CHEMICAL ENGINEERING TRANSACTIONS
VOL. 57, 2017
A publication of
The Italian Association
of Chemical Engineering
Online at www.aidic.it/cet
Guest Editors: Sauro Pierucci, Jiří Jaromír Klemeš, Laura Piazza, Serafim Bakalis
Copyright © 2017, AIDIC Servizi S.r.l.
ISBN 978-88-95608- 48-8; ISSN 2283-9216
Development of Unconventional Photocatalytic Reactors and
Processes for the Abatement of Harmful N-containing
Pollutants
Ilenia Rossettia*, Matteo Compagnonia, Gianguido Ramisb, Francesca Freyriac,
Marco Armandic, Barbara Bonellic
a Dip. Chimica, Università degli Studi di Milano and INSTM Unit Milano-Università, Milan, Italy
b Dip. Ing. Chimica, Civile ed Ambientale, Università degli Studi di Genova and INSTM Unit Genova, Genoa, Italy
c Department of Applied Science and Technology and INSTM Unit of Torino-Politecnico, Politecnico di Torino, Turin, Italy
ilenia.rossetti@unimi.it
In the present work, innovative photocatalytic processes were developed for the abatement of harmful N-
containing compounds, focusing on selectivity towards innocuous N2, to be applied for the treatment of waste
waters to meet legislative specifications.
The photocatalytic performance of the samples has been checked for the photoreduction of nitrate ions, the
photooxidation of ammonia and of organic N-containing compounds. Catalytic performance was correlated
with the properties of the nanomaterials. Two innovative semi-batch photoreactors were specifically designed
and optimized for this application.
Different photocatalysts have been compared. TiO2 has been prepared in nanosized form by using an
innovative flame pyrolysis (FP) approach, able to synthesise in one step single or mixed oxide nanoparticles,
characterized by homogeneous particle size and good phase purity. In some cases Pd has been added to
TiO2 by post synthesis impregnation. The role of the metal nanoparticles was the enhancement of the lifetime
of the photoproduced charges by electron trapping. The same materials formulations have been prepared
starting from commercial nanostructured TiO2 supports for comparison purposes. The addition of Pd revealed
the key to improve both the conversion and the selectivity to N2. The FP prepared materials were in general
more active than the commercial TiO2 P25 sample.
1. Introduction
The World Health Organisation recommended levels of nitrates, nitrite and ammonium concentration in
drinking water to be 50 ppm nitrate, 3 ppm nitrite and 0.5 ppm ammonium. A further restriction was set to 10
mg/L for nitrates in drinking water by the US Environment Protection Agency. Such limits are hard to respect
through the current technologies, and this possibly affects the health of ca. 25 million people in the US (Burow
et al., 2010), making nitrates the most ubiquitous contaminant in drinking water sources (Nolan et al., 2010).
As well, ammonia is one of the major N-based pollutants in wastewater. It contributes to eutrophication (Lee et
al., 2002) and is toxic even at very low concentration for human and animal living species. Common methods
for NH4
+/NH3 removal are based on biological nitrification, stripping, breakpoint chlorination and ion exchange,
but each of them is characterised by disadvantages. For example, the efficiency of biological treatment is
highly dependent on temperature, dissolved oxygen, the carbon source, pH, and the concentrations of toxic
substances (Christensen & Harremoes, 1978).
The economic impact of nitrogen-based pollution for the EU-27 is safely estimated between €25 billion and
€115 billion per year, reaching almost triple value when taking into account also other strictly correlated
phenomena.
Among the removal methods, biological processes are generally regarded to be the most efficient. However,
they present disadvantages, including handling difficulties and large equipment requirements. Moreover, they
are difficult to apply to treatment of wastewater that contains harmful co-existing species for bacteria.
DOI: 10.3303/CET1757054
Please cite this article as: Rossetti I., Compagnoni M., Ramis G., Freyria F., Armandi M., Bonelli B., 2017, Development of unconventional
photocatalytic reactors and processes for the abatement of harmful n-containing pollutants, Chemical Engineering Transactions, 57, 319-324
DOI: 10.3303/CET1757054
319
Therefore, the development of cheaper and easily scalable systems for their abatement is more than
welcome. This relevant topic has been recently reviewed concerning both the removal of inorganic N-
containing pollutants (Compagnoni et al., 2017) and organic contaminants such as azo-dyes, drugs and
pesticides (Freyria et al., 2017).
In this work, we propose photocatalytic process for the reduction of NO3
- and the oxidation of ammonia. In
both cases the goal is to achieve the highest possible selectivity towards innocuous N2. The set up of a new,
scalable photoreactor is described, together with the photocatalytic activity for both reactions of
nanostructured catalyst prepared by flame pyrolysis.
2. Experimental
2.1 Catalysts preparation
A commercial sample of TiO2 P25 (Evonik) was used as reference material. It is a mixture of 20 % rutile and
80 % anatase, prepared in nanostructured form, with a specific surface area ca. 50 m2/g.
TiO2 was also prepared in nanostructured form by means of a home-made flame pyrolysis apparatus (FP).
This method is suitable for the continuous one-step preparation of single or mixed oxides and is described in
detail elsewhere (Chiarello et al. 2007). The frame pyrolysis apparatus includes a nozzle burner, which is co-
fed with oxygen and with an organic solution of the oxide precursors. The organic solvent acts as fuel for the
flame, whereas oxygen is the comburent and the dispersing agent. The TiO2 sample was prepared from a
solution of Titanium(IV)-isopropylate (Aldrich, pur. 97 %) in xylene, with a 0.67 M concentration referred to
TiO2. The burner was fed with a flow rate of 2.2 mL/min of the solution and 5 L/min of O2. The pressure drop
across the nozzle was set to 1.5 bar.
In case, Pd was added as co-catalyst with a 0.1 wt% concentration by wet impregnation from a solution of
Pd(NO3)2.
2.2 Photocatalytic tests
Photocatalytic reduction of NO3
− and oxidation of NH3/NH4
+ in water were carried out in a specifically designed
Pyrex reactor. The catalysts were suspended in an aqueous solution of NH4Cl (0.2M) or NaNO3 (0.006M) and
the suspension was stirred using a magnetic stirrer. The typical photocatalyst concentration was 1 g/L, based
on preliminary screening. The reaction suspension was thoroughly degassed and then exposed to He. The
reactor was operated in semi-batch mode: the solution containing the pollutant to be photoconverted was
added at the beginning of the reaction in batch mode, whereas a gas stream continuously flowed through the
reactor. The gas was composed by He during the conditioning-outgassing phase preliminary to every
measurement. He was fed in continuous mode also during the nitrate photoreduction tests, whereas it was
substituted by synthetic air (80 vol% He + 20 vol% O2) during the ammonia photooxidation experiments. In
both cases, the N2 produced was continuously removed from the reactor. A trap for ammonia, possibly
stripped from the reactor, was placed downstream.
A medium-pressure Hg vapour lamp was placed on the top of the photoreactor with a measured specific
power of 69.5 W/m2 and emitting in the wavelength range 254 - 364 nm.
The reactor temperature was kept at 298 K by cooling water circulating in a properly designed cooling jacket
around the photoreactor.
The evolved gas was analysed on-line by gas chromatography (Agilent, mod. 5980, He was used as carrier
gas) equipped with two columns connected in series (MS and Poraplot Q) and with a thermal conductivity
detector (TCD).
Ammonium/ammonia was determined through a Perkin Elmer Lambda 35 UV-Vis spectrophotometer,
following the Nessler method at a wavelength of 420 nm.
The concentrations of NO3
−, NO2
−, and possibly other interesting anions were determined by using a ion-
chromatograph (Metrohm, 883 Basic IC plus).
In a typical run, system outgassing was carried out for ca. 1 h by flowing He and analysing the outflowing gas
by gas chromatography until elimination of O2 and N2. After conditioning, a first sampling allowed to determine
the initial concentration of the reactant, then irradiation started and repeated sampling was carried out each
hour. Conversion and selectivity were calculated as follows:
𝐶𝑜𝑛𝑣𝑒𝑟𝑠𝑖𝑜𝑛 𝑖 (%) =
(𝑛𝑖 ° − 𝑛𝑖 (𝑡))
𝑛𝑖 °
× 100 𝑖 = 𝑟𝑒𝑎𝑐𝑡𝑎𝑛𝑡 (1)
320
𝑆𝑒𝑙𝑒𝑐𝑡𝑖𝑣𝑖𝑡𝑦 𝑗 (%) =
𝑛𝑗
𝜈𝑗 (𝑛𝑖 ° − 𝑛𝑖 (𝑡))
× 100 𝑗 = 𝑝𝑟𝑜𝑑𝑢𝑐𝑡; 𝜈 = 𝑠𝑡𝑜𝑖𝑐ℎ. 𝑐𝑜𝑒𝑓𝑓. (2)
3. Results and discussion
3.1 Set up of the testing rig
A first experimental set up was tested for both the ammonia photooxidation and the nitrates photoreduction
reactions. The reactor was irradiated externally from the topside, which was closed through a quartz window,
transparent in the UV-Vis range of interest. The sketch of the reactor is reported in Figure 1.
Figure 1: Sketch of the photoreactor used. CW = cooling water.
A semi-batch reactor configuration was adopted, with a gas stream constituted by 20 vol% O2 in He
continuously bubbling in the ammonia-containing suspension of the photocatalyst during the photo-oxidation
process, whilst only He was used during the nitrate reduction tests. In the former case one of the reactants
was fed in batch mode (ammonia), oxygen continuously, and one of the products, molecular nitrogen, was
continuously removed. It should be underlined that the use of He is here suggested in order to evaluate the
formation of N2 as product. Once the process is optimized, air and N2 are considered for scalu-up. The
possible over-oxidation byproducts, nitrates and nitrites, were kept in the photoreactor, thus allowing their
possible combined post treatment. Similar approach was adopted during the tests for the photoreduction of
nitrates, where the reactant, sodium nitrate, was fed in batch mode, whereas the desired product (N2) was
continuously withdrawn, leaving in the reactor the possible over-reduction product, i.e. ammonia/ammonium.
This set up also allowed to quantify the selectivity to the desired product, nitrogen, using a simulated air in the
case of ammonia photoxidation and pure He during the photoreduction experiments. This is a very challenging
point. Indeed, usually the selectivity to N2 is evaluated from N-balance, after quantification of nitrate, nitrite and
ammonia and neglecting possible other intermediates (e.g. nitrosoamines, which are substances of concern)
(Ren et al. 2015).
The external lamp was placed on the top of the photoreactor in order to simulate a future application directly
by sunlight. This configuration allows easier scale up with respect to reactor configurations based on
immersion lamps sometimes found in the literature, which have the advantage of closer (and thus more
efficient) irradiation of the reacting suspension, but are not easily feasible for scale up, especially if the final
scope is the use of sunlight. In the present configuration with external lamp, the correct exposure of the
photocatalyst to the light source should be optimised by improved reactor geometry (proper height/diameter
ratio) and appropriate mixing of the suspension. Testing on different photoreactors geometries and fluid-
dynamic modeling are currently in progress to optimise this point.
Another issue is the possible stripping of ammonia (reactant or product) by the flowing stream, depending on
pH. Of course, this may prevent reliable ammonia quantification, but it can also damage the GC columns
when an online GC is used, as in the present case. In order to cope with both issues we have placed a
downstream removable ammonia trap. In a first design, the gas stream was sent to an absorption column,
filled with diluted H2SO4. The solution was titrated at the end of each test to quantify the possibly stripped
321
ammonia. This configuration induced an excessive pressure drop to allow the further inlet of the gaseous
stream to the online GC, therefore the absorption system was substituted with an adsorption bed, filled with
zeolites in acidic form. The possibly stripped ammonia was quantified in this case by temperature
programmed desorption (TPD) up to 800°C at the end of the test. The ammonia possibly released was
quantified through a TCD detector. During system set up we performed different testing by varying the gas
flowrate, the pH of the solution and ammonia concentration in the photoreactor. We never detected ammonia
transported downstream by stripping with both traps configurations. Therefore, in the final design we kept the
adsorption trap containing zeolites as a guard bed, only.
3.2 Photocatalytic tests
A preliminary set of activity tests for both reactions have been carried out on bare titania photocatalysts. This
semiconductor was chosen considering its well known photoactivity and photoresistance towards corrosion
(Ohtani et al. 2010). In particular, we compared the performance of a FP-prepared catalyst with the
commercial P25 benchmark. The activity of the catalysts for the photoreduction of ammonia is reported in
Figure 2.
Figure 2: Ammonia conversion versus time. Selectivity 100 % to N2.
The TiO2 sample prepared by FP exhibited higher activity with respect to the commercial sample, reaching ca.
20 % conversion after 5 h, without indication of decay or deactivation. By contrast, the commercial sample
was less active and most of all its activity was rapidly ruled out. Some induction period was observed, due to
the need of sample conditioning upon irradiation. The most interesting point is that no trace of nitrites or
nitrates was observed, confirming a full selectivity to the desired product, i.e. molecular N2.
Figure 3: Nitrate conversion versus time. Selectivity 100 % to NH3.
322
Additionally, we tested the catalysts for the photoreduction of nitrate. The results are reported in Figure 3. The
activity in this case was lower and selectivity was 100 % to ammonia, a highly undesired by-product.
Moreover, competitive reactivity was observed among the two reactions. Indeed, as soon as ammonia started
to form, it was oxidised back to nitrate with decreasing conversion. The problems to be solved are two, one
related to catalyst formulation, in order to enhance its activity and selectivity, the other related to process
design, through the addition of a hole scavenger such as and organic compound, easier to oxidise with
respect to ammonia. In the following we will deal only with the former issue due to the required brevity of this
communication.
Pd was added in very small amount to the FP-prepared catalyst. Its activity and selectivity during the
photoreduction of nitrate are reported in Figure 4. One order of magnitude higher nitrate conversion was
achieved upon the addition of a very small amount of Pd, without formation of nitrite and much more limited
selectivity to ammonia, which became a secondary byproduct, although still too relevant.
Figure 4: Nitrate conversion and selectivity to NH3 versus time. Catalyst composition: 0.1 wt% Pd/TiO2
prepared by FP.
Based on these results and given the full selectivity to N2 during the photooxidation of ammonia, a two step
process may be conceived, with the photoreduction of nitrates (and nitrites) as a first stage, followed by the
photooxidation of the originally present and newly formed ammonia.
4. Conclusions
A photoreactor has been set up allowing to achieve the photoreduction of nitrates and the photooxidation of
ammonia in semibatch mode. Its configuration, to be further optimised as for geometry and mixing, is suitable
for scale up, easily implements immobilised photocatalysts and is ready to use with direct sunlight. These first
results on both reactions let us conclude that the flame pyrolysis procedure is a viable technique for the
preparation of either bare or metal-doped semiconductors in nanosized form, to be used for the photocatalytic
abatement of inorganic N-containing pollutants in waste or drinking waters. Even if a photoreactor with
external irradiation is less efficient than one base on immersion lamps, it is much more suitable for scale up
and most of all for the direct application with solar light. Sufficiently high ammonia conversion was achieved
during photooxidation in semibatch configuration, up to ca. 20 % after 5 h, with 100 % selectivity to N2.
The same undoped catalyst was much less active during the photoreduction of nitrate, requiring the addition of
a small amount of Pd to achieve suitable conversion (ca. 14 % after 3 h, with less than 20 % selectivity to
NH3). The addition of the noble metal also allowed to improve the selectivity to N2, limiting the undesired
formation of ammonia, which was instead the only product observed with the undoped titania catalysts.
According to these results, a two step process may be designed, achieving the full removal of nitrates and
subsequently oxidising ammonia with 100 % selectivity to N2.
Acknowledgments
The financial support of Fondazione Cariplo (Italy) through the measure “Ricerca sull’inquinamento dell’acqua
e per una corretta gestione della risorsa idrica”, project “DEN - Innovative technologies for the abatement of N-
containing pollutants in water”, grant no. 2015-0186, is gratefully acknowledged. The valuable help of Paola
Marzullo, Gaia Mascetti, Veronica Praglia and Alberto Riva is gratefully acknowledged.
323
Reference
Burow K.R., Nolan B.T., Rupert M.G., Dubrovsky N.M., 2010, Nitrate in groundwater of the United States,
1991-2003, Environ. Sci. Technol., 44, 4988-4997.
Chiarello G.L., Rossetti I., Lopinto P., Migliavacca G., Forni L., 2007, Solvent nature effect in preparation of
perovskites by flame-pyrolysis. 1. Carboxylic acids, Appl. Catal. B: Environ., 72, 218-226.
Christensen M.H., Harremoes P., 1978, Nitrification and denitrification in wastewater treatment, Water Pollut.
Microbiol., 2, 391-414.
Compagnoni M., Ramis G., Freyria F.S., Armandi M., Bonelli B., Rossetti I., 2017, Photocatalytic Processes
for the Abatement of N-Containing Pollutants from Waste Water. Part 1: Inorganic Pollutants, Journal of
Nanoscience and Nanotechnology, 17, 3632–3653.
Freyria F.S., Armandi M., Compagnoni M., Ramis G., Rossetti I., Bonelli B., 2017, Catalytic and Photocatalytic
Processes for the Abatement of N-Containing Pollutants from Wastewater. Part 2: Organic Pollutants,
Journal of Nanoscience and Nanotechnology, 17, 3654–3672.
Lee J., Park H., Choi W., 2002, Selective Photocatalytic Oxidation of NH3 to N2 on Platinized TiO2 in Water,
Environ. Sci. Technol., 36, 5462-5468.
Nolan B.T., Puckett L.J., Ma L.W., Green C.T., Bayless E., Malone R.W., 2010, Predicting Unsaturated Zone
Nitrogen Mass Balances in Agricultural Settings of the United States, J. Environ. Qual., 39, 1051-1065.
Ohtani B., 2010, Photocatalysis A to Z—What we know and what we do not know in a scientific sense, J.
Photochem. Photobiol. C. Photochem. Rev., 11, 157-178.
Ren H.T., Jia S.Y., Zou J.J., Wu S.H., Han X., 2015, A facile preparation of Ag2O/P25 photocatalyst for
selective reduction of nitrate, Appl. Catal. B: Environ., 176-177, 53-61.
324