure chemical vapor deposition and annealing of
undoped polycrystalline silicon layers. Appl. Phys. Lett., 49:382–
Brown, L.M. (1971) Electron Microscopy in Material Science. U.
Valdrè, ed. Academic Press, New York, p. 377.
Burt, J. (1977) Effect of I- on interstitial silver ions in AgBr(I)
microcrystals. Photogr. Sci. Eng., 21:245–247.
Chen, S., Jagannathan, S., Mehta, R.V., Jagannathan, R., and Taddei,
A.E. (1997) Direct observation of stacking fault structure in (111)
tabular silver halide grains by high-resolution electron microscopy.
1997 Int. Symp. Silver Halide Imaging, pp. 44–48.
De Ridder, R., Van Tendeloo, G., Van Dyck, D., and Amelinckx, S.
(1976) A cluster model for the transition state and its study by
means of electron diffraction. I. Theoretical model. Phys. Stat. Sol.
(a), 38:663–674.
De Ridder, R., Van Dyck, D., Van Tendeloo, G., and Amelinckx, S.
(1977) A cluster model for the transition state and its study by
means of electron diffraction. II. Application to some particular
systems. Phys. Stat. Sol. (a), 40:669–683.
Drosd, R., and Washburn, J. (1982) Some observations on the amorphous to crystalline transformation in silicon. J. Appl. Phys.,
Fontaine, G. (1967) Calcul des énergies de fautes d’empilement dans
les composés ioniques. J. Phys. Chem. Solids 28:2553–2562.
Galvin, J.P. (1970) Sensitometric effects of iodide addition to silver
bromide emulsions. Photogr. Sci. Eng., 14:258–261.
Goessens, C., Schryvers, D., Van Landuyt, J., Amelinckx, S., Verbeeck,
A., and De Keyzer, R. (1991) Characterization of crystal defects in
mixed tabular silver halide grains by conventional transmission
electron microscopy and X-ray diffractometry. J. Cryst. Growth,
Goessens, C., Schryvers, D., Van Dyck, D., Van Landuyt, J., and De
Keyzer R. (1994) Electron diffraction evidence for ordering of
interstitial silver ions in silver bromide microcrystals. Phys. Stat.
Sol. (a) 143:277–287.
Goessens, C., Schryvers, D., Van Landuyt, J., and De Keyzer, R.
(1997a) New method to determine the parity of the number of twin
planes in tabular silver halide microcrystals from top views. J.
Goessens, C., Schryvers, D., Van Landuyt, J., and De Keyzer, R.
(1997b) Electron microscopical investigation of tetrahedral-shaped
AgBr microcrystals. J. Cryst. Growth, 172:426–432.
Goessens, C., Van Landuyt, J., Schryvers, D., Amelinckx, S., and De
Keyzer, R. (1995) Long period surface ordering of iodine ions in
mixed tabular AgBr-AgBrI microcrystals. Surface Science 337:
Granzer, F. (1989) Physical properties of phase boundaries in silver
halide crystals in relation to photography. Part I: Band structures of
abrupt phase boundaries between different silver halide crystals. J.
Im. Sci., 33:207–216.
Hamilton, J.F. (1988) The silver halide photographic process. Adv.Phys., 37:359–441.
Hamilton, J.F. and Brady, L.E. (1958) Twinning in tabular photographic grains. J. Appl. Phys., 29:994.
Hamilton, J.F. and Brady, L.E. (1964) Twinning and growth of silver
bromide microcrystals. J. Appl. Phys., 35:414–421.
Hamilton, D.R., and Seidensticker, R.G. (1960) Propagation mechanism of germanium dendrites. J. Appl. Phys., 31:1165–1168.
Herremans, W. (1983) Diffractie van elektronen aan een kristal met
stapelfouten bijna evenwijdig met de invallende bundel. Ph.D.
Thesis, University of Brussels.
Herremans, W., Gevers, R., and David, M. (1980) The electron wave
function in the vicinity of a planar defect ‘‘on edge’’. Phys. Stat. Sol.
A, 59:557–568.
Herremans, W., Van Tendeloo, G., David, M., and Gevers, R. (1985)
TEM images of a translation interface on edge. Phys. Stat. Sol. A,
Jagannathan, R. (1991) Twinned silver bromide crystals: Some insights into their formation and growth. J. Imag. Sci., 35:104–112.
Jagannathan, R. and Gokhale, V.V. (1991) Modeling of twinning
processes in silver bromide precipitation. J. Imag. Sci., 35:113–118.
Jagannathan, R., Mehta, R.V., Timmons, J.A., and Black, D.L. (1993)
Anisotropic growth of twinned cubic crystals. Phys. Rev. B, 48:13261–
Jin, J. and Ming, N. (1989) A comparison between the growth
mechanism of stacking fault and of screw dislocation. J. Cryst.
Growth, 96:442–444.
Jin, J., Ming, N., and Chernov, A. (1989) The growth mechanism and
kinetics on stacking fault reconsidered. J. Cryst. Growth, 98:341–
Kahan, A. (1977) Photohole motion in AgBr(I) microcrystals. Photogr.
Sci. Eng., 21:237–240.
Kirkland, A.I., Jefferson, D.A., Duff, D.G., and Edwards, P.P. (1989)
High resolution studies of trigonal lamellar particles of gold and
silver. Inst. Phys. Conf. Ser. No 98, Ch. 8:375–378.
Klein, E., Metz, H.J., and Moisar, E. (1964a) Zillingsbildung bei AgBrund AgCl-Kristallen in photographischen Emulsionen Teil II:
Beschreibung der vorkommenden Zwillingstypen und Folgerungen
für das Kristallwachstum. Photogr. Korr., 100:57–71.
Klein, E., Metz, H., and Moisar, E. (1964b) Mitteilingen aus den
Forschungslaboratorien der Agfa-Gevaert AG Leverkusen. SpringerVerlag, Berlin.
Li, H., Peng, X., and Ming, N. (1994a) Re-entrant corner mechanism of
fcc crystal growth of A-type twin lamella: The Monte-Carlo simulation approach. J. Cryst. Growth, 139:129–133.
Li, J., Mitsuki, T., Ozawa, M., Horiuchi, H., Kishio, K., Kitazawa, K.,
Kikuchi, K., and Achiba, Y. (1994b) Growth and morphology of C60
and C70 single crystals. J. Cryst. Growth, 143:58–65.
Ming, N., and Li, H. (1991) Twin lamella mechanism of fcc crystal
growth: The Monte-Carlo simulation approach. J. Cryst. Growth,
Ming, N., and Sunagawa, I. (1988) Twin lamellae as possible selfperpetuating step sources. J. Cryst. Growth, 87:13–17.
Ming, N., Tsukamoto, K., Sunagawa, I., and Chernov A. (1988)
Stacking faults as self perpetuating step sources. J. Cryst. Growth,
Mitchell, J.W. (1993) The silver halide photographic emulsion grain. J.
Im. Sci., 37:331–343.
Mitchell, J.W. (1994) Dislocations and donor centres in silver halide
emulsion grains. J. Photogr. Sci., 42:2–9.
Pischel, B., and Granzer, F. (1991) Physical properties of phase
boundaries in silver halide crystals in relation to photography. Part
II: The current-voltage behaviour of phase boundaries in silver
halides. J. Imag. Sci., 35:136–145.
Shechtman, D., Feldman, A., and Hutchison, J. (1993a) High-order
twin boundaries in CVD diamond films. Mater. Lett., 17:211–216.
Shechtman, D., Hutchison, J.L., Robins, L.H., Farabaugh, E.N., and
Feldman, A. (1993b) Growth defects in diamond films. J. Mater.
Res., 8:473–479.
Si-Yong, Z. (1986) Study of the properties of photographic emulsions
with varying iodide content. J. Imag. Sci., 30:16–21.
Sprackling, M.T. (1984) Disorder in tabular silver halide grains. I.
Stacking faults and twinning dislocations in rocksalt structure
crystals. J. Photogr. Sci., 32:21–24.
Steiger, R., Aebischer, J.-N., and Haselbach, E. (1991) Influence of
iodide ions on the adsorption of a cyanine dye on cubic AgBr
emulsion crystals. J. Imag. Sci., 35:1–14.
Tasker, P.W., and Bullough, T.J. (1981) An atomistic calculation of
extended planar defects in ionic crystals. Application to stacking
faults in the alkali halides. Phil. Mag. A 43(2):313–324.
Van Dyck, D., De Ridder, R., Van Tendeloo, G., and Amelinckx, S.
(1977) A cluster model for the transition state and its study by
means of electron diffraction. III. Generalisations of the theory and
relation to the SRO parameters. Phys. Stat. Sol. (a), 43:541–552.
Van Dyck, D., Van Tendeloo, G., and Amelinckx, S. (1984) Diffraction
effects due to a single translation interface in a small crystal.
Ultramicroscopy, 15:357–370.
Без категории
Размер файла
576 Кб
Пожаловаться на содержимое документа