Capacitance studies of synthetic phospholipid Langmuir films

Chemistry and Physics of Lipids 18 (1977) 49-61 © Elsevier/North-Holland Scientific Publishers Ltd.

CAPACITANCE STUDIES OF SYNTHETIC PHOSPHOLIPID LANGMUIR FILMS W.L. PROCARIONE and J.W. K A U F F M A N Department of Materials Science and Engineering, Northwestern University, The Technological Institute, Evanston, Illinois 60201, USA Received April 6th, 1976,

accepted June 25th, 1976

Synthetic phosphatidylcholine Langmuir films have been incorporated into metal-insulatormetal (MIM) thin film junctions. The capacitance characteristics of these junctions have been studied as a function of temperature, the number of lipid layers in the insulating layer, and the length of the hydrocarbon chains of the lipid molecule. The thickness of the oxide layer on the base aluminum electrode has been determined to be ;~11 A, and its effects on the capacitance characteristics have been considered in some detail. Indications of phase transitions in the temperature dependence of the capacitance imply that the basic lamellar arrangement of the lipid molecules is retained even after the samples are subjected to a dehydrating vacuum annealing process. An examination of the effects of varying the hydrocarbon chain length and salt content of the subphase during sample fabrication showed that capacitance characteristics of the M1M junction are very sensitive to small structural changes in the insulating layer.

I. Introduction There are a few reports in the literature concerning the capacitance characteristics of m e t a l - i n s u l a t o r - m e t a l (MIM) thin film junctions where the insulating layer has been composed of layers of long chain fatty acids. Handy and Scala [1 ] and Mann and Kuhn [2] studied the effect of chain length and the number o f layers on the capacitance and dielectric constant of such structures. Lundstr6m et al. [3,4] have investigated in some detail charge storage in Langmuir m e t a l - i n s u l a t o r - s e m i conductor (MIS) structures via a capacitance technique. Our interest in the MIM structure has been directed toward using it as a vehicle for studying various properties o f the material comprising the insulating layer. Capacitance measurements can be made very precisely, and the capacitance characteristics of an MIM junction should be a sensitive indicator of structural changes in the insulating layer. This report deals with samples in which the insulating film is composed of a layer or layers of synthetic phosphatidylcholine (p.c.) which form both thermotropic and lyotropic liquid crystals. There are two well-defined transition temperatures for synthetic phospho!ipids with a layered structure [5]. Naturally occurring lipids show 49

50

W.L.Procarione, J. W. Kauffman, Capacitance studies of synthetic phospholipid films

similar phase transitions, and possible roles for such transitions in biological functions have been the subject of some speculation [6]. If such phase transitions occurred in the lipid layers of our MIM structure, then we would expect changes in the insulator thickness which would be reflected as changes in capacitance values.

II. Materials

The synthetic lipids L-a-l,2-dimyristoylphosphatidylcholine (DML), L-a-1,2-dipalmitoylphosphatidylcholine (DPL) and L-a-1,2-distearoylphosphatidylcholine (DSL) were used as received from Calbiochem. In the course of the experiments, materials from several different lots were used with no apparent variation in any of the electrical properties of the samples. A mixture of dicyclohexyl and chloroform in a 3 : 2 ratio by volume was judged to be the most favorable solvent on the basis of spreading characteristics and lipid solubility. A study of the surface potential properties of phospholipid monolayers formed with chloroform and dicyclohexyl indicated that both of these solvents completely evaporated from the monolayer in a relatively short period of time (<20 min) ref. [7]. Solvents from several different lots were used with no apparent effect on the properties of the samples. Similarly, varying the ratio of dicyclohexyl to chloroform from pure dicyclohexyl to pure chloroform had no measurable effect on the electrical properties.

III. Experimental Details of the experimental procedure both with respect to sample fabrication and electrical measurements have been reported previously [8]. Ordinarily, the lipid monolayers were formed over a subphase Of doubly distilled water at pH 5.6, and then transferred to solid substrates via a dipping process. Some samples were formed over subphases with NaC1 in concentrations of 10 - 4 M to 10-2 M. The addition of the salt did not significantly affect the pH of the subphase. Capacitance measurements (at 5 kHz, 25 mV rms) were made with a General Radio 1615-A capacitance bridge using a PAR model HR8 lock-in amplifier as a null detector. Temperature dependence measurements were made in a continuous manner by using a Data Track programmed for linear 0.5°C/min temperature ramp. Unless otherwise noted, sample area was 5.6 sq mm as measured with a comparater. As will be discussed below, the specific capacitance of samples which are nominally identical may vary by several per cent. However, the precision of the total capacitance measurement itself is approximately -+0.1%. Thus it is possible to detect very small changes in capacitance, say as a function of temperature, for a particular sample.

W.L. Procarione, J. W. Kauffman, Capacitance studies o f synthetic phospholipid films

51

IV. Samples The MIM junctions were formed on 1 in. × 3 in. glass slides. Basically, the procedure was to vapor deposit a single base electrode, deposit a layer or layers of lipid molecules by the kangmuir-Blodgett [9,10] dipping process and then vapor deposit the top electrodes. This resulted in three individual MIM junctions per slide as illustrated in fig. la. Most of the samples were constructed with aluminum base and top electrodes. However, a few samples were made with copper and zinc base electrodes. Varying the base electrode metal had no apparent qualitative effect on the capacitance characteristics of the samples and all results reported here are for samples with aluminum electrodes. Ordinarily, the capacitance values of junctions on the same slide would agree within a few per cent. However, we often observed variations in the capacitance as

~

ELECTRODE E ELECTRODES

S;~ ',,~P-E

.&RE'---~

(4.8ram':,

",,~

(a) 1<--~4.0~ ' ' - ~ <-- - 7 o ~

)1

AI

BASE ELECTRODE

TOPELECTRODE

(b) Fig. 1. a: the MIM junctions are formed at the areas of intersection of the base electrodes and the top electrode. Three individual junctions are formed on each slide, b: a cross-section of the junction (i.e., perpendicular to the plane of the slide at the intersection of the base and top electrode). The arrangement of the lipid molecules that is shown is for illustrative purposes only. The precise conformation and arrangement of the molecules are unknown.

52

W.L.Procarione, J. W. Kauffman, Capacitance studies of synthetic phospholipid films

great as 50% for samples that were nominally the same, but were formed on different slides. The total capacitance of a parallel plate capacitor is determined by its area and the thickness and dielectric constant of the insulating medium. The same evaporation mask was used in the fabrication of all of the samples so that there could be no significant variation in the area of each junction. Therefore, the differences in capacitance values must have resulted froln irregularities in either or both the insulator thickness and dielectric constant. Fig. l b schematically illustrates the cross-sectional structure of the M1M junction. After the base electrode is deposited, it forms an oxide layer since it is exposed to air in the process of depositing the lipid layers. We should also note that it is possible for the top electrode to be partially oxidized by interaction with the water bound to the lipid molecules. In their experiments on fatty acid MIM junctions, Mann and Kuhn [2] estimated a 70 A oxide thickness on their aluminum electrodes. The thickness of each lipid layer is approximately 30 A. Thus the aluminum oxide with a dielectric constant 2 - 4 times greater than that expected for the lipid material forms the major part of the insulating medium for junctions with a small number of lipid layers. It is clear then that irregularities in the oxide thickness or dielectric constant could have been responsible for the variation in capacitance values for samples which were supposed to have the same number of lipid layers. In order to check this possibility, several series of samples were made simultaneously under identical conditions. Thus it was reasonable to expect that each sample in a particular series would have nearly the same oxide thickness. Reproducibility in the capacitance values was markedly improved by this procedure. Although anomalous values were still observed, samples nominally identical ordinarily showed capacitance values agreeing within 5%.

V. Results and discussion By making a series of samples with a constant oxide thickness and an increasing number of lipid layers, it becomes possible to estimate the thickness of the oxide layer and the,dielectric constant of the lipid material. In fig. 2, we have plotted 1/C versus the number of lipid layers. This series of samples was fabricated according to a consistent procedure with the lipid films being deposited immediately after the vapor deposition of the base electrode. If we extrapolate the line formed by the points through zero number of layers via a least squares fit, the remaining value of 1/C should be due to the capacitance of the oxide layer. Knowing the area of the capacitor and the dielectric constant of A1203 (9.34) we can calculate an oxide thickness of approximately 11 A. We should point out that very thin metal oxide films are often porous resulting in a dielectric constant which falls rapidly with decreasing thickness [11 ]. For oxide films below 50 A, marked departures from the bulk value of the dielectric constant may be expected [12]. Thus the value of 11 A, calculated above might be significantly underestimated and should be considered as

W.L. Procarione, J. W. Kauffman, Capacitance studies of synthetic phospholipid films

53

4.5

l

4.0

T

35

% o

3.C

:o 2.0 z

1.5 n

0.~

I

r

I

2 NUMBER

OF

i

i

3

4

LAYERS

Fig. 2. (Capacitance) -1 versus the n u m b e r of DSL layers in the MIM j u n c t i o n (~23°C). T h e error bars reflect the uncertainty in the temperature and sample area.

a minimum value. As we pointed out previously, fluctuations in tile total capacitance values of nominally identical monolayer samples indicated that the value of oxide thickness was dependent on the details of the sample fabrication procedure. Using the data from fig. 2, it becomes possible in principle to determine the dielectric constant of the lipid layers. However, there are several factors which make a precise calculation impossible. The major limitation is the uncertainty of the value of the oxide thickness. Another factor that must be considered is that in the process of depositing the top electrode the aluminum atoms should penetrate the lipid film to some depth. The maximum depth of penetration must be less than the thickness of one layer, however, since monolayer films are usually highly insulating. If it is assumed that the uncertainties in oxide thickness and penetration depth of the metal atoms are systematic errors, then their effects can be eliminated if the proper series of samples is available. For example, if we take the difference 1/C (3 layers) - 1/C (1 layer), we should then be left with a 1/C value due to a lipid bilayer alone (see fig. 4). Similarly, the difference in the 1/C values for 2 and 4 layer samples should

54

W.L.Procarione, J. W. KauJ'fman, Capacitance studies of synthetic phospholipid films

also be due to a bilayer. Using the data points from fig. 2, we obtain: _

1 (3 layers) C

-~(4 layers) -

1

(farads ~- l (1 layer) = 1.6 + 0.2 X 10 6 \ c m ] (2 layers)= 1.9 -+ 0.3 × 106 (farads t - i \cm/

The values agree within the uncertainty of the calculation. In order to determine the dielectric constant from specific capacitance values, it is necessary to know the thickness of the lipid monolayer. Although the exact conformation and arrangement of the lipid molecules in our samples are unknown, we can estimate the thickness per layer from the following values: 11 A for the phosphatidylcholine head group [13], 1.5 A per C - C bond [14], and 30 ° tilt angle for the hydrocarbon chains [14]. Numbers obtained from this scheme agree with the Xray d-spacings obtained for DPL Langmuir-Blodgett multilayers by Levine et al. [ 15]. Using the DSL bilayer thickness of 62.4 A. computed in accordance with the criteria given above, and taking an average of the 1/C values, we obtain a value of 4.1 -+ 0.9 for the dielectric constant of a DSL bilayer. This is higher than most of the reported dielectric constant values for stearic acid multilayers, but agrees with the upper limit of the values obtained for barium and calcium stearate films reported by tlandy and Scala [ 1]. Of course, we are measuring an average dielectric constant and the lipid head group may have a larger dielectric constant than the fatty acid chains. The presence of water is known to greatly affect the electrical properties of bulk lipid materials. In order to remove the water from our samples, they were subjected to a vacuum annealing process. Williams and Chapman have reported the conditions necessary for completely removing the water from bulk synthetic phosphatidylcholine [16]. On the basis of that work, our samples were heated under vacuum for a minimum of 3 hr at 65°C, 85°C, and 105°C for DML, DPL and DSL, respectively. Exceeding these temperatures resulted in a significant probability of permanently destroying the insulating nature of the junctions_The annealing process causes an overall decrease in the capacitance of 5 9% for all samples tested. The decrease is permanent as long as the samples are not exposed to air for long periods of time, and is independent of the base electrode material. Since 1/C increases because of the annealing process, we are either increasing the effective thickness of the insulating layer and/or decreasing its average dielectric constant. A permanent molecular rearrangement caused by raising the sample to relatively high temperature is not unlikely. However, X-ray studies on lipid-water dispersions indicate that the lamellar spacing decreases at higher temperatures which would imply an effect on the capacitance opposite to what we observe [17]. The environment of the lipid molecule in a dispersion is quite different from the Langmuir film where the lipid is in contact with a solid substrate. There is not presently available any study of the variation of the lamellar spacing with temperature for lipid Langmuir films.

W.L. Procarione, J. W. Kauffman, Capacitance studies of synthetic phospholipid films

55

Table 1 The known transition temperatures (°C) at varying degrees of hydration for the types of synthetic phosphatidylcholine used in this study [5,23].

At maximum hydration

Tml Tm2

Monohydrate, c~form Anhydrous, ~3form

Tc~ Tt3

DML

DPL

DSL

13.5 23.7 51 115

34.0 41.7 65 -

49.1 58.2 78 -

A second possible explanation for the decrease in the capacitance with annealing is a decrease in the average dielectric constant of the insulating layer. There is a large increase in the sample resistance (4 5 orders of magnitude) caused by the annealing. Since the presence of water increases the conductivity of bulk lipid material by several orders of magnitude [ 18], this indicates that the annealing process removes water from the lipid layers in the MIM junction. Each lipid molecule can bind 1 0 - 1 2 water molecules [19,20]. The dielectric constant of water at 20°C is ca. 78, which is approximately 8 times that of A1203 and 20 times that of the lipid. Although the dielectric constant of the water in the lipid layer may be markedly reduced due to its bound nature, it is still clear that the removal of a thin layer of water could cause a significant decrease in the total dielectric constant..There are too many unknowns, however, to make a reasonable calculation of the magnitude of such an effect. In the Introduction, we mentioned the liquid crystalline nature of phospholipids. The temperatures of the crystalline to liquid crystalline transitions of the synthetic 1,2-diacyl-L-phosphatidylcholines are listed in table 1. In analyzing the variation of capacitance characteristics with temperature, we have looked for features that might indicate the occurrence of phase transitions. In some of the following figures the temperatures of known phase transitions have been indicated by arrows. The arrows are for reference only and, unless otherwise noted, are not meant to imply that the data is indicative of a phase transition. We should also point out that the data presented in figs. 3 - 6 was obtained from selected samples which best illustrated the behavior typical of their kind. Since the samples were fabricated at different times and under slightly different conditions, quantitative comparison cannot be made among them because of varying oxide thicknesses. In addition to decreasing the overall capacitance of the samples, the annealing process also modifies the temperature dependence of the capacitance. This is illustrated in fig. 3 for a DSL 3 layer sample. On the initial heating cycle, the 1/C (T) curve shows two rather sharp discontinuities near the Tml and Tm2 transition temperatures and a continuous increase in 1/C above Tin2. After annealing, the curve is smoothed out and there is a slow decrease in 1/C above Tin2. There is little variation in the capacitance characteristics on successive runs after the annealing process.

56

T

W.L. Procarione, J. W. Kauffrnan, Capacitance studies of synthetic phospholipid films

2.75

Tml J,

Trn2

i ~'

AFTER ANNEALING

ooo ° ° ° ° ° ° ° °

¢ o OooooO°

o

o

o

o

o

o

o

"

o

o

2.70

0 x

w U Z

2.65 0

u
0 0

0 Tm2 2.60

o

0

0

INITIAL HEATING

Tmj ~ oO

CYCLE

0 0 O0~ ( ~ 0 0 0 t

i

i

i

i

i

i

30

40

50

60

70

80

90

TEMPERATURE

i

I00

°C

Fig. 3. (Capacitance) -1 versus temperature for a DSL 3-layer sample illustrating the effect of annealing on the magnitude and temperature dependence of the total capacitance value. The absolute accuracy of the data presented in this and the following figures is primarily limited by the uncertainty in the sample area which is -+6%.The precision of the total capacitance measurement is approximately +0.1%. Kopp et al. [21] recently studied, using electron microscopy, the reorganization processes in Langmuir-Blodgett layers of barium stearate, cadmium arachidate and tripalmitin. It would be reasonable to expect that similar reorganization occurs in p.c. films and that such processes would be enhanced by raising the temperature of the films. The total capacitance value of the particular sample is determined by the average thickness of the insulating layer. Therefore one cannot assume that the multilayers are intact on the basis of a regular increase of 1/C values with the number of layers. In contrast to capacitance characteristics, the conductivity of an MIM is highly sensitive to imperfections in the insulating layer, and is relatively unaffected by the presence of the oxide layer. Our conductivity measurements show a regular increase in sample resistance with the increasing length of the hydrocarbon chain in the p.c. molecule [22]. While multilayer samples remain highly insulating indefinitely, scatter in their conductance values indicates that the layers beyond the first layer do not remain intact. Given the uncertainties in the structure of the multilayer samples, we have concentrated our efforts on monolayer samples. In fig. 4, the properties of monolayers formed with lipids with different hydrocarbon chain lengths are illustrated. All three samples show a decrease in 1/C with increasing temperature. However, the change is

W.L. Procarione, J. W. Kauffman, Capacitance studies of synthetic phospholipid films

I. 72

z.a

1.68

DML

57

~2~ A

~E

z~ £xA A

"o o

o x LI.I <( I.-

1.62

0300 0o

DPL

1,58

DO

°°°

T~ITrn2

a. 1.54 o

000121000 Oo

1.98

0 0o O0 0

DSL O~~°Oo

1.94

T, Tm~ hi,' I

°°°°°~a~m~°~oo o oo oo o -4'0

' -20 ' ' 0 ' 2'0 ' ; 0 ' 6' 0 ' TEMPERATURE *C

8O ~ ' I00

Fig. 4. (Capacitance) -1 versus temperature for p.c. monolayers of different hydrocarbon chain lengths. The differences in the temperature profiles may be accounted for by varying molecular packing densities.

only a matter of a few per cent over a temperature range as great as 140°C. Since for monolayer samples the bulk of the insulating medium is the oxide layer, we should consider whether or not the oxide might be responsible for the change in capacitance with temperature. Change in thickness with temperature is described by the equation Ax = c~xAT where a is the linear coefficient o f thermal expansion. For aluminum oxide o~ = 8.0 X l 0 -6 • deg -1 in the relevant temperature range. Since c~ is positive, the oxide thickness increases with increasing temperature which would result in a trend for 1/C opposite to that shown in fig. 4. We can calculate the magnitude of the capacitance change expected from the oxide layer. Assuming a relatively thick oxide of 40 A the change in thickness for a temperature interval o f IO0°C is 0.03 A. For a dielectric constant of 9.34 and a capacitor area of 5.6 sq. mm, this translates into a

58

W.L. Procarione, J. I4/. Kauffrnan, Capacitance studies of synthetic phospholipid films

capacitance change (AC) of 7.4 × 10 -] 1 f. The value of AC for the DPL monolayer shown in fig. 4 is -1.9 × 10 - 9 f for the same temperature interval. For this probably extreme case, neglecting the effect of the oxide layer results in an error of approximately 4%. While the data for all three samples shown in fig. 4 shows a decrease in 1/C as the temperature is raised, the slope of the curve above Tin2 is clearly dependent upon the length of the hydrocarbon chains, the slope decreasing with increasing chain length. Experiments on p.c. monolayers at the air-water interface indicate that they become more tightly packed with increasing chain length (at high surface pressure) [23]. This results from an increased attractive force between molecules due to chain length dependent van der Waals forces. If the same relationship between packing density and chain length holds for monolayers transferred to a solid substrate, then the results displayed in fig. 4 might be expected. The transition occurring at Tin2 is associated with increased chain mobility [24]. The less tightly packed the molecules, the more mobility the chains will have above Tin2. This facilitates sample thinning which is reflected in the decreasing 1/C values. For the most tightly packed DSL molecules, chain mobility above Tin2 will be limited and the decrease in thickness will be the least. It is possible to check this hypothesis. While low concentrations of salt in the subphase have no effect on the packing density of the neutral phosphatidylcholine molecule at the air-water interface, higher concentrations (10 -l M) cause a 5 10%

!.93

0

%

0

H20

u

SUBPHASE

o

192 o

Tm2

o

o

°% 0 0 0 0 oCO0

0

1.91 w z I-

0

0

0

0

0

/ 1 ~ 3 M NoCI SUBPHASE

2.56

Tm2

(.9

•

•

O

2.55

Tmr

•

$

O0

i

i

L

10

20

30

.~

•

•

•

o••oO•

i

i

i

i

40

50

60

70

TEMPERATURE

i

80

aC

Fig. 5. The effect on the temperature dependence of the capacitance of a 10 - 3 M NaC1 concentration in the monolayer subphase during sample fabrication.

W.L. Procarione, J.W. Kauffman, Capacitance studies of synthetic phospholipid films

•

59

Trnl

2.52 I0 "z M

me Tm=

;$

T

NoCl

SUBPHASE

E 2.51 "o o o 2.50 0 x

w o

z 2.49 I,-

0 0

1.92

0

Tm I

Tin2

H 0

SUBPHASE

0

0

°Oo ; 000000 ,

1.91

2'0

50 ,;o 50

0

0

6'o

TEMPERATURE

~o

0

0

0

8'o '~o

0 ?

,oo

*C

Fig. 6. The effect on the temperature dependence of the capacitance of a 10-2 M NaC1concentration in the monolayer subphase during sample fabrication. expansion of the layer [25,26]. This suggests that by fabricating samples from monolayers formed over subphases containing salt at various concentrations, it should be possible to alter the packing density of the molecules and therefore the temperature dependence of the capacitance characteristics. Three DSL monolayer samples were formed over subphases with NaC1 concentrations of 10-4 M, 10 -3 M and 10 -2 M. The capacitance characteristics of these samples were compared to the data obtained from an unmodified DSL monolayer junction. The 10 -4 M and 10 -3 M samples yielded essentially the same results and the 10 -3 M sample is compared to the unmodified sample in fig. 5. The most important difference occurs above Tm2 where the slope has changed from being slightly negative to being slightly positive for the salt modified sample. An entirely different result is obtained, however, if we increase the NaC1 concentration by one order of magnitude to 10 -2 M. As is illustrated in fig. 6, the slope of 1/C versus T is now strongly negative. According to the hypothesis given above, this is the result that we would expect if the salt in the subphase had caused the monolayer to expand. Conversely, the data in fig. 5 imply that the lower salt concentrations may cause a slight

60

W.L.Procarione, .L W. Kauffman, Capacitance studies of synthetic phospholipid films

condensation of the monolayer. In figs. 3 - 6 , we have noted the temperatures of the Tml and Tin2 phase transition temperatures. In several cases the 1/C versus T cures show irregularities o5 changes of slope near these temperatures. The Tml and Tin2 temperatures were determined by differential scanning calorimetry (DSC) measurements on dilute solutions of synthetic p.c. [5]. According to other DSC data, decreasing the water content of the lipid solution below 20% causes a continuous increase in the transition temperatures [ 16]. With the exception of the lower curve in fig. 3, all of the data presented here was for samples annealed and measured under vacuum. The extremely low value of the conductivity (ca. 10 -14 ~2-1 " cm -1) o f the lipid layers after annealing suggest that no significant portion of the lipid material remains hydrated after the annealing process. Therefore, we should expect to see indications of the phase transitions at temperatures higher than we actually observed. The implication of our measurements is that removing the water after the sample has been formed has little effect on the basic lamellar structure of tile lipid material. The contact with the solid substrate is apparently responsible for maintaining the structure after the water has been removed. Of the results presented above, there are two points that should be especially emphasized. The first is that even though the lipid material is exposed to a vacuum environment and relatively high temperatures, the basic lamellar structure is apparently retained. Secondly, by examining the effects of varying the hydrocarbon chain length and the salt content of tile subphase during sample fabrication, it has been shown that the capacitance characteristics are very sensitive to small structural changes in the lipid layers. There is a potential for refining the technique so that quantitative structural determinations may be made.

Acknowledgements The authors wish to thank L. Lis for his informative comments. This work was supported in part by the Office o f Naval Research.

References [1] [2] [3] [41 [5] [6]

R.M. Handy and L.C. Scala, J. Electrochem. Soc. 113 (1966) 109 B. Mann and H. Kuhn, J. Applied Phys. 42 (1971) 4398 I. Lundstr6m and D. McQueen, Chem. Phys. Lipids 10 (1973) 181 I. Lundstr/Sm and M. Stenberg, Chem. Phys. Lipids 12 (1974) 287 H.J. Hinz and J.M. Sturtevant, J. Biol. Chem. 247 (1972) 6071 D. Chapman, in: D. Chapman and D.H.F. Wallach (Eds.), Biological Membrane, Vol. II, Academic Press, New York (1973) pp. 91-144 17] S. Simon, Ph.D. Thesis, Northwestern University (1973) [81 W.L. Procarione, J.W. Kauffman, Chem. Phys. Lipids 12 (1974) 251

W.L. Procarione, J. W. Kauffman, Capacitance studies of synthetic phospholipid films [9] [10] [11] [12] [13] [14] [15] [ 16] [ 17] [18] [19] [20] [21 ] [22] [23] [24] [25] [26]

61

K. Blodgett, J. Am. Chem. Soc. 57 (1935) 1007 K. Blodgett and J. Langmuir, Phys. Res, 51 (1937) 964 K.L. Chopra, J. Appl. Phys. 36 (1965) 655 K.L. Chopra, Thin Film Phenomena, MacGraw-Hill, New York (1969) p. 467 M.C. Phillips, E.G. Finer and H. Hauser, Biochim. Biophys. Acta 290 (1972) 347 M. Sundaralingam, Ann. N.Y. Acad. Sci. 195 (1972) 324 Y.K. Levine, A.I. Bailey and M.H.F. Wilkins, Nature (London) 220 (1968) 577 R.M. Williams and D. Chapman, Recent Prog. Chem. Fats, Lipids 11 (1970) 1 R.P. Rand and W.A. Pangborn, Biochim. Biophys. Acta 318 (1973) 299 B. Rosenberg and G. Jendrasiak, Chem. Phys. Lipids 2 (1968) 47 A.M. Gottlieb and P.T. Inglefield, Biochim. Biophys. Acta 307 (1973) 444 E.G. Finer and A. Darke, Chem. Phys. Lipids 12 (1974) 1 F. Kopp, U.P. Fringeli, K. Miihlethaler and Hs.H. Giinthard, Biophys. Struct. Mechanism 1 (1975) 75 W.L. Procarione and J.W. Kauffman, to be published. M.C. Phillips, Prog. Surf. Membr. Sci. 5 (1972) 139 J.L. Lippert and W.L. Peticolas, Proc. Nat. Acad. Sci. 68 (1971) 1572 D.O. Shah, Prog. Surf. Sci. 3 (1972) 221 D.O. Shah and J.H. Shulman, J. Lipid Res. 8 (1967) 227