The assembly of rotaxane-like dye/cyclodextrin/surface complexes on aluminium trihydroxide or goethite

Article abstract of DOI:10.1039/b518260c

Simple azo-dyes carrying phosphonic acid and arsonic acid substituents such as 4-(4′-hydroxyphenyl azo)phenylphosphonic acid (5) and 4-(4′-hydroxyphenylazo)phenylarsonic acid (6) bind more strongly to high surface area oxides such as aluminium trihydroxide and goethite than their carboxylic and sulfonic acid analogues and the phosphonate-functionalized dyes have been shown to have greater humidity fastness when printed onto commercial alumina-coated papers. Adsorption isotherm measurements provide evidence for the formation of ternary dye/cyclodextrin/surface complexes. Dyes which form such ternary complexes show higher light fastness when printed onto alumina coated papers in an ink formulation containing α-cyclodextrin. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b518260c

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
The assembly of rotaxane-like dye/cyclodextrin/surface complexes on
aluminium trihydroxide or goethite†‡
Rachel J. Cooper,a Philip J. Camp,a Ross J. Gordon,a David K. Henderson,a Dorothy C. R. Henry,a
Hamish McNab,a Sonali S. De Silva,a Daniel Tackley,b Peter A. Tasker*a and Paul Wightb
Received 23rd December 2005, Accepted 15th February 2006
First published as an Advance Article on the web 5th May 2006
DOI: 10.1039/b518260c
Simple azo-dyes carrying phosphonic acid and arsonic acid substituents such as 4-(4ꢀ-hydroxyphenyl
azo)phenylphosphonic acid (5) and 4-(4ꢀ-hydroxyphenylazo)phenylarsonic acid (6) bind more strongly
to high surface area oxides such as aluminium trihydroxide and goethite than their carboxylic and
sulfonic acid analogues and the phosphonate-functionalized dyes have been shown to have greater
humidity fastness when printed onto commercial alumina-coated papers. Adsorption isotherm
measurements provide evidence for the formation of ternary dye/cyclodextrin/surface complexes. Dyes
which form such ternary complexes show higher light fastness when printed onto alumina coated
papers in an ink formulation containing a-cyclodextrin.
isotherm measurements on the uptake of simple dyes containing
Introduction
pendant sulfonic, carboxylic, phosphonic and arsonic acid groups
The rapid development of digital cameras has been accompanied
(Fig. 1) at high surface area aluminium trihydroxide (ATH) using
by technologies which give photorealistic images by inkjet printing
this as a model for oxide treated papers. Results are compared with
onto high quality papers.1 These are usually treated with oxides
the uptake on goethite which while not used for paper treatment
such as alumina or silica to improve the quality and durability of
is an important pigmentary material.
Incorporation of chromophores into the cavities of cy-
currently used for photographic quality prints are generally similar
clodextrins (CDs) can greatly enhance their stability towards
the print and titanium dioxide to increase brightness.2,3 The inks
photobleaching.5–7 When “reactive dye” chemistry8 based on the
replacement of chlorine atoms in 1,3,5-trichlorotriazine was used
to attach a chromophore–cyclodextrin complex to cellulose, a
rotaxane was generated, “stoppered” at one end by the cellulose
surface and the other by a bulky naphthalene disulfonic acid group
(Fig. 2).5 In this paper we focus on the design requirements to form
rotaxane-like ternary complexes in which an inclusion complex of
the dye in the cyclodextrin is tethered to an ATH or goethite
surface by a ligating group on the water-soluble dye.
to those used on lower grade papers and it is important to establish
whether their performance can be improved by functionalizing the
dyes to meet the requirements of the higher grade papers. To this
end we have previously described a strategy for enhancing the
wet fastness of dyes for inkjet printing based on incorporating
ligating groups which form very stable complexes with the oxides
used in photo quality papers.4 In this paper we discuss adsorption
aSchool of Chemistry, The University of Edinburgh, Joseph Black Building,
Kings Buildings, West Mains Road, Edinburgh, UK EH9 3JJ. E-mail:
p.a.tasker@ed.ac.uk; Fax: +44 131 6506453; Tel: +44 131 6504706
bResearch Centre, Avecia, Blackley, Manchester, UK M9 3ZS
Experimental
† Based on the presentation given at Dalton Discussion No. 9, 19–21st
April 2006, Hulme Hall, Manchester, UK.
Instrumentation
‡ Electronic supplementary information (ESI) available: 1–2: Title page;
3–4: Data for isotherms of compounds 1 and 2 on ATH; 5–6: Data for
Fig. 3 – isotherms of 5 and 6 on ATH; 7–8: Data for Fig. 4 – isotherms
of 1 and 5 on goethite (plotted on a log scale); 9: Data for Fig. 5 – NMR
data for formation of 5: a-CD complex; 10: Data for Fig. 6 – UV/Vis data
for formation of 5: a-CD complex; 11–12: Data for Fig. 8 – isotherms
of compound 5 with a- and b-CD on ATH; 13–14: Data for Fig. 10 –
isotherm of compound 6 with a-CD on ATH; 15–16: Data for Fig. 11 –
isotherm of compound 5 with a-CD on goethite; 17–18: Data for Fig. 12 –
isotherm of compound 1 with a-CD on goethite; 19: Data for UV/Vis for
formation of 4: a-CD complex; 20: Data for UV/Vis for formation of 6: a-
CD complex; 21: Data for UV/Vis for formation of 7: a-CD complex; 22:
Data for UV/Vis for formation of 8: a-CD complex; 23: Data for UV/Vis
for formation of 9: a-CD complex; 24: Data for UV/Vis for formation of
10: a-CD complex; 25: ROESY NMR of 3: a-CD complex; 26: ROESY
NMR of 4: a-CD complex; 27: ROESY NMR of 5: a-CD complex; 28:
ROESY NMR of 6: a-CD complex; 29: ROESY NMR of 9: a-CD complex;
30: ROESY NMR of 10: a-CD complex; 31: Light fastness results of 3–6,
9 and 10; 32: Paper types used in inkjet tests; 33: Data for isotherm of 3
on ATH; 34: Data for isotherm of 4 on ATH; 35–36: Data for isotherms
of 7 and 8 on ATH; 37–38: Data for isotherms of 9 and 10 on ATH. See
DOI: 10.1039/b518260c
1D Nuclear magnetic resonance spectra were recorded on Bruker
AC 250 MHz and DPX Bruker 360 MHz instruments. All 2D
NMR experiments (COSY, and ROESY) were acquired from 1024
increments of 2 K data points and 16 scans each. The data were
zero-filled twice in the t1 dimension and multiplied by a squared
sine bell function (SSB) in both dimensions. NMR spectroscopic
data processing was carried out on a Bruker Silicon Graphics
O2 station with standard UXNMR software as well as on a PC
with 1D WIN NMR (960901) software (Bruker Franzen Analytic
GmbH).
Fast atom bombardment (FAB) mass spectrometry was per-
formed on a Kratos MS 50 machine. Electrospray (ESI) mass
spectra were obtained on a Thermoquest LCQ spectrometer.
FTIR spectra were recorded on a JABCO FTIR-460 spectrometer
using KBr discs. Electronic spectra were measured on an ATI
UNICAM UV/Vis spectrometer with 1 cm path length quartz
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 2785–2793 | 2785
©

View Article Online
Fig. 1 Structures of the azo-dyes and simple phosphonic acid derivatives used in this paper.
R
Goethi◦te by Bayer (Bayferroxꢀ 415). a-Cyclodextrin was dried
at 100 C under vacuum prior to use. Water was purified before
R
use on a Milli-Qꢀ water purification system.
Synthesis of the ligands shown in Fig. 1
Phenylphosphonic acid, 1, 4-(4ꢀ-aminophenylazo)phenylarsonic
acid, 7 and 4-(4ꢀ-dimethylaminephenylazo)benzenearsonic acid
hydrochloride, 8 were obtained from Aldrich.
3-Oxo-3-p-tolylpropylphosphonic acid (2). 2 was prepared in
three steps9,10 from 4-methylacetophenone. Yield (70%); mp 157–
159 ◦C. Found: C, 52.60; H, 5.72. C10H13O4P requires C, 52.64;
H, 5.74%. dH (d6-DMSO, 250 MHz) 7.85 (d, 2H, Ar–H), 7.32 (d,
2H, Ar–H), 3.15 (m, 2H, CH2) 2.30 (s, 3H, CH3), 1.90 (m, 2H,
=
CH2P). dC (d6-DMSO, 63 MHz) 197.66 (C O), 143.46 (Ar C),
133.95 (Ar C), 129.44 (Ar 2CH), 128.08 (Ar 2CH), 38.76 (CH2),
21.97 (CH2P), 21.28 (CH3). dP (d6-DMSO, 101 MHz) 27.30. MS
(FAB, NOBA) m/z 229 (MH+, 99.2%). mmax/cm−1 3500 br (OH),
Fig. 2 The proposed rotaxane structure for the attachment of a reactive
dye to mercerized cotton.5
cuvettes. Inductively coupled plasma optical emission spec-
troscopy (ICP-OES) analysis for Al, As, Fe, P and S was performed
on either a Thermo Jarrell Ash IRIS ICP-OES or a Perkin
Elmer Optima 5300 DV. C, H, N contents were obtained on a
CE-440 elemental analyzer. Melting points were recorded on a
Gallenkamp apparatus and the measurement of pH was carried
out using an ORION 410A pH meter.
Curve fitting was performed using the programs Origin 5.0 (c)
Microcal Software Inc and SigmaPlot 2000 (demo version) (c)
1986-2000 SPSS Inc.
=
=
1681 (C O), 1251 (P O).
4-(4ꢀ-Hydroxyphenylazo)benzenesulfonic acid (3). Dye 3 was
obtained from sulfanilic acid.11 Yield (58%); mp 282–285 ◦C
◦
(decomp.) (lit.,11 283 C).§ dH (D2O, 250 MHz) 7.73 (d, 2H, Ar–
H), 7.58 (m, 4H, Ar–H), 6.54 (d, 2H, Ar–H). dC (D2O, 63 MHz)
=
174.94 (C–OH), 154.58 (C–S), 142.58 (C–N N), 142.07 (C–
=
N N), 127.08 (Ar 4CH), 122.15 (Ar 2CH), 120.29 (Ar 2CH).
MS (ES, MeOH–CH3CN) m/z 277 (M−, 100%). mmax/cm−1 3244
= =
br (OH), 1187 (O S O).
4-(4ꢀ-Hydroxyphenylazo)benzoic acid (4). Dye 4 was similarly
Solvent and reagents
prepared from 4-aminophenylcarboxylic acid.11 Yield (66%);
All reagents were used as obtained from Aldrich, Lancaster
or Fluka. Solvents were used as received. High surface area
“superfine Al(OH)3” was supplied by Alcan Chemicals and
§ As is the norm for water-soluble dyes, CHN values are not included in
the paper since the dyes are isolated in their salt form.
2786 | Dalton Trans., 2006, 2785–2793
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
mp 275–277 ◦C (decomp.) (lit.,12 277–279 ◦C).§ dH (D2O, 250 MHz)
7.98 (d, 2H, Ar–H) 7.76–7.84 (m, 4H, Ar–H), 6.98 (d, 2H, Ar–H).
dC (D2O, 63 MHz) 183.01 (CO2H), 157.59 (C–OH), 157.32 (C–
were prepared by procedures developed at Avecia.13 Phenylphos-
phonic acid (25.0 g, 0.14 mol) was dissolved in concentrated
H2SO4 (102 ml) at 0 ◦C and a mixture of H2SO4 and HNO3
(10 and 13 ml) was added dropwise at 0 ◦C. After stirring
for 2 h, the reaction was quenched by the addition of ice (ca.
300 g). The precipitate was recrystallised from hot acetic acid to
give 3-nitrophenylphosphonic acid as colourless crystals. Yield
(13.3 g, 42%); mp 148–150 ◦C. Found: C, 34.86; H, 3.17; N,
6.46. C6H6NO5P requires C, 35.48; H, 2.98; N, 6.90%. dH (D2O,
250 MHz) 8.35 (d, 2H, Ar–H), 8.21 (d, 2H, Ar–H), 7.94 (q,
2H, Ar–H), 7.58 (t, 2H, Ar–H). dC (D2O, 63 MHz) 147.92 (C–
NO2), 136.88 (Ar CH), 133.14 (C–P), 130.42 (Ar CH), 126.53
(Ar CH), 125.54 (Ar CH). dP (D2O, 101 MHz) 12.47. MS
(FAB, THIOG) m/z 204 (MH+, 100%). mmax/cm−1 2873 br (OH),
=
=
N N) 145.20 (C–N N), 133.15 (Ar C–CO2H), 130.40 (Ar 2CH),
122.61 (Ar 2CH), 124.09 (Ar 2CH), 116.11 (Ar 2CH). MS (ES,
MeOH–CH3CN) m/z 241 (M−, 100%).
4-(4ꢀ-Hydroxyphenylazo)phenylphosphonic acid (5). Dye
was prepared by
procedure developed at Avecia.13 p-
5
a
Bromoacetanilide (50.0 g, 0.2◦1 mol) and NiCl2 (3.5 g, 0.03 mol)
were heated under N2 at 190 C and triethyl phosphite (45.0 ml,
0.26 mol) was added dropwise, collecting the 1-bromoethane
formed in a Dean–Stark apparatus fitted with a solid CO2
condenser. After addition was complete, the reaction was allowed
◦
=
to cool to 150 C and stirred for 1 h. The solution was allowed
1279 (P O). 3-Nitrophenylphosphonic acid (10.0 g, 0.05 mol) in
to cool to room temperature and stirred with light petroleum (bp
40–60 ◦C) overnight. The white solid which separated was recrys-
tallised from ethyl acetate to give 4-acetylaminophenylphosphonic
acid diethyl ester, as white crystals. Yield (29.4 g, 52%); mp
138–140 ◦C. Found: C, 52.33; H, 6.60; N, 5.19. C12H18NO4P
requires C, 53.14; H, 6.69; N, 5.16%. dH (CDCl3, 250 MHz)
9.44 (s, 1H, NH), 7.76 (m, 4H, Ar–H), 4.09 (m, 4H, 2CH2),
methanol (25 ml) was reduced under H2 for 6 h in the presence
of 10% Pd/C (0.3 g) to form 3-aminophenylphosphonic acid.
◦
Yield (8.5 g, 56%); mp 290 C (decomp.) (lit.,15 315 ◦C). dH
(CDCl3, 250 MHz) 7.02 (d, 1H, Ar–H), 6.60 (d, 1H, Ar–H)
6.40 (d, 2H, Ar–H). dC (CD3OD, 63 MHz) 150.11 (C–NH2),
137.89 (C–P), 129.14 (Ar CH), 121.80 (Ar CH), 117.41 (Ar CH),
112.32 (Ar CH). dP (CDCl3, 101 MHz) 25.04. MS (FAB, NOBA)
m/z 174 (MH+, 100%). 3-Aminophenylphosphonic acid was con-
verted to 3-(4ꢀ-aminophenylazo)benzenephosphonic acid (9) and
3-(4ꢀ-hydroxyphenylazo)benzenephosphonic acid (10) by coupling
with aniline or phenol respectively, following the method of
Suh et al.11
=
2.22 (s, 3H, CH3C O), 1.33 (t, 6H 2CH3). dC (CDCl3, 63 MHz)
=
169.48 (C O), 142.68 (Ar C–N), 132.55 (Ar 2CH), 122.65 (Ar
CH), 120.52 (Ar CH), 119.14 (Ar C–P) 62.09 (2 OCH2), 24.30
=
(CH3C O), 16.14 (2 CH2CH3). dP (CDCl3, 101 MHz) 20.20. MS
(FAB, NOBA) m/z 272 (MH+, 100%). mmax/cm−1 3311 br (OH),
=
=
1701 (C O), 1259 (P O). 4-Acetylaminophenylphosphonic acid
diethyl ester (20.0 g, 0.07 mol) was stirred at reflux (ca. 100 ◦C) in
concentrated HCl (500 ml) overnight. Ethanol (70 ml) was added
to aid dissolution. The solution was concentrated to 100 ml and
placed in the fridge for 48 h. The resulting white precipitate,
4-aminophenylphosphonic acid, 11 was◦collected and dried in
vacuo. Yield (8.0 g, 62%); mp 251–252 C (lit.,14 254–256 ◦C).
dH (D2O 250 MHz) 7.52 (m, 2H, Ar–H), 6.84 (m, 2H, Ar–H).
dC (D2O, 63 MHz) 147.89 (C–NH2), 131.92 (Ar 2CH), 129.66
(C–P), 115.84 (Ar 2CH). dP (D2O, 101 MHz) 14.70. MS (FAB,
9:§ dH (CD3OD, 250 MHz) 8.2 (d, 1H, Ar–H), 7.72–7.84 (m,
2H, Ar–H), 7.71 (d, 2H, Ar–H), 7.44 (t, 1H, Ar–H), 7.02 (d, 2H,
Ar–H). dP (CD3OD, 101 MHz) 16.25. MS (ES, NOBA) m/z 276
(M−, 100%).
10:§ dH (CD3OD, 250 MHz) 7.73–7.93 (m, 4H, Ar–H), 7.55 (m,
2H, Ar–H) 6.84 (d, 2H, Ar–H). dP (CD3OD, 101 MHz) 16.2. MS
(ES, NOBA) m/z 277 (M−, 100%).
Adsorption isotherm measurements‡
Aqueous stock solutions (5 × 10−3 M) of each of the ligands
shown in Fig. 1 were prepared and the pH was adjusted to
ca. 8.5 by addition of sodium hydroxide. In experiments with
cyclodextrin, a- or b-CD was added in equimolar quantities.
Accurately pre-weighed quantities of ATH or goethite, (ca. 0.4 g)
in polycarbonate centrifuge tubes were stirred with solutions of
the ligand in water (10.0 ml) of known concentrations for 2 h at
25 ◦C. The mixtures were centrifuged, filtered and analysed by
ICP-OES. The determined concentrations of phosphorus, sulfur
or arsenic defined the amounts of dye remaining in solution
and hence the amount adsorbed on the ATH or goethite. No
significant levels of aluminium or iron were detected in the
aqueous phase after equilibration with ATH or goethite. The
data, plotted using Origin 6.1, were subject to a non-linear
curve fit. By default, the maximum surface coverage (A) and
equilibrium adsorption constant (K) were obtained from the
standard Langmuir adsorption isotherm
NOBA) m/z 174 (MH+, 71%). mmax/cm−1 2865 br (OH), 1226
(P O). 4-Aminophenylphosphonic acid was converted to 4-(4 -
ꢀ
=
hydroxyphenylazo)phenylphosphonic acid, 5 using the method of
Suh et al.11 Yield (95%); mp 277–280 ◦C (decomp.).§ dH (D2O,
250 MHz) 7.72–7.83 (m, 4H, Ar–H), 7.65 (m, 2H, Ar–H) 6.75
(d, 2H, Ar–H). dC (D2O, 63 MHz) 173.91 (C–OH), 152.90 (C–P),
142.00 (C–N N), 141.09 (C–N N), 131.68 (Ar 2CH), 126.42 (Ar
2CH), 121.08 (Ar 2CH), 120.12 (Ar 2CH). dP (D2O, 101 MHz)
11.79. MS (ES, MeOH–CH3CN) m/z 277 (M−, 100%). mmax/cm−1
=
=
=
3470 br (OH), 1247 (P O).
4-(4ꢀ-Hydroxyphenylazo)phenylarsonic acid (6). Dye 6 was pre-
pared from 4-aminophenylarsonic acid using the same method11
as 3. Yield (62%); mp 288–289 ◦C (decomp.).§ dH (D2O, 250 MHz)
7.72 (d, 2H, Ar–H), 7.60–7.64 (m, 4H, Ar–H), 6.57 (d, 2H, Ar–H).
dC (D2O, 63 MHz) 174.52 (C–OH), 154.50 (C–As), 142.23 (C–
N N), 141.08 (C–N N), 131.48 (Ar 2CH), 126.93 (Ar 2CH),
121.98 (Ar 2CH), 120.37 (Ar 2CH). MS (ES, DMF–MeOH–
CH3CN) m/z 321 (M−, 22%). mmax/cm−1 3103 br (OH).
=
=
−−o
AK(c c )
/
y =
(1)
1 + K(c c−−o )
/
3-(4ꢀ-Aminophenylazo)benzenephosphonic acid (9) and 3-(4ꢀ-
hydroxyphenylazo)benzenephosphonic acid (10). Dyes 9 and 10
where y is the surface coverage, c is the residual dye concentration
−−o
in solution, and c = 1 mol dm−3. In some cases, it was necessary
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 2785–2793 | 2787
©

View Article Online
to modify this function to take into account the formation of a
second adsorbate layer. It was found that a suitable function was
of the ‘double-Langmuir’ form
Print testing
Ink preparation and printing. The dye (0.35 g) was dissolved in
a mixture of glycerol (0.75 g), thiodiglycol (0.75 g), urea (0.75 g),
Surfynol 465 (0.1 g) and distilled water (7.3 g) and was stirred for
30 min. The pH was raised to 8.5–9.5 by the addition of LiOH
and the solution was microfiltered through a 0.45 lm syringe filter.
The filtrate was placed in a nib to generate prints using a contact
method of ink lay-down. When inkjet printing was used the ink was
microfiltered through a 0.45 lm syringe filter, and injected (3–7 ml)
into a clean BC21 monochrome ink cartridge. A small volume of
ink was pulled through using a vacuum line, and the cartridge was
inserted into HP560 printer to generate prints. Reflected optical
density (ROD), L, a, b, C and h of the prints were measured using
a Gretag–Macbeth spectrodensitometer 24 h after printing.
ꢀ
ꢁ
K (c c−−o )
K (c c−−o )
/
/
1
2
y = A
+
(2)
1 + K (c c−−o ) 1 + K (c c−−o )
/
/
1
2
where K1 and K2 are the adsorption constants for the first
and second layers. These equations are all appropriate to ideal-
solution conditions, which is justified for the range of residual dye
concentrations obtained in this work (less than 10−2 mol dm−3).
UV/Vis measurements
Stock solutions (7 × 10−5 M) of each dye were prepared in a
phosphate buffer16 (ca. pH 8) and the absorption spectrum of
the dye was measured both independently and in the presence
of increasing concentrations of a-CD (1 × 10−7 to 1 × 10−2 M)
Humidity fastness. Sections of prints were sealed in a humidity
jar for 16 h at 60 ◦C. The prints were then analysed by optical
microscopy and the colour bleed into non-printed areas was eval-
uated qualitatively using a scale 1–10 (1 indicating undetectable
migration) by comparison with control prints supplied by Avecia.
◦
at a constant wavelength (ca. 300–400 nm) at 25 C. The results
from the latter experiments were analysed to yield complexation
constants (K) as follows. It is assumed that the concentrations
of dye ([dye]), cyclodextrin ([CD]), and complex ([CD-dye]) are
sufficiently low for the respective activity coefficients to be assumed
equal to unity. If the total dye concentration is denoted by [dye]0,
Light fastness. The ROD of printed cards was measured (see
above) before and after exposure to known illuminances, measured
at 420 nm, in an Atlas Ci5000 weatherometer for up to 100 h at
63 ◦C and 50% humidity.
the mole fraction of ‘free’ (uncomplexed) dye molecules by xfree
,
and the total cyclodextrin concentration by [CD]0, then we have
that [CD-dye] = [dye]0(1 − xfree) and [CD] = [CD]0 − [dye]0(1 −
Ozone fastness. The ROD of printed cards was measured (see
above) before and after 24 h exposure to 1 ppm ozone at 40 C
◦
x
free). Inserting these relations into the mass-action law
[CD − dye]c−−o
K =
and relative humidity 50% in a Hampden 903 ozone cabinet.
(3)
[CD][dye]
Results and discussion
and solving for xfree yields
ꢂ
It was assumed that a necessary condition for the formation
of a stable ternary surface/dye/CD complex analogous to the
rotaxane in Fig. 2 is that the ligating group on the dye should
afford strong binding to the oxide surface. A comparison of the
isotherms for the four dyes 3–6, which vary only in the nature
of acid group para to the azo unit, allowed us to define suitable
ligating groups. The phosphonic and arsonic acid derivatives 5
and 6 show isotherms (Fig. 3) on high surface area aluminium
trihydroxide with similar equilibrium adsorption constants and
surface coverages [800 ( 100) and 20 × 10−6 mol g−1] and [1400
( 200) and 10 × 10−6 mol g−1] respectively, based on curve-fitting
for a Langmuir model.17,18 The isotherms for 3 and 4 are poorly
defined (see ESI‡) suggesting that sulfonic and carboxylic acids
attach much more weakly to ATH.¶
ꢃ
−(1 + KD) + (1 + KD)2 + 4K[dye]0 c−−o
ꢃ
xfree
=
(4)
2K[dye]0 c−−o
−−o
where D = ([CD]0 − [dye]0)/c . The measured absorption coeffi-
cient of the dye at fixed wavelength k is a weighted sum of the free
and complex absorption coefficients:
e([CD]0, k) = xfreeefree(k) + (1 − xfree)ecomplex(k)
(5)
Substituting eqn (4) into eqn (5) yields the dependence of the
measured absorption coefficient on cyclodextrin concentration.
NMR measurements
1
In the H NMR experiments, the total dye concentration was
The phosphonic acid and arsonic acid dyes 5 and 6 bind very
strongly to goethite (see ESI‡). For such strong binding it is helpful
to present the isotherms with a logarithmic scale for the residual
concentration (see Fig. 4 and equations in the Experimental
section). When this is done the system provides evidence for double
layering. Once a coherent monolayer of 5 or 6 is formed it would
be possible for phenol groups of a second layer of dye molecules
to interact with the array of phenolic OH groups terminating
the monolayer. Such a mode of binding for the second layer is
supported by the observation that benzene phosphonic acid 1 also
gives an isotherm characteristic of very strong binding but the
fixed at 0.05 M and adjusted to ca. pH 8 by addition of
sodium carbonate. The total cyclodextrin concentration was
varied between 0.4[dye] to 2.0[dye] and the variation of the peak
position of a well resolved signal in the region of dH = 7–8 ppm was
measured at 25 ◦C. The approach outlined in the section ‘UV/Vis
measurements’ was employed to extract the dye–cyclodextrin
complexation constant (K); the only difference is that eqn (5) now
reads
d([CD]0) = xfreedfree + (1 − xfree)dcomplex
(6)
COESY and ROESY spectra (see ESI‡) were recorded in D2O,
0.03 mol of dye + 0.03 mol of a-CD, adjusted to pH 8 by addition
of NaHCO3.
¶ Some variations in the shape of the isotherms could result from different
hydration energies of the dyes but this has not been analysed.
2788 | Dalton Trans., 2006, 2785–2793
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
measurement of the variation in chemical shifts of H-atoms ortho
to the azo group of 5, using a fixed dye concentration of 0.05 M in
D2O, with various cyclodextrin concentrations (Fig. 5) gave values
K = (2 1) × 103, dfree = 7.77 0.02 ppm, and dcomplex = 7.48
0.01 ppm.
Fig. 3 The uptake of 5 and 6 on high surface area aluminium trihydroxide
as a function of residual concentration in the aqueous phase.
Fig. 5 Variation of the chemical shift (d) of the 1H signal (7–8 ppm) of 5
(0.05 M) in D2O with a-cyclodextrin concentration: pH 8 at 25 ◦C.
The incorporation of azo-dyes into the cavities of cyclodextrins
is often accompanied by a significant change in wavelength of the
chromophore22 and consequently it is also possible to use varia-
tions in UV/Vis spectra to monitor complex formation. Spectra
are also dependent on pH and concentration as azo dyes are known
to aggregate in aqueous solutions with a resultant decrease in
adsorption and a departure from the Beer–Lambert law.23 At the
dye concentrations used in this study (7 × 10−5 M) absorbances
were found to vary linearly with concentrations. Using 7 × 10−5
M
solutions at pH 8 in phosphate buffer, variations in absorption
coefficient of major adsorption peaks were measured as a function
of the concentration of added cyclodextrin. Appropriate peaks
and concentration ranges of cyclodextrin were then used to obtain
data to allow curve fitting to confirm the stoichiometry of the
dye : cyclodextrin complex and determine its formation constant.
Data for the a-cyclodextrin complex of 5 are shown in Fig. 6. The
resulting fit of absorption coefficients at increasing concentrations
of a-CD gave values of K = (3 1) × 103, efree(359 nm) = 12100
( 100) M−1 cm−1 and ecomplex(359 nm) = 16000 ( 200) M−1 cm−1.
Results for the other dyes are included in Table 1.
Fig. 4 The uptake of 1 (᭺) and 5 (᭹) on goethite, plotted on a logarithmic
scale. The equilibrium adsorption constants for the monolayer shown in
1 is [19 ( 3) × 104] and the first and second layers in 5 is [11 ( 2) × 105]
and [560 ( 220)] respectively.
surface coverage (10.9 × 10−6 mol g−1) corresponds closely to the
value for the monolayer of 5 (9.86 × 10−6 mol g−1).
The relatively weak attachment of sulfonic and carboxylic acids
to ATH and goethite suggested that dyes containing phosphonic
and arsonic acid tethering groups would be more suitable for the
assembly of ternary rotaxane-like complexes at the surface.
Another necessary condition for the formation of such
rotaxane-like structures is that the dye should form stable inclusion
complexes within the cavity of the cyclodextrin. A wide range
of techniques have been applied to the study of the formation
of such cyclodextrin complexes.19 NMR, in particular, offers the
possibility of defining the orientation of the guest in the cavity5,20–22
and of measuring formation constants for the dye–CD complex.4
An analysis of the ROESY 1H NMR spectra of 1 : 1 mixtures of
a-cyclodextrin and 5 or 10 in D2O indicates that the four H-atoms
ortho to the azo group form close contacts in the cavity with the
cyclodextrin’s H3 and H5 protons. Such an assembly with azo
dyes can lead to protection from photobleaching.5–7 In this work,
Formation constants for the para-substituted phenolic dyes
3–6 fall in the range (1.8–13.2) × 103, whilst that for the
Table 1 Equilibrium constants and absorption coefficients of the free dye
and the inclusion complexes for dyes 3–10 in phosphate buffer at 25 ◦
C
Dye
K
efree/M−1 cm−1
ecomplex/M−1 cm−1
3
1800 ( 60)
13350 ( 200)
14000 ( 60)
12100 ( 100)
12900 ( 20)
17300 ( 42)
26100 ( 40)
13000 ( 100)
14000 ( 100)
16500 ( 200)
15000 ( 50)
16000 ( 200)
14500 ( 18)
19400 ( 40)
24700 ( 40)
13000 ( 100)
37000 ( 5000)
4
8500 ( 3000)
3000 ( 700)
13200 ( 1000)
8600 ( 1000)
15700 ( 2500)
2400 ( 2000)
60 ( 100)
5
6
7
8
9
10
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 2785–2793 | 2789
©

View Article Online
dye–surface and the dye–cyclodextrin complexes will be needed
for the formation of the rotaxane structure to be favoured.
The isotherms presented in Fig. 8 follow the uptake of the
phosphonate dye 5 onto ATH in the presence of equimolar
quantities of a- or b-cyclodextrin. The initial slopes at low residual
dye concentration are comparably steep with the curve for the
dye alone, indicating that similar binding strengths are involved.
However, the surface coverages differ significantly. The increased
bulk of the dye-cyclodextrin complexes leads to a decrease in
surface coverage from 20.0 × 10−6 to 6.3 × 10−6 and 4.8 ×
10−6 mol g−1 for 5 and its a-CD and b-CD complexes, respectively.
Fig. 6 Variation of the absorption coefficient (e) of cyclodextrin–dye
complex of 5 (7 × 10−5 M) at pH 8 with concentration of a-cyclodextrin in
phosphate buffer at 25 ◦C.
meta-substituted analogue 10 is an order of magnitude lower.
1H and 13C NMR studies of the phosphonates 1 and 2 with
CD indicate much weaker binding and provided evidence for the
formation of complexes with other stoichiometries.
The adsorption isotherm data and the formation constants for
binary complex formation allow us to define which combinations
of surface, dye and cyclodextrin are most likely to form the ternary,
rotaxane-like, complex (see Fig. 7). Comparable stability of the
Fig. 8 The effect of addition of a- or b-CD on the uptake of 5 onto high
surface area aluminium trihydroxide.
Fig. 7 Schematic representation of the various equilibria associated with the formation of the ternary, rotaxane-like complex on high surface area metal
oxides.
2790 | Dalton Trans., 2006, 2785–2793
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Fig. 9 The energy minimised structure of the a-CD: 5 complex with the phosphonic acid group protruding through the secondary hydroxyl rim of the
cyclodextrin.
These data correspond to coverages of 58 ( 2), 185 ( 9) and
2
˚
242 ( 15) A per molecule for the dye and its a-CD and b-CD
complexes, based on the surface area (7 m2 g−1) of the ATH used.
The surface coverages correspond fairly closely to values which
would be predicted by close packing discs with diameters similar
to those reported24 for a- and b-CD, 13.7 and 15.3 A, respectively.
˚
For primitive cubic and hexagonal cubic close packing of a-
2
˚
CD the areas predicted are 188 and 163 A , respectively.* The
significance of the similarity between observed and calculated
surface coverages of the CD complexes needs to be interpreted
with caution because the surface area of the ATH is determined
under very different conditions and it is unlikely that the material
will have substantial areas of flat surface. Also it is unlikely
that the cross-sections of dye–cyclodextrin complexes will be
strictly circular. Molecular modelling using the Dreiding force
field method25 indicates that a-CD complexes of these simple
azo-dyes have an elliptical cross section (Fig. 9) with elongation
away from the edges and compression towards the faces of the
encapsulated benzene ring.†† The shortest and longest distances
between diametrically opposed internal oxygen atoms in the
Fig. 10 The effect of addition of a-CD on the uptake of 6 onto high
surface area aluminium trihydroxide.
the dye being reduced by the addition of cyclodextrin (see Fig. 11).
In these systems the addition of a-CD suppresses multilayering of
the dye.
˚
cyclodextrin are 7.02 and 9.60 A.
When the uptake of the comparable arsonate dye 6 onto ATH
was measured in the presence of a-CD (Fig. 10) the isotherm
indicated, as with 5 (Fig. 8) that the surface-binding of the
dye:cyclodextrin complex occurs. In this case the surface coverage
of dye is not reduced so substantially by the addition of a-CD
which is consistent with the higher solution stability of the binary
CD : 6 complex (Table 1).
In cases where the surface binding of the dye is very strong, e.g.
for the phosphonicacid and arsonic acid dyes 5 and 6 on goethite,
the uptake of the dye alone over the dye:cyclodextrin complex is
preferred and there is no evidence for the mono-layer coverage of
* Each square representing the primitive cubic packing has an area =
2
13.72 = 188 A and contains four quarters of an a-CD. Each triangle
˚
representing contacts for hexagonal packing has an area = 1/2 × 13.7 ×
13.7 sin 60◦ = 81.3 A and contains 3 × 1/6th of an a-CD.
2
˚
†† A CD–guest structure was selected from the Cambridge Structural
Database, CDEXIA01 (a-cyclodextrin–para-iodoaniline trihydrate);26 the
guest and water of solvation were removed and replaced by 5. The
energy minimised structure with the phosphonate adjacent to the rim
with secondary alcohol groups is slightly more stable (−162 kcal mol−1
(−152 kcal mol−1).
)
Fig. 11 The effect of the addition of a-CD on the uptake of 5 onto high
surface area goethite. Comparable results for 6 are included in the ESI.‡
than that with the phosphonate adjacent to the primary alcohol groups
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 2785–2793 | 2791
©

View Article Online
For the phosphonic acids 1 and 2 which do not form well defined
1 : 1 inclusion complexes with a-CD in solution (see above), no
reduction of the total uptake on goethite was expected on the
addition of a-CD in an isotherm determination and none was
found (see Fig. 12).
Fig. 12 The insignificant effect of the addition of a-CD on the uptake of
1 onto high surface area goethite. Very similar results were obtained with
2 (see ESI†).
The effects of incorporation of surface-ligating groups and of
the formation of cyclodextrin complexes on the performance of
the dyes when printed onto various papers were evaluated using
a variety of test methods. Humidity fastness of 3–5 and 9 was
compared by exposing printed text on three types of paper in sealed
humidity jars at 60 ◦C. The phosphonates 5 and 9 showed (Table 2)
a smaller degree of dye migration on alumina coated papers than
sulfonate and carboxylate analogues. Such an improvement in
wet fastness is consistent with the strong binding of 5 and 9 to
high surface area ATH, as revealed by the adsorption isotherm
studies described above. The phosphonate-functionalized dyes
also showed a greater wet fastness on silica-treated papers and it is
clear that attachment of surface ligating groups offers an effective
method of improving the performance of dyes on oxide-treated
papers.
It has been shown previously27,28 that light fastness of inkjet
dyes is very dependent on the nature of the media onto which
they are printed, and, as might be expected, the incorporation
of surface ligating groups into the simple dyes shown in Fig. 1
had no beneficial effect on their light fastness (see ESI‡). Indeed,
the phosphonate-functionalised dyes faded faster than 3 and 4
when they were printed on alumina-coated papers. However,
the light fastness of these dyes on alumina-coated papers (HG
201 and PR101 in Fig. 13) is significantly enhanced when the
ink formulation contains a-cyclodextrin. The protection of the
Fig. 13 The effect of the incorporation of a-cyclodextrin into inks on
the reflected optical density loss of 5 on exposure in an Atlas Ci 5000
weatherometer when printed on alumina-coated (HG 201 and PR 101) or
silica coated (HP premium) papers.
azo-chromophore from photo-induced bleaching reactions by
formation of ternary dye–cyclodextrin–surface complexes could
account for this effect.5–7 In contrast, the incorporation of a-
cyclodextrin into ink formulations of 5 and of some of the other
dyes in Fig. 1 gave no improvement in ozone stability. Ozone
fastness appeared to be dependent much more on the nature
of the paper than on the ability of the simple dye to form
complexes with the a-cyclodextrin and or alumina in the paper.
These results contrast with work which has shown that formation
of cyclodextrin rotaxanes on the surface of cellulose leads to
improved ozone-fastness.5,21,22
Table 2 Humidity fastness values for 3–5 and 9 on a 1–10 scale, with 1
representing an undetectable migration of dye
Conclusions
The measurement of isotherms at high surface area oxides provides
a simple method to assess the efficacy of different surface ligating
groups which might be used to attach effect molecules to solid
supports and thus generate new functional materials. Systematic
comparisons of isotherms provides evidence for the formation of
Papers/dyes
3
4
5
9
HG 201 (alumina coated)
PR 101 (alumina coated)
PGPP (silica coated)
1
2–3
2
7
7
2–3
1
1
1
1
1
1
2792 | Dalton Trans., 2006, 2785–2793
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
ternary dye–cyclodextrin–surface complexes which have rotaxane-
like structures and offer the possibility of protecting the dye
or other types of guest molecule from chemical or photolytic
degradation in such functional materials, e.g. in light harvesting
devices.
10 T. C. Myers, R. G. Harvey and E. V. Jensen, J. Am. Chem. Soc., 1955,
77, 3101.
11 J. Suh and W. J. Kwon, Bioorg. Chem., 1998, 26, 103.
12 D. Pressman, M. Siegel and L. A. R. Hall, J. Am. Chem. Soc., 1954, 76,
6336.
13 P. Wight, Internal report, Avecia, Blackley, UK.
14 L. A. Cates, V. S. Li, C. C. Yakshe, M. O. Fadeyi, T. H. Andree, E. W.
Karbon and S. J. Enna, J. Med. Chem., 1984, 27, 654.
15 S. N. L. Bennett and R. G. Hall, J. Chem. Soc., Perkin Trans. 1, 1995,
1145.
Acknowledgements
The authors wish to thank Ms Patricia Dunwoody (Avecia), Dr
Juraj Bella, Mr John Millar and Mr Wesley Kerr (University of
Edinburgh) for technical support and we thank the EPSRC (DTA),
Avecia and the University of Edinburgh for funding.
16 D. D. Perrin and B. Dempsey, Buffers for pH and Metal Ion Control,
Chapman and Hall Ltd, London, 1974.
17 K. S. W. Sing, D. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierotti,
J. Rouquerol and T. Siemieniewska, Pure Appl. Chem., 1985, 57, 603.
18 C. H. Giles, D. Smith and A. Huitson, J. Colloid Interface Sci., 1974,
47, 755.
19 J. Szejtli, Encycl. Nanosci. Nanotechnol., 2004, 2, 283.
20 C. J. Easton, S. F. Lincoln, A. G. Meyer and H. Onagi, J. Chem. Soc.,
Perkin Trans. 1, 1999, 2501.
21 M. R. Craig, T. D. W. Claridge, H. L. Anderson and M. G. Hutchings,
Chem. Commun., 1999, 1537.
References
1 P. Gregory, Chem. Br., 2000, 36, 39.
2 J. F. Suthers, Annu. Technical Conf.–Soc. Plastics Eng. 58th, 2000, 3,
2593.
3 R. A. Gill, Appl. Wet-End Pap. Chem., 1995, 54.
4 S. S. De Silva, P. J. Camp, D. K. Henderson, D. C. R. Henry, H. McNab,
P. A. Tasker and P. Wight, Chem. Commun., 2003, 1702.
5 M. R. Craig, M. G. Hutchings, T. D. W. Claridge and H. L. Anderson,
Angew. Chem., Int. Ed., 2001, 40, 1071.
6 H. L. Anderson, M. R. Craig and M. G. Hutchings, WO Pat., 0224816,
2002.
7 E. Arunkumar, C. C. Forbes and B. D. Smith, Eur. J. Org. Chem., 2005,
4051.
22 J. E. H. Buston, J. R. Young and H. L. Anderson, Chem. Commun.,
2000, 905.
23 F. Quadrifoglio and V. Crescenzi, J. Colloid Interface Sci., 1971, 35,
447.
24 S. Li and W. C. Purdy, Chem. Rev., 1992, 92, 1457.
25 S. L. Mayo, B. D. Olafson and W. A. Goddard, III, J. Phys. Chem.,
1990, 94, 8897.
26 W. Saenger, K. Beyer and P. C. Manor, Acta Crystallogr., Sect. B, 1976,
32, 120.
8 A. H. A. Renfrew, Reactive Dyes for Textile Fibres, Society of Dyes and
Colourists, Bradford, UK, 1999.
9 C. E. Maxwell, Org. Synth., 1943, 23, 30.
27 K. Vikman, J. Imag. Sci. Technol., 2003, 47, 30.
28 A. Naisby, J. Suhadolink, A. Debellis, D. Pennant and W. Renz, NIP17
Int. Conf. Digital Printing Technol., 2001, 179–185.
This journal is
The Royal Society of Chemistry 2006
Dalton Trans., 2006, 2785–2793 | 2793
©