Catalytic oxidation of formaldehyde by ruthenium multisubstituted tungstosilicic polyoxometalate supported on cellulose/silica hybrid

Article abstract of DOI:10.1016/j.apcata.2015.10.003

Cellulose/silica hybrid material produced by a sol-gel method from cellulosic fibres and silica precursors (tetraethoxysilane and 3-aminopropyltriethoxysilane) was functionalized with α-[SiW₉O₃₇Ru^III₃(H₂O)_xCl_3-x]^(10-x)- (Ru-POM), and thoroughly characterised by ¹³C and ²⁹Si solid state NMR, FTIR spectroscopy, UV-vis reflectance spectroscopy, X-ray diffraction, thermogravimetric analysis, scanning electron microscopy (SEM), and chemical analysis. The supported Ru-POM exhibited catalytic activity for the gas phase heterogeneous aerobic oxidation of formaldehyde (ca. 830 ppm in polluted air) at room temperature, in a packed-bed reactor operating at a linear velocity ca. 0.33 m/s and ca. 0.5 s residence time. Maximum formaldehyde uptake was 1.10 g/min per kg of hybrid material doped with Ru-POM (ca. 1.4% w/w). More than 400 turnovers were achieved without significant loss of catalytic activity and the only detected oxidation products of formaldehyde were carbon dioxide and water. A plausible mechanism for the catalytic oxidation of formaldehyde by supported Ru-POM has been proposed and includes formaldehyde oxidation in oxygenated ruthenium complex.

Full text of DOI:10.1016/j.apcata.2015.10.003

Accepted Manuscript
Title: Catalytic oxidation of formaldehyde by ruthenium
multisubstituted tungstosilicic polyoxometalate supported on
cellulose/silica hybrid
Author: J.A.F. Gamelas F. Oliveira M.G. Evtyugina I.
Portugal D.V. Evtuguin
PII:
S0926-860X(15)30180-0
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2015.10.003
APCATA 15579
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
30-7-2015
21-9-2015
3-10-2015
Please cite this article as: J.A.F.Gamelas, F.Oliveira, M.G.Evtyugina, I.Portugal,
D.V.Evtuguin, Catalytic oxidation of formaldehyde by ruthenium multisubstituted
tungstosilicic polyoxometalate supported on cellulose/silica hybrid, Applied Catalysis
A, General http://dx.doi.org/10.1016/j.apcata.2015.10.003
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Catalytic oxidation of formaldehyde by ruthenium multisubstituted
tungstosilicic polyoxometalate supported on cellulose/silica hybrid
J.A.F. Gamelas1* jafgas@eq.uc.pt, F. Oliveira2, M. G. Evtyugina3, I. Portugal2, D. V.
Evtuguin2* dmitrye@ua.pt
1Department of Chemical Engineering, CIEPQPF, University of Coimbra, Pólo II—R. Sílvio
Lima, 3030-790, Coimbra, Portugal.
2Chemistry Department and CICECO, University of Aveiro, 3810-193, Aveiro, Portugal.
3CESAM, University of Aveiro, 3810-193, Aveiro, Portugal.
*Corresponding author.
1

Graphical Abstract
2

Highlights




Cellulose/silica hybrid was functionalized with novel SiW9Ru3 polyoxometalate.
Supported polyoxometalate oxidises formaldehyde at room temperature
The only oxidised products of formaldehyde are water and carbon dioxide
More than 400 turnovers were achieved without loss of catalyst activity
3

Abstract
Cellulose/silica hybrid material produced by a sol-gel method from cellulosic fibres
and silica precursors (tetraethoxysilane and 3-aminopropyltriethoxysilane) was
functionalized with -[SiW9O37RuIII (H2O)3Cl3-x]7- (Ru-POM), and thoroughly
3
13
29
characterised by C and Si solid state NMR, FTIR spectroscopy, UV-Vis reflectance
spectroscopy, X-ray diffraction, thermogravimetric analysis, scanning electron
microscopy (SEM), and chemical analysis. The supported Ru-POM exhibited catalytic
activity for the gas phase heterogeneous aerobic oxidation of formaldehyde (ca. 830
ppm in polluted air) at room temperature, in a packed-bed reactor operating at a linear
velocity ca. 0.33 m/s and ca. 0.5 s residence time. Maximum formaldehyde uptake was
1.10 g/min per kg of hybrid material doped with Ru-POM (ca 1.4% w/w). More than
400 turnovers were achieved without significant loss of catalytic activity and the only
detected oxidation products of formaldehyde were carbon dioxide and water.
A
plausible mechanism for the catalytic oxidation of formaldehyde by supported Ru-POM
has been proposed and includes formaldehyde oxidation in oxygenated ruthenium
complex.
Keywords: cellulose/silica hybrids (CSH); ruthenium-substituted polyoxometalate;
formaldehyde oxidation; heterogeneous catalysis; air purification.
4

1. Introduction
Formaldehyde is the most ubiquitous indoor air pollutant with adverse effects on
human health [1, 2]. According to the International Agency for Research on Cancer
(IARC), formaldehyde is classified as carcinogenic for humans, belonging to Group 1
of hazardous substances [3]. The main sources of indoor formaldehyde are
building/finishing materials, floorings, paints, and furniture, among others [4, 5]. In the
last decades large efforts have been dedicated to avoid the use of formaldehyde-derived
thermosets in construction materials and household appliances, but indoor levels of
formaldehyde are still considerable [4, 5]. Indoor levels of formaldehyde are commonly
in the range 10 - 200 µg/m3 though most national guidelines recommend an upper-limit
concentration of 0.1 ppm (124.9 µg/m3, at 20 °C) for 8 h exposure periods [4].
Among the various methods to diminish indoor formaldehyde concentrations, both
passive (removal of sources, increased ventilation, etc.) and active (surface coating,
adsorption, etc.), the implementation of advanced catalytic systems for air purification is
widely recognized [4-6]. The most common catalytic approach to decompose volatile
organic compounds (VOCs) is the use of different ultraviolet photo-catalytic oxidation
(UVPCO) systems, where the degradation of pollutants is performed over a solid TiO2
catalyst both as thin film or ceramic substrate and with high-intensity UV light [4, 6, 7].
UVPCO systems may be economically feasible to degrade low concentration VOCs in
seconds or hours, even at room temperature. However, for VOCs concentrations as high
as 10 ppm, photo-catalysis is less effective and provides incomplete mineralization of
VOCs [6-8]. In contrast, thermal catalytic oxidation (TCO) systems using different
metal oxide particles and operating under aerobic conditions at high temperatures (100-
1200 ºC) can mineralize VOCs at relatively high concentrations. However, TCO
systems are not cost-effective for air purification at low concentration of VOCs [8, 9]
5

though recent TCO systems based on nanoscale MnO2 may compete with UVPCO [10,
11]. Much less information is available concerning the highly effective redox catalysis
using transition metal complexes mimicking enzymatic prosthetic groups. Amongst
redox catalytic systems, aerobic oxidation of formaldehyde with polyoxometalates
deserves particular attention.
Polyoxometalates (POMs) are metal-oxygen cluster anions with general formulae
m-
n-
MxOy (isopolyanions) or XzX’z’MxOy (heteropolyanions), where M is Mo, W or V,
X is mostly P, Si, B or Ge, and X’ is mainly a transition metal [12-14]. Due to their
thermal and hydrolytic stability, tunable redox properties and solubility in various
solvents, POMs have been recognized as highly attractive catalysts in homogeneous and
heterogeneous media [15-17]. For liquid- or gas-phase heterogeneous catalytic
applications POMs may be deposited on suitable organic/inorganic supports without
downstream contamination. Transition metal substituted POMs, including
[PW11CoO39]5- anchored to aminated mesoporous silica [18] and [SiW11CeO39]4- [19],
and polyvanadotungstates (PW9V3O406-, PW8V4O407-, and PW6V6O409-) [20] have been
used in aqueous solutions for the aerobic oxidation of formaldehyde to formic acid
under mild conditions (20-40 ºC). In all aforementioned catalytic systems the radical
chain mechanisms have been proposed.
Recently, cellulose/silica hybrids (CSH) doped with Keggin-type vanadium
heteropolyanions (POM-CSH) have been synthesised and examined for room
temperature oxidation of volatile organic compounds present in polluted urban air [21].
The use of biodegradable cellulosic pulp as catalyst support allowed the production of
low bulk density, porous and recyclable material which may be applied either in the
form of fluffy pulp or on a paper surface. The efficiency of VOCs oxidation catalysed
by [PVxMo10O40](3+x)- (x=1,2) was comparable to that usually reported for UVPCO
6

systems. In the course of reaction the species V(V) was reduced to V(IV) by reaction
with VOCs and, in turn, was re-oxidised to V(V) by molecular oxygen present in air
[21]. As expected from known redox properties of POMs [22], the re-oxidation of
[PV2Mo10O40]5- was faster than [PVMo11O40]4-.
Ruthenium-based complexes are promising oxidative catalysts due to their particular
redox properties. Ruthenium in monosubstituted phosphomolybdate heteropolyanion
was suggested to be the initiator of autoxidation reactions under aerobic conditions [23].
However, POMs with more than one Ru atoms typically show better catalytic
performances in oxygenation reactions than POMs with one Ru atom [24]. Thus,
―sandwich‖ type polyoxometalates containing the Ru2O2 core may act similarly to
dioxygenases in adamantane oxidation catalysis with molecular oxygen [25, 26]. The
incorporation of multiple ruthenium centres in a polyoxometalate structure enhances its
stability to oxidative decomposition and its catalytic properties due to the readily
formed peroxo or oxo complexes often active in selective oxidation reactions. In this
context, the recently synthesised ruthenium multisubstituted heteropolyanion,
[SiW9O37RuIII (H2O)3Cl3]7- [26], may be envisaged as a promising catalyst for the
4
aerobic oxidation of formaldehyde.
In the present work, -isomer of [SiW9O37RuIII (H2O)3Cl3]7- was anchored to CSH
4
and examined as catalyst for the heterogeneous oxidation of gaseous formaldehyde.
Possible routes of formaldehyde oxidation are proposed.
2. Experimental section
2.1 Materials
7

Tetraethoxysilane (TEOS, 98% w/w), 3-aminopropyltriethoxysilane (APTS, 95%
w/w), ethanol (99.5% w/w), concentrated nitric (65% w/w) and sulphuric (98% w/w)
acids, and ammonia (25% w/w aqueous solution) were supplied by Sigma-Aldrich and
used as received. Formaldehyde (37% w/w aqueous solution stabilised by methanol)
was provided by Panreac and hydroxylamine hydrochloride (>99% w/w) was supplied
by ACROS OrganicsTM. Eucalyptus globulus industrial bleached kraft pulp, used as
cellulose fibre source was supplied by Portucel-Soporcel (Portugal). This pulp, hereafter
named cellulosic pulp or cellulose, contained about 86% w/w of cellulose and 14% w/w
of glucuronoxylan (on oven-dry pulp mass). Prior to use, air dried cellulosic pulp (1.5 g)
was disintegrated (swollen) in water (150 cm3) for 48 h, filtered and then washed with
ethanol (two 50 cm3 portions) to remove the excess of water. The compound
-K6Na[SiW9O37Ru4(H2O)3Cl3]·nH2O (abbreviated as -SiW9Ru4) was synthesized and
characterised as described previously [26]. Elemental analysis data (% w/w) calculated
for -SiW9Ru4 with n=14: W, 49.4; Ru, 12.1; K, 7.0; Na, 0.69; Cl, 3.18. Experimental
results (% w/w): W, 50.0; Ru, 12.5; K, 6.9; Na, 0.83; Cl, 2.75. Molar ratio W/Ru = 2.20.
IR characteristic absorption bands (cm-1): 988 (sh); 946, as (W–Od); 888, as (W–Ob–
W); 779, 715 (sh), as (W–Oc–W).
2.2 Synthesis of cellulose/silica hybrid functionalized with ruthenium
multisubstituted polyoxometalate
A procedure similar to that reported for the synthesis of CSH functionalized with
molybdovanadophosphate anions [21] was used for the synthesis of CSH functionalized
with ruthenium polyoxoanion. Briefly, a solution of TEOS (37.5 cm3), APTS (1.55 cm3)
and ethanol (20.5 cm3) was added to water (10.2 cm3) acidified with concentrated nitric
acid (0.07 cm3) and to the resulting mixture, nitric acid (0.80 cm3) was immediately
added. The resulting solution was left to react for 40 min and thereafter dilute ammonia
8

(2.5%) was added in small portions (0.250 cm3; ca. 3.5 cm3 in total) to bring the pH to
ca. 3. The formed sol was allowed to react for 10 min. Cellulosic pulp (1.5 g)
previously swollen in water was added in small portions and the reaction extended for
more 10 min. Finally, -SiW9Ru4 dissolved in water (0.1 g in 1 cm3) was added and the
suspension allowed to react for further 20 min. The resultant brown material was
separated by filtration (G3 glass filter) and vacuum dried at room temperature.
A CSH free of ruthenium polyoxometalate (reference material) was prepared
similarly as above described, excluding the POM addition step. After adding the
cellulosic pulp the suspension was stirred for 20 min before filtering the white solid.
2.3 Analyses
29Si Magic Angle Spinning Nuclear Magnetic Resonance (MAS NMR) spectrum
was recorded at 79.49 MHz on a 9.4T Brucker Avance-400 spectrometer, with a 5 kHz
13
spinning rate, using 45º pulses, a contact time of 8 ms and a 60 s recycle delay. C-1H
cross-polarization MAS NMR (CP/MAS NMR) spectrum was also recorded on a 9.4T
Bruker Avance-400 spectrometer, with a 9 kHz spinning rate, using a contact time of
1
29
1.5 ms, a 5 s recycle delay and a H 90º pulse of 4 s. Chemical shifts in the Si and
13C NMR spectra were referenced to tetramethylsilane.
Analyses of W and Ru were carried out by Inductively Coupled Plasma - Atomic
Emission Spectrometry (ICP-AES) using a Jobin Yvon Activa M equipment. Elemental
analysis (C, H and N) was done using a Leco CHNS-932 apparatus. Thermogravimetric
(TGA) analysis was carried out under nitrogen atmosphere in the 30 - 700 ºC range at a
heating rate of 5 ºC min-1 on a TGA-50 Shimadzu thermobalance. Infrared absorption
spectra in KBr pellets were recorded on a Mattson 7000 FTIR spectrometer. UV-Vis
diffuse reflectance spectra were registered on Jasco V-560 spectrophotometer, using
9

BaSO4 as a reference. The acquisition was done in the range of 220–800 nm with a
scanning speed of 100 nm/min and a bandwidth of 5 nm. The reflectance spectra were
converted into k/s spectra using the Kubelka–Munk equation (k stands for the specific
absorption coefficient and s is the specific scattering coefficient). Scanning electron
microscopy (SEM) images were obtained on a FEG-SEM Hitachi S4100 microscope,
operating at 25 kV, using gold coated samples.
X-ray diffraction data were collected using a Philips X’Pert diffractometer operating
in the Bragg-Brentano configuration with Co-K (λ=1.79 Å) radiation (35 mA, 40 kV).
Data were collected by the step counting method (step 0.025º and time 1.0 s) in the 2θ
range of 4-50º. Circular pellets (around 1 cm diameter and 1 mm thickness) were
prepared for analysis by pressing the samples at 50 MPa during 2 min. The crystallinity
was determined as (ICr-Iam)/Icr*100, where Icr refers to the intensity of the reflection
from the 002 plane at about 26º and Iam is the intensity of the amorphous halo between
the reflection around 19º and the reflection from the 002 plane.
Mercury intrusion porosity was determined using the AutoPore IV 9500 instrument
from Micromeritics. The pressure range varied between 0.5 psi and 30000 psi, which
enabled the measurement of pore diameters between 350 µm and 0.006 µm,
respectively. The total porosity was calculated as the ratio of the total intruded volume
and the sample bulk volume. Bulk volume, corresponding to the sample volume
measured at the lowest intrusion pressure, was determined using the same technique.
The corresponding bulk density was determined by dividing the material mass by the
respective volume. Specific surface area was determined by N2 adsorption in a
Micromeritics ASAP 2000 instrument, using the BET equation.
High performance liquid chromatography (HPLC) analyses were carried out in a
ACCELA LC system (Thermo Scientific) equipped with Refractive Index and Photo
10

Diode Array detectors, using a 8 m HyperREZ XP H+ column (100 x 7.7 mm) coupled
with a 8 m HyperREZ XP H+ Guard pre-column (50 x 7.7 mm) both from Thermo
Fisher Scientific Inc. Analysis conditions were as follows: eluent, 0.010 M H2SO4; flow
rate, 0.2 mL/min; temperature, 40 ºC; sample volume, 10 L.
2.4 Experimental procedure for the formaldehyde removal/oxidation
Purified air (Air K grade) was continuously bubbled through an aqueous solution of
formaldehyde (ca. 37% w/w) at room temperature (20 ± 1 ºC) and constant flow rate
(ca. 2.5 dm3/min). The produced air/formaldehyde mixture was collected in Tedlar®
gas sampling bags (25 dm3) and had an average concentration of formaldehyde around
106 µg/m3 (ca. 833 ppm)  10%.
The CSH materials without catalyst (reference) or containing the supported catalyst
(Ru-POM-CSH) were packed in Teflon tubes (4 mm i.d., 160 mm length) sealed at both
ends with fibreglass. The weight of CSH with supported catalyst and the height of the
packed bed were similar for all the experiments (ca. 110 mg and 130 mm, respectively).
The experimental setup for catalyst testing is presented in Fig. 1. The Tedlar® bag
containing the formaldehyde/air mixture was equipped with a flow control valve and
packed-bed reactor using silicon tubing (4 mm i.d.). The outlet was connected to a
borosilicate glass flask (absorbing flask) followed by a gasometer filled with water.
Experiments were carried out at room temperature (22 ± 2 ºC). Typically, each trial
consisted in passing the formaldehyde gas mixture (5.0 dm3) through the system at an
optimized flow rate (ca. 0.25 dm3/min) followed by quantification of the formaldehyde
retained in the absorbing flask using the hydroxylamine hydrochloride method (EN ISO
9397:1997). The absorbing flask contains hydroxylamine hydrochloride (NH2OH.HCl,
11

10% w/w, 0.110 dm3) that reacts with formaldehyde to yield an oxime, water and
hydrochloric acid according to Equation (1).
NH2OH.HCl + CH2O  CH2NOH + H2O + HCl
(1)
The hydrochloric acid formed, corresponding to the reacted formaldehyde, was
quantified by potentiometric titration with 0.010 M NaOH till pH 3.6 with a relative
error lower than 5%. Blank experiments were performed in a similar manner without the
CSH materials (the Teflon tube was removed from the experimental setup, Fig. 1) to
measure the concentration of formaldehyde in the gas sampling bag. The amount of
formaldehyde retained/converted by the CSH materials (with or without POM) was
evaluated by difference, i.e., by subtracting the blank experiment results. As a rule, two
blank and two working experiments were done per each gas sampling bag.
The CSH materials (with and without POM) were tested in consecutive cycles of
formaldehyde degradation. When the catalytic activity became low the hybrid materials
were submitted to regeneration, at room temperature, by purging with purified air (ca.
0.25 dm3/min during 4 h).
In some experiments two Teflon tubes, one packed with Anhydrone® (magnesium
perchlorate desiccant) and the other with Ascarite® (sodium hydroxide-coated silica,
20-30 mesh beads) were placed in series between the absorbing flask and the gasometer
to evaluate the carbon dioxide formed by formaldehyde oxidation. The first tube acts as
water adsorbent and the second as CO2 absorbent – the weight gain of the Ascarite®
tube enables the quantification of CO2.
A Multi-Gas Infrared Analyzer MIR 9000 (Environment S.A.) was used for the
qualitative analysis of COx in the exhaust gases from the Teflon tubes packed with CSH
materials. In this case, the tube outlet was connected directly to the MIR analyser.
However, the internal pump of the equipment provided a rather unstable and relatively
12

high flow rate of gas (ca. 1.0-1.5 dm3/min) and therefore a lower residence time in the
catalytic tube in comparison with the experiments performed when using the gasometer
(gas flow of ca. 0.25 dm3/min).
3. Results and discussion
3.1 Cellulose/silica hybrid material functionalized with ruthenium polyoxometalate
CSH materials were synthesised by a sol-gel method to produce aminopropyl
modified silica sol at the surface of the cellulosic fibres which enabled anchoring the
multisubstituted ruthenium polyoxometalate to the protonated aminopropyl groups via
electrostatic interactions. The novel ruthenium polyoxometalate hybrid material (Ru-
POM-CSH) was characterised by 13C and 29Si solid state NMR, FTIR spectroscopy,
UV-Vis reflectance spectroscopy, X-ray diffraction, thermogravimetric analysis, SEM,
and chemical analysis. Comparison was made with the reference CSH material without
POM.
Characterisation by solid-state NMR provided important information regarding the
structural features of cellulose and modified silica in the new hybrid material. 13C
CP/MAS NMR spectrum (Fig. 2A) showed the characteristic six signals in the 50-120
ppm range, assigned to known carbon resonances of cellulose framework [27]. No
shifting of these signals or changes in their relative intensity was detected, in
comparison to those of the initial cellulosic fibres, suggesting that the chemical structure
and crystallinity of cellulose was maintained in the hybrid material. In addition, three
signals of low intensity were observed at 42.6, 20.9 and 8.8 ppm assigned to the C1, C2
and C3 carbon resonances, respectively, of the propylamino moieties (-H2N-C(1)H2-
C(2)H2-C(3)H2-Si) [28, 29]. The C2 chemical shift (at 21 ppm) indicates the presence of
protonated propylamino moieties [29], in agreement with the acidic conditions of the
13

synthesis. For neutral amino groups the C2 resonance would appear at 27 ppm [29]. 29Si
MAS NMR spectrum of the new hybrid material showed three major signals in the -80
to -120 ppm region, namely at -91.5 ppm (8%), -101 ppm (47%) and -111 ppm (45%)
(Fig. 2B), corresponding to (SiO)2Si(OH)2 (Q2), (SiO)3SiOH (Q3) and (SiO)4Si (Q4)
moieties, respectively [29-31]. The degree of condensation of the bound silica
framework was 0.85, which is typical for this kind of hybrid materials [21, 32]. The
resonance coming from the silicon in the ruthenium polyoxometalate structure was not
detected due to the very low amount of POM in the hybrid (<2 % w/w, Table 1) and to
the paramagnetism of the ruthenium cluster [26]. Similar 13C and 29Si MAS NMR
spectra were obtained for the reference CSH.
Please, insert Figure 2
Silica domains were dispersed on cellulosic fibres mostly in the form of a
discontinuous film, maintaining the common fibre shape (Fig. 3). The porosity of the
Ru-POM-CSH material, as measured by mercury intrusion porosimetry, was very high
(80%). Conversely, the novel material exhibited very low bulk density (0.19 g/cm3).
The high porosity of this material is an important characteristic when considering its
possible catalytic application for gas-solid heterogeneous oxidation reactions. The BET
surface area of the novel material was of 6.9 m2g-1, i.e., within the expected range for a
cellulose-based material (note that for the initial bleached pulp a value of 8.7 m2g-1 was
obtained).
Please, insert Figure 3
The FTIR absorption spectra of the novel cellulose-based hybrid material (Fig. 4b)
showed bands, not present in the spectrum of cellulosic pulp (Fig. 4a), at 445 cm-1 and
793 cm-1 (marked with * in Fig. 4) assigned to  (O-Si-O) and s (Si-O-Si), respectively
[30]. The band expected around 1080 cm-1 (as (Si-O-Si)) was overlapped with the
14

bands of cellulose in the same spectral region. In addition, a sharp band at 1385 cm-1
appeared in the spectra of the hybrid, showing the presence of nitrate anions (coming
from nitric acid used in the CSH synthesis). It should be noted that the spectra of the
new hybrid (Ru-POM-CSH) and the reference material (CSH without ruthenium
complex) were quite similar (Figs. 4b and c, respectively) showing undoubtedly that the
absorption bands due to the Ru-POM structure, at wavenumbers lower than 1000 cm-1
[26], were masked by those of the cellulosic matrix and functionalized silica in the
FTIR spectra of the Ru-POM-CSH material. Once more, this was due to the very low
amount of polyoxometalate in the latter.
Please, insert Figure 4
X-ray diffractograms (Co-K radiation) of Ru-POM-CSH (Fig. 5b) and of the
reference CSH (Fig. 5c) presented the non-resolved reflections from the 101 and 10ī
planes around 18º and the reflection from the 002 plane at about 26º, corresponding to
the cellulose I polymorph [33]. The diffraction peaks of the hybrids were weaker than
those of the initial cellulosic pulp (Fig. 5a) due to the presence of amorphous silica [21].
However, the degree of crystallinity of cellulose in the hybrids (78-80%, estimate based
on the relative intensity of the 002 reflection peak from crystalline cellulose and the
amorphous halo, as explained in the experimental section) was comparable to that of the
initial cellulosic matrix (80%).
Please, insert Figure 5
The brown coloured Ru-POM-CSH material showed significant absorption in the
visible region of the reflectance spectra, with an absorption maximum at 462 nm (Fig.
6b). The absorption of visible wavelength confirms the presence of ruthenium-
substituted polyoxometalate [23, 24], because the initial cellulosic pulp and the non-
functionalized CSH have a negligible absorption in the visible region (Figs. 6a and c,
15

respectively), i.e., both samples have a white colour. In the UV region, the spectrum of
the new ruthenium hybrid was dominated by a band with a maximum at 261 nm, being
also observed a characteristic shoulder at ca. 340 nm (Fig. 6b). When comparing the
spectrum
of
Ru-POM-CSH
with
the
one
of
initial
Ru-POM,
-K6Na[SiW9O37Ru4(H2O)3Cl3]·14H2O in aqueous solution (Fig. 6d), some differences
are easily noticeable: 1) the relative intensity of the observed bands is fairly different; 2)
the band at ca. 260 nm is very well-defined in the spectrum of the hybrid but not in that
of initial POM; and 3) in the region below 250 nm the spectra seem to have different
trends (this could not be, however, confirmed for wavelengths below 220 nm in the
reflectance spectrum of Ru-POM-CSH, because of instrumental limitations). Thus, the
data obtained by UV-Vis spectroscopy indicated the eventual strong interaction of
ruthenium polyoxometalate with supporting material. Changes in the electronic and IR
spectra of POMs anchored by electrostatic forces to silica are well demonstrated [28].
Please, insert Figure 6
The thermogram of the novel Ru-POM-CSH material showed three main intervals
of weight loss (Fig. 7b): a first step up to about 120 ºC corresponding to 6% weight loss,
mainly due to moisture release; a second step between 150 and 230 ºC (ca. 4.3% weight
loss) which is assigned to the decomposition of the protonated propylamine moieties
and nitrate anions [21]; and a third step in the 250-600 ºC range, accounting for 57%
weight loss. The latter step corresponds to cellulose carbonization and oxidation as
follows by comparison with the thermogram of cellulosic pulp shown in Fig. 7a and
may be used to determine the amount of cellulose in the materials (Table 1). Up to
about 350 ºC, the thermograms of Ru-POM-CSH (Fig. 7b) and of the corresponding
reference hybrid material, CSH (Fig. 7c), were identical. However, for temperatures as
higher as 350 ºC in the course of cellulose thermal degradation, some differences were
16

found between the thermograms suggesting that the presence of ruthenium complex
may affect the thermal degradation of hybrid material.
Please, insert Figure 7
Based on the tungsten (W) elemental analysis, the polyoxometalate content in the
novel CSH functionalized material was estimated to be 1.4 % w/w (Table 1). The W/Ru
molar ratio in the ruthenium hybrid (3.1) was higher than in the parent polyoxometalate
containing the tetra-nuclear Ru4O4 core (W/Ru molar ratio of 2.25). This result suggests
the eventual release of a labile RuCl3-x(H2O)x group during the synthesis under acidic
conditions, thus giving rise to a more stable polyoxoanion containing the Ru3O3 cluster,
[SiW9O37RuIII (H2O)xCl3-x]7-. Such a possibility was admitted previously along the
3
synthesis of -K6Na[SiW9O37Ru4(H2O)3Cl3] [26]. Unfortunately, it was impossible to
confirm the detailed structure of anchored POM since, due its very low amount in the
hybrid, it was not easily detected by NMR, FTIR, or resonance Raman spectroscopy.
The propylamine-modified silica content was found to be 35 % w/w (Table 1). The
content of propylamine moieties and nitrate anions were estimated to be 2.2 and 2.1 %
w/w, respectively. The latter were determined assuming: a) the charge of protonated
propylamine groups is balanced by the charge of ruthenium substituted polyoxoanions
plus nitrate anions, and b) the total molar amount of nitrogen (from nitrogen elemental
analysis) is the sum of the molar amounts of nitrate anions and propylamine groups.
Hence the nitrate anions as well as the propylamine groups can be determined by
solving those two equations. It should be noted that the analytical results obtained for
the reference material (CSH) were similar to those of Ru-POM-CSH except, obviously,
for the absence of POM (Table 1).
Please, insert Table 1
3.2 Formaldehyde oxidation with Ru-POM-CSH
17

Formaldehyde degradation tests were performed at room temperature by flushing an
air/formaldehyde gas mixture through Teflon tubes packed with CSH materials without
POM (reference) and doped with -SiW9Ru3 (Ru-POM-CSH). The depletion of
formaldehyde (by oxidation and/or adsorption on the CSH materials) in the outlet gas
mixture was evaluated by chemisorption methods (as described in section 2.4). Each
material was tested in consecutive cycles, each cycle corresponding to the passage of
5.0 dm3 of polluted air (ca. 830 ppm CH2O) through the catalytic packed-bed (Fig. 1).
The results obtained for CH2O removal from polluted air by Ru-POM-CSH and
CSH materials are presented in Table 2. After the first cycle, the removal of CH2O is
similar for CSH with and without POM (43-44% removal). This can be explained by
chemisorption, including strong retention of CH2O in the porous structure of the hybrids
through Lewis acid-base interactions with the hydroxyl groups of cellulose and silica,
and by the reactions of CH2O with the amine moieties of modified silica (Equation (2)):
Si(O)3-(CH2)3-NH2 + CH2O → Si(O)3-(CH2)3-N=CH2 + H2O
(2)
Formaldehyde removed by the CSH material (control) decreased almost linearly
from the 1st to the 4th cycle, reaching zero, thus suggesting CSH ―saturation‖ (Table 2).
However, after the same number of cycles, ―saturation‖ was not detected for the Ru-
POM-CSH material, which was assigned to the CH2O oxidation catalysed by Ru-POM.
Please, insert Table 2
Assuming that at fairly high concentrations of CH2O in air (> 800 ppm) the re-
oxidation of Ru-POM by oxygen could be the reaction rate limiting step, the tubes
packed with CSH and Ru-POM-CSH were ―regenerated‖ before the 5th cycle (by
purging with purified air at a flow rate ca. 0.25 dm3/min, during 4 h). After this
treatment, formaldehyde removal remained near zero for the control experiment (CSH)
and almost doubled for the Ru-POM-CSH material (5th cycle, Table 2) in agreement
18

with the postulated reactivation of Ru-POM by molecular oxygen [22-24]. Interestingly,
with an additional 4h regeneration step, before the 6th cycle, only ca. 80% of CH2O
removal capacity by Ru-POM-CSH was achieved (in comparison to that in the 5th cycle,
Table 2). Unexpectedly, regeneration with pure oxygen (2h, after the 6th cycle) re-
activated the Ru-POM-CSH material modestly (20% for CH2O conversion in the 7th
cycle, Table 2). Curiously, purging the catalytic tube with nitrogen (during 2 h) after the
7th cycle also revealed partial catalyst regeneration though less than 8% of CH2O
conversion was achieved. These results suggest that catalyst regeneration is related to
Ru-POM re-oxidation but also to CH2O desorption from the Ru-POM-CSH.
The fraction of CH2O removed by oxidation with Ru-POM-CSH material (Table 3)
was estimated by difference considering that CH2O removal occurs by chemisorption
and oxidation phenomena (Ru-POM-CSH) or only by chemisorption (CSH). The data
can also be expressed in terms of turnover number (TON) defined as the molar ratio of
oxidised CH2O per catalytic active species (Ru-POM) supported on CSH (Table 3). As
shown in Table 3, the amount of chemisorbed formaldehyde was subtracted from the
total amount of removed formaldehyde (adsorbed plus oxidised) in order to obtain only
the amount of oxidised formaldehyde. Altogether, the results confirm that the
contribution of chemisorption phenomena for CH2O removal by Ru-POM-CSH
diminishes from the 1st to the 4th cycle, whereas inversely the contribution of catalytic
oxidation reactions increases. After the 4th cycle the sorption capacity of CSH is null
and so the removal of CH2O from contaminated air was determined exclusively by
oxidation with Ru-POM (Table 3). The amount of oxidised CH2O increased
substantially after Ru-POM-CSH regeneration by purging with air.
Please, insert Table 3
19

The predominance of CH2O chemisorption in comparison to its oxidation by Ru-
POM-CSH, during the first 4 cycles, may be tentatively assigned to statistics (the
number of chemisorption sites is higher than the number of oxidation centres), to a
plausible catalyst induction period or to diffusion limitations of CH2O transport to the
reactive centres in Ru-POM-CSH.
The total TON achieved with Ru-POM-CSH after 7 cycles (sampling of ca. 35 dm3
contaminated air (830 ppm CH2O) with an average residence time in the catalyst bed
around 0.5 s) was higher than 400 thus evidencing the good performance of SiW9Ru3 in
reversible oxidative catalysis. The maximum oxidation of CH2O from the polluted air
was ca. 1.04 g/min per kg of CSH doped with Ru-POM (cycle 5, Table 3).
3.3 Oxidation reaction products
The Ru-POM catalysed oxidation of CH2O is expected to yield formic acid
(HCOOH) and carbon oxides (CO2 and CO). In fact, formic acid has been reported as
the major product of aerobic liquid-phase heterogeneous and homogeneous oxidations
of CH2O catalysed by transition-metal-substituted POMs [18, 19].
The presence of formic acid at the outlet of the catalytic packed-bed reactor (tube
with Ru-POM-CSH) was checked by bubbling the exit gas stream in an absorbing flask
filled with distilled water (instead of hydroxylamine hydrochloride solution as in the
usual experimental setup, Fig. 1) later analysed by HPLC and acid-base titration. None
of these methods revealed the presence of formic acid. Only non-reacted CH2O in its
monomeric and dimeric forms was detected by HPLC. However, it is noteworthy that in
the reference experiment (tube packed with CSH) small amounts of methanol and
formic acid were detected by HPLC in the aqueous solution from the absorber. This can
be explained by the presence of ca. 3-4% of formic acid and methanol in the CH2O
solution (37% w/w) used to generate gaseous CH2O in air.
20

The tests performed to detect the formation of carbon monoxide were not
conclusive, because the difference in CO concentration at the reactor outlet for the
experiments with and without catalyst (Ru-POM-CSH and CSH, respectively) was ca.
0.1 ppm which is close to the CO detection limit of the online gas analyser used in this
study. However, the increase of CO2 concentration in the reactor exhaust gas stream
was unequivocally confirmed in this same equipment.
The quantitative analysis of CO2 formed in the catalytic aerobic oxidation of CH2O
was carried out by a chemisorption method with Ascarite® (section 2.4) in two parallel
experiments using regenerated Ru-POM-CSH and CSH without catalyst (reference).
The amount of CH2O consumed by catalytic oxidation (i.e. the difference between the
amounts of CH2O converted by Ru-POM-CSH and chemisorbed by CSH, under similar
conditions) and the amount of CO2 formed (determined from the Ascarite® tube weight
gain) almost converged. For example, average results from the experiments described in
cycles 5 and 6 (Table 3) were as follows: 70 µmol (2.11 mg) of CH2O was consumed in
aerobic oxidation with regenerated Ru-POM-CSH and the measured amount of CO2 was
68 µmol (2.98 mg). Hence catalytic oxidation of CH2O occurred essentially to CO2.
The rigorous balance of water released by formaldehyde oxidation was hindered by
noticeable variations of water vapour content in the CH2O/air gas mixture used in the
catalytic experiments. However, the observed general tendency was an increase of water
content in the air stream exiting the catalytic tube packed with Ru-POM-CSH when
compared to the tube packed with CSH without catalyst (measured by weighting the
Anhydrone® water trap).
Overall, these results indicate that CO2 and H2O were the major products formed by
catalytic oxidation of CH2O with Ru-POM-CSH. This fact endorsed the proposal of a
predominantly non-radical mechanism for CH2O oxidation by Ru-POM/O2, because a
21

free radical oxidation mechanism leads to the formation of significant amounts of
formic acid and carbon monoxide [18-20, 34].
3.4 The proposed formaldehyde oxidation mechanism
Based on the experimental results and mechanistic hypothesis put forth in previous
studies concerning the oxidation of different substrates (including VOCs) catalysed by
POMs substituted with one or more Ru atoms [23, 24], a mechanism for CH2O
oxidation catalysed by -SiW9Ru3 supported on CSH material may be proposed, with
the necessary amendment to a gas phase reaction.
Most likely the initiation step consists in substrate oxidation by POM-Ru3III catalyst
via ligand exchange (Equations (3) and (4)):
III
III
POM-Ru3 (H2O) + CH2O  POM-Ru3 (CH2O) + H2O
(3)
(4)
POM-Ru3 (CH2O) → HPOM-Ru2 RuII + .CHO
III
III
The rate limiting step might be the formation of reduced catalyst thus explaining, at
least partially, the observed low TON in the first cycles of formaldehyde oxidation
experiments (Table 3). The partially reduced catalyst may react with molecular oxygen
(from air) forming a complex involving metal to oxygen charge transfer (Equation (5)):
III
III
III
.
HPOM-Ru2 RuII + O2 ↔ HPOM-Ru2 RuII - O2 ↔ HPOM-Ru2 RuIII - O2
(5)
.
The metal complex with activated oxygen reacts further with CHO thus
accomplishing substrate oxygenation (Equation (6)):
III
.
.
HPOM-Ru2 RuIII - O2 + CHO → POM-Ru3III + CO2 + H2O
Hence, the overall reaction (Equations (3) – (6)) is represented by Equation (7):
CH2O + O2 → CO2 + H2O
(6)
(7)
22

.
Alternatively, the possibility of CHO reaction with molecular oxygen via
conventional free radical mechanisms is not excluded:
.CHO + O2 → CO + HOO.
(8)
(9)
HOO. + CH2O → .CHO + H2O2
However, the negligible amounts of CO detected after CH2O oxidation indicate that free
radical mechanisms should not be predominant unless carbon monoxide is not further
oxidised via alternative reaction mechanisms to carbon dioxide.
Apparently, the reactions described by Equations (3)-(6) occur simultaneously
III
within one complex involving POM-Ru3 , CH2O and molecular oxygen, via non-
radical pathways. The simultaneous oxidation of CH2O, bound to one of the three RuIII
centres in the Ru3O3 trinuclear core, and the reduction of oxygen bound to another RuIII
is possible due to the fast electron exchange between two neighbouring oxygen-bridged
ruthenium atoms (RuII-O-RuIII↔RuIII-O-RuII). Oxygen transfer to substrate takes place
III
in the HPOM-Ru2 RuII-O2-CHO complex. However, this scenario presumes the
formation of mixed CH2O and O2 ligands within Ru3O3 clusters, not always attainable
due to the competition between CH2O and O2. Hence, when the concentration of CH2O
is fairly high, catalyst inactivation is possible due to complete occupation of the active
sites by CH2O. Analysis of catalysts using UV-Vis reflectance spectroscopy provided a
further insight into this matter.
The comparison of UV-Vis reflectance spectra of initial Ru-POM-CSH and the
same material after the 7th cycle of use in the polluted air oxidation (Figs. 9a and b,
respectively) revealed a relative decrease in the characteristic absorption at ca. 460 nm
(with apparent shift of 10 nm to lower wavelength) and a concomitant increase in
absorption at ca. 400 nm. Such spectral features were previously observed for
23

ruthenium-substituted POM in the presence of electron donating ligands and attributed
to the strong interaction of ruthenium-oxygen cluster with these ligands substituting the
water molecule [24, 35]. Assuming the aforementioned spectral features are due to the
H2O→CH2O ligand exchange in Ru3O3 core, the inverse diminishing of absorption at
ca. 400 nm and the increment at 460 nm after the catalyst regeneration step (i.e. purging
the catalyst bed with purified air) (Fig. 9c) was interpreted in terms of reversible
CH2O→H2O exchange. Hence, during the purging of catalyst bed by air, a significant
part of bound CH2O was removed allowing the coordination of oxygen in the Ru3O3
core while the oxidation of the remaining CH2O took place. This assumption is
supported by the fact that purging of the catalyst bed with pure nitrogen, after the
formaldehyde oxidation step, also led to partial catalyst regeneration meaning that at
least part of CH2O bound to Ru3O3 core was removed.
Please, insert Figure 8
4. Conclusions
Multisubstituted
ruthenium
tungstosilicic
polyoxometalate,
-
[SiW9O37RuIII (H2O)3Cl3-x]7- (Ru-POM), was supported (ca. 1.4 % w/w) on
3
cellulose/propylamine-modified silica hybrid material obtained by sol-gel synthesis.
This new catalytic material (Ru-POM-CSH) has a significantly low bulk density (0.19
g/cm3) and high porosity (80%) due to the preservation of the physical structure of
cellulose and the general fibre shape of cellulosic pulp. Modified silica with attached
multisubstituted ruthenium POM was presented as a dense film on the surface of
cellulose fibres.
Multisubstituted ruthenium tungstosilicic polyoxometalate demonstrated catalytic
activity in the gas phase heterogeneous aerobic oxidation of formaldehyde at room
temperature. The ruthenium polyoxometalate proved to be a reversible redox catalyst
24

without significant loss of catalytic activity after more than 400 turnovers. The major
oxidation products were carbon dioxide and water indicating the possibility of complete
mineralization of formaldehyde in polluted air. However, the continuous purification of
air using Ru-POM-CSH material will be dependent on the level of CH2O concentration
and requires further studies. The extreme concentrations of CH2O used in this study (ca.
830 ppm of CH2O in air, i.e. three orders of magnitude (103) higher than the maximum
concentrations typically found in indoor sources and almost four orders of magnitude
higher than recommended indoor concentrations) allowed proof-of-concept but led to
undesirable catalyst saturation and hence decrease in the catalytic activity. However
catalyst regeneration was proven possible by air purging at room temperature.
Noteworthy, the proposed mechanism of CH2O oxidation catalysed by ruthenium
multisubstituted POM needs further detailed studies. In particular, the contribution of
autoxidation pathways to the CH2O oxidation should be clarified.
ACKNOWLEDGEMENTS
This work was carried out at CICECO – Aveiro, Institute of Materials within the scope
of grant financed by FCT (PEst-C/CTM/LA0011/2013).
25

References
[1] T.Godish, Sick Buildings: Definition, Diagnosis and Mitigation, Lewis Publishers,
Boca Raton, FL, 1995.
[2] SCHER (Scientific Committee on Health and Environmental Risks). Opinion on risk
assessment on indoor air quality, European Commission: Brussels, 2007.
[3] IARC (International Agency for Research on Cancer - World Health Organization).
IARC classifies formaldehyde as carcinogenic to humans, press release no.153,
International Agency for Research on Cancer, Lyon, 2004.
[4] T. Salthammer, S. Mentese, R. Marutzky, Chem. Rev. 110 (2010) 2536-2572.
[5] T. Salthammer, Angew. Chem. Int. Ed. Engl. 52 (2013) 3320-3327.
[6] A.T. Hodgson, H. Destaillats, D.P. Sullivan, W.J. Fisk, Indoor Air. 17 (2007) 305-
316.
[7] M. Pelaez, N. T. Nolan, S. C. Pillai, M. K. Seery, P. Falaras, A.G. Kontos, P.S.M.
Dunlop, J.W.J. Hamilton, J.A. Byrne, K. O’Shea, M. H.Entezari, D.D. Dionysiou, Appl.
Cat. B, 125 (2012) 331-349.
[8] L. Lin, Y. Chai, B. Zhao, W. We, D. He, B. He, Q. Tang, Open J. Inorg. Chem. 3
(2013) 14-25.
[9] K. Everaert, J. Baeyens, J. Hazard. Mater. 109 (2004) 113-139.
[10] Q. Xu, Y. Zhang, J. Mo, X. Li, Environ. Sci. Technol. 45 (2011) 5754-5760.
[11] R. Averlant, S. Royer, J.-M. Giraudon, J.-P. Bellat, I. Bezverkhyy, G. Weber, J.-F.
Lamonier, ChemCatChem, 6 (2014) 152 – 161.
[12] M. T. Pope, A. Muller, Angew. Chem. Int. Ed. Engl. 30 (1991) 34-48.
[13] M. T. Pope, A. Müller, Polyoxometalate Chemistry: From Topology Via Self
Assembly to Applications, Kluwer, Dordrecht, 2001.
[14] D. L. Long, E. Burkholder, L. Cronin, Chem. Soc. Rev. 36 (2007) 105-121 .
26

[15] N. Mizuno, M. Misono, Chem. Rev. 98 (1998) 199-218.
[16] R. Neumann, Prog. Inorg.Chem. 47 (1998) 317-370.
[17] M. Vazylyev, D. Sloboda-Rozner, A. Haimov, G. Maayan, R. Neumann, Topics in
Catal. 34 (2005) 93-99.
[18] O.A. Kholdeeva, M.P. Vanina, M.N. Timofeeva, R.I. Maksimovskaya, T.A.
Trubitsina, M.S. Melgunov, E.B. Burgina, J. Mrowiec-Bialon, A.B. Jarzebski, C.L. Hill,
J. Catal. 226 (2004) 363-371.
[19] O. A. Kholdeeva, M. N. Timofeeva, G. M. Maksimov, R. I. Maksimovskaya, W.
A. Neiwert, C. L. Hill, Inorg. Chem. 44 (2005) 666-672.
[20] W. Guo, Z. Luo, H. Lv, C. L. Hill, Aerobic Oxidation of Formaldehyde Catalyzed
by Polyvanadotungstates, ACS Catalysis, 4 2014, 1154-1161.
[21] J. A. F. Gamelas, M. G. Evtyugina, I. Portugal, D. V. Evtuguin, RSC Adv. 2 (2012)
831-839.
[22] I. V. Kozhevnikov, Chem. Rev. 98 (1998) 171-198.
[23] R. Neumann, M. Dahan, Polyhedron 17 (1998) 3557-3564.
[24] R. Neumann, M. Dahan, J. Amer. Chem. Soc. 120 (1998) 11969-11976.
[25] A. M. Morris, O. P. Anderson, R. G. Finke, Inorg. Chem. 48 (2009) 4411-4420.
[26] J. A. F. Gamelas, H. M. Carapuça, M. S. Balula, D. V. Evtuguin, W. Schlindwein,
F. G. Figueiras, V. S. Amaral, A. M. V. Cavaleiro, Polyhedron 29 (2010) 3066-3073.
[27] K.Wickholm, P.T. Larsson, T. Iversen, Carbohydr. Res. 312 (1998) 123-129.
[28] J. A. F. Gamelas, D. V. Evtuguin, A. P. Esculcas, Transit. Metal Chem. 32 (2007)
1061-1067.
[29] G. S. Caravajal, D. E. Leyden, G. R. Quinting, G. E. Maciel, Anal. Chem. 60
(1988) 1776-1786.
[30] B. J. S. Johnson, A. Stein, Inorg. Chem. 40 (2001) 801-808.
27

[31] R. C. Schroden, C. F. Blanford, B. J. Melde, B. J. S. Johnson, A. Stein, Chem.
Mater. 13 (2001) 1074-1081.
[32] S. Sequeira, D. V. Evtuguin, I. Portugal, A. P. Esculcas, Mat. Sci. Eng. C-Bio S 27
(2007) 172-179.
[33] A. C. O’ Sullivan, Cellulose 4 (1997) 173-207.
[34] H-X. Ding, A-M. Zhu, X-F Yang, C-H Li ,Y Xu. J. Zhang, X-F. Yang,. J. Phys. D:
Appl. Phys. 39 (2006) 3603–3608.
[35] C. Rong, M.T. Pope, J. Am. Chem. Soc. 114 (1992) 2932-2938.
28

Figure captions
Fig. 1 Experimental setup used for catalyst testing and formaldehyde determination in
the air mixture.
Fig. 2 13C CP/MAS NMR spectrum (A) and 29Si MAS NMR spectrum (B) of the newly
synthesised Ru-POM-CSH material.
Fig. 3 SEM image of novel Ru-POM-CSH material.
Fig. 4 FTIR spectra of cellulosic pulp (a), Ru-POM-CSH (b) and non-functionalized
CSH (c). The asterisk denotes the bands assigned to  (O-Si-O) and s (Si-O-Si), at 445
cm-1 and 793 cm-1, respectively.
Fig. 5 X-ray diffractograms of cellulosic pulp (a), Ru-POM-CSH (b) and non-
functionalized CSH (c).
Fig. 6 UV-Vis (k/s) spectra of initial cellulosic pulp (a), Ru-POM-CSH (b) and non-
functionalized CSH (c) and absorption spectrum of initial Ru-POM in aqueous solution
(30 mol dm-3) (d).
Fig. 7 Thermogravimetric curve of initial cellulosic pulp (a) Ru-POM-CSH (b) and non-
functionalized CSH (c).
Fig. 8 UV-Vis (k/s) spectra of Ru-POM-CSH before (a) and after (b) 7 cycles of
reaction with CH2O in polluted air and after the successive purging of the catalyst bed
with purified air for 6 h (c).
29