Critical temperature and le/g phase transitions in monolayer films of the amphiphilic tempo derivatives at the air/water interface

Article abstract of DOI:10.1021/jp0122878

Phase behavior of five members of a homologous series of 4-alkaneamido-2,2,6,6-tetramethyl-1-piperidinyloxy (C_nTEMPO with n = 14, 16, 18, 20, 22) was investigated with classical Langmuir monolayer techniques and with Brewster angle microscopy and 2D voltammetry. Surface potential data, BAM images, and 2D voltammetry showed that only C₂₀- and C₂₂TEMPO monolayers have a 2D LE/G critical point above room temperature and that the remaining three compounds with shorter alkane chains form supercritical monolayers. The unusually low values of the critical temperature for these amphiphiles was explained in terms of the molecular structure of their headgroup featuring two polar groups (the nitroxy and the amide moieties) located in the opposite positions of a bulky piperidine ring. The presence of the two polar groups causes a change of orientation of the headgroup during monolayer expansion and results in a substantial increase (ca. 50-100 Aì?²/molecule) of the cross sectional area of these amphiphiles. As a result, in the expanded monolayers, the van der Waals attraction between the alkane chains is weakened as they can only partially align. The decreased cohesion within the monolayers leads to the observed decrease of their 2D critical temperature. Analysis of the plots of BAM reflected light intensity as a function of amphiphile's surface concentration obtained at different temperatures yielded the C₂₀TEMPO critical temperature of 28 ?°C. BAM and 2D voltammetry can both be used to precisely determine the position of the C₂₂TEMPO LE/G phase transition (98.5 Aì?²/molecule at 23.5 ?°C). The proximity of the C₂₀TEMPO critical point to the room-temperature results in a less precise value of this phase transition (127-130 Aì?²/molecule at 23.5 ?°C). A remarkably consistent ability of fresh line microelectrodes used in 2D voltammetry to nucleate the 2D gas phase makes this method less ambiguous and thus superior to BAM in determining the position of the LE/G phase transitions.

Full text of DOI:10.1021/jp0122878

9002
J. Phys. Chem. B 2001, 105, 9002-9010
Critical Temperature and LE/G Phase Transitions in Monolayer Films of the Amphiphilic
TEMPO Derivatives at the Air/Water Interface†
Michael J. Johnson,‡ Chinmay Majmudar,‡ Janusz J. Skolimowski,§ and Marcin Majda*,‡
Department of Chemistry, UniVersity of California at Berkeley, Berkeley, California 94720-1460, and
Department of Organic Chemistry, UniVersity of Lo´dz, Narutowicza 68 90-136 Lo´dz, Poland
ReceiVed: June 18, 2001
Phase behavior of five members of a homologous series of 4-alkaneamido-2,2,6,6-tetramethyl-1-piperidinyloxy
(CnTEMPO with n ) 14, 16, 18, 20, 22) was investigated with classical Langmuir monolayer techniques and
with Brewster angle microscopy and 2D voltammetry. Surface potential data, BAM images, and 2D
voltammetry showed that only C20- and C22TEMPO monolayers have a 2D LE/G critical point above room
temperature and that the remaining three compounds with shorter alkane chains form supercritical monolayers.
The unusually low values of the critical temperature for these amphiphiles was explained in terms of the
molecular structure of their headgroup featuring two polar groups (the nitroxy and the amide moieties) located
in the opposite positions of a bulky piperidine ring. The presence of the two polar groups causes a change of
orientation of the headgroup during monolayer expansion and results in a substantial increase (ca. 50-100
Å2/molecule) of the cross sectional area of these amphiphiles. As a result, in the expanded monolayers, the
van der Waals attraction between the alkane chains is weakened as they can only partially align. The decreased
cohesion within the monolayers leads to the observed decrease of their 2D critical temperature. Analysis of
the plots of BAM reflected light intensity as a function of amphiphile’s surface concentration obtained at
different temperatures yielded the C20TEMPO critical temperature of 28 °C. BAM and 2D voltammetry can
both be used to precisely determine the position of the C22TEMPO LE/G phase transition (98.5 Å2/molecule
at 23.5 °C). The proximity of the C20TEMPO critical point to the room-temperature results in a less precise
value of this phase transition (127-130 Å2/molecule at 23.5 °C). A remarkably consistent ability of fresh
line microelectrodes used in 2D voltammetry to nucleate the 2D gas phase makes this method less ambiguous
and thus superior to BAM in determining the position of the LE/G phase transitions.
Introduction
While several in depth investigations of this amphiphile exist
in the literature, the reported values of Tc range from 26 to
(I)
The relationship between structure and properties of Langmuir
monolayer films has been a subject of paramount importance
in essentially all investigations of these two-dimensional mo-
lecular systems. Recent rapid progress1 in the characterization
of the condensed phase structures and phase transitions revealing
rich polymorphism of the ordered phases of systems such as
fatty acids can be related to new experimental techniques such
as epifluorescence microscopy,2-4 Brewster angle microscopy
(BAM),5,6 and grazing incidence X-ray diffraction.7 While
structural issues naturally dominate research concerned with
ordered condensed phases, the questions addressing liquid and
gaseous phases (LE, also referred to as L1, and G) of Langmuir
and the related Gibbs monolayers concern existence of the LE/G
phase transition and the values of critical temperature. The
relationship of the LE/G critical temperature (Tc(I)) to the
molecular structure of amphiphiles has not been systematically
addressed. Precise determination of Tc(I) has been difficult and
is known only for a few aliphatic acids and their derivatives. In
the early literature, Adam and Jessop determined that at 16 °C,
lauric acid, ethyl laurate, dodecyl alcohol, and myristic nitrile
were above their critical point.8,9 To date, there is no agreement
70 °C.2,10,11 Recent ellipsometric experiments allowed Casson
and Bain to determine a LE/G phase transition in Gibbs
monolayers of decanol.12 These measurements and the related
results of Riegler and co-workers13 concerning hexane and
heptane monolayers on the water surface clearly indicate that
the relationship of the most rudimentary elements of an
amphiphile’s structure and its 2D liquid-gas critical temperature
should be systematically investigated as it is not well understood.
We have recently reported measurements suggesting an
unusually low LE/G critical temperature of Langmuir mono-
layers of 4-tetradecaneamido-2,2,6,6-tetramethyl-1-piperidin-
yloxy (C14TEMPO).14 In this report, we address the molecular
basis of this finding and present results obtained with four
additional members of this homologous series with longer alkyl
chains (C16, C18, C20, and C22). Our measurements involving
surface pressure and surface potential measurements as well as
Brewster angle microscopy (BAM) and 2D electrochemistry15
allowed us to determine that at room temperature only C20-
TEMPO and C22TEMPO monolayers exhibit LE/G phase
transitions and that the derivatives with shorter alkyl chains are
supercritical. We also expand on our earlier observation that
2D electrochemical techniques can serve to determine the
position of the LE/G phase transition. To this end, we compare
the capabilities of the surface pressure and surface potential
measurements, Brewster angle microscopy, and 2D electro-
(I)
in the measurements of the Tc value for pentadecanoic acid.
† Part of the special issue “Royce W. Murray Festschrift”.
* Corresponding author. E-mail: majda@socrates.berkeley.edu.
‡ University of California at Berkeley.
§ University of Lo´dz.
10.1021/jp0122878 CCC: $20.00 © 2001 American Chemical Society
Published on Web 09/13/2001

Amphiphilic TEMPO Derivatives
J. Phys. Chem. B, Vol. 105, No. 37, 2001 9003
Figure 1. Schematic diagram of the experimental setup used in the characterization of the TEMPO monolayers. C, counter electrode; R, reference
electrode; W, working line microelectrode. The inset shows a line microelectrode as it touches the water surface.
TABLE 1: Results of CHN Analysis of the TEMPO
chemistry to determine the position of this phase transition. We
find that both BAM and 2D voltammetry are quite precise.
However, we show that, at least in the case of the C22TEMPO
monolayers, the latter technique offers a distinct advantage.
Amphiphiles
%C
%H
%N
compound mp (°C) calcd found calcd found calcd found
C14TEMPO
C16TEMPO
C18TEMPO
C20TEMPO
C22TEMPO
49.5
57.5
61.0
69.0
74.5
72.4 72.5 11.9 11.9
73.3 73.2 12.1 12.2
74.1 74.1 12.2 12.3
74.8 74.5 12.4 12.6
75.4 75.4 12.5 12.7
7.3
6.8
6.4
6.0
5.7
7.3
6.8
6.4
6.0
5.8
Experimental Section
Reagents. House-distilled H2O was passed through a four-
cartridge Millipore purification train and a 0.2 µm hollow-fiber
final filter. The resistivity of the resulting water (DI water) was
>18.3 MΩ cm. Octadecyltrichlorosilane (OTS) and 3-mercap-
topropyltrimethoxysilane (MPS) were from Aldrich. OTS was
vacuum-distilled into sealed glass ampules, which were opened
as needed immediately prior to the individual experiments.
Octadecylmercaptan (OM) (Tokyo Kasei, Tokyo, Japan) was
used without further purification. Concentrated HNO3 (Fisher,
ACS Certified Plus), chloroform (Fisher, ACS certified spec-
tranalyzed), methanol (Fisher, spectroscopic grade), and all the
other reagents were used as received.
Synthesis of CnTEMPO Amphiphiles. The synthesis of the
C14-C22TEMPO amphiphiles involved a carbodiimide assisted
coupling of 4-amino-2,2,6,6-tetramethyl-1-piperidinyloxy radical
(4-aminoTEMPO) with a desired aliphatic acid. 0.011 mol of
4-aminoTEMPO and 0.01 mol of a long chain carboxylic acid
were dissolved in 50 mL of methylene chloride. The solution
was purged with nitrogen and stirred. After 5 min, 0.01 mol of
1-[3-(Dimethylamino)propyl]-3- ethylcarbodiimide hydrochlo-
ride coupling agent was added. The cloudy orange mixture,
while stirred under nitrogen overnight, slowly converted to a
clear orange solution. The resulting solution was washed twice
with 50 mL saturated NaHCO3 and twice with 50 mL of brine
and then dried with MgSO4. The organic layer was collected
and yielded orange crystals upon removal of the solvent. As
expected, the color of the crystals becomes less intense as the
length of the alkane chain in this homologous series of
compounds increases, reflecting a decrease of the weight fraction
of the colored nitroxide moiety in the crystals. TLC tests
indicated no other major products in these samples. Column
chromatography was used for further purification. The crystals
were dissolved in methylene chloride, loaded onto a neutral
alumina column, and eluted with methylene chloride. A single
orange band moved through the column and the main part of
the band was collected. Only the leading and following edge
of the product band were discarded, resulting in a high yield
(∼70%) of a very pure product. A UV-vis spectrum of a
chloroform solution of these products features are broad band
with the maximum absorbance at 470 nm. The purity, a critical
component in the subsequent studies of monolayer films of these
compounds at the air/water interface, was ascertained by melting
point measurements, CHN analysis, and pressure-area iso-
therms. A plot of the melting points versus alkane chain length
was linear with r2 ) 0.991. The individual values listed in Table
1 agreed well with the available literature. Table 1 also lists
the results of CHN elemental analysis. The pressure-area (π-
MMA) isotherms (discussed in more details in the next section)
may be used as a sensitive, qualitative test of sample’s purity.
Even small amounts of impurities cause irregularities such as
kinks along a sharply rising section of π-MMA curves.
Following their purification, our samples did not exhibit any
such irregularities.
Experiments on the Water Surface and Monolayer Tech-
niques. The experimental setup combining several measurement
techniques addressing a monolayer film at the air/water interface
in a Langmuir trough is shown schematically in Figure 1.
C14-22TEMPO monolayers were spread from 1.0 mM chloro-
form solutions at 150 Å2/molecule and compressed at 11 Å2/
molecule per min. To minimize mechanical vibration of the
water surface, Brewster angle microscopy (BAM) was per-
formed on a small active vibrational table (MOD-2, Halcyonics
GmbH, Goettingen, Germany). These instruments were enclosed
in a Plexiglas box on a large table within a thermostated, laminar
flow enclosure operated under positive pressure with filtered
air. While, persistent building vibrations beyond our control
resulted in some BAM reflected light fluctuations shown in our

9004 J. Phys. Chem. B, Vol. 105, No. 37, 2001
Johnson et al.
data, we believe that these vibrations did not affect the mean
behavior of the monolayers presented in this report.
The Brewster angle microscope (BAM2plus, Nanofilm Tech-
nologie GmbH, Goettingen, Germany) is equipped with a 50
mW Nd:YAG laser. It was typically operated at 40% power.
BAM observations involving monolayer surface concentrations
near phase transitions were done in short, ca. 1-3 s exposures
interspersed by ca. 15 s combining to a total observation time
of ca. 3 min at each surface concentration of an amphiphile.
More frequent or longer exposure times were avoided under
these conditions to prevent local laser beam induced tempera-
ture increases which could offset the position of phase transi-
tions. In other experiments, the exposure time was typically 30-
90 s.
Surface potential versus mean molecular area (MMA) mea-
surements were done with a Kelvin probe (vibrating plate
method, Spot1 by KSV Instruments, Ltd., Helsinki, Finland) in
a large Langmuir trough equipped with two symmetrically
positioned barriers (KSV 2000). Cleanliness of the air/water
interface was checked before each series of measurements and
was judged to be satisfactory if the change of the surface
potential was less than (5 mV after changing the position of
the barriers to sweep the entire trough area and thus to compress
surface impurities into a small area occupied by the Kelvin
probe. Subsequently, the barriers were returned to the initial
position, the surface potential reading was zeroed again, and a
monolayer to be investigated was spread on the water surface.
Figure 2. Set of surface pressure-MMA isotherms recorded for Cn-
TEMPO amphiphiles on 50 mM HNO3 at 23.5 °C. Curves 1-5
correspond to n ) 14-22, respectively.
CHART 1: Chemical Structure of C14TEMPO
2D Electrochemical Measurements. Electrochemical mea-
surements at the air/water interface required specially designed
“line” microelectrodes that can be positioned exactly in the plane
of the air/water interface.16 They were produced by creating a
sharp gradient of wettability along a fracture line of ∼800 Å
thick gold films, vapor-deposited on microscope glass slides
(ca. 8 × 20 mm2). The pattern of the deposited gold film
includes two circular areas, later used as electrical contact pads,
and a strip of gold (0.50 mm in width) running between them.
Pretreatment of the glass substrates (Corning 2947) and their
subsequent treatment with 3-mercaptopropyltrimethoxysilane
(MPS) to induce adhesion of the vapor-deposited gold film were
described recently.17 Following gold vapor deposition, mono-
layers of octadecane mercaptan (OM) and octadecyltrichloro-
silane (OTS) were formed on gold and on glass surfaces,
respectively, by self-assembly to render all the surfaces of the
substrate hydrophobic. By breaking such an electrode substrate
in half (along a line drawn with a diamond pencil on the reverse
side of the glass substrate perpendicular to the gold strip), one
exposes a clean, and thus hydrophilic, edge surface of glass
and gold and create two identical microelectrodes. These
electrodes can then be positioned at the air/water interface by
touching the water surface with the newly created edge surface,
as shown in Figure 1. Thus, a line of wettability is formed along
the edge of the gold microband between the hydrophilic gold
cross sectional area and the hydrophobic (OM-coated) front face
of the gold strip. The electrodes were always positioned at the
air/water interface following monolayer spreading and solvent
evaporation. Their performance at the air/water interface is
typically reproducible for ∼30 min. Later, contamination of the
microband deteriorates the electrochemical signal and the
electrode must be discarded. Cyclic voltammetry was done with
an Ensman Model 852 bipotentiostat (Bloomington, IN) in a
three-electrode configuration under computer control. The
reference electrode (SCE) and a Pt counter electrode were
immersed in the subphase in a Langmuir trough behind the
barrier, where their presence did not interfere with a monolayer
compression.
Results and Discussion
Surface Pressure and Surface Potential Isotherms. The
chemical structure of C14TEMPO, the shortest chain homo-
logue of the series, is shown in Chart 1. These amphiphiles
feature two polar groups in the headgroup region, the nitroxy
(>N•-O) and amide groups. These are separated by a bulky,
lipophilic piperidine ring with four side methyl groups. These
structural features are critical to the observed phase behavior
reported recently and discussed in this report.14 Specifically,
we postulate that in the high-pressure region, these compounds
assume a vertical orientation of the TEMPO headgroup,
minimizing their cross sectional area (see Figure 2). However,
as their monolayers are expanded, the tetramethyl-1-piperi-
dinyloxy ring tilts its N-C4 axis and ultimately assumes a
roughly horizontal orientation which maximizes hydration of
both of its polar fragments. In this orientation, its mean
molecular area is substantially larger, 90-100 Å2/molecule. As
shown below, this postulate is consistent with the surface
pressure observed in this region.
A set of π-MMA isotherms for all five-members of this
series is shown in Figure 2. A very gradual approach of the
surface pressure to π ) 0 in the expansion experiments is
characteristic for all these compounds. Two additional features
are noteworthy. The isotherm for C14TEMPO is shifted to lower
mean molecular areas at higher pressures, suggesting partial
solubility of this shortest chain derivative. Indeed, an increas-
ingly larger shift was observed as a function of the decreasing
compressing rate supporting this conclusion.18 A similar shift

Amphiphilic TEMPO Derivatives
J. Phys. Chem. B, Vol. 105, No. 37, 2001 9005
Only in the case of the C20- and C22TEMPO monolayers does
µZ not appreciably change when these monolayers are spread
on the water surface. In other words, the surface potential
continues to be dominated by the background value of the clean
water surface (defined as 0 D) and indicates the LE-G
coexistence of these TEMPO monolayers. The sharp increase
of µZ during monolayer compression clearly indicates comple-
tion of the LE/G phase transition. The smaller value of the MMA
of this transition, which is reproducible to (10 Å2/molecule,
recorded for C20TEMPO is not clear to us. The supercritical
state of the three shorter chain TEMPO monolayers is clearly
evidenced by the high values of µZ at all MMA’s. While surface
potential measurements are notoriously difficult to interpret
quantitatively,9,20 the conspicuous, nonlinear decrease of the
effective molecular dipole moment observed for all five mono-
layers at MMA < 100 Å2/molecule is consistent with the
hypothesis of the gradual rotation of the piperidine headgroup
upon compression from approximately horizontal to vertical.21
The supercritical nature of the C14-C18TEMPO monolayers
may seem surprising when viewed in comparison with aliphatic
alcohols, acids, esters, and nitriles mentioned in the Introduction.
In those cases, only the amphiphiles with the chain length not
exceeding C12 were shown to exhibit Tc(I) below room temper-
ature. We note that the recent results of Casson and Bain actually
Figure 3. Set of µZ vs MMA data obtained from the surface potential
measurements according to eq 1 on the 50 mM HNO3 subphase at 23.5
°C. Curves 1-5 correspond to CnTEMPO amphiphiles with n ) 14-
22, respectively.
(I)
place Tc of dodecanol above room temperature.12 While the
van der Waals chain-chain interactions constitute the major
component of monolayer cohesiveness, a uniform correlation
of LE/G critical temperature with the chain length can be
expected only within a group of structurally similar amphi-
philes.9 The conspicuously large cross sectional area of the
headgroup (particularly in a region of low surface pressures)
of the TEMPO group relative to that of the alkane chain results
in a decrease in the van der Waals interactions between the
alkane chains due to their stearic inability to align over their
full length. In effect, the large ratio of the headgroup to chain
cross sectional areas renders the effective alkane chain length
of the TEMPO amphiphiles to be significantly shorter than their
physical length.
was also observed when an increasing time was allowed to
incubate the C14TEMPO monolayers at a large MMA before
their constant rate compression was initiated. The other observa-
tion is the inverse dependence of the collapse pressure on the
chain length, which reflects the increasing rigidity of the liquid
monolayers with the increasing chain length.
During monolayer expansion, the mean molecular area at
which pressure assumes a value independent of further expan-
sion (approximately π ) 0) is usually associated with the onset
of the liquid/gas (LE/G) phase transition.9 Alternatively, the so-
called pressure takeoff point recorded during monolayer com-
pression is interpreted as the end of the LE-G coexistence region.
Clearly, in the case of the CnTEMPO isotherms, this method
of determining the existence and the position of the LE/G phase
transition is neither precise nor accurate. Specifically, the C20
and C22 derivatives appear to exhibit the same position of this
phase transition, while the gradual approach of the isotherms
of the other three derivatives to π ) 0 makes this determination
very ambiguous, if not impossible.
Brewster Angle Microscopy. The capability of Brewster
angle microscopy to image the morphology of Langmuir
monolayers and to determine phase transitions is well estab-
lished.22-24 Naturally, observation of a texture featuring coexist-
ence of the liquid and gaseous phases at a particular MMA
constitutes a proof that a LE/G phase transition did occur. Three
examples of such images obtained for C22TEMPO are shown
in Figure 4. However, in cases when BAM images of homo-
geneous texture are obtained, even in a range of zero surface
pressures, the opposite conclusion cannot be made with
certainty. In an attempt to remove this ambiguity, in the
experiments reported below, we measured the reflected light
intensity of the BAM laser beam integrated over the entire
illuminated area (ca. 320 × 430 µm2) of the water surface. These
data are plotted in Figures 5, 6, and 7 as a function of the Cn-
TEMPO surface concentrations. The data representing individual
amphiphiles were collected by increasing their surface concen-
tration over the reported range. The latter corresponds to MMA’s
of ca. 450-50 Å2/molecule. Each data point represents the
integrated light intensity (“the gray level”) at a particular
concentration averaged over 30-90 s. The bars reflect the full
range of random light intensity fluctuations recorded during that
time. Despite active vibrational isolation, these fluctuations are
caused by the vibrations of the water surface induced by
mechanical vibrations beyond our control. The integrated-light
The surface potential data shown in Figure 3 provide the first
unambiguous evidence that only two longest chain derivatives
(C20- and C22TEMPO) exhibit a critical point above room
temperature. Using the Helmholtz equation
µZ ) (∆V)(MMA)ꢀꢀ0
(1)
the surface potential (∆V) measurements were interpreted in
terms of the normal component of the effective molecular dipole
moment (µZ) and plotted versus MMA.9,19,20 In a seminal paper,
Vogel and Mo¨bius demonstrated that the terminal methyl group
contributes +0.35 D to the effective dipole model in monolayers
with ordered, perpendicularly oriented alkane chains such as
stearic acid.19 In our case, however, the large cross sectional
area of the TEMPO headgroup relative to that of the chain (even
at small MMA’s) is likely to allow the terminal methyl group
sufficient conformational flexibility to minimize the contribution
of its dipole moment to the measured µZ. Thus, the magnitude
of µZ in Figure 3 is likely associated only with the TEMPO
headgroup.

9006 J. Phys. Chem. B, Vol. 105, No. 37, 2001
Johnson et al.
Figure 5. Plots of the average, reflected BAM light intensities
integrated over the area of illumination of 320 × 430 µm2 vs surface
concentration of C14TEMPO (closed circles), C16TEMPO (open circles),
and C18TEMPO (closed squares). The error bars correspond to the full
range of random light intensity fluctuations observed at each MMA
over ca. 30-90 s exposures. The other conditions are given in Figure
2.
their LE/G critical temperatures appear to be below 23.5 °C.
This conclusion is consistent with the surface potential measure-
ments shown in Figure 3. Needless to say, exhaustive BAM
monitoring of these three monolayers in the range of concentra-
tions explored in Figure 5 have never yielded a biphasic texture
that would be expected for a LE-G coexistence. We should
point out that the somewhat nonlinear behavior of C18TEMPO
in these experiments is not clear to us. As discussed below,
presence of a weak inflection point in the C18TEMPO data set
(I)
could suggest proximity of Tc for this compound. However,
linear plots of light intensity versus surface concentration which
did not exhibit this small inflection were also obtained for C18-
TEMPO at 6 and 0 °C. The surface potential and 2D electro-
chemistry data (see below) also did not give any evidence of a
phase transition for C18TEMPO. Consequently, we include this
amphiphile together with the other two in Figure 5 in the group
of 2D supercritical fluids.
Figure 6 shows the results of the same type of experiments
for the C20- and C22TEMPO monolayers. Clearly, the behavior
of these two compounds at 23.5 °C is substantially different.
At surface concentrations below ca. 1.7 × 10-10 mol/cm2
(corresponding to MMA > 100 Å2/molecule), a two-phase
region of LE and G phases is always observed, with the area
fraction covered by the 2D gas phase increasing as the
monolayers are expanded (see also Figure 4). The fluctuations
of the reflected light intensity in this region are due to a usual
monolayer flow and correspond to the variable fractional surface
area covered by each of the LE and G phases within the
illuminated area of 320 × 430 µm2. While only two data points
are marked for each surface concentration in this region, we
observed and recorded essentially all levels of reflected light
intensities between the high and low values shown in Figure 6.
These fluctuations gradually disappear when the apparent surface
concentration of an amphiphile decreases below ca. 8 × 10-11
mol/cm2. In that region, as the size of the gas domains and the
Figure 4. Brewster angle micrographs of C22TEMPO monolayer
recorded at (from top to bottom) 99, 105, and 130 Å2/molecule with
the other experimental conditions as those in Figure 2. The length of
the horizontal edges of these images corresponds to 430 µm.
intensity at 5 × 10-11 mol cm-2 and at lower concentrations is
the same as the value measured for the clean water surface.
Figure 5 presents the plots of the light intensity obtained for
C14-18TEMPO monolayers. The linearity of these plots reveals
the linearly varying density of the TEMPO monolayers over
the full range of their apparent surface concentrations from a
high-pressure region to ca. 450 Å2/molecule. On the basis of
this evidence, we conclude that these three members of the
TEMPO group do not undergo a phase transition. In other words,

Amphiphilic TEMPO Derivatives
J. Phys. Chem. B, Vol. 105, No. 37, 2001 9007
Figure 6. Plots of the average, reflected BAM light intensities
integrated over the area of illumination of 320 × 430 µm2 vs surface
concentration of C20TEMPO (open circles) and C22TEMPO (closed
circles). The pairs of the light intensities each shown for a given MMA
correspond to the minimum and maximum values of the reflected light
recorded at that MMA. These result from the fluctuations in the density
and size of the G and LE domains observed in the illuminated area
due to monolayer flow. The bars correspond to the full range of random
light intensity fluctuations observed for each measurement over ca. 30-
90 s exposures. The other conditions are given in Figure 2.
Figure 7. Plots of the average, reflected BAM light intensities
integrated over the area of illumination of 320 × 430 µm2 vs surface
concentration of C20TEMPO recorded at 27 °C (open circles) and 29
°C (closed circles). The pairs of the light intensities shown for the same
MMA in the experiment done at 27 °C correspond to the minimum
and maximum values observed in the LE-G coexistence region. The
other details and conditions are given in Figure 6.
position of the LE/G phase transition. Unlike what we reported
for the C22TEMPO system, an attempt to determine the position
of the LE/G phase transition during a gradual expansion of the
C20TEMPO LE phase is associated with a large histeresis of
ca. 40-60 Å2/molecule. Clearly, in this case, the kinetic barrier
to nucleate the gas phase is substantial. The existence of such
barriers and related phenomena are well-known and are related
to a decreased equilibrium vapor pressure (P) of a substance in
a cavity of small radius (r) relative to the equilibrium vapor
pressure measured above a flat surface (P0) (see eq 2).25
Similarly, the Kelvin equation predicts the equilibrium vapor
pressure to be higher for a droplet of liquid of small radius:
fraction of the surface area covered by the gas phase continue
to increase, the probability of observing simultaneously LE and
G coexisting within the small laser illuminated area decreases
and becomes negligible. We note that since the compositions
of the G and LE phases in the LE-G two-phase region are
constant, the reflected light intensity is expected to be ultimately
constant. The three data points at the lowest surface concentra-
tions in Figure 6 lead to that region.
Experiments such as those of Figure 6 do not provide an
accurate means of measuring the onset of the LE/G phase
transition. This is due to the small size of the monolayer area
probed by BAM. For example, if during monolayer compression
the small “bubbles” of the 2D gas phase are no longer seen in
the area probed by BAM but persist elsewhere and are thus
unnoticed, the MMA corresponding to the end of the two-phase
region may be overestimated. This difficulty notwithstanding,
if the BAM monitoring is extended to a sufficiently long time
(ca. 3 min; see Experimental Section) at each new MMA, the
position of the phase transition can be measured reproducibly.
In these types of experiments, we did not observe any histeresis
for the C22TEMPO system. In other words, the position of the
phase transition did not depend on whether it was measured
during slow, stepwise monolayer expansion or compression.
Specifically, C22TEMPO has a phase transition at 99 ( 1 Å2/
molecule. The precision of these measurements reflects largely
the uncertainty in the determination of the surface area. Since
these measurements involve times of several minutes, factors
such as the small solubility of the amphiphiles are responsible
for this uncertainty.
2γVm
P
RT ln
) (
(2)
( )
P0
r
Here γ is the surface tension of the liquid/vapor interface, and
Vm is the molar volume of the substance. The plus and minus
signs apply to the cases of small droplets of liquid and small
cavities of vapor, respectively, as encountered near G/L and
L/G phase transitions. In the former case, the increased
equilibrium vapor pressure of a small droplet of liquid results
in a commonly observed supersaturation of the vapor phase.25
The fact that the expansion of C20TEMPO monolayer does not
result in a LE/G phase transition at the same MMA as that
associated with the collapse of the G-LE coexistence region
during compression is a pronouncement of the instability of a
vapor phase in small 2D cavities described by the 2D version
of the Kelvin equation. What is surprising is the substantial
discrepancy in the behavior of the two rather similar monolayer
systems C20TEMPO and C22TEMPO. We believe that this is
related to the low value of the C20TEMPO critical temperature.
Indeed, consideration of the higher-temperature BAM data in
Figure 7 confirms this. While the plot of the reflected light
intensity vs surface concentration obtained at 27.0 °C clearly
The position of the phase transition of C20TEMPO, 127 ( 3
Å2/molecule is more difficult to determine. The value given here
was obtained in the compression experiments in which disap-
pearance of the LE-G coexistence region is taken as the

9008 J. Phys. Chem. B, Vol. 105, No. 37, 2001
Johnson et al.
exhibits a coexistence region consistent with the LE/G phase
transition at ca. 130 Å2/molecule, the same experiment carried
out at 29 °C yielded a linear dependence of the reflected light
intensity on the surface concentration over its entire range.
Needless to say, BAM images never revealed coexistence-type
textures at this temperature. Thus, the critical temperature for
C20TEMPO can be estimated to be 28 °C. We note that phase
nucleation near the critical temperature was treated theoretically
by McGraw and Reiss.26 Their prediction of an increased barrier
for nucleation under such conditions is reflected, albeit in 2D,
in the data presented above. We further note that preliminary
measurements of the reflected light intensities vs surface
concentration done for C18TEMPO yielded linear plots at
temperatures as low as 0 °C. Thus, the difference between C20-
vibration isolation is turned off are consistent with this explana-
tion. The oscillations and the average magnitude of the Faradaic
current at this MMA decrease slowly to zero with time and
upon further expansion of the monolayer. This behavior is
reversible upon recompression of the monolayer as shown in
Figure 8A, curve 4.
The same experiment can be initiated in a LE-G coexistence
region. Initially, only the capacitive background is recorded (see
Figure 8B, curves 1 and 2). When the monolayer compression
reaches the point of the phase transition (the end of the
coexistence region), the oscillatory behavior of the Faradaic
current is observed as shown in Figure 8B, curve 3. Further
compression yields a fully developed 2D voltammogram. In
these types of experiments, the position of the phase transition
can be accurately measured if the MMA of the initial positioning
of a fresh electrode is close to the phase transition and if the
expansion or compression is carried out in small steps. The
experiments of Figure 8A,B were done under such conditions
and yielded the correct (see below) position of the phase
transition of 98 Å2/molecule. However, if these conditions are
not met, the 2D electrochemical results can lead to an erroneous
conclusion. For example, consider the experiment in Figure 8C.
After recording the first voltammogram at 97 Å2/molecule which
indicated correctly the LE state of the monolayer, the monolayer
was expanded to 105 Å2/molecule. At this point, the monolayer
is certainly in the coexistence region. Yet voltammogram 2 in
Figure 8C exhibits the same oscillatory behavior as what we
associated above with the onset of the phase transition. Clearly,
such interpretation of this result would result in an error.
(I)
and C18TEMPO Tc values is greater than 28 °C, suggesting
that this homologous series of surfactants does not reflect a
commonly accepted rule of “equivalent states” which states that
a one carbon addition to the alkane chain changes monolayer
properties of an amphiphile in nearly the same fashion as a 5-10
°C decrease of temperature.9,27
2D Electrochemistry. In our previous paper, we reported a
discontinuity in the 2D voltammetry of tetradecane ferrocene-
ketone (C14Fc) upon slow expansion of its monolayer on the
water surface.14 Our observations indicated that the 2D volta-
mmetric “line” microelectrode which resides by self-positioning
exactly in the plane of the air/water interface functions as a
nucleating site for the monolayer gas phase. Once nucleated at
the line microelectrode, the gas phase isolates this sensor from
the LE phase, and the current signal abruptly vanishes, as the
surface concentration of the amphiphile in the G phase is
substantially lower and below detection limit of 2D voltammetry.
We were able to reproducibly determine (within (2 Å2/
molecule) the position of the LE/G phase transition by 2D
voltammetry both during C14Fc monolayer expansion and
compression.14 In the latter case, the appearance of the current
signal coincided with the final collapse of the coexistence region.
The performance of the line microelectrodes described in
Figure 8C is revealing in that it implies an affinity of the
microelectrode for the LE phase that could lead to a delayed
(erroneous) reporting of the LE/G phase transition. Additional
investigations of this and related phenomena showed that fresh
line microelectrodes (those never exposed to the air/water
interface) exhibit a remarkable and unmistakable ability to
differentiate between the LE and LE-G states of the monolayer.
Consider the following experiment. After an approximate
position of the phase transition is determined using 2D volta-
mmetry during monolayer expansion or compression experi-
ments as described above, the definitive test involves simply
placing a fresh line microelectrode at a monolayer covered
interface at a specific MMA and recording a 2D voltammogram.
This is illustrated in Figure 8D. In the case of C22TEMPO
monolayers at 99 Å2/molecule or higher, the response of a fresh
electrode is always limited to a capacitive background, indicating
that the electrode is in contact exclusively with the G phase.
At 98 Å2/molecule (or smaller), the same experiment with a
fresh microelectrode invariably yields a well developed Faradaic
signal, indicating that the electrode is in contact exclusively with
the LE phase of the monolayer. This behavior was observed
consistently with over 100 microelectrodes examining the state
of the C22TEMPO monolayers (in several independently spread
monolayers) around the LE/G phase transition. We observed
no exceptions from this behavior of fresh line microelectrodes.
This remarkable consistency in the behavior of fresh line
microelectrodes allows us to conclude that at 23.5 °C the LE/G
phase transition must occur at 98.5 ( 0.5 Å2/molecule.
We now consider similar experiments done with the
C14-22TEMPO monolayers. As reported previously for C14-
TEMPO, the three monolayers with the shorter chains, those
(I)
with Tc below room temperature, did not show any disconti-
nuities in 2D voltammetric experiments even following pro-
longed equilibration of their monolayers at 450 Å2/molecule.
In the case of the two longer chain derivatives, C20-22TEMPO,
2D electrochemistry can be used to determine the position of
their LE/G phase transition. Two types of experiments can be
done to accomplish this. Both are illustrated in Figure 8 using
C22TEMPO. The first experiment is analogous to that described
above for the C14Fc case. A line microelectrode is positioned
at the water surface in the LE region of the monolayer and a
2D voltammogram is recorded. The monolayer is then expanded
in small steps and current voltage curves are recorded after each
expansion. Initially (see Figure 8A, curve 1), a well-developed
Faradaic current is observed as expected for the LE state. Once
the MMA of the phase transition is reached, the current begins
to oscillate (curve 2) and then decays to a background level
(curve 3). The oscillations can be associated with the presence
of small 2D cavities of the gas phase newly generated at the
line microelectrode. Since compressibility of the gas phase is
naturally larger than that of the LE phase, small mechanical
vibrations in the monolayer result in the observed fluctuations
in the length of the microelectrode remaining in contact with
the LE phase. The observed frequency of these fluctuations of
7-10 Hz, which corresponds to the frequency of the mechanical
laboratory noise, and the increase of their amplitude when
The consistent ability of fresh microelectrodes to unambigu-
ously determine the state of the monolayer within just 1 Å2/
molecule around the point of the LE/G phase transition is even
more remarkable when one compares the size of a line electrode
(500 µm in length) with the average size and density of the 2D
gas cavities. The latter two can be assessed from the BAM

Amphiphilic TEMPO Derivatives
J. Phys. Chem. B, Vol. 105, No. 37, 2001 9009
Figure 8. 2D voltammetry of C22TEMPO at the air/water interface (50 mM HNO3 subphase). All curves in each of the panels A-C were recorded
at the same line electrode. In panel D, each 2D voltammogram was recorded at a fresh line microelectrode. The 2D voltammograms were recorded
at the following MMA’s: A, curves 1-4 correspond to 97, 99, 101, and 98 Å2/molecule, respectively; B, curves 1-3 correspond to 102, 100, and
98 Å2/molecule, respectively; C, curves 1 and 2 correspond to 97 and 105 Å2/molecule, respectively.; D, curves 1-3 correspond to 100, 99, and
98 Å2/molecule, respectively.
images of the monolayer at 99 Å2/molecule such as that shown
in Figure 4. Examination of these images yields the average
diameter of a gas cavity of 5-20 µm. While their density
fluctuates from one region of the monolayer to the next, it is
relatively low. Since the consistent absence of the Faradaic
signal under these conditions (Figure 8D, curves 1 and 2) must
be associated with the fact that the entire 500 µm length of a
line electrode is in contact with the 2D gas phase, it is clear
that the line electrodes function as nucleating sites of the G
phase.
Similarly to the diminished capabilities of BAM to detect
LE/G phase transition in C20TEMPO monolayers, the ability
of the line microelectrodes to accomplish this is also less
effective. Just as in the BAM experiments, nucleation of the G
phase at a microelectrode during stepwise expansion of these
monolayers is burdened with a large histeresis of as much as
30 Å2/molecule. When a C20TEMPO monolayer in the LE-G
coexistence region is slowly compressed, the appearance of the
well developed 2D signal was first observed in the range of
130 to 125 Å2/molecule. This is consistent with the BAM result
reported above. The behavior of fresh line microelectrodes to
detect the 2D G phase is also less precise. The point nearest to
the phase transition at which LE phase was detected was not as
sharply defined as in the C22TEMPO case and involved a range
of MMA’s of 127-130 Å2/molecule. Naturally, the proximity
of the critical temperature to the room temperature of this
monolayer and the expected diminished differences between the
physical properties of the LE and G phases of C20TEMPO are
responsible for these results. An interesting pronouncement of
the diminished differences between LE and G phases is an
increased equilibrium concentration of the C20TEMPO in the
gas phase. While the surface concentration of the C22 derivative
in the G phase is clearly below the detection limits of the 2D
voltammetry, the higher G phase concentration of the C20
derivative allows us to record a measurable Faradaic current in
this phase. Its magnitude gives an estimate of the C20TEMPO

9010 J. Phys. Chem. B, Vol. 105, No. 37, 2001
Johnson et al.
gas-phase concentration of 1.1 ( 0.4 × 10-11 mol/cm2 (corre-
sponding to a MMA of 1500 Å2/molecule). Characterization of
the properties of the LE and G phases near the critical
temperature is a subject of current investigations.
Interestingly enough, microelectrodes first exposed to the 2D
gas phase also exhibit a somewhat delayed response when
monolayer compression results in elimination of the LE-G
coexistence region. To reiterate the main observation: only fresh
line microelectrodes can consistently show the high/low type
current signal when first contacting the LE or LE-G states of
the monolayer, respectively.
Conclusions
We focused on the comparison of BAM and 2D electro-
chemistry as two independent techniques capable of determining
the position of the LE/G phase transition and the value of the
2D critical temperature of C20-22TEMPO monolayers. We
showed that by constructing plots of the integrated intensity of
the reflected BAM laser light versus the amphiphile’s surface
concentration for several values of temperature, one observes
linear plots above the critical point, and that below the critical
temperature, such plots exhibit a distinct inflection with multiple
values of the reflected light intensities at each surface concentra-
tion below the LE/G phase transition. Using this approach, we
determined the C20TEMPO LE/G critical temperature to be 28
°C.
Acknowledgment. We are grateful to Professor Rebecca
Braslau for suggesting significant modifications to the procedure
of the CnTEMPO synthesis and for her hospitality at UC Santa
Cruz, where the modified synthesis was first carried out. This
research was supported financially by the U.S. National Science
Foundation under Grant CHE-0079225.
References and Notes
(1) Kaganer, V. M.; Mo¨hwald, H.; Datta, P. ReV. Mod. Phys. 1999,
71, 779-819.
(2) Knobler, C. M. Science 1990, 249, 870-874.
(3) Knobler, C. M.; Rashmi, C. D. Annu. ReV. Phys. Chem. 1992, 43,
207-236.
BAM and 2D voltammetry can also be used to determine
the position of the LE/G phase transition. Using C22TEMPO
monolayers as a test case, we found that both techniques are
equally accurate and precise. BAM requires a somewhat time-
consuming set of measurements in the immediate vicinity of
the phase transition. Since in this technique only the images
featuring a LE-G coexistence texture can be unambiguously
interpreted, one must minimize a possibility of “false negative”
type results. Contrary to this feature of BAM, the very nature
of the phenomena allowing 2D voltammetry to determine the
phase transition circumvents such uncertainties offering unam-
biguous results at MMA’s both below and above the phase
transition. We find this feature of the 2D voltammetric method
to be clearly superior relative to BAM. To expand on this
point: we discovered that fresh line microelectrodes exhibit a
remarkably consistent ability to nucleate the 2D gas phase and
become fully contacted by the gas phase. Consequently,
whenever a fresh microelectrode is placed at the air/water
interface covered with a C22TEMPO monolayer in a LE-G
coexistence region, the 2D gas phase is immediately nucleated
along its entire length. This then yields only a small capacitive
current signal in this region. A 2D voltammetric signal featuring
a large Faradaic current is always observed if a fresh electrode
is placed in contact with a monolayer in the LE state. Thus,
within a 1 Å2/molecule change in MMA around the phase
transition, the 2D current switches between the on/off levels
with a ratio of ∼50. This consistent ability of line electrodes to
nucleate the 2D gas phase (also observed in the case of C14Fc
monolayers14) is perhaps related to the hydrophilic character
of clean gold, which is newly generated when a line electrode
is created by fracturing a gold-coated glass slide (see Experi-
mental Section). Thus, at the first contact with the monolayer
on the water surface, the clean surface of the gold edge is
hydrated and, following nucleation of the 2D gas phase, prefers
to stay in contact with that phase rather than to contact a more
lipophilic LE phase of the monolayer. We note, however, that
the line electrodes, once exposed to the LE phase, no longer
exhibit the affinity for the G phase that the fresh electrodes do.
This leads, in experiments such as that in Figure 8C, to a delayed
nucleation of the 2D gas phase upon monolayer expansion.
(4) McConnell, H. M. Annu. ReV. Phys. Chem. 1991, 42, 171-195.
(5) Henon, S.; Meunier, J. ReV. Sci. Instrum. 1991, 62, 936-939.
(6) Ho¨nig, D.; Mobius, D. J. Phys. Chem. 1991, 95, 4590-4592.
(7) Als-Nielsen, J.; Jackuemain, D.; Kjaer, K.; Leveiller, F.; Lavav,
M.; Leizerowitz, L. Phys. Rep. 1994, 246, 251-313.
(8) Adam, N. K.; Jessop, G. Proc. R. Soc. London, Ser. A 1926, 110,
423-441.
(9) Gaines, G. L. J. Insoluble Monolayers at Liquid-Gas Interfaces;
Wiley-Interscience: New York, 1966.
(10) Knobler, C. M. Recent Developments in the Study of Monolayers
at the Air-water Interface. In AdVances in Chemical Physics; Progogine, I.,
Rice, S., Eds.; Wiley: New York, 1990; Vol. 77, pp 397-449.
(11) Moore, B. G.; Knobler, C. M.; Akamatsu, S.; Rondelez, F. A. J.
Phys. Chem. 1990, 94, 4588-4595.
(12) Casson, B. D.; Bain, C. D. J. Am. Chem. Soc. 1999, 121, 2615-
2616.
(13) Pfohl, T.; Mo¨hwald, H.; Riegler, H. Langmuir 1998, 14, 5285-
5291.
(14) Johnson, M. J.; Anvar, D. J.; Skolimowski, J. J.; Majda, M. J. Phys.
Chem. B 2001, 105, 514-519.
(15) Majda, M. Translational Diffusion and Electron Hopping in
Monolayers at the Air/Water Interface. In Organic Thin Films nad Surfaces;
Ulman, A., Ed.; Academic Press: San Diego, 1995; Vol. 20, pp 331-347.
(16) Charych, D. H.; Landau, E. M.; Majda, M. J. Am. Chem. Soc. 1991,
113, 3340-3346.
(17) Wittek, M.; Mo¨ller, G.; Johnson, J. J.; Majda, M. Anal. Chem. 2001,
73, 870-8777.
(18) Johnson, M. J. Ph.D. Dissertation, UC Berkeley, Berkeley, CA,
2001.
(19) Vogel, V.; Mo¨bius, D. J. Colloid Interface Sci. 1988, 126, 408-
420.
(20) Taylor, D. M. AdV. Colloid Interface Sci. 2000, 87, 183-203.
(21) The dependence of the effective dipole moment of the TEMPO
amide headgroup on its orientation with respect to the water surface was
investigated numerically using the QChem density functional package. In
general, these calculations support our hypothesis. However, since several
assumptions have to be made concerning plausible conformations of the
two polar groups of this amphiphile and the headgroup orientation with
respect to the water surface, we cannot credibly offer these results as a
quantitative proof of this hypothesis.
(22) Overbeck, G. A.; Mo¨bius, D. J. Phys. Chem. 1993, 97, 7999-
8004.
(23) Rivier, S.; Henon, S.; Meunier, J.; Schwartz, D. K.; Tsao, M.-W.;
Knobler, C. M. J. Chem. Phys. 1994, 101, 10045-10051.
(24) Melzer, V.; Vollhardt, D.; Brezesinski, G.; Mo¨hwald, H. J. Phys.
Chem. B 1998, 102, 591-597.
(25) Adamson, A. W.; Gast, A. P. Physical Chemistry of Surfaces, 6th
ed.; Wiley: New York, 1997.
(26) McGraw, R.; Reiss, H. J. Stat. Phys. 1979, 20, 385-413.
(27) Peterson, I. R.; Brzezinski, V.; Kenn, R. M.; Steitz, R. Langmuir
1992, 8, 2995-3002.