Non-covalent interactions in host-guest complexes with fluorinated phenyl compounds

Article abstract of DOI:10.1002/(SICI)1099-1395(199705)10:5<305::AID-POC874>3.0.CO;2-5

Complexation constants with the macrocyclic azoniacyclophane CP44 and phenyl guest compounds with at least four fluorine atoms or alternatively protons at the ring were obtained by NMR shift titrations in water. The fluorinated compounds show free energie

Full text of DOI:10.1002/(SICI)1099-1395(199705)10:5<305::AID-POC874>3.0.CO;2-5

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 305–310 (1997)
NON-COVALENT INTERACTIONS IN HOST–GUEST COMPLEXES WITH
FLUORINATED PHENYL COMPOUNDS†
1
1
2
2
¨
¨
XIAO FEI, YONG-ZHENG HUI, VOLKER RUDIGER AND HANS-JORG SCHNEIDER *
1Shanghai Institute of Organic Chemistry, Academia Sinica, Shanghai 200032, China
2FR Organische Chemie, Universita¨t des Saarlandes, D 66041 Saarbru¨cken, Germany
Complexation constants with the macrocyclic azoniacyclophane CP44 and phenyl guest compounds with at least four
fluorine atoms or alternatively protons at the ring were obtained by NMR shift titrations in water. The fluorinated
compounds show free energies of complexation which are smaller by ⌬⌬G=3·4–7·7 kJ molϪ1 in comparison with the
protonated compounds. The NMR shifts induced upon 100% complexation (CIS values) were obtained simultaneously
from non-linear least-squares fitting and indicate intra-cavity inclusion in all cases. The CIS values agree roughly with
screening constants calculated from aromatic ring current and linear electric field effects, the latter resulting from the
permanent charges at the host compound. Molecular mechanics calculations (CHARMm) indicate that intracavity
inclusion is possible with all compounds with negligible strain induced (<1 kJ molϪ1) in the macrocycle upon
complexation. In contrast, ␣-cyclodextrin can accommodate fluorinated phenyl compounds only at the rim of the cavity
without larger strain. Preliminary data with ␣-cyclodextrin, obtained by competitive UV–visible titration with methyl
orange, indicate again a smaller association free energy (⌬⌬G=1·-7 kJ molϪ1) for pentafluorphenol compared with
normal phenol as guest. © 1997 by John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 10, 305–310 (1997) No. of Figures: 3 No. of Tables: 3 No. of References: 19
Keywords: non-covalent interactions; host–guest complexes; fluorinated phenyl compounds
Received 3 August 1996; revised 22 November 1996; accepted 27 November 1996
INTRODUCTION
constants are still difficult to assess.7 Depending on the
method used they are reported to be higher5 than those of
protonated analogues, or similar.8 Information on the
association modes or complex geometries is, to the best of
our knowledge, virtually absent, as also is experimental
insight into the relevant binding mechanism. Host–guest
complexes with arene–fluorine interactions have not been
studied hiterto, and were the main target of the present
investigation. The importance for several applications9 of
non-covalent interactions with fluoro compounds has been
stressed recently.
Hydrophobic or lipophilic interactions in and with fluori-
nated compounds show puzzling behaviour. Thus,
perfluorinated surfactants show lower critical micelle con-
centrations than the protonated analogues, but their
sub-micellar aggregation numbers seem to be similar.2
Protonated and fluorinated surfactants do not form mixed
micelles, indicating less attractive C–H . . . F—C inter-
actions.3 With cyclodextrins (CDs) as host compounds,
surface tension measurements show stronger association
only with ␤-CD and only weak interactions with ␣-CD,4,5
although the cavity of the latter is, in contrast to what was
originally believed,4,5a not too small for intracavity inclusion
(see computer-aided simulations below, Figure 3). Associa-
tions of fluorinated surfactants with ␤-CD have also been
studied kinetically,6 but the corresponding association
RESULTS AND DISCUSSION
As host compound we used the azoniacyclophane CP44
(Scheme 1), which was synthesized by a modification of the
original protocol of Odashima et al.10 The advantages of
such water-soluble cyclophane hosts are that their com-
plexes with a large variety of organic substrates in aqeous
solution are experimentally well characterized by strong
NMR shift changes,11 and that the relevant binding mecha-
nisms have been studied in detail.1,12
* Correspondence
to:
H.-J.
Schneider.
E-mail:
CH12hsvr@rz.uni-sb.de.
† This is Part 70 of the Saarbrücken series Supramolecular
Chemistry for Part 69, see Ref. 1.
Contract grant sponsor: Volkswagen-Stiftung.
Contract grant sponsor: Deutsche Forschungsgemein-
schaft.
Molecular mechanics calculations with the aid of the
CHARMm force field by Bru¨nger and Karplus13 indicated
that the CP44 cavity is just wide enough to accommodate
© 1997 by John Wiley & Sons, Ltd.
CCC 0894–3230/97/050305–06 $17.50

306
X. FEI ET AL.
Scheme 1 (cont.)
non-linear curve fitting reliable association constants K and
also the NMR shifts of the totally complexed material (CIS
values, Table 1). Only for the very lipophilic toluene and
pentafluorotoluene was it necessary to add methanol to
water (all solvents deuterated). The measurements with the
latter and with the anionic host compounds V1–V4 were
performed with the permethylated host CP44 at pH 7.0 (pD
7.4), and those with the other electroneutral derivatives V7–
V12 with the protonated CP44 at pH 2·0 (pD 2·4) in order
to secure protonation of the acidic groups. Measurements
with related CPnn hosts have been shown previously15 to
lead to deviations of free complexation energies with
protonated or permethylated cyclophanes of usually below
2 kJ molϪ1
.
The experimentally observed complexation NMR shield-
ing effects on the guest compounds (Table 1) are typical of
many intracavity inclusion structures in such cyclophanes.
After a recent reparametrization of aromatic ring current
effects,16 we were able to calculate the expected CIS values
based on force field (CHARMm) optimized complex
geometries of such complexes (Figures 1 and 2). As pointed
out earlier, one needs to take into account the sizeable
electric field effects exerted by the N+ charges of the
cyclophanes in order to obtain realistic NMR shielding
variations.11c Although the deviations in the calculated
complexes are sizeable (Table 2), the generally found
approximate agreement is believed to be the first proof of
intracavity inclusion of the kind illustrated in Figures 1 and
2. This is of particular significance as the alternative method
of intermolecular NOE observations completely failed here.
Even application of the ROESY spin-lock technique,17
which can resolve the problem of unfavourable correlation
times of complexes with molecular weights around 1000,
did not lead to any usuable cross peaks, owing to too many
rapidly exchanging symmetric protons with on average only
Scheme 1
either a protonated or a fluorinated phenyl ring. The
CHARMm/QUANTA-calculated molecular volumes of the
fluorinated rings are about 10% larger than those of the
protonated analogues, but for a tight fit do not require
widening of the CP44 cavity (see appendix). Calculation of
the total steric energy of the host CP44 alone before and
after complexation, with both the free and the complex
structure energy-minimized, showed the upper limit of the
possible strain induced in the host to be on average below
1 kJ molϪ1
.
The host compounds (Scheme 1) were chosen so that at
least one proton was available for NMR shift measurements
as a function of guest concentration and vice versa, which
then were evaluated as described previously14 to obtain from
© 1997 by John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 305–310 (1997)

NON-COVALENT INTERACTIONS IN HOST–GUEST COMPLEXES
307
Table 1. Complexation constants K (mol lϪ1) and CIS values (ppm)
Guest
o-H
m-H
p-H
CH2
CH3
K
K(H/F)a
[mol/l]
V1
V2
V3
V4
Ϫ0·60
–
Ϫ1·11
–
Ϫ0·83
–
Ϫ1·19
–
Ϫ0·51
Ϫ0·12
Ϫ0·74
–
–
–
–
–
–
–
768
119
886
75
6·5
11·8
16
Ϫ0·46
Ϫ0·07
V5
V6
V7
V9
V10
V11
V12
Ϫ1·63
–
Ϫ1·85
Ϫ1·90
–
Ϫ1·42
–
Ϫ1·45
Ϫ1·65
–
Ϫ0·95
–
Ϫ0·85
Ϫ1·00
–
–
–
Ϫ0·77
Ϫ0·28
59
3·6
201
238
11
66
16
–
Ϫ0·90
Ϫ0·52
–
–
–
–
–
–
22
4
Ϫ1·66
–
Ϫ1·68
–
Ϫ0·95
Ϫ0·49
–
aK(H/F)=ratio of complexation constant of protonated for that of fluorinated guest com-
pound
weak intermolecular contacts.
tions with non-fluorinated compounds generally lead to
better agreement.11c
As is obvious from the association constants (Table 1)
and the corresponding free enthalpies ⌬⌬G (Scheme 1), all
the fluoro compounds show distinctly smaller interactions
with the cyclophane. The ⌬⌬G for toluene, measured in the
same solvent and with the same permethylated CP44 host,
reaches nearly a difference of 10 kJ molϪ1. The smaller CIS
value observed for the methyl signal of pentafluorotoluene
V6 (Table 1) points to a less deep immersion of the
fluorinated compound in the cavity. Similarly, smaller CIS
values with fluorinated derivatives are seen, e.g., with the
para-protons of V1/V2 and V11/V121 as well as with the
CH2 signals of V3/V4 and pf V9/V10. Obviously, in spite of
the sufficiently wide cavity, the fluorinated compounds are
less attracted into the ␲-electron-rich CP44 cavity. Inspec-
tion of the force field-generated geometries with the
fluorinated compounds immersed in the cavity (Figures 1
and 2) illustrate that here all fluorine atoms would be in
close contact at the centres of the ␲-moieties of the CP44
phenyl rings. The negative charges in all these locations
must lead to electrostatic repulsions, explaining the reason
for the lowered affinity (cf. the ⌬⌬G values) and the less
pronounced intra-cavity inclusion (cf. the CIS values) with
the fluorinated structures.
With cyclodextrins only a preliminary study was carried
out with ␣-cycloamylose (six glucose units) and phenol and
pentafluorophenol as guests, using methyl orange (MO) in
competitive titrations by UV–visible spectrophotometry.
Measurements at eight different wavelengths between 480
and 550 nm yielded association constants for MO with only
small deviations between 700 and 736
Ϫ1, in agreement
M
with literature data.19 Competitive titrations gave K=28 MϪ1
or ⌬⌬G=8·3 kJ molϪ1 for phenol, and K=14 MϪ1 or
⌬⌬G=6·55 kJ molϪ1 for pentafluorophenol. The lowered
affinity of the fluorinated guest here is understandable in
view of the narrow ␣-CD cavity, as visible in the CHARMm
calculated geometry, which shows only partial immersion of
the pentafluorphenol at the rim of the host (Figure 3). It
should be noted, however, that the ␣-CD according to
molecular mechanics calculations is just wide enough to
accommodate an unfolded (all-trans) perfluoroalkane chain.
The heptaamylose ␤-CD would be wide enough to accom-
modate also a perfluorophenyl host. Unfortunately, the
formation of precipitates inhibited the determination of
association constants with, e.g., pentafluorophenol and ␤-
CD (at concentrations around 0.01
M
for the phenol and
Another contribution to the diminished binding of the
fluorinated compounds is expected from the induced dipole
effect of the positively charged host ammonium centres on
the aromatic moiety of the guest compounds, which has
been shown earlier to provide most of the driving force in
such complexations12a, c, d, 18 Fluorine atoms in the guest
molecules strongly deplete the electron density in the
phenyl rings, and therefore also lower the stabilization by a
cation ␲–effect between the host and guest. It should be
noted that both the electrostatic effect from the permanent
partial charges and the induced dipoles are not described in
the force field used, which therefore cannot be expected to
lead to completely realistic complex geometries. This is
believed to be the major reason for the deviations between
calculated and experimental NMR shieldings, as calcula-
0·0035 for ␤-CD). However, before precipitation the
absorption of the added dye MO decreased markedly,
indicating strong complexation.
M
ACKNOWLEDGMENT
This work was generously supported by the Volkswagen-
Stiftung, Hannover, and the work in Saarbru¨cken also by the
Deutsche Forschungsgemeinschaft, Bonn.
EXPERIMENTAL AND COMPUTATIONAL DETAILS
1,6,20,25-Tetra[6.1.6.1]paracyclophane
(CP44)10. The
cylophane in the amine form was obtained via reaction of
N,NЈ-bis(p-tolylsuphonyl)-4,4Ј-diaminophenylmethane and
1,4-dibromobutane similarly to the literature procedure10 in
© 1997 by John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 305–310 (1997)

308
X. FEI ET AL.
35% yield after chromatography in the cyclization step, and
79% yield on detosylation with phenol–HBr.
and the last precipitate was dried in vacuum at 50 ЊC,
yielding 2·11 g (93%) of NMR-pure material. H NMR
1
Octamethyl-CP44 (permethylation). A mixture of 1.50 g
(2.97 mmol) of CP44 (described above), 2·3 (23 mmol)
CaCO3 and 10 g (70 mmol) of methylodide was stirred
under reflux in 30 ml methanol for 17 h. After removal of
methanol and methyl iodide on a round evaporator, 8 ml of
water were added and the solution was acidified with HCl.
The resulting solid was filtered off, air-dried and stirred
under reflux for 2 h with 80 ml of concentrated HCl and
methanol. After removal of the solvent and HCl, the residue
was dissolved in 40 ml of water and extracted with
chloroform until the chloroform became colourless. The
aqueous solution was concentrated to about 5 ml; after
addition of excess ethanol and acetone, the precipitate
formed was filtered off. The last procedure was repeated,
(D2O), ␦ (ppm): 7·60 (d, o-ArH, 8 H), 7·36 (d, m-ArH, 8 H),
4·18 (s, ArCH2, 4 H), 3·80 (broad, NCH2, 8 H), 3·53
(s, NMe, 24 H), 1·34 (broad, NCH2CH2, 8 H).
NMR measurements. NMR measurements were carried
out on a Bruker AM 400 system. To obtain association
constants, 7–9 samples were measured at concentrations
giving a complexation range between about 20 and 80%, as
described previously.14
UV–visible titrations. The titrations were carried with
methyl orange as described previously,19 using a Kontron
Uvikon instrument.
Molecular mechanics calculations. The calculations were
performed on a Silicon Graphics workstation under
QUANTA (Polygen/MSI) with the force field CHARMm13
(version 22), with Gasteiger derived charges using ␧=3·0 as
Figure 1. CHARMm simulations of host–guest structures: (a)
CP44 (permethyl)+toluene (side view); (b) CP44 (permethyl)+to-
luene (top view)
Figure 2. (a) CP44 (permethyl)+pentafluorotoluene (side view);
(b) CP44 (permethyl)+pentafluorotoluene (top view)
© 1997 by John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 305–310 (1997)

NON-COVALENT INTERACTIONS IN HOST–GUEST COMPLEXES
309
Table 2. Calculated CIS values (ppm) for complexes with ␣-CDa
Guest
Proton No.
⌬
␹
LEF ⌺(⌬ +LEF)
Exp. CIS Difference (⌺Ϫexp.)
␹
V2
V9
1
1
2
3
4
1
2
3
4
1
1
0·05
1·98
1·55
0·63
0·49
1·30
1·68
1·74
0·85
0·35
0·28
0·17
0·06
0·07
0·11
0·13
0·03
0·08
0·13
0·19
0·09
0·12
0·11
1·92
1·49
0·52
0·36
1·26
1·59
1·61
0·67
0·27
0·17
0·12
1·90
1·65
1·00
0·90
0·77
1·63
1·42
0·95
0·28
0·49
0·23
0·02
Ϫ0·16
Ϫ0·48
Ϫ0·54
0·49
Ϫ0·04
0·19
Ϫ0·28
0·01
V5
V6
V12
Ϫ0·32
a Positive aromatic ring anistropic effects (⌬ ) denote upfield shifts: positive linear electric field effects
␹
(LEF) denote downfield shifts; positive differences (between calculated and experimental shifts) denote
upfield deviation.
the dielectric constant. Steepest decent minimization was
used with stop criteria of 5000 steps of 0.001 kcal molϪ1
and of ␣-cyclodextrin with pentafluorophenol, 0·25 kcal-
,
molϪ1 (for structures see Figures 1–3).
after which adopted basis set Newton Raphson minimiza-
tions were carried out under same conditions.
Strain energies of hosts calculated from energy-mini-
mized structures of the free host and the host in the
host–guest complex11c were as follows: of CP44 (per-
REFERENCES
1. H.-J. Schneider, C. Kasper, V. Palyuikin and V. V. Samoshin,
Supramol. Chem., in press.
2. (a) D. J. Jobe, B. D. Skalski and R. E. Verrall, Langmuir 9,
2814 (1993); (b) N. J. Turro and P. C. C. Lee, J. Phys. Chem.
86, 3367 (1982); (c) S. Kato, S. Hararda, H. Nakashima and H.
Nomura, J. Colloid Interface Sci. 150, 305 (1992) and
references cited therein.
methylated)
upon
complexation
with
toluene,
0·33 kcal/molϪ1, with pentafluorotoluene, 0·19 kcal molϪ1
;
3. V. C. Reinsborough, T. D. Schultz and X. Xiang, Aust. J. Chem.
43, 11 (1990).
4. X.-K. Jiang, J.-H. Gu, X.-E. Cheng and Y.-Z. Hui, Acta Chim.
Sini. 159, 165 (1987).
5. (a) R. Palepu and V. C. Reinsborough, Can. J. Chem. 65, 1550
(1989); (b) R. Palepu, J. E. Richardson and V. C. Reinsborough
Langmuir 5, 218 (1989).
6. D. J. Jobe and R. E. Verrall, J. Phys. Chem. 98, 10814 (1994),
and references cited therein.
7. H. Mwakibete, R. Cristantino, D. M. Bloor and E. Wyn Jones,
Langmuir 11, 57 (1995).
8. E. Junquera, G. Tardajos and E. Aicart, Langmuir 9, 1213
(1993).
9. S. Gaentzler and P. Vierling, New J. Chem. 17, 585 (1993)
(emulsifying and biological applications); L. Soleviolan, B.
Devallez, M. Postel and J. G. Riess, New J. Chem. 17, 581
(1993) (emulsifying properties); T. Utsuki, F. Hirayama and K.
Uekama, J. Chem. Soc. Perkin Trans. 2 109 (1993), (photo-
dimerization); J. D. Druliner and E. Wasserman, J. Fluorine
Chem. 72, 75 (1995) (perfluoralkane isolation with CDs).
10. K. Odashima, A. Itai, Y. Iitake and K. Koga, J. Org. Chem. 50,
4478 (1985).
Figure 3. ␣-Cyclodextrin+pentafluorophenol (with electrostatics).
Without electrostatics the same result is obtained: guest leaves the
cavity upon energy minimization
11. (a) K. Odashima and K. Koga in Cyclophanes, edited by P. M.
Keehn and S. M. Rosenfeld, Vol. 2, p. 629. Academic Press,
New York (1983); (b) F. Diederich in Cyclophanes (Mono-
© 1997 by John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 305–310 (1997)

310
X. FEI ET AL.
graphs in Supramolecular Chemistry) edited by J. F. Stoddart.
APPENDIX
Royal Society of Chemistry, Cambridge (1991); (c) H.-J.
Schneider, Recl. Trav. Chim. Pays-Bas 112, 412 (1993).
12. (a) H.-J. Schneider, Angew. Chem. 103, 1419 (1991); Angew.
Chem. Int. Ed. Engl. 30, 1417 (1991); (b) H.-J. Schneider, T.
Blatter, B. Palm, U. Pfingstag, V. Ru¨diger and I. Theis, J. Am.
Chem. Soc. 114, 7704 (1992); (c) D. A. Dougherty Science 271,
163 (1996); (d) P. C. Kearny, L. S. Mizoue, R. A. Kumpf, J. E.
Forman, A. McCurdy and D. A. Dougherty, J. Am. Chem. Soc.
115 9907 (1993); (e) T. H. Webb and C. S. Wilcox Chem. Soc.
Rev. 383 (1993); (f) C. S. Wilcox; T. H. Webb, F. J. Zawacki,
N. Glagovich and H. Schu, Supramol. 1, 129 (1993).
13. A. T. Bru¨nger and M. Karplus Acc. Chem. Res. 24, 54 (1991).
14. H.-J. Schneider, R. Kramer, S. Simova and U. Schneider J. Am.
Chem. Soc. 110, 6442 (1988).
15. I. Theis, PhD Dissertation, Universität des Saarlandes (1990).
16. H.-J. Schneider, V. Ru¨diger and U. Cuber, J. Org. Chem. 60,
996 (1995).
17. D. Neuhaus and N. P. Williamson, The Nuclear Overhauser
Effect in Structural and Conformational Analysis. Verlag
Chemie, New York (1989).
Molar volumes of protonated and fluoroinated guest
compounds
Rates of
Compound
RX
volumes,
RF/RH
Volume (Å3)
V1
V2
V3
V4
V5
V6
V7
V8
401
428
457
495
382
426
413
433
462
499
354
368
1·07
1·08
1·12
1·05
1·08
1·04
1·16
18. H.-J. Schneider and T. Blatter, Angew. Chem. 100, 1211 (1988);
Angew. Chem. Int. Ed. Engl. 27, 1163 (1988).
19. R. I. Gelb, L. M. Schwartz, R. F. Johnson and D. A. Laufer, J.
Am. Chem. Soc. 10, 1869 (1979); R. I. Gelb, L. M. Schwartz,
M. Rades, R. B. Edmonds and D. A. Laufer, J. Am. Chem. Soc.
104, 6283 (1982).
V9
V10
V11
V12
F/H alone
© 1997 by John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 305–310 (1997)