Assembly and disassembly of Langmuir-Blodgett films on poly[1- (trimethylsilyl)-1-propyne]: The uniqueness of calix[6]arene multilayers as permeation-selective membranes

Article abstract of DOI:10.1021/ja970980w

A series of calix[6]arene-based surfactants have been synthesized, which contain amide oxime head groups on their upper rim and 5,5-dimethylhexyl (I), n-octyl (II), n-dodecyl (III), and n-hexadecyl (IV) groups on their lower rim. Composite membranes that were fabricated from Langmuir-Blodgett (LB) multilayers derived from each surfactant plus poly[1-(trimethylsilyl)-1- propyne] (PTMSP) cast film showed a significant increase in their selectivity toward helium and nitrogen, relative to bare PTMSP. In sharp contrast, analogous composites that were prepared with LB multilayers of conventional single chain surfactants [arachidic acid (AA), cadmium arachidate (AA/Cd²⁺) and stearoamideoxime (SA)] exhibited permeation properties that were identical with those of bare PTMSP. When a polymeric surfactant [poly(1- octadecene-co-maleic anhydride), POM] is used for LB film construction, a modest increase in selectivity was observed. These findings, together with an analysis of representative composites by X-ray photoelectron spectroscopy, provide compelling evidence for the presence of intact, calix[6]arene-based LB assemblies on the surface of PTMSP; with the conventional single chain surfactants, however, disassembly and absorption into the bulk phase of the support is favored. The results of this study highlight the need for having individual surfactant molecules span individual pores on the surface of the support and strong intermolecular forces between neighboring surfactants to produce relatively defect-free LB films.

Full text of DOI:10.1021/ja970980w

J. Am. Chem. Soc. 1997, 119, 6909-6918
6909
Assembly and Disassembly of Langmuir-Blodgett Films on
Poly[1-(trimethylsilyl)-1-propyne]: The Uniqueness of
Calix[6]arene Multilayers as Permeation-Selective Membranes
Robert A. Hendel, Eisaku Nomura, Vaclav Janout, and Steven L. Regen*
Contribution from the Department of Chemistry and Zettlemoyer Center for Surface Studies,
Lehigh UniVersity, Bethlehem, PennsylVania 18015
ReceiVed March 27, 1997X
Abstract: A series of calix[6]arene-based surfactants have been synthesized, which contain amide oxime head groups
on their upper rim and 5,5-dimethylhexyl (I), n-octyl (II), n-dodecyl (III), and n-hexadecyl (IV) groups on their
lower rim. Composite membranes that were fabricated from Langmuir-Blodgett (LB) multilayers derived from
each surfactant plus poly[1-(trimethylsilyl)-1-propyne] (PTMSP) cast film showed a significant increase in their
selectivity toward helium and nitrogen, relative to bare PTMSP. In sharp contrast, analogous composites that were
prepared with LB multilayers of conventional single chain surfactants [arachidic acid (AA), cadmium arachidate
(AA/Cd2+), and stearoamideoxime (SA)] exhibited permeation properties that were identical with those of bare PTMSP.
When a polymeric surfactant [poly(1-octadecene-co-maleic anhydride), POM] was used for LB film construction,
a modest increase in selectivity was observed. These findings, together with an analysis of representative composites
by X-ray photoelectron spectroscopy, provide compelling evidence for the presence of intact, calix[6]arene-based
LB assemblies on the surface of PTMSP; with the conventional single chain surfactants, however, disassembly and
absorption into the bulk phase of the support is favored. The results of this study highlight the need for having
individual surfactant molecules span individual pores on the surface of the support and strong intermolecular forces
between neighboring surfactants to produce relatively defect-free LB films.
Introduction
We have previously shown that Langmuir-Blodgett (LB)
films derived from certain calix[6]arenes exhibit significant
permeation selectivity when supported on poly[1-(trimethyl-
silyl)-1-propyne] (PTMSP).1-3 For example, we have shown
that multilayers of 1, deposited onto PTMSP cast film, are much
more permeable toward helium than nitrogen.1 The fact that
this selectivity is greater than what would be expected based
on Graham’s law has provided compelling evidence that the
pores through which these permeants pass are of molecular
dimensions. Although the precise nature of these pores remains
to be established, our working hypothesis has been that the
“molecular pores” of the calix[6]arenes are of primary impor-
tance. In principle, transient gaps between the calix[6]arenes
(resulting from thermal motion), as well as defects, may also
contribute to the overall pore structure of such assemblies.4 Of
special significance is the fact that these membranes represent
the only known examples of LB films in which permeation
through the assemblies (as opposed to defects) represents the
major diffusional pathway.5-11
Our success in fabricating permeation-selective LB films from
calix[6]arenes was found to be critically dependent upon the
specific support material that was used.2 Thus, when mi-
croporous PTMSP cast films were employed, significant
selectivity was observed; when macroporous polymer supports
were used (i.e., Celgard and Nuclepore membranes), however,
no significant selectivity was detected. Our presumption has
been that the relatively contiguous surface of PTMSP film
minimizes the formation of defects within the LB overlayer.
Specifically, we hypothesized that by minimizing the need for
“bridging the gaps” on the surface of the support, the LB film
is more likely to remain intact. Recent positron annihilation
studies of PTMSP indicate that these cast films contain pores
that are ca. 10 Å in diameter.12 Given the size of the
calix[6]arene framework (a cylindrical conformation maintains
an outer diameter of ca. 14 Å), one would expect that indiVidual
calix[6]arene-based surfactants should be able to span indi-
Vidual pores on the surface of PTMSP. In contrast, Celgard [a
X Abstract published in AdVance ACS Abstracts, July 1, 1997.
(1) Conner, M. D.; Janout, V.; Regen, S. L. J. Am. Chem. Soc. 1993,
115, 1178.
(2) Conner, M. D.; Janout, V.; Kudelka, I.; Dedek, P.; Zhu, J.; Regen,
S. L. Langmuir 1993, 9, 2398.
(3) Dedek, P.; Webber, A. S.; Janout, V.; Hendel, R. A.; Regen, S. L.
Langmuir 1994, 10, 3943.
(4) Stern, S. A. J. Membr. Sci. 1994, 94, 1.
(5) Rose, G. D.; Quinn, J. A. Science 1968, 636.
(6) Rose, G. D.; Quinn, J. A. J. Colloid Interface Sci. 1968, 27, 193.
(7) Albrecht, O.; Laschewsky, A.; Ringsdorf, H. Macromolecules 1984,
17, 937.
(8) Stroeve, P.; Spooner, G. J. R.; Bruinsma, P. J.; Coleman, L. B.;
Erdelen, C.; Ringsdorf, H. Am. Chem. Soc., Symp. Ser. 1990, 447, 177.
(9) Kigashi, N.; Kunitake, T.; Kajiyama, T. Polym. J. 1987, 19, 289.
(10) Stroeve, P.; Coelho, M. A. N.; Dong, S.; Lam, P.; Coleman, L. B.;
Fiske, T. G.; Ringsdorf, H.; Schneider, J. Thin Solid Films 1989, 180, 291.
(11) Stroeve, P.; Bruinsma, P. J. Thin Solid Films 1994, 244, 958.
(12) Yampol’ski, Y. P.; Shantorovich, V. P.; Cherrynyakovski, F. P.;
Kornilov, A. I.; Plate, N. A. J. Appl. Polym. Sci. 1993, 47, 85.
S0002-7863(97)00980-3 CCC: $14.00 © 1997 American Chemical Society

6910 J. Am. Chem. Soc., Vol. 119, No. 29, 1997
Hendel et al.
stretched form of poly(propylene)] would require ca. 6 × 103
calix[6]arenes to cover a rectangular pore that is typically 2000
Å × 500 Å; for Nuclepore membranes that contain 300 Å
diameter cylindrical pores, ca. 400 calix[6]arenes would be
needed.
Although our studies to date have demonstrated the utility
of PTMSP as support material for LB films, the question of
uniqueness of the calix[6]arenes, themselves, has not been firmly
established. In particular, no direct comparisons have yet been
made between calix[6]arene-based LB films and those made
from conventional single chain or polymeric surfactants.
Moreover, the question of whether or not the permeation
properties of calix[6]arene-based LB films can be fine-tuned
by altering their molecular structure and/or composition remains
unanswered. The primary aim of the work that is described
herein was to address both of these issues.13
appeared (at ca. -10 °C), and then cooled to -78 °C. To this cuprate
solution was added 20 mL of a THF solution that contained 9.05 g
(40.0 mmol) of 3-butenyl-1-p-toluenesulfonate over a period of 30 min.
The mixture was stirred for 3 h while the temperature was gradually
raised to -10 °C, and then poured into 100 mL of a saturated solution
of NH4Cl that was cooled to 0 °C and stirred for 10 min. The organic
layer was separated, washed with saturated NH4Cl until appearing
colorless, further washed with saturated NaCl, and then dried over
anhydrous MgSO4. Concentration and purification by distillation
1
afforded 3.05 g (68%) of 5,5-dimethyl-1-hexene having bp 82 °C; H
NMR (CDCl3) δ 0.87 (s, 9 H), 1.23-1.27 (m, 2 H), 1.96-2.03 (m, 2
H), 4.88-5.01 (m, 2 H), 5.76-5.87 (m, 1 H).
To a mixture of 0.95 g (25 mmol) of NaBH4 and 1.32 g (5 mmol)
of 18-crown-6 in 50 mL of anhydrous THF was added 0.77 g (5 mmol)
of TiCl3.15 After the mixture was stirred for 1 h at 25 °C under a
nitrogen atmosphere, a solution made from 2.86 g (25 mmol) of 5,5-
dimethyl-1-hexene and 10 mL of anhydrous THF was added in a
dropwise manner. The mixture was then stirred for 18 h at 25 °C,
followed by sequential addition (dropwise) of 17.5 mL of 25 wt %
CH3ONa in CH3OH (w/w) and 45 mL of 30% H2O2. The mixture
was stirred for an additional 1 h at room temperature and extracted
with CH2Cl2 (3 × 50 mL), and the combined organic extract was
washed with water (3 × 30 mL) and dried (anhydrous MgSO4).
Removal of solvent under reduced pressure afforded 2.85 g (88%) of
Experimental Section
General Methods. Unless stated otherwise, all reagents and
chemicals were obtained from commercial sources and used without
further purification. EMS silica gel 60 was used for column chroma-
trography; preparative thin layer chromatograpy employed EM Science
silica gel 60 F-254. Detection by TLC was made by a combination of
10% sulfuric acid in water, I2, and UV (254 and 365 nm). House-
deionized water was purified with a Millipore Milli-Q-filtering system
containing one carbon and two ion-exchange stages. Chloroform and
methanol, which were used as spreading solvents for monolayer
1
5,5-dimethyl-1-hexanol as a colorless oil having H NMR (CDCl3) δ
0.85 (s, 9 H), 1.14-1.19 (m, 2 H), 1.27-1.29 (m, 2 H), 1.35 (s, 1 H),
1.50-1.54 (m, 2 H), 3.63 (t, 2 H). HRMS for MNH4+(C8H22NO)+:
calcd, 148.1701; found, 148.1691.
To 50 mL of cooled (-10 °C) anhydrous CH2Cl2 was added,
sequentially, 1.21 g (12 mmol) of triethylamine, 1.0 g (7.7 mmol) of
5,5-dimethyl-1-hexanol, and 0.97 g (8.5 mmol) of methanesulfonyl
chloride. The mixture was stirred for 2 h at -10 °C, and then quenched
by addition of 10 mL of ice water. Separation of the organic layer,
followed by washing (10 mL of 1 M HCl, 10 mL of saturated NaHCO3,
and 10 mL of brine), drying (anhydrous MgSO4), concentration under
reduced pressure, and purification by column chromatography (silica,
CH2Cl2), afforded 1.38 g (86%) of 5,5-dimethylhexyl-1-methane-
1
formation, were HPLC grade (Burdick and Jackson, Baxter). All H
NMR spectra were recorded on a Bruker 360-MHz instrument; chemical
shifts are reported in ppm and were referenced to residual solvents.
High resolution mass spectra (fast atom bombardment) were obtained
at the Mass Spectrometry Facility of the University of California,
Riverside. Elemental analyses were obtained from Midwest Microlab
(Indianapolis, IN). 37,38,39,40,41,42-Hexahydroxycalix[6]arene and
37,38,39,40,41,42-hexakis(1-octyloxy)calix[6]arene were prepared by
using procedures similar to those that have been previously described.14
All monolayer measurements, film casting, and membrane fabrications
were performed in a clean room that was maintained under a positive
pressure via a Hepa filtration system. The polymeric surfactant, poly(1-
octadecene-co-maleic anhydride) (POM), was obtained from a com-
mercial source (Polysciences, MW 30 000 to 50 000) and was used as
received. Eicosanoic (arachidic) acid was purchased from Aldrich
Chemical Co.
1
sulfonate as a colorless oil having H NMR (CDCl3) δ 0.85 (s, 9 H),
1.15-1.20 (m, 2 H), 1.32-1.35 (m, 2 H), 1.68-1.72 (m, 2 H), 2.99
(s, 3 H), 4.21 (t, 2 H). HRMS for (C9H21O3S)+: calcd, 209.1211; found,
209.1207.
37,38,39,40,41,42-Hexakis(5,5-dimethyl-1-hexyloxy)-
calix[6]arene (1a). To a solution of 0.50 g (0.785 mmol) of
37,38,39,40,41,42-hexahydroxycalix[6]arene in 20 mL of anhydrous
DMF was added 0.47 g (11.78 mmol) of 60% NaH in mineral oil.
After the mixture was stirred for 30 min at 60 °C under a nitrogen
atmosphere, 1.18 g (5.65 mmol) of 5,5-dimethylhexyl-1-methane-
sulfonate was then added. After additional stirring for 24 h, 30 mL of
water was added, and the product mixture was allowed to cool to room
temperature. The resulting precipitate was filtered and washed with
water and methanol to give a yellow powder. Precipitation from CHCl3/
CH3OH gave 0.50 g (48%) of 37,38,39,40,41,42-hexakis(5,5-dimethyl-
1-hexyloxy)calix[6]arene as a white powder. Recrystallization from
CHCl3/CH3OH (1/1, v/v) afforded 0.26 g of crystalline material
(needles) having mp 235-237 °C; 1H NMR (CDCl3) δ 0.88 (s, 54 H),
1.14-1.35 (br m, 36 H), 3.24 (br s, 12 H), 3.92 (br s, 12 H), 6.70-
6.74 (m, 6 H), 6.90 (br s, 12 H). Anal. Calcd for C90H132O6: C, 82.52;
H, 10.15. Found: C, 82.59; H, 10.30.
5,5-Dimethylhexyl-1-methanesulfonate. A solution of 5.00 g (69.3
mmol) of 3-buten-1-ol in 30 mL of anhydrous pyridine was stirred
and cooled in an ice bath.
p-Toluenesulfonyl chloride (16.5 g, 86.6 mmol) was then added in
small portions, while maintaining the temperature of the mixture at
0-5 °C. The resulting slurry was allowed to stand for 16 h, while
keeping the temperature below 10 °C, and then poured into 100 g of
ice water. Subsequent extraction with CH2Cl2 (4 × 30 mL), washing
of the combined extract with 1 M sulfuric acid (4 × 30 mL), 10 mL
of saturated NaHCO3, and water (2 × 10 mL), drying (anhydrous
MgSO4), and concentration under reduced pressure afforded 14.4 g
1
(92%) of 3-butenyl-1-p-toluenesulfonate as a colorless oil having H
5,11,17,23,29,35-Hexabromo-37,38,39,40,41,42-hexakis(5,5-di-
methyl-1-hexyloxy)calix[6]arene (2a). A solution of 0.50 g (0.381
mmol) of 37,38,39,40,41,42-hexakis(5,5-dimethyl-1-hexyloxy)-
calix[6]arene and 1.02 g (5.72 mmol) of N-bromosuccinimide in 20
mL of 2-butanone was stirred at room temperature with access toward
light for 40 h. To this solution was added 50 mL of a 10% aqueous
solution of NaHSO3 (w/w), and the resulting mixture then stirred for
30 min at room temperature. The organic components were extracted
with CH2Cl2 (3 × 30 mL), and the combined extract was washed with
water (2 × 20 mL) and dried (anhydrous MgSO4). Removal of solvent
NMR (CDCl3) δ 2.34-2.40 (m, 2 H), 2.42 (s, 3 H), 4.04 (t, 2 H),
5.02-5.07 (m, 2 H), 5.61-5.68 (m, 1 H), 7.32 (d, 2 H), 7.76 (d, 2 H).
A dispersion of 8.96 g (100.0 mmol) of CuCN in 140 mL of
anhydrous THF was stirred under a nitrogen atmosphere and cooled to
-78 °C. To this dispersion was added, dropwise, 118 mL of 1.7 M
(CH3)3CLi in pentane over the course of 1 h. The mixture was then
stirred and warmed, gradually, until a homogeneous orange solution
(13) A preliminary account of this work has previously appeared: Lee,
W.; Hendel, R. A.; Dedek, P.; Janout, V.; Regen, S. L. J. Am. Chem. Soc.
1995, 117, 6793.
(14) Markowitz, M. A.; Janout, V.; Castner, D. G.; Regen, S. L. J. Am.
Chem. Soc. 1989, 111, 8192.
(15) Lee, H. S.; Isagawa, K.; Toyoda, H.; Otsuji, Y. Chem. Lett. 1984,
673.

Poly[1-(trimethylsilyl)-1-propyne]
J. Am. Chem. Soc., Vol. 119, No. 29, 1997 6911
under reduced pressure afforded an oily crude product, which was
triturated with hot isopropyl alcohol to give 0.38 g (56%) of
5,11,17,23,29,35-hexabromo-37,38,39,40,41,42-hexakis(5, 5-dimethyl-
1-hexyloxy)calix[6]arene as a colorless powder. Recrystallization from
CHCl3/CH3OH (1/1, v/v) yielded 0.28 g of crystalline product having
direct alkylation of 37,38,39,40,41,42-hexahydroxycalix[6]arene with
1-bromododecane afforded a 69% yield of 1c having mp 76-77 °C;
1H NMR (CDCl3, 60 °C) δ 0.88 (t, 18 H), 1.28-1.41 (m, 120 H), 3.29
(s, 12 H), 3.91 (s, 12 H), 6.72 (t, 6 H), 6.92 (d, 12 H). Anal. Calcd
for C114H180O6: C, 83.16; H, 11.01. Found: C, 81.11; H, 11.02.
5,11,17,23,29,35-Hexabromo-37,38,39,40,41,42-hexakis(1-
dodecyloxy)calix[6]arene (2c). By using procedures similar to those
described for the preparation of 2a, a 58% yield of 2c was obtained
(starting with 1c) having mp 181-183 °C; 1H NMR (CDCl3, 60 °C) δ
0.87 (t, 18 H), 1.26 (s, 108 H), 1.63 (br s, 12 H), 3.64 (br s, 12 H),
3.87 (br s, 12 H), 6.95 (br s, 12 H). Anal. Calcd for C114H174O6Br6:
C, 64.59; H, 8.27; Br, 22.61. Found: C, 64.31; H, 8.23; Br, 22.46.
5,11,17,23,29,35-Hexacyano-37,38,39,40,41,42-hexakis(1-
dodecyloxy)calix[6]arene (3c). By using procedures similar to those
described for the preparation of 3a, a 36% yield of 3c was obtained
1
mp 264-267 °C; H NMR (CDCl3, 60 °C) δ 0.87 (s, 54 H), 1.21-
1.40 (br m, 24 H), 1.69 (br s, 12 H), 3.71-3.75 (m, 12 H), 3.88 (br s,
12 H), 6.88 (br s, 12 H). Anal. Calcd for C90H126O6Br6: C, 60.61; H,
7.12; Br, 26.88. Found: C, 60.77; H, 7.09, Br, 27.14.
5,11,17,23,29,35-Hexacyano-37,38,39,40,41,42-hexakis(5,5-dimethyl-
1-hexyloxy)calix[6]arene (3a). A solution of 0.37 g (0.207 mmol) of
5,11,17,23,29,35-hexabromo-37,38,39,40,41,42-hexakis(5, 5-dimethyl-
1-hexyloxy)calix[6]arene and 1.11 g (12.40 mmol) of CuCN in 15 mL
of anhydrous 1-methyl-2-pyrrolidinone was refluxed with stirring under
a nitrogen atmosphere for 4 h. The temperature of the mixture was
then cooled to 100 °C, and a solution made from 3.35 g (12.4 mmol)
of FeCl3.6H2O plus 10 mL of concentrated HCl and 50 mL of water
was then added. The mixture was stirred at 100 °C for 20 min, followed
by additional stirring at room temperature for 16 h. The resulting
precipitate was filtered and washed, sequentially, with 1 M HCl (20
mL), water (20 mL), and 5 mL of CH3OH to give 0.36 g of crude
product. Purification by column chromatography (silica, hexane/CHCl3/
acetone, 20/20/1, v/v/v) afforded 201 mg (66%) of 5,11,17,23,29,35-
hexacyano-37,38,39,40,41,42-hexakis(5,5-dimethyl-1-hexyl-
oxy)calix[6]arene as a white powder. Recrystallization from CHCl3/
CH3OH (1/1, v/v) gave 36 mg of crystalline product having mp 250-
251 °C; 1H NMR (CDCl3) δ 0.77-0.86 (m, 54 H), 1.05-1.86 (m, 36
H), 3.58-4.22 (m, 24 H), 6.15-7.58 (m, 12 H). Anal. Calcd for
C96H126N6O6: C, 78.97; H, 8.69; N, 5.76. Found: C, 78.92; H, 8.65;
N, 5.75.
1
(starting with 2c) having mp 198-199 °C; H NMR (CDCl3) δ 0.82
(m, 54 H), 1.15-1.85 (m, 120 H), 3.58-4.20 (m, 24 H), 6.12-7.57
(m, 12 H). Anal. Calcd for C120H174O6N6: C, 80.22; H, 9.76; N, 4.68.
Found: C, 80.26; H, 9.80; N, 4.72.
5,11,17,23,29,35-Hexaamidoxime-37,38,39,40,41,42-hexakis(1-
dodecyloxy)calix[6]arene (III). By using procedures similar to those
described for the preparation of II, a 22% yield of III was obtained
(starting with 3c) having mp >233 °C; 1H NMR (DMSO-d6, 60 °C) δ
0.84 (s, 18 H), 1.24 (br s, 108 H), 1.75 (br s, 12 H), 3.49 (t, 6 H), 3.72
(br s, 12 H), 4.48 (d, 6 H), 5.09-5.43 (m, 12 H), 6.88-7.67 (m, 24
H), 9.31 (s, 6 H). Anal. Calcd for C120H192O12N12: C, 72.25; H, 9.70;
N, 8.43. Found: C, 72.36; H, 9.61; N, 6.93.
37,38,39,40,41,42-Hexakis(1-hexadecyloxy)calix[6]arene (1d). By
using procedures similar to those described for the preparation of 1a,
direct alkylation of 37,38,39,40,41,42-hexahydroxycalix[6]arene with
1-bromohexadodecane afforded a 71% yield of 1d having mp 88-89
°C; 1H NMR (CDCl3, 60 °C) δ 0.88 (t, 18 H), 1.27-1.41 (m, 168 H),
3.29 (br s, 12 H), 3.91 (s, 12 H), 6.72 (t, 6 H), 6.92 (d, 12 H). Anal.
Calcd for C138H228O6: C, 83.58; H, 11.58. Found: C, 83.44; H, 11.41.
5,11,17,23,29,35-Hexabromo-37,38,39,40,41,42-hexakis(1-
hexadecyloxy)calix[6]arene (2d). By using procedures similar to those
described for the preparation of 2a, a 43% yield of 2d was obtained
5,11,17,23,29,35-Hexaamidoxime-37,38,39,40,41,42-hexakis(5,5-
dimethyl-1-hexyloxy)calix[6]arene (I). A solution of 0.146 g (0.100
mmol) of 5,11,17,23,29,35-hexacyano-37,38,39,40,41,42-hexakis(5,5-
dimethyl-1-hexyloxy)calix[6]arene, 0.42 g (6.0 mmol) of hydroxylamine
hydrochloride, and 0.24 g (6.0 mmol) of sodium hydroxide in a mixture
of 2 mL of water, 10 mL of tert-butyl alcohol, and 75 mL of ethanol
was stirred at 70 °C for 2 days. After being cooled to room temperature,
the mixture was concentrated under reduced pressure and water (5 mL)
was then added. The resulting mixture was extracted with CHCl3 to
give a turbid solution. Concentration under reduced pressure and
purification by preparative TLC (Rf ) 0.56; silica, CHCl3/CH3OH/H2O,
65/25/4, v/v/v) afforded 43 mg (24%) of 5,11,17,23,29,35-hexaami-
doxime-37,38,39,40,41,42-hexakis(5,5-dimethyl-1-hexyloxy)-
calix[6]arene as a colorless powder having mp >250 °C; 1H NMR
(DMSO-d6) δ 0.73-0.89 (m, 54 H), 1.22-1.31 (m, 12 H), 1.44 (br s,
12 H), 1.84 (br s, 12 H), 3.47 (d, 6 H), 3.78 (br s, 12 H), 4.46 (d, 6 H),
5.36-5.66 (m, 12 H), 6.99-7.80 (m, 12 H), 9.38-9.54 (m, 6 H).
HRMS for (C96H145O12N12)+: calcd, 1658.1105; found, 1658.1093.
5,11,17,23,29,35-Hexabromo-37,38,39,40,41,42-hexakis(1-
octyloxy)calix[6]arene (2b). By using procedures similar to those
described for the preparation of 2a, a 56% yield of 2b was obtained
1
(starting with 1d) having mp 155-156 °C; H NMR (CDCl3, 60 °C)
δ 0.88 (t, 18 H), 1.26-1.42 (m, 156 H), 1.62 (br s, 12 H), 3.63 (br s,
12 H), 3.86 (br s, 12 H), 6.95 (br s, 12 H). Anal. Calcd for
C138H222O6Br6: C, 67.47; H, 9.10; Br, 19.52. Found: C, 67.35; H,
8.90; Br, 19.64.
5,11,17,23,29,35-Hexacyano-37,38,39,40,41,42-hexakis(1-
hexadecyloxy)calix[6]arene (3d). By using procedures similar to those
described for the preparation of 3a, a 51% yield of 3d was obtained
1
(starting with 2d) having mp 162-165 °C; H NMR (CDCl3) δ 0.85
(t, 18 H), 1.15-1.84 (m, 168 H), 3.55-4.20 (m, 24 H), 6.12-7.57 (m,
12 H). Anal. Calcd for C144H222O6N6: C, 81.08; H, 10.48; N, 3.94.
Found: C, 80.89; H, 10.70; N, 3.71.
5,11,17,23,29,35-Hexaamidoxime-37,38,39,40,41,42-hexakis(1-
hexadecyloxy)calix[6]arene (IV). One hundred and fifty milligrams
of a solution (0.42 mmol of nitrile units, 0.07 mmol calix[6]arene) and
0.09 g (2.25 mmol) of sodium hydroxide, in a mixture of 30 mL of
tert-butyl alcohol and 90 mL of ethanol, was stirred at 70 °C for 1
week. After being cooled to room temperature, the solvent was then
removed under reduced pressure. Water (30 mL) was then added to
the mixture, followed by extraction with CHCl3 (3 × 10 mL). The
combined CHCl3 extract appeared as a turbid solution. Removal of
solvent under reduced pressure, followed by recrystallization (four
times) from CHCl3/CH3OH/H2O (100/25/2, v/v/v), afforded 50 mg of
fine white needles having Rf ) 0.6 (silica, CHCl3/CH3OH,H2O, 65/
1
(starting with 1b) having mp 218-219 °C; H NMR (CDCl3, 60 °C)
δ 0.95 (t, 18 H), 1.25-1.60 (m, 72 H), 3.55 (s, 12 H), 3.85 (s, 12 H),
6.95 (s, 12 H). Anal. Calcd for C90H126O6Br6: C, 60.61; H, 7.12; Br,
26.88. Found: C, 60.36; H, 6.95; Br, 27.07.
5,11,17,23,29,35-Hexacyano-37,38,39,40,41,42-hexakis(1-
octyloxy)calix[6]arene (3b). By using procedures similar to those
described for the preparation of 3a, a 18% yield of 3b was obtained
1
(starting with 2b) having mp 213-214 °C; H NMR (CDCl3) δ 0.85
(t, 18 H), 1.15 (m, 72 H), 3.55-4.40 (m, 24 H), 6.15-7.15 (m, 12 H).
Anal. Calcd for C96H126O6N6: C, 78.97; H, 8.70; N, 5.76. Found: N,
5.65.
1
25/3, v/v/v); mp >300 °C; H NMR (DMSO-d6, 80 °C) δ 0.85 (t, 18
5,11,17,23,29,35-Hexaamidoxime-37,38,39,40,41,42-hexakis(1-
octyloxy)calix[6]arene (II). By using procedures similar to those
described for the preparation of I, a 56% yield of II was obtained
H), 1.25 (s, 156 H), 1.72 (br s, 12 H), 3.46 (d, 6 H), 3.70 (br s, 12 H),
4.48 (d, 6 H), 5.05-5.30 (m, 12 H), 7.03-7.45 (m, 24 H), 9.21 (br s,
6 H). Anal. Calcd for C144H240O12N12: C, 74.12; H, 10.30; N, 7.21.
Found: C, 74.19; H, 10.30; N, 6.92.
1
(starting with 3b) having mp 259-260 °C; H NMR (DMSO-d6, 60
°C) δ 0.85 (t, 18 H), 1.28 (br s, 60 H), 1.82 (br s, 12 H), 3.46 (d, 6 H),
3.75 (br s, 12 H), 4.46 (d, 6 H), 5.30-5.77 (br s, 12 H), 6.85-7.81
(m, 12 H), 9.62 (br s, 6 H). HRMS for (C96H145O12N12)+: calcd,
1658.1105; found, 1658.1102.
37,38,39,40,41,42-Hexakis(1-dodecyloxy)calix[6]arene (1c). By
using procedures similar to those described for the preparation of 1a,
Stearoamidoxime (SA). A solution composed of stearonitrile (357
mg, 1.35 mmol), hydroxylamine (468 mg, 6.63 mmol), sodium
hydroxide (269 mg, 6.73 mmol), water (4 mL), tert-butyl alcohol (20
mL), and ethanol (150 mL) was stirred at 70 °C for 48 h. After being
cooled to room temperature, the solution was concentrated under
reduced pressure and extracted with 3 × 10 mL of acetone. Subsequent

6912 J. Am. Chem. Soc., Vol. 119, No. 29, 1997
Hendel et al.
Chart 1
concentration under reduced pressure and purification by preparative
thin layer chromatography [silica, CHCl3/CH3OH (10/1), Rf ) 0.41]
and recrystallization from ethanol afforded 117 mg (29%) of stearoa-
flange size (Kurt J. Lesker Co., Allentown, PA), and five (2.10 in. o.d.
× 1.64 in. i.d. × 0.010 in.) low carbon steel washers (Boker’s, Inc.,
Allentown, PA). The Pyrex glass square, ring seal, and five washers
were cleaned with chloroform, methanol, and acetone, with the aid of
kimwipes. The ring seal was then adhered to the glass square by using
a 5% toluene solution (HPLC grade) of PTMSP, which acted as a
“glue”. The steel washers were placed symmetrically within the ring
seal/glass square casting unit. A PTMSP/toluene casting solution (ca.
180 mg/30 mL) was then poured into the ring seal and covered with a
large piece of filter paper (Whatman qualitative circles, 18.5 cm) in
order to keep the casting unit dust-free. The toluene was allowed to
evaporate for at least 15 h, leaving a PTMSP film across the steel
washers and glass square. A surgical blade (S/P Surgical Blades, Baxter
Diagnostics) was used to cut out the individual washers. Water was
poured into the ring seal to help separate the washers. The PTMSP
cast films were placed between several large pieces of filter paper and
allowed to dry for at least 24 h. The resulting membranes, having a
typical thickness of ca. 15 µm, were placed in antistatic bags for at
least 15 min prior to use in composite membrane fabrication.
Fabrication of Composite Membranes. PTMSP/surfactant com-
posite membranes were prepared by conventional vertical Langmuir-
Blodgett (LB) “dipping” with use of an MGW Lauda film balance.
All transfers were performed in a positive pressure clean room. The
subphase for each experiment was pure water, which was maintained
at 25 °C. When CdCl2 was used, 0.5 mM solutions were prepared
from 99.99+% salt (Aldrich). Spreading solvents (see Surface Pres-
sure-Area Isotherms section for details) were allowed to evaporate
for at least 30 min prior to compression. A nitrogen flow of ca. 175
cm3/min at the gas-subphase interface was maintained throughout each
experiment. The dipping speed that was used for each monolayer
transfer was 4 mm/min. Monolayers of I, II, III, and SA were
transferred at a constant pressure of 40 dyn/cm after a single
compression. Monolayers of POM and AA were transferred with a
constant surface pressure of 30 dyn/cm after a single compression.
Drying of the LB films between bilayers of I, II, and III was carried
out above the air-water interface under a flow of nitrogen for at least
60 min; for bilayers that were made from SA, AA, AA/Cd2+, and POM,
a delay (drying) time of 30 min was used. Transfer ratios of I, II, and
III onto PTMSP cast films were typically 1.0 and 1.2 for down- and
up-trips, respectively. Transfer ratios for SA onto PTMSP cast films
were typically 0.9 and 1.1 for down- and up-trips; for POM, the typical
transfer ratios were 1.0 and 1.2 for the down- and up-trips. Transfer
ratios for AA on PTMSP were 1.2 and 0.7 for down- and up-trips,
respectively; for AA that was spread over a 0.5 mM CdCl2 solution,
transfer ratios were typically 1.2 and 1.2 for down- and up-trips,
respectively. Monolayers of IV were transferred at a constant surface
pressure of 30 dyn/cm after three compressions and two expansions.
Drying between bilayers of IV was carried out under ambient conditions
(23 °C, 22 h, away from the air/water interface). Transfer ratios of IV
1
midoxime as a white solid having mp 75-77 °C; H NMR (acetone-
d6, 40 °C) δ 0.88 (t, 3 H), 1.30 (m, 30 H), 1.55 (m, 2 H), 2.02 (t, 2 H),
4.87 (br s, 2 H), 7.80 (br s, 1 H). HRMS for (C18H38ON2)+: calcd,
299.3062; found, 299.3067.
Surface Pressure-Area Isotherms. Surface pressure-area iso-
therms were recorded by use of an MGW Lauda film balance, which
was equipped with a computerized data acquisition station. All
isotherms were measured at 25 °C. Water (ca. 1 L), which was used
as a subphase, was purified Via a Milli-Q filtration system and purged
with nitrogen for 15 min. Before addition to the film balance, the
surface of this degassed water was removed Via aspiration in order to
remove surface-active contaminants. All surfactant solutions were
spread onto the aqueous subphase having a surface area of 600 cm2,
using a gas-tight, 50 µL Hamilton syringe. Solutions of I, II, and III
were made by using a mixed solvent [CHCl3/CH3OH, 8/3 (v/v)]; typical
concentrations that were used were ca. 1.2 mg/mL. A spreading
solution of SA (1.2 mg/mL) was made by using CHCl3/CH3OH (2/3,
v/v) as solvent; with IV, a solution (0.17 mg/mL) made from CHCl3/
CH3OH (5/1, v/v) was used. POM and AA were spread as a
chloroform solution, using a concentration of 1.0 mg/mL. Actual
concentrations were determined by direct weighing of aliquots after
evaporation of solvent with a Cahn 27 electrobalance. In all cases,
spreading solvents were allowed to evaporate for at least 30 min prior
to compression under a flow of nitrogen. Limiting areas were
determined by extrapolating from the condensed region to zero surface
pressure.
Surface Viscosity Measurements. For surface viscosity experi-
ments, a home-built canal viscometer (192 mm × 40 mm solid Teflon
block, having a centrally located 6.5 mm slit) was placed in front of
the compressing barrier and the monolayer was compressed at a rate
of 25 cm2/min. When the surface pressure began to rise, the
compression speed was decreased to 13 cm2/min to allow for more
precise control over the surface pressure. Compressions were stopped
when a target pressure of 20 dyn/cm was reached. The resulting
monolayer was then allowed to equilibrate at this pressure for 3 h.
After this period of time, the moving barrier was expanded at the
maximum speed of 120 cm2/min, leaving the canal viscometer at its
original position. The resulting surface pressure was then recorded as
a function of time. The rate of surface pressure decrease was taken as
a measure of the surface viscosity of the monolayer.
Supports. Supports that were made from PTMSP were prepared
by using a casting technique. All of the PTMSP that was used in this
work was generously supplied to us by Air Products and Chemicals,
Inc. (Allentown, PA), having MW ) 8 × 105 (intrinsic viscosity
method). A typical casting apparatus consisted of a Pyrex glass square
(8 in. × 8 in × 1/8 in.), an aluminum centering ring seal, 160 ISO

Poly[1-(trimethylsilyl)-1-propyne]
J. Am. Chem. Soc., Vol. 119, No. 29, 1997 6913
onto PTMSP were typically 0.95 and 1.2 for initial down- and up-
trips, and 1.0 and 1.2 for subsequent down- and up-trips. All composite
membranes were carefully placed in tin specimen boxes (Fisher
Scientific) which were lined with Whatman 5.5 cm qualitative filter
papers (LB film side up). Each composite membrane was allowed to
remain in the laboratory ambient (clean room) for a minimum of 12 h
before gas permeation measurements were carried out. For prolonged
storage, the PTMSP/LB composite membranes were placed in a
desiccator that contained anhydrous CaSO4, which was maintained
under an argon atmosphere.
were measured to obtain a clear SA spectrum after subtracting the bare
PTMSP spectrum as a “background”.
Results
Surfactant Design and Synthesis. Specific calix[6]arenes
that were chosen as synthetic targets for the present study were
I, II, III, and IV. The amide oxime moiety was viewed as an
attractive head group based on its strong hydrophilicity and its
potential for monolayer stabilization via intermolecular hydrogen
bonding. Calix[6]arenes I and II, which are structural isomers,
differ only in the bulkiness of their alkyl chains and their
potential for closing off their “molecular pores”. Examination
of I by CPK space-filling models indicates that the bulky tert-
butyl groups should prevent closure of its molecular pore, when
held in a cylindrical conformation such that all of the amide
oxime groups define one face of the molecule; with II, however,
closure of these pores by having the alkyl chains fold into the
void of the calix[6]arene is a distinct possibility. The key issue,
here, is whether attractive van der Waals forces among the alkyl
chains are maximized through intramolecular or intermolecular
interactions in the monolayer state. If intramolecular forces
were dominant, then pore closure should be favored; if
intermolecular forces were more important, however, then the
pores should remain open. Since van der Waals forces decrease
with hydrocarbon branching, one might expect to observe a more
cohesive monolayer with II if intermolecular interactions were
dominant. In order to probe this question further, calix[6]arenes
III and IV were chosen as additional synthetic targets. Here,
elongation of the alkyl chains would be expected to enhance
the cohesiveness of a monolayer if intermolecular forces were
of major importance.
To place the barrier properties of these calix[6]arene as-
semblies into perspective, the permeation properties of analogous
composites that were prepared from conventional single chain
surfactants [i.e., multilayers of arachidic acid (AA), cadmium
arachidate (AA/Cd2+), and a single chain surfactant bearing an
amide oxime head group (stearoamideoxime, SA)] and a
polymeric surfactant [poly(1-octadecene-co-maleic anhydride)
(POM)] were defined. In principle, these single chain surfac-
tants have the potential for “slithering” through the micropores
of PTMSP and for entering its bulk phase. Such disassembly
should then result in defect formation and a reduction in
permeation selectivity. The polymeric surfactant (POM) was
of interest since individual alkyl chains have the potential for
penetrating the micropores of PTMSP, but the entire polymer
molecule should remain at its surface. How such penetration
would affect the permeation properties of the LB assembly was
an additional question that we sought to answer.
Gas Permeation Measurements. Gas permeation measurements
were made with a home-built stainless steel permeation apparatus
(Figure 4). The gases studied were He (Airco, grade 4.7) and N2 (JWS
Technologies, prepurified grade). A membrane to be measured was
placed in the permeation cell between two Viton rubber O-rings (3.45
cm i.d., Scientific Instrument Services, Inc.) with a support screen (4.70
cm, Millipore Corp.) and held securely with a quick flange clamp
(Scientific Instrument Services, Inc., not shown). Polymer/surfactant
composites were always placed in the cell such that the LB film faced
the high pressure side of the pressure gradient. The permeant gas
travelled from the gas cylinder to an inline filter (15 µm) and elastomer
diaphragm regulator (Brooks Instrument, 8601D), which was connected
to a Heise gauge port. Another plastic tubing line connected from the
Heise gauge outlet to the permeation cell. The pressure gradient that
was applied to each membrane was 0.7 atm (10 psig). After passing
through the membrane, the gaseous permeant was directed into a 40
cm long glass U-tube flowmeter (2 mm i.d.). The volumetric flow
rate of the gas was then measured by recording the time (tf - ti) that
was required for a methyl isobutyl ketone solution containing crystal
violet to travel a set distance (df - di), thereby sweeping out a defined
volume. Measurements were taken until steady-state values were
achieved (typically 1 to 2 h). At least ten volumetric flow rates were
recorded for each membrane, and the mean and standard deviations
were determined. The normalized flux was calculated with use of the
mean volumetric flow rate, the area that was available for flow (9.36
cm2), and the pressure gradient (0.7 atm) that was employed. This
procedure was repeated for the next permeant gas. Selectivity ratios
were simply the ratio of the normalized fluxes for He and N2 (see eqs
1 and 2). In general, the permeation properties were first measured
for He and then for N2.
X-ray Photoelectron Spectroscopy. Surface analyses of bare
PTMSP, PTMSP/I, and PTMSP/SA composites were performed with
a Scienta ESCA-300 spectrometer (Scienta Instruments AB, Uppsala,
Sweden). Composite membranes consisted of four LB monolayers that
were deposited onto ca. 20 µm thick PTMSP films, which were coated
onto silicon wafers (ca. 2 cm2). Transfer ratios were typically 1.0 (
0.2 for each down- and up-trip. The ESCA-300 spectrometer was
equipped with an X-ray source, a monochromator, an imaging lens, a
hemispherical analyzer, and a multichannel detector. The X-ray source
was a rotating titanium alloy anode. An electron beam directed on the
aluminum band of the anode produced Al KR radiation, which was
monochromized with seven toroidally bent R-quartz crystals. The
X-rays were focused on the composite membrane surface at 45° from
the analyzer. Photoelectrons from the composite were collected by a
multielement lens and focused on a slit aperture pair at the entrance
plane of a hemispherical (radius ) 30 cm) electron energy analyzer
(HMA). The multichannel detector, placed at the exit of the HMA,
consisted of two microchannel plates, a P10 phosphor, and a CCD TV
camera. Analysis was carried out at a pressure range of 10-9 to 10-10
Torr; degassing typically required ca. 30 min. A take-off angle of 15°
was used for each experiment. The XPS analysis focused on C, Si, N,
and O that were present on the surface of the composite.
The synthetic approaches that were used to prepare
calix[6]arenes I, II, III, and IV are outlined in Scheme 1. In
brief, alkylation of 37,38,39,40,41,42-hexahydroxycalix[6]arene
with a suitable alkylating agent, followed by bromination with
N-bromosuccinimide, cyanation, and addition of hydroxylamine,
afforded the desired amphiphiles. A branched mesylate, which
was required for the synthesis of I, was prepared by sequential
tosylation of 3-buten-1-ol, coupling with tert-butyl cuprate,
hydroxylation, and mesylation (Scheme 2). The single-chain
surfactant, SA, was prepared by direct addition of hydroxyl-
amine to stearonitrile.
Monolayer Properties. Surface pressure-area isotherms
that were recorded for I, II, III, and IV over a pure water
subphase at 25 °C are shown in Figure 1; corresponding
isotherms for SA, AA, AA/Cd2+, and POM are presented in
Figure 2. Limiting areas that were determined by extrapolation
from the condensed region of each isotherm to zero surface
Transmission FTIR Spectroscopy. Analyses of bare PTMSP cast
film and PTMSP/SA composites were made with a Mattson Polaris
FTIR spectrometer, equipped with WinFirst Fourier software tools for
Microsoft systems. A freshly prepared PTMSP cast film was first
analyzed (1200 scans) under a nitrogen atmosphere. Four monolayers
of SA were then deposited onto this PTMSP surface with the LB vertical
dipping method; transfer ratios were 1.0 ( 0.2. After allowing the
composite to air-dry for 24 h in the clean room, additional FTIR spectra
(1200 scans) were recorded. Several different regions of the composite

6914 J. Am. Chem. Soc., Vol. 119, No. 29, 1997
Hendel et al.
Scheme 1
Figure 2. Surface pressure-area isotherms for AA, SA, and POM
over a pure water subphase at 25 °C; the isotherm observed for AA
spread over 0.5 mM CdCl2 is also shown.
Figure 1. Surface pressure-area isotherms for I, II, III, and IV over
a pure water subphase at 25 °C. The isotherm that is shown for IV
represents the third compression. For I, II, and III, no difference
between the first and third compression was observed. Inset: Isotherm
for IV, which shows hysteresis between the first, second, and third
compressions.
and POM were also examined. Qualitatively, monolayers that
were formed from I, SA, and POM showed relatively low
viscosity; in each case, there was a precipitious drop in surface
pressure when the monolayer was exposed to a 6.5 mm slit
(Figure 3). In contrast, monolayers that were assembled from
II and III were considerably more viscous. Extraordinarily high
surface viscosity was found in the case of IV, where only a
small pressure drop (<0.5 dyn/cm) was recorded after 2 h.
Fabrication and Permeation Properties of Composite
Membranes. By using standard LB vertical dipping methods,
composite membranes were fabricated with even numbers of
monolayers of surfactant and 15 µm thick cast films of PTMSP
as support material. The specific protocols that were followed
are described in the Experimental Section. In all cases, efficient
transfers were observed during down-trips (into water) and up-
trips (into air) as judged by the transfer ratios. Specifically,
the decrease in monolayer area at the air-water interface divided
by the geometerical surface area of the substrate that passed,
vertically, through the interface was generally 1.0 on the down-
trips and 1.2 for the up-trips.
Scheme 2
pressure were 190, 175, 175, and 162 Å2/molecule for I, II,
III, and IV, respectively; the limiting areas for SA, AA, and
AA/Cd2+ were 19, 20 and 20 Å2/molecule, respectively. A
repeat unit limiting area for POM was estimated to be 36 Å2.
With the exception of IV, all of these surfactants produced
monolayers with no detectable hysteresis. In the case of IV,
stable monolayers were formed but significant hysteresis was
clearly evident (Figure 1). After two consecutive compressions
and expansions, however, this hysteresis was eliminated.
To judge the cohesiveness of these calix[6]arenes in the
monolayer state, we examined their relative surface viscosities
at the air-water interface by use of a canal viscometer (see
Experimental Section). For purposes of comparison, the relative
surface viscosities of monolayers that were prepared from SA
The permeation properties of each composite, with respect
to He and N2, were then determined by use of a home-built
apparatus (Figure 4). A detailed description of this apparatus,
as well as the protocols for obtaining volumetric flow rates (F),
are given in the Experimental Section. Normalized fluxes

Poly[1-(trimethylsilyl)-1-propyne]
J. Am. Chem. Soc., Vol. 119, No. 29, 1997 6915
Table 1. Flux of He and N2 through Conventional Surfactant/
PTMSP Compositesa
106P/l (cm3/cm2‚s‚cm Hg)
He/N2 (R)
surfactant monolayers
He
N2
(composite)
none
0
0
4
4
4
4
2
4
4
2
4
6
8
530
474
519
537
532
528
521
541
499
452
374
288
264
218
579
530
592
592
603
586
585
592
571
197
0.91
0.90
0.88
0.91
0.88
0.90
0.89
0.91
0.88
2.3
5.2
7.2
10.6
12.6
AA
AA/Cd2+
SA
POM
71.1
40.0
24.8
17.3
10
Table 2. Flux of He and N2 through Calix[6]arene/PTMSP
Compositesa
Figure 3. Surface pressure of monolayers formed from I (4), II (O),
III (1), and IV (2) at 25 °C on pure water as a function of time of
exposure to a 6.5 mm slit opening (canal viscometer); the initial surface
pressure was 20 dyn/cm. The inset shows the results of similar
experiments for SA (9) and POM (3).
106P/l (cm3/cm2‚s‚cm Hg)
He/N2 (R)
(composite)
surfactant monolayers
He
N2
none
0
0
2
2
4
4
6
8
10
2
530
474
469
352
317
226
264
171
79
350
269
176
167
158b
586c
289
211
154
136
65.3
56
579
530
113
58.3
12.4
6.9
5.8
3.7
1.3
0.91
0.90
4.1
6.0
26
33
46
46
61
18
15
I
II
20.0
17.7
4.4
2
4
40
46
4
3.6
4b
3.04b
51.9b
0.9c
13
4c
2
2
4
4
6
8
650c
III
22.0
17.9
3.0
2.5
0.77
0.76
0.56
11.3
7.97
0.46
<0.29
12
51
56
85
74
73
Figure 4. Schematic representation of the home-built permeation
apparatus (not drawn to scale).
10
40.6
235
202
78.3
49.6
IV
2
2
4
4
20.7
25.5
170
>171
(P/l) were then calculated according to eq 1, where P represents
the permeability coefficient that characterizes the composite
membrane-permeant combination, l is the thickness of the
composite membrane, A is the membrane’s cross-sectional area,
and ∆p is the pressure gradient that was employed.16 Specific
normalized flux values that were measured for bare PTMSP,
and for each of the composite membranes that were fabricated
in this study, are listed in Tables 1 and 2.
a Normalized flux (P/l) equals the observed flow rate, divided by
the area of the membrane (9.36 cm2) and by the pressure gradient (0.7
atm) employed. Each set of data corresponds to a separate composite
membrane made by using an ca. 15 µm thick PTMSP support plus
multiple layers of amphiphile. All measurements were made at ambient
temperatures. Values were obtained from 5-10 independent measure-
ments; the error in each case was (5%. b Prior to gentle wiping of the
surface with a clean tip swab. c After gentle wiping of the surface with
a clean tip swab.
P
l
F
)
(1)
A(∆p)
that were fabricated with larger numbers of monolayers,
however, showed an increase in permeation selectivity.
Deposition of up to four monolayers of AA, AA/Cd2+, and
SA onto PTMSP did not significantly alter the barrier properties
of the support; in all cases, the He and N2 normalized flux values
remained essentially unchanged. In contrast, deposition of two
layers of POM resulted in a moderate decrease in the normalized
flux of He and a greater decrease in the normalized flux or N2.
If we define permeation selectivity, R, as the normalized flux
of He divided by the normalized flux of N2 (eq 2), then a bilayer
composite (2 monolayers of POM on PTMSP) showed a
selectivity that was similar to what would be predicted based
on Graham’s law [i.e., R(He/N2) ) 2.6]. Analogous membranes
(P/l)He
R )
(2)
(P/l)N2
In sharp contrast, those composite membranes that were
fabricated from each of the calix[6]arene-based LB films showed
permeation selectivities that exceeded Graham’s law, even when
only a single bilayer of surfactant was used (Table 2).
Analogous composite membranes that were fabricated with
larger numbers of monolayers generally showed further increases
in permeation selectivity. The most notable membranes were
(16) Henis, J. M. S.; Tripodi, M. K. J. Membrane Sci. 1981, 8, 223.

6916 J. Am. Chem. Soc., Vol. 119, No. 29, 1997
Hendel et al.
Figure 5. X-ray photoelectron spectrum of PTMSP/II (four mono-
layers).
Figure 7. FTIR spectrum of (A) PTMSP and (B) PTMSP/SA (four
monolayers) after background subtraction of PTMSP.
ically, the resulting barrier properties were almost identical with
that of bare PTMSP (Table 2).
Discussion
Monolayer Properties. The surface pressure-area isotherm
behavior of SA, AA, AA/Cd2+, and POM was as expected.
Arachidic acid showed its characteristic liquid condensed to
ordered-like solid state transition at ca. 28 dyn/cm.17 When
compressed over a cadmium subphase, the absence of the liquid
phase was clearly evident.18,19 Monolayers that were formed
from SA appeared to be somewhat more compressible than that
of AA; in this case, no sharp phase transition was apparent.
The limiting areas that were estimated for SA, AA, and AA/
Cd2+ are consistent with surfactant assemblies that contain a
near-maximum in hydrocarbon density. In the case of POM,
the maleic anhydride groups produce a more loosely packed
assembly, as reflected by a limiting area of 36 Å2 per repeat
unit.
Examination of the surface pressure-area isotherm behavior
for I, II, III, and IV reveals similar compressibility but slightly
different limiting areas. If one assumes that all of the amide
oxime groups are in contact with the water surface, that the
calix[6]arene framework adopts a near-cylindrical conformation,
and that all of the alkyl chains are extended into air, then the
limiting areas that would be expected for II, III, and IV would
be ca. 160 Å2/molecule (CPK models). Although this value is
in excellent agreement with the limiting area that has been
Figure 6. X-ray photoelectron spectrum of PTMSP/SA (four mono-
layers).
those that were made with four monolayers of IV, where the
He/N2 selectivity approached 200!
Surface Analysis by X-ray Photoelectron Spectroscopy.
To confirm the presence of II on the surface of PTMSP after
LB deposition, a sample was analyzed by X-ray photoelectron
spectroscopy (XPS). As can be seen in Figure 5, the clear
presence of C, N, and O on the surface of the composite and
the absence of Si provides strong evidence for an intact
calix[6]arene film. Similar analysis of PTMSP/SA, however,
showed no eVidence of the surfactant on the polymer’s surface;
only the carbon and silicon atoms of the PTMSP were apparent
(Figure 6). That SA was, in fact, transferred to the PTMSP
support was confirmed by transmission FTIR spectroscopy.
Specifically, the C-H bands of the SA at 2926 and 2855 cm-1
were observed after background subtraction of the PTMSP
(Figure 7).
Robustness of Calix[6]arene/PTMSP Composites. To
judge the robustness of a calix[6]arene/PTMSP composite, a
typical four-monolayered membrane was fabricated from II plus
PTMSP and allowed to remain in a desiccator under an argon
atmosphere for 4 months. Analysis of its permeation properties
before and after this storage period showed no significant
differences. However, gentle wiping of the surface with a clean
tip swab substantially altered its permeation properties. Specif-
(17) Gabrielli, G.; Gvarni, G. G. T.; Ferroni, E. J. Colloid Interface.
Sci. 1976, 54, 424.
(18) Chollet, P. A. Thin Solid Films 1978, 52, 343.
(19) Rabolt, J. F.; Burns, F. C.; Schlotter, W. E. J. Chem. Phys. 1983,
78, 946.

Poly[1-(trimethylsilyl)-1-propyne]
J. Am. Chem. Soc., Vol. 119, No. 29, 1997 6917
observed for calix[6]arene IV, the limiting areas for II and III
appear to be slightly larger. Repetitive examination of each of
these surfactants (including different preparations) suggests that
this difference is, in fact, real. One possible explanation for a
lower limiting area for IV would be that stronger intermolecular
hydrophobic interactions between the alkyl chains induce a more
cylindrical conformation of the calix[6]arene; with II and III,
a partial splay of the calix[6]arene framework could lead to
larger areas. Nonetheless, the general similarity of the limiting
areas of II, III, and IV, together with the significant differences
in their alkyl chain lengths, provide a compelling “classic
Langmuir” argument that these chains are extended into air.20
The larger limiting area that has been observed for I is fully
consistent with a predicted value of 190 Å2/molecule, based on
CPK models. In this case, the bulky tert-butyl groups increase
the overall cross-sectional area of the molecule.
evidence for disassembly of the LB film and absorption of the
surfactant into the bulk polymer phase.
Assembly of Langmuir-Blodgett Films On PTMSP.
Permeation and spectroscopic results that have been obtained
with II indicate that this surfactant is capable of forming intact
LB films on the surface of PTMSP. Thus, deposition of only
a single bilayer of II onto PTMSP affords a He/N2 permeation
selectivity that is ca. 5 times greater than that predicted by
Graham’s law. The fact that this selectivity increases with
additional bilayers of the calix[6]arene further indicates that
some “sealing” within the LB film has taken place. The
observation that simple wiping of such a surface with a clean
tip swab converts its permeation characteristics to those of bare
PTMSP provides further evidence that the assembly resides at
the surface of the support. Finally, the XPS results that have
been obtained for a four-monolayer composite of II plus PTMSP
provides direct evidence for an intact calix[6]arene film on the
surface of PTMSP.
The most notable feature among these calix[6]arene-based
surfactants, with respect to their suface pressure-area isotherm
behavior, is the hysteresis that has been observed with IV. The
simplest explanation for such hysteresis, we believe, is that
strong intermolecular hydrophobic interactions lead to partially
separated “islands” of surfactant during the first compression.
As the monolayer is then expanded and recompressed, the
monolayer is mechanically forced into a more densely-packed
state. Although partial retention and subsequent release of the
spreading solvent could also account for this behavior, the
unusually high surface viscosity that is associated with IV (Vide
ante) strongly suggests that this island model is more probable.
Very similar permeability results have been obtained for LB
films that were prepared from I, III, and IV. In the case of I,
however, the relatively low selectivity that is associated with a
single bilayer composite may be accounted for by a reduced
cohesiveness within the assembly and a greater tendency for
defect formation (Vide ante). Deposition of additional mono-
layers of I does, however, help to “patch up” these defects.
Given the strong cohesiveness of IV in the monolayer state, as
judged by surface viscosity, the high permeation selectivity that
has been found with only four monolayers is not so surprising.
Presumably, this strong cohesiveness translates into a reduced
tendency for defect formation and a higher percentage of
permeation through the bulk of the LB film. Composite
membranes that were fabricated from multilayers of the
polymeric surfactant, POM, showed significantly lower per-
meation selectivities relative to those that were made from the
calix[6]arenes. However, these selectivities generally exceeded
that which is predicted by Graham’s law.
Surface viscosities that have been observed for monolayers
of II, III, and IV support a model in which intermolecular
hydrophobic interactions between neighboring alkyl chains
dominate over intramolecular interactions. The relatively high
viscosity that is associated with monolayers of IV clearly reflects
a more cohesive assembly and stronger intermolecular forces.
Since the only difference among II, III, and IV is the length of
their alkyl chains, and since IV contains the longest chains, we
conclude that intermolecular interactions between neighboring
alkyl chains must be very strong. The fact that monolayers of
II and III exhibit significantly greater cohesiveness at the air/
water interface, compared with those of I, can also be accounted
for in terms of stronger hydrophobic interactions; i.e., branching
of the alkyl chains would be expected to reduce the magnitude
of the van der Waals attractions. If intermolecular hydrogen
bonding between neighboring amide oxime groups also con-
tributes to the cohesiveness of monolayers of II, III, and IV,
then the relatively low surface viscosity of I may also be due,
in part, to a decrease in such bonding as a result of the greater
splay of the molecule.
On the Uniqueness and Tunability of Calix[6]arene-Based
LB Films as Permeation-Selective Membranes. Early studies
from our laboratories have established the uniqueness of PTMSP
as support material for calix[6]arene-based LB films in creating
permeation-selective membranes. The present findings show
that these calix[6]arene-based amphiphiles, when used in
combination with PTMSP, are also unique relative to conven-
tional single chain surfactants. In contrast to single chain
surfactants, which are able to “slither” through the pores of
PTMSP and enter the bulk phase, the calix[6]arenes are
sufficiently large such that similar disassembly is prevented. It
would appear, therefore, that a key factor in creating permeation-
selective LB films is to use indiVidual surfactants that can span
indiVidual holes on the surface of the support. Thus, a strategy
that relies upon strong intermolecular forces between large
numbers of surfactants to “bridge the gaps” of the support is
unlikely to be successful. The ability of POM to produce
permeation-selective LB films on PTMSP is consistent with this
view. Here, a single POM molecule that can cover a micropore
of PTMSP is capable of producing selectivities that exceed
Graham’s law. Whether or not other polymeric surfactants can
achieve the level of permeation selectivity that is associated
with these calix[6]arenes remains to be determined.
Disassembly of Langmuir-Blodgett Films on PTMSP.
The permeation properties that have been observed for com-
2+, and SA plus PTMSP are
posites derived from AA, AA/Cd
in sharp contrast to those that have been made from I, II, III,
IV, and POM. On the basis of these permeability data alone,
it would appear as if the single chain surfactants were not even
present on the surface of the support. Analysis of a PTMSP/
SA (four monolayers) composite by X-ray photoelectron
spectroscopy, in fact, provides clear and direct evidence for the
absence of SA on the surface of PTMSP. The fact that very
good Y-type transfers of SA were observed for the down- and
up-trips, however, implies that transient LB films were formed
on the surface of the support. The appearance of SA within
the film, as seen by transmission FTIR spectroscopy, together
with these permeation and XPS results, constitutes strong
Comparison of the calix[6]arenes that have been investigated
in the present study indicates that modest changes in molecular
structure can have a significant influence on the permeation
properties of corresponding LB films. Most importantly, the
need for strong intermolecular interactions within the assembly
to minimize defect formation appears to be of primary impor-
(20) Langmuir, I. J. Am. Chem. Soc. 1917, 39, 1848.

6918 J. Am. Chem. Soc., Vol. 119, No. 29, 1997
Hendel et al.
tance. Thus, the strong cohesiveness of monolayers of IV can
be directly correlated with its high He/N2 selectivity in the form
of an LB film. Judging from its cohesiveness, it would also
appear that closure of the calix[6]arene pores through intramo-
lecular hydrophobic interactions of the alkyl chains is unlikely
and that intermolecular forces are dominant. Finally, it should
be noted that defining the precise pore structure within these
calix[6]arene-based LB films remains as one of the most
challenging aspects that lies ahead. The principal uncertainties
that await clarification are (i) the registry of the calix[6]arene
pores between the layers (i.e., whether or not they align
themselves to produce contiguous channels), (ii) the contribu-
tions that transient gaps between the calix[6]arenes make to the
overall pore structure, and (iii) the presence or absence of
residual defects.
impossible to create relatively defect-free LB films that rely
upon intermolecular association to bridge gaps on the surface
of a support. Our presumption is that most of the failures that
have been reported in the literature in forming defect-free LB
films, in combination with macroporous supports, are due to
this bridging problem.
Strong Attractive Forces between Nearest Neighbors
Helps To Minimize Defect Formation within Polymer-
Supported LB Films. Comparison of the permeation properties
of the series of calix[6]arene-based LB films reported herein
has shown that modest changes in molecular structure can have
a significant influence on the permeation properties of corre-
sponding LB films. In particular, our results have shown that
by strengthening the attractive van der Waals forces between
the calix[6]arenes through elongation of the alkyl chains, defect
formation within these LB assemblies can be minimized.
Conclusions
Calix[6]arene-based LB composites, of the type reported
herein, represent the thinnest membranes that have been
fabricated to date for gas separations. A single bilayer of II,
for example, represents a maximum film thickness of only 38
Å. The fact that the flux of a permeant across a membrane is
inversely proportional to the membrane’s thickness strongly
suggests that LB films such as these should have considerable
potential as high performance membranes that are capable of
extraordinarily high throughputs and high permeation selectivi-
ties. Efforts aimed at exploring such possibilities are continuing
in our laboratories.
Bridging the Gaps of a Support with Surfactant Mono-
layers Is Unlikely To Produce Defect-Free LB Films Unless
Individual Surfactants Can Span Individual Pores. The
results of this study have shown that LB films derived from
calix[6]arene-based surfactants as well as poly(1-octadecene-
co-maleic anhydride) (POM), when supported on poly[1-
(trimethylsilyl)-1-propyne] (PTMSP) cast film, possess perme-
ation selectivities that are not dominated by defects. In contrast,
those LB films that were assembled from conventional single-
chain surfactants showed no significant permeation selectivity
and appeared to undergo spontaneous disassembly. These
striking differences have been found to correlate with the
surfactant’s ability to remain at the surface of PTMSP; i.e., only
those surfactants that can span individual pores of PTMSP
remain on its surface and lead to permeation-selective mem-
branes. We conclude, therefore, that it will be difficult if not
Acknowledgment. We are grateful to the Divison of Basic
Energy Sciences and the Department of Energy (DE-FG02-
85ER-13403) for support of this research.
JA970980W