Building block syntheses of gallic acid monomers and tris-(O-gallyl)-gallic acid dendrimers chemically attached to graphite powder: A comparative study of their uptake of Al(III) ions

Article abstract of DOI:10.1021/la902497s

A synthesis of graphite powder covalently modified with, gallic acid (3,4,5-trihydroxybenzoic acid), via a 1,2diaminoethane "linker" molecule, to form gallylaminoethylaminocarbonyl graphite (gallic-carbon) is reported. The synthesis was used as a model for a "ground-upwards building-block" approach to a primary dendrimer of gallic acid covalently attached to the surface of graphite powder, tris-(O-gallyl)- gallylaminoethylaminocarbonyl graphite (TGGAcarbon). The resulting modified carbon materials were characterized, at each stage of the syntheses using X-ray photoelectron spectroscopy (XPS) analysis, The effects of increasing the modifier's structural complexity from monomelic gallic-carbon to the analogous primary dendrimer TGGA-carbon were explored by comparing each, material's efficacy toward, the adsorption of Al(III) ions from water. The uptake of Al(III) ions by gallic-carbon and TGGA-carbon was measured using UV-vis spectroscopy. In comparison to the case of monomelic gallic-carbon, the rate of adsorption of A1(III) ions by the TGGA-carbon was found to be 2,3 times more rapid. Furthermore, the total uptake of Al(III) ions was greater (reducing the concentration of 1000 ppb Al(III) solutions to below the WHO legal limit of 100 ppb in less than 5 min) and irreversible, in contrast to the gallic-carbon where the adsorption was found to be under thermodynamic control and to follow a Freundlich isotherm

Full text of DOI:10.1021/la902497s

pubs.acs.org/Langmuir
© 2009 American Chemical Society
Building Block Syntheses of Gallic Acid Monomers and Tris-(O-gallyl)-gallic
Acid Dendrimers Chemically Attached to Graphite Powder: A Comparative
Study of Their Uptake of Al(III) Ions
Junju Ye,† Poobalasingam Abiman,† Alison Crossley,‡ John H. Jones,§ Gregory G. Wildgoose,*,†
and Richard G. Compton*,†
†Department of Chemistry, Physical and Theoretical Chemistry Laboratory, University of Oxford, South Parks
Road, Oxford, OX1 3QZ United Kingdom, ‡Materials Department, University of Oxford, Parks Road, Oxford,
OX1 3PH United Kingdom, and §Chemical Research Laboratory, University of Oxford, Parks Road, Oxford,
OX1 3TA United Kingdom
Received July 10, 2009. Revised Manuscript Received September 2, 2009
A synthesis of graphite powder covalently modified with gallic acid (3,4,5-trihydroxybenzoic acid), via a 1,2-
diaminoethane “linker” molecule, to form gallylaminoethylaminocarbonyl graphite (gallic-carbon) is reported. The
synthesis was used as a model for a “ground-upwards building-block” approach to a primary dendrimer of gallic acid
covalently attached to the surface of graphite powder, tris-(O-gallyl)-gallylaminoethylaminocarbonyl graphite (TGGA-
carbon). The resulting modified carbon materials were characterized at each stage of the syntheses using X-ray
photoelectron spectroscopy (XPS) analysis. The effects of increasing the modifier’s structural complexity from
monomeric gallic-carbon to the analogous primary dendrimer TGGA-carbon were explored by comparing each
material’s efficacy toward the adsorption of Al(III) ions from water. The uptake of Al(III) ions by gallic-carbon and
TGGA-carbon was measured using UV-vis spectroscopy. In comparison to the case of monomeric gallic-carbon, the
rate of adsorption of Al(III) ions by the TGGA-carbon was found to be 2.3 times more rapid. Furthermore, the total
uptake of Al(III) ions was greater (reducing the concentration of 1000 ppb Al(III) solutions to below the WHO legal
limit of 100 ppb in less than 5 min) and irreversible, in contrast to the gallic-carbon where the adsorption was found to be
under thermodynamic control and to follow a Freundlich isotherm.
1. Introduction
begun to explore methods of covalently attaching different
organic molecules, such as amino acid derivatives,13-16 polypep-
tides,17 aromatic compounds,18,19 and amino-terminated poly-
ethers (“Jeffamines”)20 to graphite microparticles. The resulting
materials have enabled us to explore various physicochemical
properties such as factors influencing the observed changes in
surface pKa of the modifying molecules,20-22 and to ascertain
kinetic23 and thermodynamic parameters24,25 controlling the
adsorption of several metal ions of environmental concern
such as As(III), Cu(II), Cd(II), and Hg(II) by these chemically
Currently, there is much interest in using chemically modified
carbon-based materials for a range of applications including, but
not limited to, sensors1-3 and biosensors,2-4 catalyst supports,5
and environmental remediation.6-8 The surface chemistry of
graphitic carbon materials is rich, with many synthetically useful
functional groups (e.g., carboxyl, hydroxyl or quinonyl moieties)
either naturally present or readily introducible.9-12 The presence
of such groups makes carbon a potentially useful, inexpensive
support material for solid-phase syntheses. Recently, we have
(14) Sljukic, B.; Wildgoose, G. G.; Crossley, A.; Jones, J. H.; Jiang, L.; Jones,
T. G. J.; Compton, R. G. J. Mater. Chem. 2006, 16, 970–976.
(15) Abiman, P.; Wildgoose, G. G.; Crossley, A.; Compton, R. G. J. Mater.
Chem. 2008, 18, 3948–3953.
(16) Abiman, P.; Wildgoose, G. G.; Crossley, A.; Compton, R. G. Electro-
analysis 2009, 21, 897–903.
*Corresponding authors. E-mail: gregory.wildgoose@chem.ox.ac.uk
(G.G.W.); richard.compton@chem.ox.ac.uk (R.G.C.). Telephone: þ44 (0)1865
275406 (G.G.W.); þ44 (0)1865 275413 (R.G.C.). Fax: þ44 (0)1865 275410
(1) Wildgoose, G. G.; Banks, C. E.; Leventis, H. C.; Compton, R. G. Microchim.
Acta 2006, 152, 187–214.
(2) Gooding, J. J. Electroanalysis 2008, 20, 573–582.
(17) Wildgoose, G. G.; Leventis, H. C.; Davies, I. J.; Crossley, A.; Lawrence,
N. S.; Jiang, L.; Jones, T. G. J.; Compton, R. G. J. Mater. Chem. 2005, 15, 2375–
2382.
(18) Leventis, H. C.; Streeter, I.; Wildgoose, G. G.; Lawrence, N. S.; Jiang, L.;
Jones, T. G. J.; Compton, R. G. Talanta 2004, 63, 1039–1051.
(19) Abiman, P.; Wildgoose, G. G.; Compton, R. G. J. Phys. Org. Chem. 2008,
21, 433–439.
(3) Wring, S. A.; Hart, J. P. Analyst 1992, 117, 1215–1229.
(4) Compton, R. G.; Wildgoose, G. G.; Wong, E. L. S. In Biosensing Using
Nanomaterials; Merkoc-i, A., Ed.; Wiley: Hoboken, NJ, 2009.
(5) Wildgoose, G. G.; Banks, C. E.; Compton, R. G. Small 2006, 2, 182–193.
(6) Jiang, L.; Liu, L.; Luo, X.; Niu, J.; Lu, G. Huagong Shikan 2007, 21, 64–67.
(7) Moore, E.; Sanvicens, N.; Pravda, M.; Guilbault, G. G. Int. J. Environ. Anal.
Chem. 2003, 83, 545–553.
(20) Abiman, P.; Wildgoose, G. G.; Crossley, A.; Jones, J. H.; Compton, R. G.
(8) Yang, B.-Y.; Mo, J.-Y.; Lai, R. Gaodeng Xuexiao Huaxue Xuebao 2005, 26,
227–230.
(9) Wildgoose, G. G.; Abiman, P.; Compton, R. G. J. Mater. Chem. 2009, 19,
4875–4886.
(10) Masheter, A. T.; Xiao, L.; Wildgoose, G. G.; Crossley, A.; Jones, J. H.;
Compton, R. G. J. Mater. Chem. 2007, 17, 3515–3524.
Chem.;Eur. J. 2007, 13, 9663–9667.
(21) Abiman, P.; Crossley, A.; Wildgoose, G. G.; Jones, J. H.; Compton, R. G.
Langmuir 2009, 25, 7768.
(22) Masheter, A. T.; Abiman, P.; Wildgoose, G. G.; Wong, E.; Xiao, L.; Rees,
N. V.; Taylor, R.; Attard, G. A.; Baron, R.; Crossley, A.; Jones, J. H.; Compton,
R. G. J. Mater. Chem. 2007, 17, 2616–2626.
(11) Thorogood, C. A.; Wildgoose, G. G.; Crossley, A.; Jacobs, R. M. J.; Jones,
J. H.; Compton, R. G. Chem. Mater. 2007, 19, 4964–4974.
(12) Thorogood, C. A.; Wildgoose, G. G.; Jones, J. H.; Compton, R. G. New J.
Chem. 2007, 31, 958–965.
(13) Wildgoose, G. G.; Leventis, H. C.; Simm, A. O.; Jones, J. H.; Compton,
R. G. Chem. Commun. 2005, 3694–3696.
(23) Chevallier, F. O. G.; Sljukic, B.; Wildgoose, G. G.; Jiang, L.; Jones, T. G. J.;
Compton, R. G. ChemPhysChem 2006, 7, 697–703.
(24) Sljukic, B.; Wildgoose, G. G.; Crossley, A.; Jones, J. H.; Li, J.; Jones,
T. G. J.; Compton, R. G. J. Mater. Chem. 2006, 16, 970–976.
(25) Xiao, L.; Wildgoose, G. G.; Crossley, A.; Knight, R.; Jones, J. H.;
Compton, R. G. Chem.;Asian J. 2006, 1, 614–622.
1776 DOI: 10.1021/la902497s
Published on Web 09/24/2009
Langmuir 2010, 26(3), 1776–1785

Ye et al.
Article
modified carbon materials from aqueous laboratory solutions
and from “real” river and borehole water samples.
not less than 18.2 MΩ cmat25 °C was used. Nonaqueoussolvents
were dried over 5 A molecular sieves prior to use. A 100 mg L
-1
˚
aluminum(III) stock solution was prepared by dissolving alumi-
num ammonium disulfate dodecahydrate (<99.5%, AnalaR) in
deionized water. The stock solution was diluted as required.
Voltammetric measurements were performed using a type II
μ-Autolab (EcoChemie, Utrecht, The Netherlands) computer-
controlled potentiostat. All experiments were conducted in a
thermostatted (22 °C) three-electrode cell with a solution volume
of 20 cm3 and a three-electrode configuration. The working
electrode consisted of a basal-plane pyrolytic graphite electrode
(bppg, Le Carbon, U.K.). The reference electrode was a saturated
calomel reference electrode (SCE, Radiometer, Copenhagen,
Denmark), and a clean bright platinum coil (99.99% Goodfellow,
U.K.) acted as the counter electrode.
X-ray photoelectron spectroscopy (XPS) was performed on a VG
Clam 4 MCD analyzer system at the OCMS Begbroke Science
Park, University of Oxford, U.K. using X-ray radiation from the
Mg KR band (hυ=1253.6 eV). All XPS experiments were recorded
using an analyzer energy of 100 eV for survey scans and 20 eV for
detail scan with a takeoff angle of 90°. The base pressure in the
analysis chamber was maintained at not more than 2.0 ꢀ 10-9 mbar.
Each carbon sample studied was mounted on a stub using double
sided adhesive tape and then placed in the ultrahigh vacuum analysis
chamber of the spectrometer. Analysis of the resulting spectra was
performed using Microcal OriginPro 8. Assignment of spectral
peaks was performed using the UKSAF32 and NIST33 databases.
pH measurements were made using a pH213 pH meter (Hanna
instruments) calibrated using reference buffer solutions of pH
4.01 ( 0.01 and pH 7.00 ( 0.01 (Hamilton).
The focus of this report is twofold. First, we seek to further
develop a “building block” synthetic strategy whereby carbon
surfaces are covalently modified with “linker” molecules from
which the adsorbent molecules can be attached and built in a
“ground upwards” approachto formever morecomplex chemical
architectures. In doing this, we seek to demonstrate the potential
utility of using graphite materials as a support in complex solid-
phase synthesis. Second, we seek to examine what are the effects
of moving from a monomeric adsorbent attached to an inert
graphite carrier to the more complex analogous polymeric or, in
this case, primary dendritic structure. How are the thermo-
dynamics and kinetics of adsorption affected? Can we predict
the adsorption behavior on moving from a monomeric adsorbent
to a dendritic one as has been suggested for certain specific cases
by, for example, Crooks and co-workers?26,27 How do the
monomeric and dendritic adsorbent materials compare with
one another?
The model modifying compound chosen for this work is gallic
acid (3,4,5-hydroxybenzoic acid). Gallic acid is a polyphenolic
antioxidant found in many natural products, and like many
structurally related polyphenolic compounds, such as pyrocate-
chol violet (PCV) and tannic acid,28,29 we demonstrate herein that
it is capable of forming complexes with Al(III) ions in aqueous
media. Thus, it fulfills our requirements in that the gallic acid
molecule can be attached to the carbon surface by coupling the
carboxyl group to a 1,2-diaminoethane linker molecule to form
gallylaminoethylaminocarbonyl carbon (gallic-carbon). The re-
maining hydroxyl groups on the gallic acid moiety then allow for
further attachment of gallyl moieties to build up a primary
dendritic structure (tris-(O-gallyl)-gallylaminoethylaminocarbo-
nyl carbon, TGGA-carbon) which is structurally related to tannic
acid. The ability of both the gallic-carbon and TGGA-carbon to
bind Al(III) ions allows us to then ascertain and compare the
factors affecting each material’s adsorption behavior. Note that,
although we demonstrate that both materials are capable
of rapidly reducing relatively high concentrations of Al(III)
ions in water to below the World Health Organization (WHO)
legal limits of 100 ppb and 200 ppb set for large and small
scale drinking water treatment plants, respectively,30 we are not
claiming to have developed a material for water treatment
applications. The laboratory samples used herein are ideal and
are unrepresentative of the complex speciation and aquatic
chemistry of Al(III) contaminants found in ground and waste-
water samples.
UV-vis spectroscopy was performed using a model Evolution
60 UV-vis spectrophotometer with VISIONlite Quant version 2.2
software (Thermoscientific). Fourier transform infrared (FT-IR)
spectroscopy was performed on a Paragon 1000 (Perkin-Elmer)
€
1
instrument using KBr plates and nujol as the solvent. H NMR
spectroscopy was performed on a Bruker Spectrospin 300
(300 MHz) instrument.
2.2. Synthetic Methods. 3,4,5-Triacetoxybenzoic Acid.
The synthesis of acetyl-protected gallic acid was adapted from the
method of Turner et al.34 In a 250 mL round-bottom flask, with a
magnetic stirrer, were combined gallic acid (5.0 g, 29 mmol) and
aceticanhydride (17mL, 176 mmol, excess). The slurrywasstirred
as a catalytic amount of sulfuric acid (32 μL) was added. The
temperature rose rapidly from 21 to 75 °C over about 1 min, and
the slurry became a clear yellow solution. The mixture was stirred
and allowed to cool to room temperature over 20 min. Next
100 mL of water was added to the flask to remove any excess
aceticanhydride. Afterfurtherstirringfor 2.5h, a whitecrystalline
product was isolated by filtration and further washed with 3 ꢀ
20 mL aliquots of water. The acetyl-protected gallic acid product
was dried in a stream of air for 10 min and then vacuum-dried
overnight. From 5.0 g (29 mmol) of gallic acid starting material,
we obtained acetyl-protected gallic acid (8.0 g, 93%) as a white
crystalline solid, mp 166 °C (lit. 166-168 °C);34 FT-IR (KBr)
2. Experimental Section
2.1. Reagents and Equipment. All chemicals were of analy-
tical grade and used as obtained from Sigma-Aldrich (U.K.),
Fluka (U.K.), or AnalaR (U.K.) without further purification.
Synthetic graphite powder, consisting of irregularly shaped par-
ticles between 2 and 20 μm in size (measured along the largest
axis),31 was purchased from Aldrich (U.K.). All aqueous solu-
tions were prepared using deionized water from Millipore
(Vivendi, U.K.) A UHQ grade water system with a resistivity of
ν
max/cm-1 1788, 1694, and 1592 (CO);35 δH (300 MHz; CDCl3;
Me4Si) 2.32 (9 H, s, 3 ꢀ OAc), 7.90 (2H, s, 2 ꢀ ArH);34 the proton
adjacent to the carboxylic group gave a broad peak >12.5.34
Note that the gallic acid starting material does not dissolve
in CDCl3.
Methyl Gallate. The synthesis of methyl gallate was adapted
from the work of Alam et al.36 and Zhao et al.37 Gallic acid
(26) Niu, Y.; Sun, L.; Crooks, R. M. Macromolecules 2003, 36, 5725–5731.
(27) Sun, L.; Crooks, R. M. J. Phys. Chem. B 2002, 106, 5864–5872.
(28) Haslam, E.; Mills, S. D.; Armitage, R.; Haworth, R. D.; Rogers, H. J.;
Searle, T. J. Chem. Soc. 1961, 1836.
(29) Wilson, A. D.; Sergeant, G. A. Analyst 1963, 88, 109.
(30) Aluminium in Drinking-water; W.H.O.: 1998; http://www.who.int/water_
sanitation_health/dwq/chemicals/en/aluminium.pdf.
(32) http://www.uksaf.org/.
(33) http://srdata.nist.gov/xps/.
(34) Turner, S. R.; Voit, B. I.; Nielsen, R. B. U.S. Patent 5196502; 1993:496477;
CAN 119:96477, 1993.
(35) Kutani, N. Chem. Pharm. Bull. 1960, 8, 72–76.
(36) Alam, A.; Takaguchi, Y.; Ito, H.; Yoshida, T.; Tsuboi, S. Tetrahedron 2005,
61, 1909–1918.
(31) Wildgoose, G. G.; Wilkins, S. J.; Williams, G. R.; France, R. R.; Carnahan,
(37) Zhao, Y.; Hao, X. J.; Lu, W.; Cai, J. C.; Yu, H.; Sevenet, T.; Gueritte, F.
D. L.; Jiang, L.; Jones, J. H.; Compton, R. G. ChemPhysChem 2005, 6, 352–362.
J. Nat. Prod. 2002, 65, 902–908.
Langmuir 2010, 26(3), 1776–1785
DOI: 10.1021/la902497s 1777

Article
Ye et al.
(5.0 g, 29mmol) wasdissolved in dry methanol (150mL), towhich
a catalytic amount (0.3 mL) of concentrated sulfuric acid was
carefully added. The solution was heated under reflux for 24 h at
65 °C, after which time the solution was cooled to room tempera-
ture and 2.0 M sodium hydroxide was added dropwise until the
excessacidhad been neutralized. The solvent was removed using a
rotary evaporator to yield crude methyl gallate as a white crystal-
line solid. In order to separate the methyl gallate from
any unreacted gallic acid impurities, the product was dis-
solved in a mixture of ethyl acetate (60 mL) and pure water
(20 mL). The organic layer containing the methyl gallate was
separated, further washed with pure water (2 ꢀ 20 mL) and brine
(40 mL), and then dried over magnesium sulfate. The solvent was
removed under vacuum, and the product was recry-
stallized from hot water, before being dried on a high vacuum
line to give methyl gallate (2.9 g, 53.5%) as a white crystalline
NMR (DMSO-d6) δH (300 MHz; DMSO-d6; Me4Si) 9.18 (9 H, br
s, 9 ꢀ OH), 7.80 (2 H, s, 2 ꢀ ArH), 6.63-7.27 (6 H, m, 6 ꢀ ArH),
3.73 (3 H, s, CO2CH3).40,41
1,2-Diaminoethane-Modified Graphite Powder. Graphite
powder (1.2 g, 0.1 mol) was stirred in concentrated nitric acid
(21 mL, 0.315 mol) and sulfuric acid (7 mL, 0.126 mol), that is, a
3:1 ratio, for 12 h, washed with a sufficient quantity of pure water
until the washings ran neutral, and dried under vacuum. The
surface carboxyl groups, 1, were then converted to more reactive
carboxylic acid chloride intermediates, 2, by stirring the oxidized
graphite powder (1.0 g) in thionyl chloride (15 mL, 0.206 mol) at
room temperature for 90 min under a nitrogen atmosphere. The
excess thionyl chloride was removed under reduced pressure in a
rotary evaporator. Next 1,2-diaminoethane (1 mL, 15 mmol) and
ethyl diisopropylamine (1.5 mL, 9 mmol) were added to a beaker
containing dry dioxane (25 mL), and the solution was then added
to the carboxylic acid chloride modified carbon powder in the
rotary evaporator (note that the reaction was performed at room
temperature to avoid the evaporation of dioxane). After 15 min,
the reaction mixture was transferred into a round-bottomed flask
with gentle stirring for 18 h under an argon atmosphere to
form the 1,2-diaminoethane-modified carbon 3. This was then
washed with dry dioxane and deionized water and dried under
vacuum.
solid, mp 185-188 °C (lit. 188-191 °C38); FT-IR (KBr) νmax
/
cm-1 3340br (OH) 1700, 1618, and 1533(CO);39 δH (300 MHz;
DMSO-d6; Me4Si) 9.28(3 H, brs, 3 ꢀ OH), 6.91(2 H, s, 2 ꢀ ArH),
3.75 (3 H, s, CO2CH3).36,37
Methyl Tris-O-(tri-O-acetylgallyl)-gallate. The method des-
cribed herein is adapted from the work of Haslam et al.28 and
Nomura et al.40 3,4,5-Triacetoxybenzoic acid (2.5 g, 8.43 mmol)
was dissolved in chloroform (40 mL), to which excess thionyl
chloride (10 mL, 140 mmol) was added. The solution was then
heated under reflux at 62 °C for 90 min, after which the chloro-
form solvent was removed on a rotary evaporator. Any residual
thionyl chloride was removed by washing the crude product with
dry toluene (2 ꢀ 20 mL). The residue thionyl chloride was
removed with toluene on a rotary evaporator to give 3,4,5-
triacetoxybenzoic chloride as a white crystalline product (2.65 g,
100%). Next, excess 3,4,5-triacetoxybenzyl chloride (3.15 g,
10 mmol) was dissolved in dry dioxane (60 mL) under a nitrogen
atmosphere together with methyl gallate (0.46 g, 2.5 mmol) and
pyridine(3 mL, 0.037 mmol). The reactionsolution was stirred for
4 days, after which time the solvent was removed using a rotary
evaporator, and any residual pyridine was removed by washing
with dry toluene. The crude product was dissolved in chloroform
(60 mL) and washed with 2.0 M NaHSO4 (20 mL) followed by
pure water (2 ꢀ 10 mL) and finally brine (40 mL). The organic
layer was separated from the aqueous layer and dried over
anhydrous sodium sulfate. The solvent was removed under
reduced pressure to give methyl tris-O-(tri-O-acetylgallyl)-gallate
(2.83 g, 85%) as a pale orange crystalline solid,28,40 mp 55-58 °C
(from chloroform); FT-IR (KBr) νmax/cm-1 1772 (CO); δH (300
MHz; CDCl3; Me4Si) 7.78-8.04 (8 H, m, 8 x ArH), 3.73 (3 H, s,
OCH3), 2.26-2.35 (27 H, m, 9 ꢀ OAc).
Gallic-Carbon (Gallylaminoethylaminocarbonyl Carbon).
The acetyl-protected gallic acid (3,4,5-triacetoxybenzoic acid,
0.5 g, 1.69 mmol) was converted to the reactive acid chloride
derivative using thionyl chloride (15 mL, 0.21 mol) with gentle
stirring at room temperature for 90 min under a nitrogen
atmosphere. The excess thionyl chloride was again removed
under reduced pressure in a rotary evaporator to yield 3,4,5-
triacetoxybenzoyl chloride as a pale yellow crystalline solid.
Next the 3,4,5-triacetoxybenzoyl chloride, obtained above, was
dissolved in dry dioxane (25 mL) in a round-bottomed flask,
and then ethyl ethyldiisopropylamine (1.5 mL, 9 mmol) and
1.0 g of the 1, 2-diaminoethane-modified carbon powder was
added. The reaction mixture was stirred under argon for 18 h,
before being filtered, and the solid powder was then washed
with dry dioxane followed by deionized water and finally dried
under vacuum to yield acetyl-protected gallic-carbon, 6, as the
product.
Finally the acetyl-protectinggroupswere removed bytreating6
with 0.25 M NaOH solution (25 mL) for 1 h at 58 °C. The product
was filtered under suction and washed with copious quantities of
water and then acetonitrile, before being dried under vacuum to
yield the product, gallylaminoethylaminocarbonyl carbon (gallic-
carbon) 7.
Methyl Tris-(O-gallyl)-gallate. This procedure was modified
from the work of Nomura et al.40 and Ramesh et al.41 Methyl tris-
O-(tri-O-acetylgallyl)-gallate (2.0 g, 2 mmol) was added to aqu-
eous methanol (1:4 v/v water/methanol, 20 mL) containing
ammonium acetate (618 mg, 8.0 mmol). After 6 h, the methyl
tris-O-(tri-O-acetylgallyl)-gallate completely dissolved, and the
solution was stirred for a further 21 h. After removal of the
solvent, the crude product was extracted using aliquots of
ethyl acetate (3 ꢀ 20 mL). The combined extracts were then dried
over anhydrous sodium sulfate, and the solvent removed under
reduced pressure to yield pale yellow crystals together with
an orange oil. FT-IR and NMR characterization revealed that this
product contained the desired product, methyl tris-(O-gallyl)-gallate
(Me-TGG), together with some residual impurities in the orange
Tris-O-(tri-O-acetylgallyl)-gallylaminoethylaminocarbonyl
Carbon. A total of 2.5 g (8.43 mmol) of 5 was activated by
converting the carboxylic acid group to the corresponding acid
chloride as described above. The yellow crystalline product of
3,4,5-triacetoxybenzoyl chloride was dissolved in dry dichloro-
methane (20 mL). Gallic-carbon (10 g) was suspended in dry
dichloromethane (40 mL) containing ethyldiisopropylamine
(1.5 mL) under an inert nitrogen atmosphere for 10 min, after
which the solution of 3,4,5-triacetoxybenzoyl chloride was
slowly added. The reaction suspension was stirred at room
temperature for 4 days, after which the modified carbon powder
was filtered off and washed with copious quantities of dry
dichloromethane, water and acetonitrile, beforebeing dried under
vacuum overnight.
1
oil.40,41 FT-IR (KBr) νmax/cm-1 3400br (OH), 1698 (CO); H
Tris-(O-gallyl)-gallylaminoethylaminocarbonyl Modified
Carbon (TGGA-carbon). A total of 6 g of the resulting tris-
(38) Kurkin, V. A.; Zapesochnaya, G. G.; Krivenchuk, P. E.; Yurenik, A. Y.;
Artamonova, L. P. Chem. Nat. Prod. 1984, 20, 367–368.
O-(tri-O-acetylgallyl)-gallylaminoethylaminocarbonyl
carbon
(39) Kuroyanagi, M.; Yamamoto, Y.; Fukushima, S.; Ueno, A.; Noro, T.;
Miyase, T. Chem. Pharm. Bull. 1982, 30, 1602–1608.
(40) Nomura, E.; Hosoda, A.; Morishita, H.; Murakami, A.; Koshimizu, K.;
Ohigashi, H.; Taniguchi, H. Bioorg. Med. Chem. 2002, 10, 1069–1075.
(41) Ramesh, C.; Mahender, G.; Ravindranath, N.; Das, B. Tetrahedron 2003,
59, 1049–1054.
was treated with ammonium acetate to remove the O-acetyl
protecting groups as described for the solution-phase synthesis
of Me-TGG above. The powder was thenfilteredoff, washedwith
copious quantities of acetonitrile, chloroform, and pure water so
as to remove any unreacted species from the carbon surface, and
1778 DOI: 10.1021/la902497s
Langmuir 2010, 26(3), 1776–1785

Ye et al.
Article
in section 2.2. In order to improve the coverage of the linker
groups on the graphite microparticles, carboxyl groups were first
introduced onto the graphite surface via the standard method of
acid oxidation.9,44 This pretreatment is known to increase the
surface coverage of carboxyl groups on the graphite powder by as
much as between 6 and 10 times that naturally present, corre-
sponding to between ca. 1 ꢀ 10-11 and 2 ꢀ 10-11 mol cm-2 9,10
.
Before coupling the gallic acid groups to the 1,2-diamino-
ethane-modified carbon, it is first necessary to protect the hydro-
xyl groups in the gallic acid molecule to prevent unwanted
polymerization and side reactions. To this end gallic acid, 4,
was converted to 3,4,5-triacetoxybenzoic acid, 5, as described in
section 2.2. The acetyl-protected gallic acid (0.5 g, 1.69 mmol) was
converted to the reactive acid chloride derivative using thionyl
chloride and coupled to the 1,2-diaminoethane-modified carbon
powder before the acetyl protecting groups were removed to yield
gallylaminoethylaminocarbonyl carbon (gallic-carbon, 7,), as
detailed in section 2.2 and Scheme 1.
3.2. Characterization of Gallic-Carbon Powder Using
X-ray Photoelectron Spectroscopy. In order to confirm that
each stage of the synthesis of gallic-carbon was successful, X-ray
photoelectron spectroscopy(XPS) was carriedout on the graphite
power prior to any modification, the oxidized graphite 1, the 1,2-
diaminoethane-modified carbon 3, acetyl-protected gallic acid
modifiedcarbon6, and finally the gallic-carbonproduct 7. In each
case, a wide survey scan was first performed over the range
0-1200 eV for each sample. Next one detailed scan was per-
formed over the C1s region, 10 scans over the N1s region, and
finally 10 scans over the O1s region. In some cases, small peaks
could be observed corresponding to the trace presence of Ca and
Si impurities, most likely arising from the borosilicate glassware
during the synthesis. The atomic percentages of C, O, and N
elements at each stage of the synthetic procedure are listed in the
Table 1.
Figure 2a shows the XPS spectrum of unmodified graphite
powder. The two large peaks could be clearly observed at 285 and
533 eV corresponding to the C1s and O1s emissions, respectively.
Repeat scans over the O1s region (not shown) revealed a broad
peak (530-536 eV) which could not be quantitatively deconvo-
luted but which qualitatively includes contributions from hydro-
xyl, quinonyl, and carboxyl surface groups on the carbon
surface.9 A quantitative analysis revealed that the atomic percen-
tage of oxygen on the oxidized graphite surface (Figure 2b)
increased to 5.6%, compared to 3.2% in the unmodified graphite
(Figure 2a). After modification with 1,2-diaminoethane, a new
peak corresponding to emission from the N1s level could be
observed which was not present in either the blank or oxidized
graphite samples. The atomic percentage of nitrogen on the
surface was found to be 3.7%.
After coupling the tri-O-acetyl-protected gallic acid 6 onto the
carbon surface (Figure 2c), a corresponding decrease of the
relative atomic percentage of nitrogen was observed, with a
corresponding increase in the relative atomic percentage
of oxygen atoms, consistent with the stoichiometry of the
tri-O-acetyl-protected gallic acid molecule. After deprotecting
the acetyl groups, one would expect a corresponding decrease in
the percentage of oxygen on the surface. Note that Table 1 only
reports the absolute atomic percentage of each element, but it is
apparent that the percentage of oxygen relative to nitrogen has
indeed decreased, which is again consistent with the stoichiometry
expected from the final product, gallic-carbon.
Figure 1. Calibration plot made using standard additions of Al-
(III) of the log of the adsorption intensity (relative to pure water)
recorded using UV-vis spectroscopy at a wavelength of 580 nm.
dried to yield the tris-(O-gallyl)-gallylaminoethylaminocarbonyl-
modified carbon (TGGA-carbon).
2.3. UV-Vis Spectroscopic Analytical Protocols for
Measuring the Concentration of Al(III) Ions. The concentra-
tionofaluminum(III) ionswasdeterminedusing UV-visspectro-
metry by adding excess pyrocatechol violet (PCV) as a
colorimetric complexing agent in pH 6.1 buffer, according to
literature methods:29,42,43 The pH 6.1 buffer solution used for
UV-vis experiment was prepared by adjusting the pH of a 1.0 M
ammonium acetate solution with acetic acid. A stock solution of
0.15% w/w pyrocatechol violet (PCV) indicator for the UV-vis
spectroscopic experiments was prepared by dissolving 0.015 g of
PCV in 100 mL of deionized water.
A total of 500 μL of PCV and 2.5 mL of buffer solution were
added to a 25 mL volumetric flask. Aliquots of either standard
additions of known Al(III) concentrations or the sample solutions
containing an unknown concentration of Al(III) ions were added,
and the solution diluted to 25 mL using pure water, shaken, and
allowed to stand for 2 h at 25 °C. Note that the sample solutions
were diluted by a factor of 2-25 times depending on the initial
Al(III) concentration used so that the concentration of Al(III)
ions fell within the linear detection range.29
The maximum absorption intensity in the UV-vis spectrum
was observed at 580 nm for the Al3þ-PCV complex and at
450 nm for the free PCV. The adsorption intensity of the
Al(III)-PCV complex at 580 nm was measured relative to
deionized water in a 1 cm3 plastic cuvette and was found to obey
the Beer-Lambert Law. The Al(III) ion concentration was then
determined by comparison to a calibration plot made using
known standard concentrations of Al(III) (Figure 1). The inten-
sity of adsorption was found to be proportional to the Al(III)
concentration over the range 0-400 ppb, with a limit of detection
(determined from 3σ) of <1 ppb. The values of Al(III) ion
concentration reported herein that were determined using this
method are given as the mean of three repeated measurements in
all cases.
3. Results and Discussion
The following sections describe the building block syntheses
and subsequent characterization of gallic-carbon and TGGA-
carbon before each material’s behavior toward the adsorption of
Al(III) ions is compared.
3.1. Chemical Modification of Graphite Powder Surface
with Gallic Acid. The first step in the building block synthesis of
gallic-carbon (shown in Scheme 1) is to functionalize the graphite
support with linker molecules of 1,2-diaminoethane, as described
(42) Anton, A. Anal. Chem. 1960, 32, 725–726.
(43) Zhu, X. S.; Bao, L.; Guo, R.; Wu, J. Anal. Chim. Acta 2004, 523, 43–48.
(44) Marcolino, L. H.; Janegitz, B. C.; Lourencao, B. C.; Fatibello, O. Anal.
Lett. 2007, 40, 3119–3128.
Langmuir 2010, 26(3), 1776–1785
DOI: 10.1021/la902497s 1779

Article
Ye et al.
Scheme 1. Synthesis of Gallic-Carbon Powder
Table 1. Atomic Percentage Elemental Composition on the Surface of
Graphite Powders at Each Stage of the Synthesis of Gallic-Carbon,
Determined Using XPS
0.37 V versus SCE, in agreement with the literature.45 In the case
of the gallic-carbon, a reassuringly similar, irreversible oxidation
wave is observed upon the first scan at 0.32 V, suggesting that we
have successfully modified the carbon surface with gallic acid groups
as proposed. Note that the voltammetry of 6 either in solution or
attached to the carbon surface did not give rise to any voltammetric
signal. The appearance of the irreversible oxidation wave at 0.32 V
in the case of gallic-carbon after removal of the acetyl protecting
groups provides further evidence to suggest that the deprotection
step was successful and resulted in the formation of gallic-carbon.
3.4. Chemical Modification of Graphite Powder with
% element surface composition
sample
N
O
C
unmodified graphite
oxidized graphite
0.0
0.0
3.2
5.6
5.1
7.2
7.9
96.9
94.4
91.2
91.9
90.3
1,2-diaminoethane-modified graphite 3.7
tri-O-acetyl gallic modified graphite
gallic-carbon
1.0
1.8
Tris-(O-gallyl)-gallylaminoethylaminocarbonyl Using
a
3.3. Characterization of Gallic-Carbon Powder Using
Cyclic Voltammetry. In addition to performing XPS characteri-
zation, we also carried out characterization of the gallic-carbon,
abrasively immobilized onto a bppg working electrode, by compar-
ing its voltammetric response to that of 5 mM gallic acid in pH 4.5
acetate buffer solution (Figure 3). In the case of gallic acid in
solution (Figure 3b), the main voltammetric feature of interest is a
large, electrochemically irreversible oxidation wave observed at
“Building Block” Strategy. In order to synthesize the primary
dendrimer of gallic acid, as derivatives of tris-(O-gallyl)-gallic acid
(TGGA), it is again necessary to use O-acetyl protecting groups
on the phenolic groups of the incoming gallyl substrate, and to
protect the carboxylic acid group of the starting gallic acid
(45) Kilmartin, P. A.; Zou, H. L.; Waterhouse, A. L. J. Agric. Food Chem. 2001,
49, 1957–1965.
1780 DOI: 10.1021/la902497s
Langmuir 2010, 26(3), 1776–1785

Ye et al.
Article
molecule. This is necessary in order to prevent any unwanted or
uncontrolled self-condensation reactions and polymerization to
form dendrimers closely related to tannic acid derivatives. The
difficulty then lies in selectively removing the O-acetyl protecting
groups while leaving the gallyl-gallate linkages intact. We there-
fore decided to check the selectivity of our deprotection method in
a solution phase synthesis of methyl tris-(O-gallyl)-gallate (Me-
TGG), where characterization of the obtained product using
1
standard H NMR and FT-IR techniques could be performed.
The synthesis of Me-TGG is shown in Scheme 2 and detailed in
section 2.2. Characterization of the resulting Me-TGG product
(section 2.2) confirmed that we had successfully synthesized Me-
TGG, and that the mild hydrolysis conditions used were selective
for the removal of the O-acetyl protecting groups in preference to
cleavage of the gallyl-gallate ester linkages.41
Having confirmed that our proposed synthetic strategy allows
us to obtain primary dendrimers of gallic acid in the form of Me-
TGG, we next attempted to chemically modify the surface of
graphite powder with TGGA. However, instead of first synthe-
sizing TGGA and then attempting to couple this large, sterically
hindered molecule to the 1,2-diaminoethane linker attached to the
graphite powder (which is itself severely sterically hindered on the
graphite particle surface and present at a relatively low surface
concentration), we decided to synthesize the TGGA-carbon
powder starting from gallic acid modified carbon powder
(gallic-carbon), that we successfully synthesized above, and
further extending the “building-block” synthesis as shown in
Scheme 2 and detailed in section 2.2. As the gallic acid derivatives
used in this synthesis are much moresoluble in the chosen solvents
than TGGA, any unreacted material is likely more easily removed
from the carbon surface than would be the case if we attempted to
couple TGGA onto the surface directly. Thus, we can be
confident that only chemically attached TGGA dendrimers will
be formed on the carbon surface without any physisorbed
materials giving rise to possible misinterpretation of our experi-
mental data.
Figure 2. XPS spectra (wide scans 0-1200 eV) of (a) unmodified
graphite powder, (b) oxidized graphite powder, and (c) tri-O-acetyl
gallic-carbon powder.
Again, the deprotection of the tri-O-acetylgallyl groups, con-
tained in the tris-O-(tri-O-acetylgallyl)-gallylaminoethylamino-
carbonyl carbon intermediate (Scheme 2), was achieved using
the mild hydrolysis conditions successfully developed in the
solution phase synthesis of MeTGG described above.
3.5. Characterization of TGGA-Carbon Using X-ray
Photoelectron Spectroscopy. In order to confirm that each
stage of the synthesis of TGGA-carbon had been successful, XPS
analysis was performed on the tri-O-acety-protected gallic-
carbon, gallic-carbon, tris-O-(tri-O-acetylgallyl)-gallylaminoethyl-
aminocarbonyl carbon, and finally the TGGA-carbon product.
A wide survey scan was performed for each sample from 0 to
1200 eV as shown in Figure 4, and detailed scans over the C1s, O1s,
and N1s regions of interest. As expected, C, O, and N were the
Figure 3. First scan cyclic voltammograms (scanning from -1 V
to þ1 V versus SCE; scan rate, 100 mV s-1) recorded in pH 4.5
acetate buffer solution of (a) gallic-carbon powder abrasively
immobilized on the surface of a bppg electrode and (b) 5 mM
gallic acid in pH 4.5 buffer solution.
Scheme 2. Synthetic Route to Me-TGG (Solution Phase) or Alternatively TGGA-Carbon (Solid-Phase Building-Block Synthesis)
Langmuir 2010, 26(3), 1776–1785
DOI: 10.1021/la902497s 1781

Article
Ye et al.
only elements observed to be present on the surface with the
occasional exception of trace (<0.2 atomic %) Si impurities
arising from the borosilicate glassware. Table 2 details the
absolute atomic percentage of each element, C, O, and N, on
the surface of each sample determined from the area under each
peak. The chemical environment of carbon and oxygen atoms
within the gallyl moieties of the dendrimer are very similar to
those atoms in surface oxo-groups decorating the edges of the
graphene sheets,9-12 and therefore prevents detailed deconvolu-
tion and analysis of either the C1s or O1s peak. Furthermore, as
each stage of the building-block synthesis involves changing the
number of both carbon and oxygen atoms to the surface, we
cannot follow the synthesis using the absolute surface percentage
corresponding to each element. Fortunately, the number of
nitrogen atoms on the surface is not affected by the modification
chemistry. Therefore, the area of the N1s signal was used as a
“marker” and, knowing the modifying molecules’ stoichiometry
(and remembering to correct for the relative atomic sensitivity
factors32), can be used to decouple the percentage of the C1s and
O1s peak area attributable to the modifying molecules from the
background oxygen and carbon signals arising from the graphite
support. The resulting values are detailed in Table 2.
Having ascertained the relative atomic surface percentage of
each element at every stage of the synthesis, it is instructive to
compare the ratios of carbon and oxygen atoms relative to
gallic-carbon predicted from the stoichiometry of each moiety
as we proceed through the synthesis with the experimentally
obtained ratios (Table 3). As an illustratory example of this, the
gallic-carbon, to which we refer each ratio to, contains 2
nitrogen atoms, 10 carbon atoms (including the carbonyl
carbon to which the 1,2-diamino linker forms an amide bond
to the carbon surface and the 2 carbons in the ethyl chain of this
linker group), and 5 oxygen atoms. Consider the tri-O-acetyl-
protected precursor of the gallic-carbon. This now has 16
carbon atoms and 8 oxygen atoms, resulting in a stoichometric
ratio of 1.6 for both of these elements. Also listed in Table 3 are
the stoichometric ratios predicted if we had formed mono- and
bis-(O-gallyl)-gallylaminoethylaminocarbonyl groups on the
graphite surface. In comparison to the possible mono- and
bis-(O-gallyl)-gallylaminoethylaminocarbonyl structures, and
within the limit of experimental error, the XPS data give ratios
for both the change in carbon and oxygen atoms relative to
gallic-carbon that are not inconsistent with the formation of
the desired TGGA-carbon product together with the possi-
ble formation of some bis-(O-gallyl)-gallylaminoethylamino-
carbonyl carbon.
Figure 4. Wide survey XPS spectrum of TGGA-carbon.
Table 2. Absolute Elemental Percentage of C, O, and N Atoms on the Surface of Each Modified Carbon Sample Determined Using XPSa
percentage of the absolute
signal attributable to
atoms in the modifier
absolute elemental percentage
O%
material
tri-O-acetyl gallic-carbon
gallic-carbon
tris-O-(tri-O-acetylgallyl)-
gallylaminoethylaminocarbonyl-carbon
TGGA-carbon
C%
93.5
93.9
92.3
N%
0.5
0.7
0.6
C% O%
6.0
5.3
7.0
8.0
6.0
22.3
1.5
1.1
4.6
93.0
5.8
1.0
24.6
5.1
a Note that trace Si impurities (<0.2%) occasionally observed are not included. Also included is the decoupled percentage of the absolute C and O
signal attributable to atoms within the modifying molecules (determined using the N1s peak as a “marker”, see text).
Table 3. Comparison of the Predicted and Experimentally Observed Stoichometric Ratios of C and O Atoms for Each Material Relative to the
Number of C and O Atoms in Gallic-Carbon As Determined from XPSa
experimental stoichometric
atomic ratio determined from
XPS
predicted stoichometric atomic
ratio relative to gallic-carbon
material
C
O
C
O
gallic-carbon
tri-O-acetyl gallic-carbon
tris-O-(tri-O-acetylgallyl)-
1.0
1.6
4.9
1.0
1.6
5.2
1.0
1.3
3.7
1.0
1.4
4.2
gallylaminoethylaminocarbonyl -carbon
TGGA-carbon
bis-O-(tri-O-acetylgallyl)-
3.1
3.6
3.4
3.8
4.1
4.6
gallylaminoethylaminocarbonyl -carbon
bis-(O-gallyl)-
gallylaminoethylaminocarbonyl -carbon
mono-O-(tri-O-acetylgallyl)-
gallylaminoethylaminocarbonyl -carbon
mono-(O-gallyl)-
2.4
2.3
1.7
2.6
2.4
1.8
gallylaminoethylaminocarbonyl -carbon
a Also included are the predicted stoichometric ratios for the related bis- and mono-(O-gallyl)-gallylaminoethylaminocarbonyl carbon structures.
1782 DOI: 10.1021/la902497s
Langmuir 2010, 26(3), 1776–1785

Ye et al.
Article
Table 4. Quantitative Analysis of the Concentration of Al(III) Ions
Remaining in Solution after Exposure to 200 mg of Gallic-Carbon
Powder for 30 min
initial concentration of
Al(III)/μg L-1
Al(III) ion concentration
remaining in solution/μg L-1
270
675
945
1350
2700
51
107
183
333
1671
Table 5. Quantitative Analysis of the Concentration of Al(III) Ions
Remaining in Solution after 30 min of Stirring with 200 mg of TGGA-
Carbon Powder
Initial concentration of
Al(III)/μg L-1
Al(III) ion concentration remaining in
solution/μg L-1
200
500
1000
2700
<L.O.D.
6
416
1443
after 5 min of stirring was 1671 and 51 μg L-1 for an initial
concentration of 2700 and 200 μg L-1, respectively, it is apparent
that the TGGA-carbon is more effective at removing Al(III)
ions, especially from solutions close to the WHO legal limit of 200
ppb, where effectively all of the Al(III) ions are removed. In a
series of control experiments using unmodified graphite powder
under identical conditions, no adsorption of Al(III) ions was
observed.
Furthermore, if we compare the rate of adsorption by the
TGGA-carbon with that of gallic-carbon, using theory developed
by Chevallier et al.,23 the minimum rate of adsorption of Al(III) by
TGGA-carbon was found to be 3.86 ꢀ 10-4 cm s-1 compared to
1.72 ꢀ 10-4 cm s-1 determined for gallic-carbon. We note that the
rate of adsorption in the Chevallier model is first order with
respect to the number of surface sites available for complexation
and that the minimum rate of adsorption of TGGA-carbon is 2.25
times of that of gallic-carbon. This is therefore consistent with an
increased number of gallic acid moieties available in the dendritic
TGGA molecule to complex the Al(III) ions, and the ratio of ca.
2.3:1 for TGGA-carbon/gallic-carbon is consistent with the
results of XPS characterization in that the TGGA-carbon likely
consists of a mixture of tris- and bis-(O-gallyl)-gallylaminoethyl-
aminocarbonyl groups on the carbon surface.
3.6.2. Investigating the Adsorption of Al(III) Ions of Dif-
ferent Initial Concentrations with a Constant Mass of Either
Gallic-Carbon or TGGA-Carbon. As stated in the Introduction,
we are interested in comparing how the adsorption efficacy and
the parameters controlling the uptake of adsorbates vary between
monomeric and dendritic forms of adsorbents. Therefore, 25 mL
aliquots of Al(III) solutions with concentrationsranging from 200
to 2700 μg L-1 Al(III) ions were prepared by diluting the stock
solution appropriately. Each sample was separately stirred with
200 mg of either gallic-carbon or TGGA-carbon powder for a
period of 30 min, after which the modified carbon powder was
removed by filtration and the Al(III) concentration remaining in
the filtrate was determined, the results of which are shown in
Tables 4 and 5 for gallic-carbon and TGGA-carbon, respectively.
The capacity of the adsorbent and the equilibrium relationships
between adsorbate and adsorbent can be described by adsorption
isotherms, of which the Langmuir and Freundlich isotherms were
the earliest and simplest relationships used. The data obtained in
Tables 4 and 5 were therefore analyzed using linearized forms of
both the Langmuir isotherm and the Freundlich isotherm. None
Figure 5. Plot of the concentration of Al(III) ions remaining in
solution after treatment for varying lengths of time with (a) 200 mg
of TGGA-carbon powder with an initial Al(III) concentration of
2700 and 200 μg L-1 (inset: enlarged view of the data starting from
an initial Al(III) concentration of 200 μg L-1) and (b) 200 mg
of gallic-carbon powder with an initial Al(III) concentration of
2700 μg L-1
.
3.6. A Comparison of the Adsorption of Al(III) Ions by
Gallic-Carbon and TGGA-Carbon. In order to evaluate and
compare the efficacy of gallic-carbon and TGGA-carbon toward
the removal of Al(III) ions from aqueous media, the concentra-
tion of Al(III) ions present in any given sample, before and
after exposure to gallic-carbon, was determined using UV-vis
spectroscopy, as described in section 2.3. The kinetic and, where
appropriate, thermodynamic parameters controlling the adsorp-
tion of Al(III) ions by each material studied were ascertained as
follows.
3.6.1. Investigating the Kinetics of Al(III) Adsorption. In
order to ascertain the rate of adsorption of Al(III) ions by either
gallic-carbon or TGGA-carbon, 25 mL of either a 2700 μg L-1 or
a 200 μg L-1 Al(III) solution was separately stirred with 200 mg of
each modified carbon powder for varying lengths of time. The
modified carbon powders were then removed by filtration, and
the concentration of Al(III) ions remaining in the solution was
determined using UV-vis spectroscopy. Figure 5 reveals that the
adsorption of Al(III) ions from the solution occurred rapidly
within the first 5 min for both gallic-carbon and TGGA-carbon,
after which no further adsorption occurred. In the case where the
initial Al(III) concentration was 2700 μg L-1, the concentration
was reduced to ca. 1400 μg L-1 within the first 5 min of stirring
with TGGA carbon, while for the 200 μg L-1 solution the
concentration of Al(III) ions was reduced to below the detection
limit in the same time. Comparing this to the performance of
gallic-carbon, where the concentration of Al(III) ions remaining
Langmuir 2010, 26(3), 1776–1785
DOI: 10.1021/la902497s 1783

Article
Ye et al.
This is not consistent with the adsorption process following
a thermodynamically controlled adsorption isotherm, and it
appears that the TGGA-carbon is simply “titrating” the Al(III)
ions from solution. Repeat experiments at an elevated tempera-
ture of 40 °C gave similar results for the amount of Al(III) ions
removed by TGGA-carbon, again suggesting that the adsorption
is not under thermodynamic control and is, in effect, an irrever-
sible titration. To confirm this, 25 mL of two Al(III) solutions
with an initial concentration of 1000 and 2000 μg L-1 were
separately exposed to increasing amounts of TGGA-carbon
powder (25, 50, 200, 400, and 600 mg). The corresponding plot
of concentration of Al(III) remaining in the solution after
30 min of stirring with TGGA-carbon powder versus the mass
of TGGA-carbon added is shown in Figure 7. From this, it is
clearly apparent that the TGGA-carbon powder is simply
“titrating” the Al(III) ions from solution and that the adsorp-
tion behavior of the primary dendrimer adsorbent, at least
toward Al(III) ions under the conditions used herein, differs
markedly and perhaps unexpectedly from that of the mono-
meric form of the adsorbent material, that is, gallic-carbon.
Note that, in the case where the initial concentration of Al(III)
was 1000 μg L-1, exposure to 400 mg of TGGA-carbon
powder reduced the concentration to 56 μg L-1. This is lower
than the upper WHO legal limit (200 ppb) for small-scale water
treatment plants and indeed lower than the more stringent
legal limit set for drinking water supplies and large-scale water
treatment facilities (100 ppb), with the caveat that these are
laboratory samples not “real” sample matrices where the
speciation of Al(III) ions is more complex as mentioned in
the Introduction. Note that the titration data above corre-
spond to the removal of ca. 2.2 μmol of Al3þ ions per gram of
TGGA-carbon and imply a loading of ca. 0.7 μmol of dendron
per gram of modified carbon material or approximately
Figure 6. Freundlich plot for the uptake of Al(III) ions from
solutions of different initial concentration by 200 mg of gallic-
carbon powder.
1 ꢀ 10-12 mol cm-2
.
Figure 7. Plot of the concentration of Al(III) remaining in the
solution after 30 min of stirring with TGGA-carbon powder versus
the mass of TGGA-carbon added for two initial concentrations of
4. Conclusions
Al(III) of 1000 and 2000 μg L-1
.
Synthetic protocols for the covalent attachment of gallic acid
molecules to the surface of graphite microparticles, the solution
phase synthesis of TGGA (a primary dendrimer of gallic acid),
and the solid-phase synthesis of TGGA covalently attached to
graphite microparticles in a ground-upward “building-block”
fashion are reported. Characterization of the successful synthesis
of TGGA in the solution phase was performed using standard 1H
NMR and FT-IR characterization techniques. The synthesis of
gallic-carbon and TGGA-carbon powders was followed at
each stage of the procedure using XPS. In the case of the
TGGA-carbon material, XPS characterization suggested that
the material consisted of a mixture of both tris- and some
bis-(O-gallyl)-gallylaminoethylaminocarbonyl moieties on the
carbon surface.
The efficacy of the gallic-carbon and TGGA-carbon powders
toward the removal of Al(III) ions from aqueous solutions was
then examined and compared using UV-vis spectroscopy. Both
gallic-carbon and TGGA-carbon were found to exhibit rapid
kinetics for the uptake of Al(III) ions, fully occurring within 5 min
of exposure of the sample to the modified carbon powders. The
kinetics of adsorption were investigated, and the adsorption rate
constants were found to be3.86ꢀ 10-4 and 1.72 ꢀ 10-4 cms-1 for
TGGA-carbon and gallic-carbon, respectively. Thus, TGGA-
carbon adsorbs Al(III) ions 2.3 times faster, consistent with the
increased number of adsorption sites in the dendritic form of
gallic acid and confirming the results of XPS characterization in
that the TGGA-carbon may actually consist of a mixture of
of the data obtained for either gallic-carbon or TGGA-carbon
were found to conform to the Langmuir isotherm.
In the case of gallic-carbon, the data in Table 4 were found to fit
to the linearized form of the Freundlich equation given by eq 1:
1
ð1Þ
log Nads ¼ log Kþ log C
n
where Nads is the metal ions uptake, described by the mass of
Al(III) ions adsorbed per milligram of gallic-carbon powder, C is
the concentration of Al(III) remaining in the solution, and K and
n are Freundlich constants relating to the maximum adsorption
capacity and adsorption intensity, respectively. The larger the
value of K and the smaller the value of n (which typically takes
values from 1 to 10), the higher the affinity of the adsorbent
toward the adsorbate. Fitting the experimentally determined data
to the Freundlich isotherm in the case of gallic-carbon gave a
linear relationship (R2=0.98), shown in Figure 6. From this plot,
the values of K and n were determined to be K=0.38 L g-1 and n=
0.99, which is comparable to our previous studies involving
the adsorption of other metal ions such as Cu(II) and Cd(II)
using carbon powders chemically modified with monomeric
adsorbents.14,16
In contrast to the adsorption behavior of gallic-carbon, it is
immediately apparent from Table 5 that 200 mg of TGGA-
carbon reduces the concentration of Al(III) ions by ca.
500 μg L-1 regardless of the initial Al(III) ion concentration.
1784 DOI: 10.1021/la902497s
Langmuir 2010, 26(3), 1776–1785

Ye et al.
Article
both tris- and some bis-(O-gallyl)-gallylaminoethylaminocarbo-
nyl moieties on the carbon surface.
synthesis, and we have demonstrated that relatively complex
chemical architectures can be constructed, and selective chemi-
stries can be readily performed on their surface. Issues with
relative loading of modifying molecules on these support
materials are yet to be overcome, and this is an ongoing area
of our research. The TGGA-carbon material shows promise as
an inexpensive and easy way to make material. Polyphenolic
compounds, such as gallic acid and TGGA, may complex other
trivalent metal ions of greater environmental concern than
Al(III), and this is an area of ongoing investigation in our
laboratory.
While the adsorption of Al(III) ions bygallic-carbon was found
to be reversible and to obey a Freundlich isotherm, adsorption by
the TGGA-carbon was irreversible. The TGGA-carbon simply
titrates Al(III) ions from solution, with Al(III) concentrations
below the WHO limits obtainable from laboratory solutions with
initial concentrations of 500 μgL-1 within 5 min of exposure. It
would appear that forming dendrimers or polymers of adsorbent
materials on solid supports may, in certain cases, such as the
comparison made herein between gallic acid and TGGA groups
on graphite powder, drastically alter their adsorption behavior.
We believe that graphitic carbon materials show great
promise as relatively inexpensive supports for solid-phase
Acknowledgment. G.G.W. thanks St. John’s College for
support via a Junior Research Fellowship.
Langmuir 2010, 26(3), 1776–1785
DOI: 10.1021/la902497s 1785