Biophysical Journal Volume 72 June 1997 2651-2659 2651

Order in Phospholipid Langmuir-Blodgett Monolayers Determined by

Total Internal Reflection Fluorescence

Xin Zhai and J. Mieke Kleijn*

Department of Physical and Colloid Chemistry, Wageningen Agricultural University, Dreijenplein 6, 6703 HB Wageningen,
The Netherlands

ABSTRACT Orientational order parameters of two diphenylhexatriene (DPH)-based fluorescent probes, 2-(3-(diphenyl-
hexatrienyl)propanoyl)-1-hexadecanoyl-sn-glycero-3-phosphocholine (DPHpPC) and 1-(4-trimethylammoniumphenyl)-6-
phenyl-1,3,5-hexatriene (TMA-DPH), in dipalmitoylphosphatidylcholine (DPPC) Langmuir-Blodgett monolayers on quartz
have been determined by total internal reflection fluorescence (TIRF). From these order parameters orientation distributions
were reconstructed by the maximum-entropy method. For monolayers transferred from the liquid-condensed phase,
preferential tilt angles with respect to the substrate normal around 140 in the tail region and 50 near the glycerol-acyl chain
linkage were found, as reflected by the DPHpPC and TMA-DPH probes, respectively. The degree of ordering near the
headgroup region seems to be larger than that further away from the surface. A substantial fraction of the TMA-DPH probes
have a flat orientation and are probably located between the phospholipid headgroups and the substrate surface. Monolayers
transferred from the liquid-expanded phase show a more random ordering, and most of the probe molecules (DPHpPC) are
more or less flat on the surface. The results are consistent with earlier atomic force microscopy measurements on identical
monolayers and are in reasonable agreement with previously published data on other organized phospholipid systems.

INTRODUCTION
Biological membranes are dynamic structures in which the solid substrate resembles half the profile of a lamellar
molecular orientational order is generally believed to play bilayer membrane (Leermakers and Nelson, 1990; Has-
an essential role in the maintenance of membrane functions. monay et al., 1994). It has been found experimentally that
Apart from its significance for other membrane properties electron diffraction patterns obtained from supported phos-
such as lipid-bound protein mobility, molecular organiza- pholipid monolayers appear to be the same as those ob-
tion is a crucial factor in membrane permeability. This is tained from free-standing bilayers (Hui et al., 1974).
illustrated, for example, by the finding that the permeability It is our intention to investigate the relationship between
of one-component phospholipid bilayers to small water- the molecular organization and ion permeability in phos-
soluble molecules exhibits a maximum in the coexistence pholipid mono- and bilayers on solid substrates. One of the
region of gel and fluid domains (Clerc and Thompson, experimental techniques involved is total internal reflection
1995). It has been shown that for phospholipid monolayers fluorescence (TIRF), by which it is possible to determine
adsorbed on a mercury electrode surface, the permeability the orientation distribution of fluorophores in an adsorption
for metal ions dramatically depends on the applied electrical layer on optically transparent substrates (Thompson and
potential (Nelson and van Leeuwen, 1989). Using a theo- Burghardt, 1986; Bos and Kleijn, 1995a,b). By using opti-
retical model for the phospholipid monolayer, it was dem- cally transparent conductive films deposited on quartz slides
onstrated that this potential-dependent behavior can be ex- as the substrates, TIRF can be combined with electrochem-
plained in terms of structural variations (Leermakers and ical techniques. This allows simultaneous determination of
Nelson, 1990). the order in phospholipid layers and their permeability for
Phospholipid films on solid supports are attracting in- metal ions and other (electroactive) compounds.
creasing attention as model systems for biological mem- In this paper we present the first results of TIRF orien-
branes, as well as for their application in biosensor devices tation measurements on dipalmitoylphosphatidylcholine
(Sharma and Rogers, 1994; Nikolelis et al., 1995). The (DPPC) monolayers transferred on quartz by the Langmuir-
significance of such model systems is suggested by, e.g., Blodgett (LB) technique, using the fluorescent probes
theoretical models and numerical simulations which show 2-(3-(diphenylhexatrienyl)propanoyl)- 1 -hexadecanoyl-sn-
that the segment density profile for a lipid monolayer on a
glycero-3-phosphocholine (DPHpPC) and 1-(4-trimethyam-
moniumphenyl)-6-phenyl- 1 ,3,5-hexatriene (TMA-DPH).
To our knowledge, this is the first time that the order in
Received for publication 25 October 1996 and in final form 24 February single phospholipid monolayers on solid substrates has been
1997.
Address reprint requests to Dr. J. Mieke Kleijn, Department of Physical
measured. Diphenylhexatriene (DPH) and its derivates are
and Colloid Chemistry, Agricultural University, P.O. Box 8038, 6700 EK commonly used for the determination of orientational order
Wagenigen, The Netherlands. Tel.: 31-317-482145/482178; Fax: 31-317- and dynamics in biomembrane studies. The DPHpPC mol-
483777; E-mail: [email protected]. ecule is a phospholipid with the DPH moiety covalently
© 1997 by the Biophysical Society attached to the sn-2 position in one of the lipid chains (see
0006-3495/97/06/2651/09 $2.00 Fig. 1), and therefore it provides structural information on
2652 Biophysical Journal Volume 72 June 1997

1987; Deinum et al., 1988). Kooyman et al. (1983) and

N- Lentz (1989) have discussed this phenomenon and ex-
plained it as being caused by the coupling of two excited
states.
In TIRF, fluorescent molecules at or near an optically
transparent solid interface are selectively excited by an
evanescent wave, created by the total internal reflection of
a laser beam at this interface. By variation of the polariza-
TMA
N-D tion angle T of the incident light beam, the direction of the
electric field component of the evanescent field is changed.
As a result, the interaction between the transition dipole
moments of the fluorophores and the evanescent field is
modified, which, in turn, gives rise to a change in the
fluorescence signal. It has been shown (Bos and Kleijn,
1995a) that by measuring both the fluorescence intensity
and its polarization as a function of the polarization angle of
the incident light, all of the principal information available
on the orientation distribution of the fluorophores enclosed
in the fluorescence signal (i.e., the order parameters (P2)
and (P4)) can be retrieved.
Assuming that the orientation of the molecules does not
change significantly on the time scale of fluorescence and
that energy transfer between the fluorophores is negligible,
the parallel and perpendicular polarized components of the
detected fluorescence are given by
DPPC DPHpPC TMA-DPH Fl = C((,u *E)2f.(v)) (la)

FIGURE 1 Structure formulas of the fluorescence probes DPHpPC and F1 = Q(IL * E)2j_(P)) (lb)
TMA-DPH and their locations in a DPPC bilayer.
where the vectors IL and v stand for the directions of the
absorption and emission transition dipole moments of the
excited molecule, respectively; E represents the direction of
the lipid chain order. TMA-DPH is expected to be anchored the electric field vector of the evanescent field; andfi(v) and
with its charged group in the headgroup region of phospho- fi(v) are the collection efficiencies for the parallel and
lipid mono- and bilayers, and therefore reflects molecular perpendicular components of the fluorescence, respectively.
order near the glycerol-acyl chain linkage (Parente and The constant C incorporates the magnitude of the absorption
Lentz, 1985; Lentz, 1989). First we derive expressions to and emission dipole moments, the quantum yield, and ex-
obtain the second- and fourth-rank order parameters (P2) perimental parameters such as the surface concentration of
and (P4) of the orientation distributions of the probes from fluorophores, the intensity of the evanescent field, and prop-
TIRF measurements. Orientation distributions are calcu- erties of the detection system. The angle brackets ( ) in Eq.
lated from (P2) and (P4N using the maximum-entropy 1 denote a time and ensemble average over all abundant
method. Besides the orientation of the probes, the angle orientations of the fluorescent groups in the sample.
between their absorption and emission dipole moments is Both assumptions made to arrive at Eq. 1 seem to be
also evaluated. reasonable for the systems studied here. It can be shown that
in the limiting case of large rotational movements with
correlation times much smaller than the fluorescence life-
THEORY time, polarization of the fluorescence is completely lost, and
The theory underlying orientation measurements using only the second-rank order parameter (P2) can be retrieved.
TIRF has been described before in general terms (Bos and This is not the case here. In the fluorescence time region, the
Kleijn, 1995a). Here we treat the case where the fluorophore only mode that gives rise to fluorescence depolarization is
is a cylindrically symmetrical moiety, in which the absorp- rotational diffusion along directions perpendicular to the
tion dipole moment is parallel with the long molecular axis, long symmetry axis of the DPH moiety, usually described
as is the case for DPH-based probes (Kooyman et al., 1983; for lipid bilayers as wobbling motion (Cheng, 1989). Liter-
Lentz, 1989). The direction of the emission dipole moment ature values for the rotation correlation times for DPHpPC
of DPH and its derivates is less well defined and has been and TMA-DPH in gel-phase lipid bilayers, as determined
reported to depend on the molecular environment (Kooy- from time-resolved fluorescence anisotropy measurements,
man et al., 1983; Van Ginkel et al., 1986; Van Langen et al., vary between 1 and 8 ns (Mulders et al., 1986; Cheng, 1989;
Zhai and Kleijn Order in DPPC Monolayers Studied by TIRF 2653

Muller et al., 1994a,b; Bemsdorff et al., 1995), whereas the

fluorescence lifetimes of these probes are in exactly the /
same range (Parente and Lentz, 1985; Van Ginkel et al.,
1986; Cheng, 1989; Lentz, 1989; Bernsdorff et al., 1995). In
the gel phase wobbling motions are fairly restricted, with
maximum deviations of 10-20° from the mean orientation
angle (Parente and Lentz, 1985; Lentz, 1989; Florine-
Casteel, 1990). It is to be expected that for phospholipid x.

monolayers on solid substrates the rotational diffusion is

even more restricted. Because the absorption and emission FIGURE 2 Definition of the orientation angles of the transition dipole
spectra of DPH probes show very little overlap (Shinitzky moments of the DPH group. On the left is the laboratory frame, in which
and Barenholz, 1974; Lentz, 1989), energy transfer between the xy plane corresponds to the substrate surface. The direction of the
absorption dipole moment, ,u, is given by angles 4 and 0. On the right is
the probes, which would cause fluorescence depolarization, the molecular frame, in which the z' axis represents the long molecular axis
is negligible. of DPH (,u is parallel to this axis). In this coordinate system the direction
Defining an orthogonal coordinate system in which the xy of the emission dipole moment, v, is defined by angles a and f3.
plane corresponds to the interface and the xz plane to the
plane of incidence, the direction of E can be given in terms cos 4 cos 0 cos a sin ,B - sin 4) sin a sin ,3\
of the relative magnitudes of its x, y, and z components. + cos 4 sin 0 cos 13
These vary with the polarization of the incident light beam
in the following way (Harrick, 1967): v = sin 4)cos0cosasin13+cos4)sinasin1, (4b)
+ sin 4) sin 0 cos ,B
- sin 0 cos a sin 13 + cos 0 cos (3
EX = Excos P (2a)
By using Eqs. 3 and 4, Eq. 1 can be elaborated into
Ey =cysin T (2b) expressions for Fl, and F, in terms of the orientation angles
of the transition dipole moments, a, 1, 4, and 0. These
Ez= Ecos T (2c) expressions can be simplified by assuming that the interface
is isotropic in the x and y directions, so that all values for 4
The values of Ex, Ey, and E, depend on the refractive indices and a have the same probability. This assumption seems to
of the media at both sides of the interface and on the angle be justified for the substrate used here: there is no prefer-
of incidence, as described by Harrick (1967). ential direction on the quartz surface. Therefore, the angles
For a pulsed excitation of the fluorophores and detection 4 and a can be eliminated from the expressions for Fl and
along the normal of the interface, here defined as the z axis, F1 by integration over a and 4 from 0 to 2-r. Furthermore,
the collection efficiencies of the two polarized components the angle 13 between the absorption and emission dipole
of the fluorescence are given by (Axelrod, 1979) moments is taken as a constant, i.e., in a particular phos-
pholipid monolayer 1 is assumed to be the same for all DPH
fJ(v) = V2 + 1/2(1- y)V2 (3a) probe molecules. Thus the orientation distribution of the
molecules is given by a normalized orientation distribution
f±(v) = 2+ ½(12- y)V2 (3b) function in 0 only, N(0), and the expressions for the fluo-
rescence intensities Fl and F1 become
where -y is the dichroic factor, depending on the aperture
angle of the detection system (Burghardt and Thompson, 1
1984). Fl, = ygC{(3 - 2y)E2 + (5 - 2y)E 2
The direction of the absorption dipole moment ,i of the + (cos20)[(-2 + 4-y)Ex2 + (-6 + 4,y)E 2 + (8 - 4y)Ez2]
fluorescent group in the xyz coordinate system is described + (cos40)[(-1 - 2y)Ex2 + (1 2y)E2 + 4,yE,2]
-

by a polar angle 0 and an azimuthal angle 4, as depicted in + cos213[(3 + 2,y)Ex2 + (-3 + 2,y)E 2]
Fig. 2. To describe the direction of the emission dipole + (cos20)cos213[(-6 - 8y)Ex2 + (6 - 8y)EY + 4 yE2]
moment v, we define a second coordinate system x'y'z' in + (cos40)cos213[(3 + 6'y)Ex2 + (-3 + 6,y)EY2- 12yEz2]}
which the z' axis is parallel to the absorption dipole mo- (5a)
ment. In this coordinate system the direction of v is given by
a polar angle 13 (i.e., the angle between the absorption and 12
F1 = 16C{(5 - 2y)Ex -4 (3 - 2y)EY2
emission dipole moments) and an azimuthal angle a (see
Fig. 2). The expressions for the two vectors ,u and v in the + (cos20)[(-6 4,y)Ex2 (-2 + 4,y)EY + (8 - 4y)Ez2]
+ +
xyz coordinate system are + (cos40)[(1 - 2y)Ex2 + (-1 2y)EY + 4,yEz2]
-

+ cos213[(-3 + 2y)Ex2 + (3 + 2,y)EY2]

cos 4) sin 0 + (cos20)cos2,3[(6 - 8'y)Ex2 + (-6 - 8y)EY + 4yEz2]
sin 4 sin 0 (4a) + (cos40)cos213[(-3 + 6-y)Ex2 + (3 + 6,y)EY2- 12Ez2]}
cos 0 (5b)
2654 Biophysical Journal Volume 72 June 1997

in which (molar ratio 20:1) dissolved in chloroform, or DPPC and TMA-DPH

(molar ratio 10:1) dissolved in chloroform/methanol (volume ratio 1:1)
were spread on a subphase of pure water ("nanopure," resistivity 18.3 Mfl,
(x)-| x N(6)sin 0 dO (6) pH -6). After solvent evaporation the monolayer was compressed at a rate
of 80 mm2/s (- 1.5 X 10-3 nm2/s per molecule) to a prespecified surface
0=0
pressure. Subsequently, the monolayer was allowed to relax for -60 min.
By using Eq. 2 it can be shown that the fluorescence Usually equilibrium was attained (i.e., no detectable change in surface
pressure over a time period of 15 min) between 30 and 60 min after
intensities depend on the polarization angle T of the inci- compression.
dent light as Transfer to a quartz substrate (50 X 20 X 1 mm, Suprasil 2; Heraeus
GmbH, Hanau, Germany) was performed by lifting the substrate vertically
F11(P) = C(All + Bllcos2p) (7a) out of the subphase at a speed of 0.8 mm2/s. Before transfer the quartz
plates were cleaned by overnight immersion in a chromic acid solution,
F_(T) = C(A_ + BLcos2T) (7a) followed by immersion in an alcoholic base solution for a few hours, and
then rinsed with water before use. DPPC/DPHpPC monolayers were trans-
in which All, Bll, Al, and B1 are functions of the orientation ferred at two different surface pressures, 30 and 6.5 mN/m, respectively,
distribution of the fluorophores, the dichroic factor y, and whereas DPPC/TMA-DPH was transferred only at 30 mN/m. All of the
EX9 Ey, and Ez. From Eq. 7 it follows that four measurements film transfers were performed at a temperature of 21 ± 1°C, and in all
suffice to disclose all available information concerning the cases the transfer ratios were larger than 95%. TIRF and fluorescence
orientation distribution as reflected in the fluorescence re- microscopy measurements confirmed incorporation of the probe molecules
in the supported monolayers, although their final concentration is not
sponse of the system. The constant C, two 0-related param- known (TMA-DPH, for example, is slightly soluble in water; Huang and
eters ((cos20) and (cos4 )), and the angle ,3 between absorp- Haugland, 1991). Fluorescence microscopy images gave no indication of
tion and emission moments can be determined by clustering or phase separation of the probes.
experimentally measuring the parallel and perpendicular The molecular structure of DPPC monolayers transferred on quartz at
polarized components of the fluorescence at P = O and various surface pressures has previously been investigated by atomic force
P = 90°. From (cos20) and (cos40) the order parameters microscopy (AFM) (Zhai and Kleijn, 1997). AFM measurements were
carried out using a NanoScope III system (Digital Instruments, Santa
(P2) and (P4) can be calculated. These are defined as Barbara, CA), in the contact mode at room temperature in air. Molecular
scale images were captured at minimal force, i.e., in the attractive part of
(P2) = ½/2(3(cos20)- 1) (8a) the force-distance curve. Therefore, the tip is not pushed into the mono-
layer, but stays on the surface as a result of adhesion forces. To further
(P4) = 1/8(35(Cos40)-30(Cos20)+ 3) (8b) minimize artefacts resulting from scanner drift and sample deformation, the
height profiles captured in the trace and retrace directions were compared,
To obtain an approximation of N(O) we use the maxi- and scanning was repeated in various directions. Typical AFM images
mum-entropy method, as explained in a previous paper (Bos obtained for monolayers transferred at 30 mN/m and 6.5 mN/m are shown
and Kleijn, 1995a). In this approach no a priori assumptions in Fig. 3. It was found that incorporation of 5 mol% DPHpPC in the 30
are made concerning the shape of the distribution function mN/m monolayers does not affect the molecular organization of these
(except that it is continuous). The underlying idea is that the layers.
A detailed description of the TIRF set-up has been given before (Bos
most probable distribution function is the one that can be and Kleijn, 1995b). Here only modifications made for the measurements
realized in the greatest number of distinct ways, subject to described in this paper are mentioned. A sealed head pulsed nitrogen laser
the known constraints (Bevensee, 1983). The result of such (model VSL-337ND; Laser Science, Cambridge, MA) with an emission
an analysis is a Maxwell-Boltzmann distribution with or- wavelength of 337 nm was utilized for excitation. A horizontally polarized
dering energy U(O) (Van Langen et al., 1989): beam was selected using a Glan-Laser polarizing prism (Melles Griot,
Zevenaar, The Netherlands). Subsequently, the polarization angle 'I was
N(O) = exp(-U(O)/kT) adjusted by passage of the beam through a Berek polarization compensator
(model 5540; New Focus, Mountain View, CA). In the TIRF cell (filled
- Noexp{X2P2(cos 0) + X4P4(cos O)} with water) the quartz substrate with phospholipid monolayer was optically
coupled to a quartz prism using immersion liquid. The reflection spot at the
No is the normalization factor. A uniquely defined combi- was approximately 1 mm2. For orientation measure-
solid/liquid interface
nation of A2 and A4 can be obtained by fitting the experi- ments detection of the fluorescence took place at a wavelength of 478 nm.
mentally accessible values (P2) and (P4) using a least- In the wavelength range from 400 to 500 nm the detection efficiency of the
squares method. spectrograph (model 1233; EG&G Princeton Applied Research, Princeton,
NJ) is a factor 1.10 better for vertically polarized light than for horizontally
polarized light. (Horizontal polarization corresponds to the direction par-
allel to the plane of incidence.)
MATERIALS AND METHODS All TIRF measurements were carried out at ambient room temperature.
1,2-Dihexadecanoyl-sn-glycero-3-phosphocholine (DPPC) was purchased Before each series of measurements of F11(0), F11(90o), F1(00), and
from Fluka. 2-(3-(Diphenylhexatrienyl)propanoyl)-1-hexadecanoyl-sn- F1(90') on a particular monolayer, a controlled background experiment
glycero-3-phosphocholine (DPHpPC) and 1-(4-trimethylammoniumphe- was performed to exclude contributions to the fluorescence by parts of the
nyl)-6-phenyl-1,3,5-hexatriene, p-toluenesulfonate (TMA-DPH) were ob- TIRF cell excited by scattered radiation. Using a bare quartz substrate
tained from Molecular Probes (Eugene, OR). All chemicals were used under experimental conditions that were otherwise the same, both the
without further treatment. horizontally and vertically polarized components of the fluorescence were
A home-built Langmuir trough (area 600 cm2), equipped with a mi- determined for T = 00 and T = 900. Before further data analysis these
crobalance for surface pressure measurement by the Wilhelmy-plate background values were subtracted from the corresponding values of
method, was used for monolayer transfer. Mixtures of DPPC and DPHpPC Fl(00), F,1(90'), F1(00), and F1(90').
Zhai and Kleijn Order in DPPC Monolayers Studied by TIRF 2655

mN/m), a number of orientation measurements have been

performed. The results in terms of order parameters are
6.00 0 depicted in Fig. 4. Before discussing these results in detail,
it should be noted that the TIRF measurements show that
the various monolayers stay stably on the quartz substrates
when brought into contact with water: the observed fluo-
4.00O rescence signals are not from the aqueous bulk phase, be-
cause the fluorescent probes are not (DPHpPC) or hardly
(TMA-DPH) water soluble, whereas TMA-DPH is not flu-
orescent in aqueous solution (Lentz, 1989). It was found
that the total fluorescence intensity is practically the same at
every position on the surface.
One might ask how it is possible that a monolayer with
hydrophobic tails facing away from the substrate is stable in
0
1 1 Bringing such a monolayer into contact with water
,water.
A 0°2 00 4 00 6.00 results in unfavorable interactions between the lipid tails
nm and water (for a compact monolayer restricted to interac-
tions between the methyl ends of the tails and water).
However, the substrate surface is hydrophilic, i.e., not par-
ticularly attractive for the hydrophobic tails either. From
AFM measurements (Zhai and Kleijn, 1997) we concluded
-6.00 that there is some water between the headgroups and the
solid support (the surface roughness of the bare quartz is
much higher than that of quartz covered with a compact
_*
_ = t_ _ _phospholipid monolayer). For DPPC LB bilayers on oxi-
-4.00 dized silicon, Tamm and McConnell (1985) have drawn the
same conclusion from lateral diffusion measurements. In
contact with water the quartz surface is negatively charged.
Although DPPC bears no net charge, there is a net electro-

141",~~~~~~~~~~~~

B 0 ~~~2.00 4.00 6.00

30 mN/rn (a) and 6.5 mN/rn (b). The bright spots in image (a) coffespond 0.5
to themethyl ends of the phospholipid chains (Zhai and Kleijn, 1997).
VV

For analysis of the data in terms of order parameters of the DPH probes,
for the components of the evanescent field E0,, EY, and E, values of 0.4787, #
0.9873, and 1.014 were used, respectively. These values were obtained
using the equations given by Harrick (1967), with a refractive index of 0
1.577 for quartz in the UV region, 1.333 for water, and an angle of
incidence of 700. The dichroic factor -y was taken as 0.98, corresponding to
the aperture angle of the detection system (100) (Burghardt and Thompson,
1984). For a variety of values for 13, the constant C and parameters (cos2O)
and (COS4O) were calculated according to Eq. 5 using a least-squares
numerical fit. Subsequently, for f3 the value was taken that yielded realistic
values for (P2) and (P4), i.e., both located within their physical boundaries, -0.5
which follow from their definition given in Eq. 8. -0.5 0 0.5 1
P2>
RESULTS AND DISCUSSION FIGURE 4 Combinations of order parameters (P2) and (P4) obtained for
ech 'ththre monolayers of DPPCIDPHpPC transferred to quartz at 30 mN/in (@),
ForHeaChForornftefthee
o
atye
m
DPHpPCtansferrd at
tyes
30
N/in,
o
of

DPC/DPHpP trans-
onolayers (ie.,
monlayes (i.e., PPC/
DPPC/
DPPC/DPHpPC transferred at 6.5 mN/in (V), and DPPC/TMA-DPH trans-
ferred at 30 mN/in (LI). The solid curves indicate the physical boundaries
ferred at 6.5 mN/m, and DPPC/TMA-DPH transferred at 30 of (P2) and (P4).
2656 Biophysical Journal Volume 72 June 1997

static attraction between the surface and the lipid head- tilamellar dispersions of dioleoylphosphatidylethanolamine
groups (Tamm and McConnel, 1985). (DOPE), Cheng (1989) found a of 290, whereas Muller et
As can be seen from Fig. 4, there is quite a lot of scatter al. (1994a) reported values between 12° and 25° for this
in the obtained order parameters for each type of monolayer, probe in small unilamellar vesicles of various lipids. For
especially in the case of DPHpPC as the probe molecule. TMA-DPH in multibilayer lipid systems, Van Ginkel et al.
This scatter results mainly from two factors. In the first (1986) and Deinum et al. (1988) found values for B in the
place, part of the scatter is due to stochastic errors in the range of 12-20°; for this probe in vesicles of various lipids
orientation measurements. This is illustrated by the fact that Bernsdorff et al. (1995) and Muller et al. (1994b) obtained
generally the spread in (P4) is larger than in (P2). The order values between 00 and 190. For the parent probe DPH, it has
parameters are calculated from the experimentally accessi- been found that (3 increases with bilayer water content (Van
ble parameters (cos20) and (cos40), and from Eq. 8 it is easy Ginkel et al., 1986; Van Langen et al., 1987) and is lowered
to see that any experimental errors in these quantities give by the presence of cholesterol (Kooyman et al., 1983; Van
rise to larger errors in (P4) than in (P2). Secondly, the data Ginkel et al., 1986; Deinum et al., 1988). For DPH the range
points in Fig. 4 per monolayer type are results of measure- of values reported for (3 (0-35°) is somewhat larger than
ments on various freshly prepared LB films and at different those for DPHpPC and TMA-DPH (Kooyman et al., 1983;
spots. It is quite conceivable that the structural organization Mulders et al., 1986; Van Ginkel et al., 1986; Muller et al.,
varies from film to film and from spot to spot. A similar 1996).
observation, albeit on a smaller length scale, has been made For the interpretation of our TIRF orientation measure-
in AFM measurements: although the DPPC monolayers ments in terms of order parameters, ,B was initially consid-
transferred at 30 mN/m have a regular, compact structure ered a free variable and then determined, so that the ob-
(see Fig. 3 a), corresponding to a distorted hexagonal lattice, tained (P2) and (P2) were within their physical meaningful
the exact lattice parameters show local variations. This parameter space. The small spread in the obtained B values
probably results from variations in substrate topology and for each type of monolayer strengthens our confidence in
discontinuities in the transfer process. this analysis. (However, this small spread does not imply
Despite the scatter in the results, there is no overlap in the that the assumption of a constant (B for all of the DPH
regions in which the combinations of (P2) and (P4) are probes in a particular monolayer is correct; we will come
found for the different monolayers. In Table 1 the averages back to this later.) Furthermore, the values for found here
of the obtained values for (P2) and (P4) per monolayer type (0-36°) are in reasonable agreement with previous findings.
are listed, together with the corresponding parameters for It should be noted that no literature data are available for the
the orientation distributions as calculated following the angle between the absorption and emission dipole moments
maximum-entropy method. of DPH probes in phospholipid monolayers on solid
For the DPPC/DPHpPC monolayers transferred at a sur- supports.
face pressure of 30 mN/m, all of the experimental data Fig. 5 a shows the orientation distribution functions N(O)
yielded (P2) and (P4) combinations within their physical obtained by using the maximum-entropy method and the
boundaries, provided that , was taken between 00 and 10. average order parameters listed in Table 1. Fig. 5 b gives the
Apparently, under these conditions the absorption and emis- corresponding number density functions, defined as N(O)
sion dipole moments of the DPH probe are colinear. For the sin 0.
DPPC/DPHpPC monolayers transferred at a lower pressure The number density curve for the DPPC/DPHpPC mono-
of 6.5 mN/m, it was found that physical meaningful com- layers transferred at 30 mN/m, i.e., from the liquid-con-
binations of (P2) and (P4) were obtained for = 36 ± 4°. densed (LC) state at the air/water interface, shows that the
Finally, for the DPPC/TMA-DPH monolayers for ,3, a value orientation of the long molecular axis of the DPH groups
of 32 ± 30 was found. exhibits a broad distribution around 14° with respect to the
In the literature a variety of values have been reported for normal of the substrate surface; most DPH groups have a tilt
the angle between the absorption and emission dipole mo- angle in the range 5-25°. The DPH containing tail of
ments of DPH probes, obtained from fluorescence depolar- DPHpPC is assumed to be aligned with the hydrocarbon
ization experiments. As already mentioned in the Theory tails in the DPPC monolayer. Thus the probe order param-
section, it has been established that (3 is sensitive to the eters would reflect the overall molecular order in the phos-
environment of the probe molecule. For DPHpPC in mul- pholipid tail region. This assumption seems to be reason-

TABLE 1 Average order parameters, angle between absorption and emission dipole moments of the DPH group, and
coefficients of the orientation distribution function as obtained by the maximum-entropy method, for the various monolayers
studied
Monolayer type (P2) (P4) f3 A2 A4 No
DPPC/DPHpPC, 30 mN/m 0.60 0.37 00 2.02 1.42 0.219
DPPC/DPHpPC, 6.5 mN/m 0.09 0.25 360 -1.67 1.78 0.321
DPPC/TMA-DPH, 30 mN/m 0.27 0.63 320 -4.15 14.3 1.25 X 10-3
Zhai and Kleijn Order in DPPC Monolayers Studied by TIRF 2657

using infrared reflection-absorption spectroscopy (IRRAS).

N(O) Hasegawa et al. (1996) determined a tilt angle of 17° in a
10-monolayer DPPC LB film transferred at 40 mN/m, using
UV absorption spectroscopy and chlorprozamine as the UV
probe. Although these tilt angles seem to be in fair agree-
ment with our results, the problem is that they are obtained
from one order parameter only, i.e., from (P2). A given
value for (P2) can correspond to different orientation distri-
butions, and is in fact insufficient to determine a collective
tilt angle (Kooyman et al., 1983; Lafrance et al., 1995).
When a tilt angle is calculated from the average (P2) ob-
tained here (0.60), a value of 310 is obtained; (P2) values for
the above-mentioned systems studied by Okamura et al. and
Hasegawa et al. were found in the range 0.8-0.9. The
N(O)sinO discrepancy between these and our (P2) value might be
explained by the fact that in a single monolayer the influ-
ence of the substrate on the ordering is of course much
greater than in multilayer LB films.
From IR spectroscopy studies it has been concluded that
the conformational and orientational order in the LC phase
of a phospholipid monolayer at the air/water interface can
.'1 be compared with the gel phase of phospholipid bulk sys-
tems (Mitchell and Dluhy, 1988). In both cases the confor-
TTS
.T
II
. mation of the hydrocarbon tails is mostly all-trans. For
15
11
30 45 60 75 90 DPPC in the gel phase, IR attenuated total reflection (IR-
ATR) experiments revealed (P2) values between 0.7 and 0.8
orientation angle 0 (degrees)
(Okamura et al., 1990, 1991).
FIGURE 5 (a) Orientation distributions N(O) as calculated from the Synchrotron x-ray diffraction on DPPC monolayers in the
average order parameters in Table 1 according to the maximum-entropy LC phase at the air-water interface revealed preferential tilt
method for monolayers of DPPC/DPHpPC transferred to quartz at 30 angles of 25-30°, depending on surface pressure (Brezesin-
mN/m ( ), DPPC/DPHpPC transferred at 6.5 mN/m (.), and DPPC/ ski et al., 1995; Mohwald et al., 1995). X-ray diffraction
TMA-DPH transferred at 30 mN/m (--- ). (b) Corresponding number
density functions, N(6) sin 0; the error bars are based on the standard experiments yield directly the orientation without relying on
deviations in the various N(0) sin 0 curves calculated separately from each order parameters. These relatively high tilt angles are con-
of the (P2), (P4) combinations given in Fig. 4. tributed to hydration of the phosphatidylcholine headgroup,
resulting in a bulky head that does not allow for a projected
area on the air/water interface below 0.22-0.23 nm2 per tail.
able, because the presence of DPHpPC does not However, for DPPC monolayers on quartz transferred at 30
significantly affect the molecular organization of the com- mN/m, our AFM measurements revealed an average area of
pact monolayer as observed by AFM (Zhai and Kleijn, 0.20 nm2 per hydrocarbon chain, corroborating the lower
1997). The question of perturbation of phospholipid bilay- preferential tilt angles found here. For oriented stacks of
ers by DPHpPC has been addressed in a review by Lentz bilayers, deposited by pipetting solutions of phospholipids
(1989). Calorimetric measurements on DPPC vesicles have on a substrate, x-ray diffraction revealed orientation angles
shown that the presence of DPHpPC lowers the phase of 20-35°, depending on the degree of hydration of the
transition temperature only slightly, demonstrating that the films (Blaurock and McIntosh, 1986; Katsaras et al., 1992,
probe does not substantially perturb the overall phase struc- 1995; Tristam-Nagle et al., 1993).
ture of the bilayer membrane. On the other hand, results Determination of the orientational order by fluorescence
from differential scanning calorimetry experiments have depolarization experiments using DPH-based probes has
suggested that DPHpPC disrupts bilayer order in its vicinity. been performed only on bulk systems, i.e., multibilayer and
Because we report here for the first time on the order in vesicle systems. For phosphatidylcholines, preferential tilt
single monolayers of phospholipids on solid substrates, angles of 10-30° with respect to the bilayer normal have
comparison with previously published data is not straight- been obtained (Kooyman et al., 1983; Van Langen et al.,
forward. In the literature various DPPC tilt angles for dif- 1989; Florine-Casteel, 1990). In some cases (Mulders et al.,
ferent systems, obtained with different techniques, have 1986; Deinum et al., 1988) for phosphatidylcholine multibi-
been reported. For thin DPPC LB films (nine monolayers) layers in the gel phase practically the same combinations of
transferred at 30 mN/m, Okamura et al. (1991) evaluated an (P2) and (P4) values have been found as obtained here for
orientation angle of 180 with respect to the substrate normal DPPC monolayers transferred from the LC phase.
2658 Biophysical Journal Volume 72 June 1997

At a surface pressure of 6.5 mN/m a DPPC monolayer at much larger than for the DPPC/TMA-DPH monolayers
the air-water interface is in the liquid-expanded (LE) phase. transferred at the same surface pressure. Apparently, varia-
AFM images of such monolayers transferred to quartz show tion in the ordering of the TMA-DPH molecules from film
a loosely packed, irregular structure (Fig. 3 b). The combi- to film and from spot to spot is much less than that of the
nation of order parameters for such layers obtained here is DPHpPC molecules. This is another indication that ordering
for some of the investigated locations on the LB films very near the headgroup region is stronger and better defined
near that for a random distribution ((P2) = (P4) = 0; see than in the outer tail region of the DPPC monolayer.
Fig. 4). From Fig. 5 b it can be seen that-averaged over With respect to the assumption that the angle 3 between
various films and spots-most of the DPH groups are at an absorption and emission dipole moment is constant, we
angle of 700 or more with respect to the normal of the expect it to be justified for the DPHpPC molecules in the
substrate surface (i.e., lying more or less flat on the surface), monolayers transferred at 30 mN/m. These layers have a
whereas only a small part of the DPH-containing phospho- very compact, regular structure, and the DPH groups are all
lipid tails is standing up. IRRAS measurements (Mitchell embedded in the tail region of the monolayer. For the
and Dluhy, 1988) have shown that DPPC monolayers at the monolayers transferred at 6.5 mN/m, the structure as ob-
air/water interface in the LE phase have a highly disordered served by AFM is much more open and almost random.
conformation of the hydrocarbon chains, with a large num- This implies that the environment of the individual DPH
ber of gauche conformers. Therefore, it does not seem probes varies and that there is some distribution in 3.
appropriate to speak in terms of tilt angles of the lipid Furthermore, for the DPPC/TMA-DPH monolayers trans-
chains, and the low values of the order parameters merely ferred at 30 mN/m, the assumption of a constant 13 appears
reflect the disorder in the system. to be less applicable. Because the TMA-DPH molecules can
For TMA-DPH in DPPC monolayers transferred at 30 be divided into two populations (molecules that are lying
mN/m, a bimodal distribution is clearly found. In this case flat, probably between substrate surface and phospholipid
about half of the TMA-DPH molecules have their long headgroups, and molecules that are aligned between the
molecular axes normal to the substrate surface (modal tilt DPPC molecules), there probably is also a bimodal distri-
angle is -5°), and the other half are lying more or less flat bution in P3. Provided that 3 is not directly correlated to the
on the surface (see Fig. 5 b). We suspect that this latter molecular tilt angle, this has no consequences for our anal-
population is localized between the headgroups of the phos- ysis. However, it would be more appropriate to give values
pholipids and the negatively charged substrate surface for (cos2p3) than for P3: from Eq. 5 it follows that this is in
(TMA-DPH itself is positively charged). Bimodal distribu- fact the experimentally obtainable quantity.
tions have also been reported for TMA-DPH in lipid bilayer
systems. In a fluorescence polarization microscopy study on CONCLUSIONS
DPPC vesicles, Florine-Casteel (1990) has found that in the
gel phase about 5% of the TMA-DPH molecules are parallel Using TIRF and two kinds of DPH-based probes, DPHpPC
to the bilayer surface, whereas the rest of the probe mole- and TMA-DPH, we have determined the second- and
cules exhibit an orientation angle around 300 with respect to fourth-rank orientational order parameters of DPPC Lang-
the bilayer normal. Neutron scattering experiments of Pe- muir-Blodgett monolayers on quartz. Furthermore, the an-
bay-Peyroula et al. (1994) revealed that in oriented multi- gle between the absorption and emission dipole moments of
layers of gel-phase DPPC on glass no less than 40% of the the probes has been evaluated. For monolayers transferred
TMA-DPH molecules are located close and parallel to the from the liquid-condensed phase, tilt angles were broadly
bilayer surface, whereas the rest were found to be at angles distributed around 140 in the lipid tail region and more
around 250 relative to the bilayer normal. sharply distributed around 50 near the glycerol-acyl chain
As for the part of the TMA-DPH molecules exhibiting linkage. It was found that a large fraction of the TMA-DPH
orientation angles around 50, these are expected to reflect probe molecules are oriented more or less parallel to the
the order near the glycerol-acyl chain linkage of the phos- monolayer, probably between the phospholipid headgroups
pholipids (Parente and Lentz, 1985; Lentz, 1989). Apart and the substrate surface. Monolayers transferred from the
from having a smaller tilt angle, their orientation distribu- liquid-expanded phase show a more random distribution.
tion is much narrower than that of the DPHpPC probes in All of the results are-as far as comparison can be
similar monolayers (Fig. 5 b). This points to a stronger made-in reasonable agreement with previously published
ordering near the headgroup region than further away from data and show that TIRF combined with maximum-entropy
the substrate surface. IRRAS measurements on DPPC analysis is a good instrument for determining the order and
monolayers at the air/water interface have shown compara- preferential orientation in very thin phospholipid layers
ble results, i.e., the chains exhibit more conformational (monolayers, bilayers, ... ) on solid substrates.
order adjacent to the interface than at their tails (Gericke et
al., 1996) REFERENCES
Comparing the sets of data points for the different mono- Axelrod, D. 1979. Carbocyanine dye orientation in red cell membrane
layers in Fig. 4, it is striking that the scatter in the results for studied by microscopic fluorescence polarisation. Biophys. J. 26:
the DPPC/DPHpPC monolayers transferred at 30 mN/mn is 557-573.
Zhai and Kleijn Order in DPPC Monolayers Studied by TIRF 2659

Bernsdorff, C., R. Winter, T. L. Hazlett, and E. Gratton. 1995. Influence of Lentz, B. R. 1989. Membrane "fluidity" as detected by diphenylhexatriene
cholesterol and ,B-sitosterol on the dynamic behaviour of DPPC as probes. Chem. Phys. Lipids. 50:171-190.
detected by TMA-DPH and PyrPC fluorescence: a fluorescence lifetime Mitchell, M. L., and R. A. Dluhy. 1988. In situ FT-IR investigation of
distribution and time-resolved anisotropy study. Ber. Bunsen-ges. Phys. phospholipid monolayer phase transitions at the air-water interface.
Chem. 99:1479-1488. J. Am. Chem. Soc. 110:712-718.
Bevensee, R. M. 1983. Maximum Entropy Solutions to Scientific Prob- Mohwald, H., H. Baltes, M. Schwendler, C. A. Helm, G. Brezesinski, and
lems. Prentice Hall, Englewood Cliffs, NJ. H. Haas. Phospholipid and protein monolayers. 1995. Jpn. J. Appl. Phys.
Blaurock, A. E., and T. J. McIntosh. 1986. Structure of the crystalline 34:3906-3913.
bilayer in the subgel phase of dipalmitoylphosphatidylglycerol. Bio- Mulders, F., H. van Langen, G. van Ginkel, and Y. K. Levine. 1986. The
chemistry. 25:299-305. static and dynamic behaviour of fluorescent probe molecules in lipid
Bos, M. A., and J. M. Kleijn. 1995a. Determination of the orientation bilayers. Biochim. Biophys. Acta. 859:209-218.
distribution of adsorbed fluorophores using TIRF. I. Theory. Biophys. J. Muller, J. M., E. E. van Faassen, and G. van Ginkel. 1994a. The interpre-
68:2566-2572. tation of the time-resolved fluorescence anisotropy of diphenyl-
Bos, M. A., and J. M. Kleijn. 1995b. Determination of the orientation hexatriene-phosphatidylcholine using the compound motion model. Bio-
distribution of adsorbed fluorophores using TIRF. II. Measurements on chem. Biophys. Res. Commun. 201:709-715.
porphyrin and cytochrome c. Biophys. J. 68:2573-2579. Muller, J. M., E. E. van Faassen, and G. van Ginkel. 1994b. Experimental
Brezesinski, G., A. Dietrich, B. Struth, C. B6hm, W. G. Bouwman, K. support for a novel compound motion model for the time-resolved
Kjaer, and H. Mohwald. 1995. Influence of ether linkages on the fluorescence anisotropy decay of TMA-DPH in lipid vesicle bilayers.
structure of double-chain phospholipid monolayers. Chem. Phys. Lipids. Chem. Phys. 185:393-404.
76:145-157. Muller, J. M., G. van Ginkel, and E. E. van Faassen. 1996. Effect of lipid
Burghardt, T. P., and N. L. Thompson. 1984. Effect of planar dielectric molecular structure and gramicidin A on the core of lipid vesicle
interfaces on fluorescence emission and detection. Evanescent excitation bilayers. A time-resolved fluorescence depolarization study. Biochem-
with high-aperture collection. Biophys. J. 46:729-737. istry. 35:488-497.
Cheng, K. H. 1989. Fluorescence depolarization study of lamellar liquid Nelson, A., and H. P. van Leeuwen. 1989. Metal voltammetry at dioleoyl
crystalline to inverted cylindrical micellar phase transition of phosphati- lecithin (di-O-PC) coated mercury electrodes: Cd, Cu, Eu, Pb, V and Zn
dylethanolamine. Biophys. J. 55:1025-1031. and the effect of electrolyte composition. J. Electroanal. Chem. 273:
183-199.
Clerc, S. G., and T. E. Thompson. 1995. Permeability of dimyristoyl
phosphatidylcholine/dipalmitoyl phosphatidylcholine bilayer mem- Nikolelis, D. P., C. G. Siontorou, V. G. Andreou, K. G. Viras, and U. J.
branes with coexisting gel and liquid-crystalline phases. Biophys. J. Krull. 1995. Bilayer lipid membranes as electrochemical detectors for
68:2333-2341. flow injection immunoanalysis. Electroanalysis. 7:1082-1089.
Deinum, G., H. van Langen, G. van Ginkel, and Y. K. Levine. 1988. Okamura, E., J. Umemura, and T. Takenaka. 1990. Orientation studies of
Molecular order and dynamics in planar lipid bilayers: effects of unsat- hydrated dipalmitoylphosphatidylcholine multibilayers by polarized
uration and sterols. Biochemistry. 27:852-860. FTIR-ATR spectroscopy. Biochim. Biophys. Acta. 1025:94-98.
Florine-Casteel, K. 1990. Phospholipid order in gel- and fluid-phase cell- Okamura, E., J. Umemura, and T. Takenaka. 1991. Fourier transform
size liposomes measured by digitized video fluorescence polarization infrared/attenuated total reflection study on subtransition of hydrated
microscopy. Biophys. J. 57:1199-1215. dipalmitoylphosphatidylcholine multibilayers. Vib. Spectrosc. 2:95-100.
Gericke, A., D. J. Moore, R. K. Erukulla, R. Bittman, and R. Mendelsohn. Parente, R. A., and B. R. Lentz. 1985. Advantages and limitations of
1996. Partially deuterated phospholipids as IR structure probes of con- 1-palmitoyl-2-[[2-[4-(6-phenyl-trans-1, 3, 5-hexatrieny)phenyllethyl]
formational order in bulk and monolayer phase. J. Mol. Struct. 379: carbonyl]-3-sn-phosphatidylcholine as a fluorescent membrane probe.
227-239. Biochemistry. 24:6178-6185.
Pebay-Peyroula, E., E. J. Dufourc, and A. G. Szabo. 1994. Location of
Harrick, N. J. 1967. Internal Reflection Spectroscopy. John Wiley/ diphenyl-hexatriene and trimethylammonium-diphenyl-hexatriene in di-
Interscience, New York. palmitoylphosphatidylcholine bilayers by neutron diffraction. Biophys.
Hasegawa, T., Y. Ushiroda, M. Kawaguchi, Y. Kitazawa, M. Nishiyama, Chem. 53:45-56.
A. Hiraoka, and J. Nishijo. 1996. UV absorption spectroscopic analysis Sharma, A., and K. R. Rogers. 1994. Biosensors. Meas. Sci. .Technol.
of the molecular orientation of a drug penetrated into a DPPC mem- 5:461-472.
brane. Langmuir. 12:1566-1571.
Shinitzky, M., and Y. Barenholz. 1974. Dynamics of the hydrocarbon layer
Hasmonay, H., A. Hochapfel, C. Betrencourt, A. Tahir, and P. Peretti. in liposomes of lecithin and sphingomyelin containing dicetylphosphate.
1994. Lasalocid and biomimetic membranes: insertion in Langmuir J. Biol. Chem. 249:2652-2657.
films of lipids. Biochim. Biophys. Acta. 1193:287-292.
Tamm, L. K., and H. M. McConnell. 1985. Supported phospholipid bilay-
Huang, Z., and R. P. Haugland. 1991. Partition coefficients of fluorescent ers. Biophys. J. 47:105-113.
probes with phospholipid membranes. Biochem. Biophys. Res. Commun.
181:166-171. Thompson, N. L., and T. P. Burghardt. 1986. Total internal reflection
fluorescence measurement of spatial and orientational distributions of
Hui, S. W., D. F. Parsons, and M. Cowden. 1974. Electron diffraction of fluorophores near planar dielectric interfaces. Biophys. Chem. 25:91-97.
wet phospholipid bilayers. Proc. Natl. Acad. Sci. USA. 71:5068-5072.
Tristam-Nagle, S., R. Zhang, R. M. Suter, C. R. Worthington, W. J. Sun,
Katsaras, J., V. A. Raghunathan, E. J. Dufourc, and J. Dufourcq. 1995. and J. F. Nagle. 1993. Measurement of chain tilt angle in fully hydrated
Evidence for a two-dimensional molecular lattice in subgel phase DPPC bilayers of gel phase lecithins. Biophys. J. 64:1097-1109.
bilayers. Biochemistry. 34:4684-4688. Van Ginkel, G., L. J. Korstanje, H. van Langen, and Y. K. Levine. 1986.
Katsaras, J., D. S. C. Yang, and R. M. Epand. 1992. Fatty-acid chain tilt The correlation between molecular orientational order and reorienta-
angles and directions in dipalmitoyl phosphatidylcholine bilayers. Bio- tional dynamics of probe molecules in lipid multibilayers. Faraday
phys. J. 63:1170-1175. Discuss. Chem. Soc. 81:49-61.
Kooyman, R. P. H., M. H. Vos, and Y. K. Levine. 1983. Determination of Van Langen, H., D. Engelen, G. van Ginkel, and Y. K. Levine. 1987.
orientational order parameters in oriented lipid membrane systems by Headgroup hydration in egg-lecithin multibilayers affects the behaviour
angle-resolved fluorescence depolarization experiments. Chem. Phys. of DPH probes. Chem. Phys. Lett. 138:99-104.
81:461-472. Van Langen, H., C. A. Schrama, G. van Ginkel, G. Ranke, and Y. K.
Lafrance, C. P., A. Nabet, R. E. Prud'homme, and M. Pezolet. 1995. On the Levine. 1989. Order and dynamics in the lamellar L,,, and in the hexag-
relationship between the order parameter (P2(cos 0)) and the shape of onal H,, phase: dioleoylphosphatidyl-ethanolamine studied with angle-
orientation distribution. Can. J. Chem. 73:1497-1505. resolved fluorescence depolarization. Biophys. J. 55:937-947.
Leermakers, F. A. M., and A. Nelson. 1990. Substrate-induced structural Zhai, X., and J. M. Kleijn. 1997. Molecular structure of dipalmitoylphos-
changes in electrode-adsorbed lipid layers: a self-consistent field theory. phatidylcholine Langmuir-Blodgett monolayers studied by atomic force
J. Electroanal. Chem. 278:53-72. microscopy. Thin Solid Films. In press.