Liquid-like interfacial correlation in LB films

Article abstract of DOI:10.1016/S0921-4526(98)00235-X

We present here results of X-ray reflectivity and diffuse scattering study of three Langmuir-Blodgett films, all having nine monolayers. These systems are the cadmium arachidate film grown using the conventional method, the arachidic acid film with trapped CdS layers at the interfaces formed by exposing the cadmium arachidate film to hydrogen sulphide and ferric stearate film grown from a monolayer of the pre-formed salt on water. Self-consistent analysis of all the scattering data measured, viz., in specular, in transverse and in longitudinal directions, for each film shows that the interfaces are conformal in nature and the logarithmic in-plane correlation, characteristic of capillary waves on liquid surfaces, exists in all the three films.

Full text of DOI:10.1016/S0921-4526(98)00235-X

Physica B 248 (1998) 217—222
Liquid-like interfacial correlation in LB films
M.K. Sanyal*, J.K. Basu, A. Datta
Surface Physics Division, Saha Institute of Nuclear Physics, 1/AF, Bidhannagar, Calcutta 700 064, India
Abstract
We present here results of X-ray reflectivity and diffuse scattering study of three Langmuir—Blodgett films, all having
nine monolayers. These systems are the cadmium arachidate film grown using the conventional method, the arachidic
acid film with trapped CdS layers at the interfaces formed by exposing the cadmium arachidate film to hydrogen sulphide
and ferric stearate film grown from a monolayer of the pre-formed salt on water. Self-consistent analysis of all the
scattering data measured, viz., in specular, in transverse and in longitudinal directions, for each film shows that the
interfaces are conformal in nature and the logarithmic in-plane correlation, characteristic of capillary waves on liquid
surfaces, exists in all the three films. ( 1998 Elsevier Science B.V. All rights reserved.
PACS: 68.18.#p; 61.46.#w; 68.10.!m; 68.55.!a
Keywords: Langmuir—Blodgett films; X-ray reflectivity; Diffuse scattering
1
. Introduction
diffuse intensity peaks near the specular ridge. In
these cases analysis of measured data is done by
calculating total scattering profiles in various direc-
tions in reciprocal space (q , q and q ) [3,5].
Systematic analysis of specular reflectivity data
provides us the electron density profile, and in turn
compositional profile as a function of depth [1].
Diffuse scattering studies, on the other hand, gives
us information regarding correlation of various in-
terfaces present in the system under study [2]. In
most cases, one can separate out [3,4] specular and
off-specular components of scattering and analyse
them separately. But for systems having large in-
plane correlation length, it becomes impossible to
separate out these components [2] because the
V
W
X
Langmuir—Blodgett (LB) films are deposited
by a simple process, that involves the sequential
transfer of an organic monolayer, spread normally
on water, by repeated dipping of a substrate [6,7].
The quality of interfaces are generally excellent
and one can measure X-ray reflectivity up to a
very large scattering angle. LB films of fatty
acid salts are generally formed by spreading a
corresponding fatty acid monolayer on water con-
taining dissolved inorganic salt. In this method
the number of metal ions and the number of
acid molecules available are not equal. This intro-
duces a large uncertainty in the composition of the
*
Corresponding author. Tel.: #91 37 534549; fax: #91 33
3
74637; e-mail: milan@cmp.saha.ernet.in.
0
921-4526/98/$19.00 ( 1998 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 1 - 4 5 2 6 ( 9 8 ) 0 0 2 3 5 - X

2
18
M.K. Sanyal et al. / Physica B 248 (1998) 217—222
monolayer, and, in general, the actual mechanism
of salt formation during the transfer of the mono-
layer to the substrate is not clear so far. However, if
an LB film is deposited from a monolayer of
a preformed salt of a fatty acid, such as ferric
stearate (FeSt) on water, at least the uncertainty in
the monolayer composition is much reduced.
Our understanding regarding the nature of cor-
relation in the interfaces [8] and the nature of
molecular stack [9] in these films are also far from
complete. The stability of the interfaces present in
these films are intriguing, too. It was shown re-
cently that by exposing a cadmium arachidate
ly alkaline (pHK8.0). Fourier transform infra red
spectra of the purified bulk FeSt sample, as well as
a 11 ML LB film of this FeSt on calcium fluoride,
could not detect the presence of hydroxy fatty acid
salts. A 9 ML LB film of this preformed and purifi-
ed FeSt was deposited on a Si(1 0 0) substrate at a
surface pressure, temperature and subphase (water)
pH of 30 mN m\ꢂ (surface tension 42 mN m\ꢂ),
25°C and 5.6, respectively. The deposition rate was
3 mm/min. Details of the sample preparation, puri-
fication and LB deposition have been provided
elsewhere [12].
(
CdA) LB film in an hydrogen sulphide (H S) envi-
ꢀ
ronment, cadmium sulphide (CdS) nanostructures
can be formed [10] without much disturbance to
the embedding LB structure. We have performed
an X-ray scattering study of three LB films to
understand the nature of interfaces present in these
systems and to probe molecular arrangements in
these films. In the next section we have described
these films and in the third section we have outlined
the experimental details and the formalism used to
analyse the measured X-ray scattering data. In the
last section the main results are discussed.
3. X-ray scattering study
X-ray specular reflectivity measurements were
performed on these films by keeping the angle of
incidence (a) equal to the angle of reflection (b)
while the diffuse scattering data were taken by
keeping these angles unequal [3,5]. The data was
collected using a triple axis spectrometer (optix
X-rays obtained
microcontrole) with Cu K?
ꢂ
from a rotating anode generator (FR 591, Enraf
Nonius) and a Si(1 1 1) monochromator. The inci-
dent beam was collimated by slits having horizon-
tal and vertical apertures of 100 and 5000 lm,
respectively, with the horizontal aperture determin-
ing the width of the incident beam in the scattering
plane. In case of diffuse scattering measurements
the scattered beam was defined in the scattering
plane by a horizontal slit of aperture 100 lm,
placed in front of the detector. The vertical aperture
of the detector slit was kept wide open to effectively
integrate [3,5] the out-of-plane component of the
scattering.
2
. The LB films
Two LB films of cadmium arachidate (CdA),
each 9 monolayer (ML), were deposited on quartz
substrates from a spread monolayer of arachidic
acid (AA) on an aqueous solution of cadmium
chloride (4;10\ꢁM) at a surface pressure, temper-
ature and pH of 30 mN m\ꢂ (surface tension
4
2 mN m\ꢂ), 10°C and 6.4, respectively. The depos-
ition rate was 10 mm/min. Of these films one was
The specular reflectivity data indicate that the
intensity of the multilayer Bragg peaks in all
the films increase with increasing horizontal de-
tector slit width even beyond the beam divergence
width. Moreover in a log—log plot the scattered
intensity was found to decay linearly, with a slope
(g!1), as a function of the in-plane scattering
vector q , where g depends on q . These results
clearly indicate that we are dealing with scatter-
ing which has a strong diffuse component around
the specular ridge similar to that observed in scat-
tering from liquid surfaces [5,13] with logarithmic
exposed to H S gas for 60 mins. The details of the
ꢀ
experimental arrangement used to prepare these
LB films has been described earlier [11]. In the
UV—VIS absorbance data taken on these films the
absorption edge for the H S exposed film was
ꢀ
found to be at 455$5 nm indicating the formation
of nanocrystalline CdS in this film [10,11].
V
X
Ferric stearate (FeSt) was prepared by adding
a measured amount of ferric sulphate solution to
freshly prepared sodium stearate solution in hot,
distilled water until the residual solution was slight-

M.K. Sanyal et al. / Physica B 248 (1998) 217—222
219
height—height correlation due to the presence of
capillary waves. It is also evident that the longitudi-
nal off-specular scans follow the specular profiles
quite closely. So the multiple interfaces present in
the system are highly conformal [3,4]. It is known
ing technique [18,19] with the effective roughness,
(p ) given by
%
&& G
(
p )ꢀ"pꢀ#ꢂBc !ꢂB ln(2p/i¸).
(3)
%&& G
G
ꢀ
#
ꢀ
Here the actual roughness of an interface, p , is
G
more than the measured effective roughness, (p ) .
%&& G
that beyond a certain critical in-plane length, r ,
#
interfaces in organic multilayer films behave as
a correlated system [14,15], exhibiting a single log-
arithmic height—height correlation for all the
4. Results and discussions
interfaces. The value of r depends on the compress-
#
ibility of the film and decreases with decreasing
The results of the analysis of a CdA film is pre-
sented in Fig. 1. In the lower panel the specular
compressibility [14]. Hence in analogy with these
theoretical formulations and using our earlier ex-
pression [5] for liquid surfaces, we can write down
g(r) for the interfaces present in our LB films as
reflectivity data (q "0) along with the fit (solid
V
line) is shown as (a). It was observed that a simplifi-
ed model (shown in the inset) can mimic the main
features of the reflectivity data. In this simplified
model the number of slices used to represent the
g (r)"pꢀ#pꢀ#Bc #B ln (ir/2),
(1)
GH
G
H
#
where p , p are the rms roughnesses of the ith and
G
H
jth interfaces, B is given by i ¹/pc, and i is the low
wave vector cutoff for capillary waves in LB films
decided by van der Waals interactions and can be
written as [16,17] iꢀ"A/2pcdꢁ, where A is the
effective Hamaker constant, c is the interfacial ten-
sion and d is the total film thickness.
In our experiments, as mentioned earlier, we
have kept the out-of-plane wave vector resolution
(
*q ) coarse so that the q integration in the struc-
W
W
ture factor is effectively performed during data col-
lection [3,5]. Using a gaussian resolution function
for q , obtained from the direct beam profile, we
V
can write the expression [5] for the detected scat-
tered intensity, I(q ,q ), in terms of the Kummer
V X
function and gamma function as
R(q )q
I(q , q )"I
X X
ꢄ2k sin a
ꢄ
V X
1
1!g
1!g 1 !qꢀ¸ꢀ
"
C
F
; ;
V
, (2)
A
2
B
ꢂ ꢂ
A
2
2
4pꢀ
B
(
n
where R(q ) is the specular reflectivity as a function
Fig. 1. Lower panel: (a) specular reflectivity, (b) off-specular
longitudinal and obtained electron density (electrons/ A> ꢃ) pro-
X
of q for a multilayer and the other terms are as
X
explained in Ref. [5].
files (in inset) for the unexposed cadmium arachidate LB film (9
monolayers) are shown. The solid line in (a) and (b) are cal-
culated profiles obtained from the model shown in inset. Upper
panel: transverse diffuse scattering data measured at the first
three Bragg peak positions (1, 2 and 3 in the lower panel) shown
along with the fitted curves (solid lines). Refer to the text for
details.
In the specular condition (q "0), the Kummer
V
function becomes unity and 2k sin a cancels with
ꢄ
q ; Eq. (2) gives us the effective reflectivity as ob-
X
tained earlier [5] for liquid—vapour interfaces. The
reflectivity profile, R(q ) was calculated using a slic-
X

2
20
M.K. Sanyal et al. / Physica B 248 (1998) 217—222
electron density profile (EDP) is minimized. Al-
though one can get a perfect fit to the reflectivity
data by increasing the number of slices to represent
the details of the EDP, the essential features of the
EDP remains unchanged [20]. The obtained EDP
(shown in the inset of Fig. 1) indicates the increase
in the absolute values of electron densities in each
bilayer as a function of depth, possibly resulting
from an increased disorder and formation of
patches due to incomplete film coverage, towards
the top of the film (z"0). This feature is common
to all the three LB films presented.
In the EDP the Cd positions are indicated by the
higher electron density boxes. The total film thick-
ness was found to be 247.5 A
times the obtained bilayer spacing of 55 A
pected for the 9 ML film. The fitted values obtained
for p were 2.1$0.1A for the air-film interface
>
which is equal to 4.5
>
, as ex-
>
%
&&
and 2.0$0.2 A for all the other interfaces.
The upper panel of Fig. 1 shows three sets of
>
transverse scans taken at the 1st, 2nd and 3rd
Fig. 2. Lower panel: (a) specular reflectivity, (b) off-specular
longitudinal and obtained electron density (electrons/Aꢃ) pro-
Bragg peaks, at q values of 0.108, 0.228, 0.34 A
>
\ꢂ,
>
X
respectively. All the curves were fitted using Eq. (2)
solid line). Only B and the additive constant back-
files (in inset) for cadmium archidate LB film (9 monolayers)
exposed to hydrogen sulphide (60 min) are shown. The solid line
in (a) and (b) are calculated profiles obtained from the model
shown in inset. Upper panel: transverse diffuse scattering data
measured at the first three Bragg peak positions (1, 2 and 3 in the
lower panel) shown along with the fitted curves (solid lines).
Refer to the text for details.
(
ground, arising mainly due to dark counts, were
allowed to vary in order to obtain a consistent fit of
all these data. The value of B obtained from this
analysis was 2.1$0.2 A
>
ꢀgiving an effective surface
tension of 62.8 mN m\ꢂ (at 27°C).
Along with the specular data, we have shown the
longitudinal diffuse data taken with a constant
angular offset of 1.0mrad (as (b) in the lower panel).
The parameters obtained from fitting the transverse
diffuse and specular data, namely B, (p ) and
background counts, were used to calculate longitu-
dinal diffuse scattering profiles (solid line) to check
self-consistency of the analysis scheme. The excel-
lent agreement of the calculated profile with the
measured data demonstrates the assumed conform-
ality of the interfaces.
air—film interface and 2.0$0.2A
> for the other
interfaces. We should mention here that calcu-
lations based on a simple model for the EDP with
slices of size 14 A> , which is a quarter of the bilayer
%
&& G
thickness, around the metal sites in the film cor-
rectly brings out the most prominent feature in
the measured reflectivity profile, i.e. vanishing
of the fourth Bragg peak. However, the quality of
the overall fit to the data is significantly improved
by using the detailed model (shown in the inset).
The observed rounding of EDP around metal sites
in the obtained model of the film is mainly due to
variation of size and orientation of CdS nanopar-
ticles (as projected in the z direction) and due to the
presence of capillary waves at the interfaces. These
details will be presented elsewhere [20].
For the exposed CdA film the specular reflectiv-
ity data and fit (solid line) is shown in the lower
panel of Fig. 2 and the obtained EDP is shown in
the inset. We have used 36 slices, each having
a thickness of 7 A
film thickness and the bilayer spacing was found to
be 252 A and is 56 A respectively. The obtained
values for p%&& for this film were 2.1$0.1 A for the
> , to represent the EDP. The total
>
>
The upper panel of Fig. 2 shows three set of
transverse scans measured at the 1st, 2nd and 3rd
>

M.K. Sanyal et al. / Physica B 248 (1998) 217—222
221
Bragg peaks, at q values of 0.107, 0.22, 0.33 A
>
\ꢂ,
to represent this EDP. In the upper panel of this
figure we have shown the asymptotic tails of the
transverse diffuse scattering data taken at the 1st,
X
respectively. The data were fitted (solid lines) in
a manner similar to that for the unexposed film and
the obtained value for B was 2.1$0.2 A
>
ꢀ same as
2nd 3rd Bragg peaks, at q values of 0.118, 0.249,
X
for the former.
0.375A
>
\ꢂ, respectively. Unlike the other two films,
Along with the specular data (a), we have shown
the longitudinal diffuse data (b) taken with a con-
stant angular offset of 0.8 mrad. Self-consistent
analysis to the data was performed as in the earlier
case and the fit is shown (solid line) as (b) in Fig. 2.
The agreement of the calculated profile with the
measured data demonstrates the assumed conform-
ality of the interfaces inspite of the formation of
CdS at the interfaces.
we could not analyse the transverse data upto the
central region due to a split observed in the specu-
lar peak. We feel that this split is due to misoriented
domains in the film. It should be noted here that the
Kummer function of Eq. (2) not only represents the
slope of the asymptotic tail but also the branching
point of this tail from the central Gaussian-like
function. Presence of multiple central peaks pro-
hibit us to detect the proper branching point.
Nevertheless, the proper value of B could be ob-
tained by fitting all asymptotic tails together. The
solid lines indicate the fits and the value of B ob-
The lower panel of Fig. 3 shows the specular
data for the FeSt LB film. The solid line is the
reflectivity profile calculated from the EDP shown
in the inset. We have used 45 slices of 5 A
>
thickness
tained was 3.0$0.4 A>
ꢀ.
It is interesting to note that in case of the FeSt
film the estimated value of the effective interfacial
tension (c"K44 mN m\ꢂ) turns out to be quite
close to the value of the surface tension of the
monolayer from which the film was deposited. This
indicates that no major structural rearrangement
has taken place in this film after deposition, as is
expected for a preformed salt. But for the CdA film
the obtained value of effective interfacial tension
(
cK63 mN m\ꢂ) differs significantly from that of
the acid monolayer used for LB deposition. This
result indicates a possibility of structural rearrange-
ments during deposition, due to the salt formation
and the transfer process. However, we note that the
effective interfacial tension of CdA film remains
unchanged even after the CdS formation. We also
observe that the molecular stacking of the LB film
remains almost unchanged even after the chemical
reaction, confining the formed CdS nanoparticles
in the interfaces.
Acknowledgements
Fig. 3. Lower panel: (a) specular reflectivity and obtained elec-
tron density (electrons/Aꢃ) profile (in inset) for ferric stearate LB
film (9 monolayers) are shown. The solid line is the calculated
profile obtained from the model shown in inset. Upper panel:
asymptotic tails of transverse diffuse scattering data measured at
>
It is our pleasure to thank Prof. S.S. Major, Dr.
A. Dhanabalan and Mr. N. Prashanth Kumar of
the Department of Physics, Indian Institute of
Technology, Mumbai, and Prof. J. Dutta of the
Institute of Wetland Management and Ecological
Design for their support in the sample preparation.
the first three Bragg peak positions (q values indicated) shown
X
along with the fitted curves (solid lines). Refer to the text for
details.

2
22
M.K. Sanyal et al. / Physica B 248 (1998) 217—222
[10] E.S. Smotkin et al., Chem. Phys. Lett. 152 (1988) 265; N.J.
References
Geddes et al., J. Phys. Chem. 97 (1993) 13 767; P. Facci
et al., Proc. Natl. Acad. Sci. USA 93 (1996) 10 556.
[11] A. Dhanabalan et al., Solid State Commun. 99 (1996) 859.
[12] A. Datta, M.K. Sanyal, A. Dhanabalan, S.S. Major,
J. Phys. Chem. B101 (1997) 9280.
[
1] M.K. Sanyal et al., Europhys. Lett. 36 (1996) 265; S. Baner-
jee et al., Phys. Rev. B 54 (1996) 16 377.
2] S.K. Sinha et al., Phys. Rev. B 38 (1988) 2297.
3] M.K. Sanyal et al., Mat. Res. Soc. Symp. Proc. 237 (1992)
[
[
3
93; S.K. Sinha et al., Physica B 198 (1994) 72.
[13] A. Braslau et al., Phys. Rev. A 38 (1988) 2457; D. Schwartz
et al., Phys. Rev. A 41 (1990) 5687; B.M. Ocko et al., Phys.
Rev. Lett. 72 (1994) 242.
[14] J. Daillant, O. Belorgey, J. Chem. Phys. 97 (1992) 5824; 97
(1992) 5837.
[
[
[
4] V. Holy, T. Baumbach, Phys. Rev. B 49 (1994) 10668.
5] M.K. Sanyal et al., Phys. Rev. Lett. 66 (1991) 628.
6] J. Zasadzinski et al., Sci. 263 (1994) 1726; G. Roberts
(Ed.), Langmuir—Blodgett Films, Plenum, New York,
1
990.
[15] R. Holyst, Phys. Rev. A 44 (1991) 3692.
[
[
[
7] I. Langmuir, Trans. Faraday Soc. 15 (1920) 62; K.B.
Blodgett, J. Am. Chem. Soc. 57 (1935) 1007.
8] V. Nitz et al., Phys. Rev. B 54 (1996) 5038; A Gibaud et al.,
Phys. Rev. Lett. 74 (1995) 3205.
[16] I.M. Tidswell et al., Phys. Rev. Lett. 66 (1991) 2108.
[17] M. Sferraza et al., Phys. Rev. Lett. 78 (1997) 3693.
[18] M. Piecuch, L. Nevot, Mat. Sci. Forum 59/60 (1990) 93.
[19] M.K. Sanyal et al., J. Synchrotron Rad. 4 (1997) 185.
[20] J.K. Basu, M.K. Sanyal, Phys. Rev. Lett. 79 (1997) 4617.
9] A. Malik et al., Phys. Rev. B 52 (1995) 11654.