Adsorption and oxidation of sulfur dioxide on particles

Adsorption and oxidation of sulfur dioxide on particles


648KB Sizes 0 Downloads 18 Views




A. LIBERTI,D. Bxocco and M. POSSANZINI Laboratorio Inquinamento Atmosferico C.N.R., Via Montorio Romano, 36 Rome, Italy (Received




Abstract-The role of atmospheric particulated matter in afiecting atmospheric SO2 and its reactions has been investigated. A variety of dusts of various sources (urban particulated matters and stack emissions of industrial plants) have been characterized in terms of their physical and chemical properties and submitted to a SO1 adsorption process at room temperature and desorption at 175°C. The interaction between particles and SO2 can occur through two processes: adsorption and conversion to sulphate. The extent of these processes depends upon the particles, chemical composition and their nature, which can be defined in terms of pH, titratable acidity, surface area, humidity and degree of surface coverage by adsorbed components. SO2 adsorption by particles is the primary process which occurs in two steps, only the first one being apparently of environmental significma. Humidity has an important role in the adsorption. the higher its value the higher results the amount of adsorbed SO,. The behaviour of atmospheric dusts collected in di!Terent areas and seasons is very similar, the reaction constant of the first order process being 4-5 x lo-’ min-‘. Both ‘fresh’ particles coming from stacks, which do not carry S02. as well as ‘aged’ atmospheric particles adsorb SO2 the relative extent of the process being mainly determined by the dust reaction. The former in most cases do not release SOI by heating, this behaviour being taken as an example of chemisorption, whereas the latter lose SO2 by heating, SO2 being retained by only a physical bond. Conversion to sulphate occurs with a very high rate on particles coming from industrial emissions, the alkaline reaction being the determining factor, whereas it does not take plaa on urban atmospheric dusts to an appreciable extent. This mechanism is supported by measurements by differential thermal analysis and by X-ray photoelectron spectroscopy. Though, in the atmosphere, it is impossible to discriminate various effects due to homogeneous and heterogeneous reactions, the main interaction between SO2 and particulated matter is adsorption. most catalytic reactions occurring at high temperature and most probably at the chimney outlet.

Particulate matter is the most abundant pollutant and its composition and effect on the environment has

been the object of many investigations. The behavior of particulate matter is usually related to its concentration and to date little weight has been put on reactions which either might occur on it or might be al&&d by its presence. The role of atmospheric particulate matter vs SO1 has been long debated and it has been observed that in the presence of several finely ground particles the sulfur dioxide concentration decreases; a catalytic effect is usually attributed to this material. Reactions of sulfur dioxide with particles have been assumed to yield sulfate and sulfuric acid (Cheag et a!., 1971; Corn and Cheng, 1972). A theoretical treatment of particle-catalyzed oxidation of atmospheric pollutants including SO1 has been given to assess the possible importance to the overall pollution problem (Judeikis and Siegel, 1973). However, in an open atmosphere it is impossible to discriminate between homogeneous and heterogeneous reactions as well as gas-to-particle conversion, including heteromolecular nucleation, condensation and thermal coagulation,.as all processes occur

simultaneously and are greatly influenced by humidity and by meteorological conditions. It must be stressed that it is incorrect to discuss atmospheric particulate matter as a unique system with identical properties in all areas, as its origin can be quite different as a result of dispersion and condensation. The former process occurs by grinding or atomization of solids and liquids and by the transfer of powders in suspension through the air stream, whereas the latter takes place when supersaturated vapour condenses to give formation of non-volatile products. Both processes occur simultaneously and it is obvious that specific industrial emissions can affect the nature and composition of atmospheric particulate matter. On account of its complex composition and nature, it seemed essential to define the characteristics of particulate matter and to submit dusts of various sources to an SOI-adsorption and dcsorption process to obtain information of general validity on the interaction of SO1. The aim of this investigation was the study of the behavior of particulate matter collected from the urban environment as well from stacks, in terms of




B~occo and M P~SSANZINI

Water pump


“H”m!d,f aw


Fig. 1. Experimental set-up for SO2 adsorption sulfur dioxide temperature.


and oxidation

at ambient


The experimental set-up is shown in Fig. 1. A dust sample (1ODmg) was placed into a reactor consisting of a glass tube 15 cm length, 1 cm i.d., with a sintered glass soldered in the middle part. The reactor was set in a vertical position and air flowing through the fritted glass kept the dust floating giving conditions similar to a fluid bed. Purified air at a constant Bow (SOml min-‘) at atmospheric pressure was passed into a hygrostat to acquire a constant humidity. Sulfur dioxide was added as the gas stream flowed through a thermostated condenser at 20°C containing a permeation tube. In most experiments the rate of permeation was l.l51(~min-’ and the SOI concentration in the gas stream became 8.6 ppm. Measurements were carried out either with a continuous monitoring of SO2 with a flame photometric analyzer (Monitor Labs. mod. 8450) or discontinuously by adsorp tion of SO2 into a sodium tetrachloromercurate solution usmg the West and Gaeke method. With the former procedure the intensity of the current versus time was recorded and the amount of SO2 obtained by integration. With the second method the amount of SO, adsorbed was calculated as the difference between the amount measured and that obtained by running a blank with the reactor empty. A typical graph obtained in the adsorption process with a continuous monitoring of SO1 is shown in Fig.

by dusts.

2. Plots of this type have been used to obtain kinetic mformation on the adsorption process. In all experiments, the SO1 concentration was kept constant whereas gas streams with various humidities were obtained by placing solutions of various composition in the hygrostat at 20°C (35% Hum: sat. solution CaCII. 6H,O-559; Hum: sat. solution Ca(NOs)2~4H&--76% Hum: sat. solution NaCl). To obtain information on the release of SO2 by atmospheric particles. desorption experiments were carried out by using the set-up shown in Fig. 3. Nitrogen was made to flow (50 ml min-‘) through a heater kept at 175°C and then mto the absorbing solution where SO, was determmed. Airborne samples collected from various sources were investigated by evaluating their behaviour in terms of the above described adsorption-desorption process. They can be classified as: (a) urban particulate matter collected directly from the atmosphere by means of an electrostatic precipitator or from ventilation plants of buildings in various seasons; (b) black emissions from various sources: fly ash from oil and coal electric power furnaces. from a cement factory, and from iron and steel plants. Some of them were sampled in the chimney and some from electrostatic filters before the chimney. The above materials were characterized in terms of the following parameters: pH; titratable acidity or alkalinity; surface area and relative porosity; humidity; degree of coverage of the surface by other adsorbed components or reaction products of former reactions. The pH of a dust can be measured with a glass electrode by stirring a sample of dust (150 mg in lOOmI de-ionized water). The titratable acidity or alkalinity was measured by electrometric titration of the water suspension using Gran’s plots according to Brocco et al. (1976). The surface area could be determined by nitrogen adsorption (Devitofranccso and Liberti, 1966). No general

> E








Fig. 2. Measurement of SO, adsorption by an atmospheric dust (sample 1) vs time, SO2 being analyzed by flame photometric detector. Operating conditions: mfluent SOI cont. 8.6ppm; gas flow rate-50 ml min-‘; temperature 20-C: r.h.-35?,. Curve A: empty reactor. curve B: reactor containing 1OOmg of atmospheric dust.




Fig. 3. Apparatus used for desorption measurements.


Adsorption and oxidation of sulfur dioxide Table 1. pli, titratable acidity, surface area. S”/,. C?; and free SO,

H’ lmeqg-‘)


1 Atmoipbenc 2 3 4 5 6 7 8 9 10

dust (Rome Sumriter 74) Atmospbenc dust (Rome Wrntef 74-75) Atmospheric dust (Mdut STMI 139) Atmospbmc dust (Turm Wtnter 75) Dust from cement fwtory stack (fksct~) bust from ttmn~ac rttdurtry strclr (Yugoslnvm) Dust from blast ftuttwx PHF> (Fmtta! Dust from lion ~~lplomer~tton pIsnt (Francet Dust from LD fumra. PAC (Fmna) Soot from cxpmmcntal oil furnace (Prsa)

I I Ash tom od burner m clectr~c powct *t,tmn (Monf&oa) 12 Ash from naphtha combusttoo (Rome) 13 Fly srh from o nixpbtha burner (Remet 14 Flv ash from coal fired wwr clattt (Mannbeam Df

_ -

1.27 668 600 550 10 77 a66 993 8.10 1024 341

0 357 0.350 0142 0.066


0088 0.210 0152 0238 -

207 1 so 680 640

1.880 7.100 _ -

0.015 -




procedure was foilowed to evaluate the degree. of coverage of the surface: solvent extraction was used to determine organics of a dust and gas eiution in a closed system lo strip off volatile compounds.

RESULTS AND DISCUSSION Analytical results are collected in Table 1. With the exception of samples 7, 8 and 9, which have a high iron content, other metals such as manganese and aluminum, which are usually catalytically active, are present in the dusts only as minor components. Atmospheric dusts collected in summer time are almost neutral whereas those sampled in winter have an acid reaction. The pH and the surface activity of particulates from stack emissions differ noticeably from each other. They can be definitely basic as cement and LD furnace dusts or acid as most fly ashes. Adsorption&sorption



Results for the adsorption process described collected in Table 2. They refer to standard, trolled experimental conditions where air with a tive humidity of 55% at 20°C containing 8.6 ppm flowed for 30 min through a 100 mg sample.

are conrelaSO2

surfux Prcl (m’g-‘)

S”. lW/WI

c. (W/WI

81 12 II I2 27 I.8 8.0 38 20 190 .

3.3 2.8 47 20 1.3 1.3 05 0.6 04 7.4

226 224 42.0 _ 57 150 3.5

06 14 138






SO2 lmgg II 07 0.9 1.2 14 0 0 0 0 02 24 14 0.7 0t -

The pH and the particuiates’ acidity are the most important factors which a&et SO2 adsorption, dust with an alkaline reaction not being saturated in the experimental conditions described. However, all materials adsorb SO*, the amount adsorbed increasing with increase in the humidity. A plot of SO* adsorption on winter atmospheric dust measured at various humidities, is shown in Fig. 4. Measurements carried out with the same dust after benzene extraction at different humidities are also shown in the same graph. In both cases the higher the humidity, the higher is the amount of SOz adsorbed. Desorption experiments indicate that urban atmospheric dust desorbed SO1 with an almost quantitative yield, whereas in most dusts collected from stacks no desorption occurs on heating at 175°C under a nitrogen flow. A comparison between the results of Table 2 and the content of free SO2 allows particulate matter to be classified in terms of its ‘aged’ or ‘fresfi’ nature. A stack emission particulate can be defined as a ‘fresh system which is going to equilibrate in the atmosphere, whereas atmospheric particulate matter. unless

TabIe 2. Adsorption and desorption of SO;! by particulate matter


2 3 4 5 6 7 8 9 10 II 12 13 14 *

75 lS2 127 10s 345 96 34s 345 345 18 7 3:: 60

ARer 30 min flowing

r h 59 ’ i DLkion

23 44 37 3&t 100 28 100 loo Ial 6 2 4 96 18 air


carried out at 12YC.

100 100 100 100 0 0 0 0 25 0 0

8.6 ppm SO2 and

Fig. 4. SO2 adsorption by atmospheric dust (sample 2) at various humidity .levcis: _ (a) _ untreated sampie; (b) the same sample after benzene extraction.









Fig. 5. SOI adsorption rate by various atmospheric dusts; conditions as in Fig. 2.

directly affected by a specific emission, is an ‘aged’ system where equilibrium with air components has been reached. Particulate matter sampled from a stack does not show the presence of SOr, which is always found in ‘aged’ systems. Both dusts might be reactive but reactions in a ‘fresh’ system proceed faster and are much more extensive. The extent of SO2 adsorption depends mainly upon the titratable alkalinity and, accordingly, samples 5-9. which are ‘fresh’ dust, consisting mainly of various metal oxides, have a high capacity for SO* uptake which is taken directly from the atmosphere. However, adsorption also occurs on industrial dust having a weak alkalinity or an acid surface. Fly ashes should in no sense be considered homogeneous material and they can be assumed to consist of acid and alkaline centers. The former do not take part in the adsorption process, whereas the latter, though comparatively very few. should be considered responsible for the reaction with SOz. Samples 10-14 exhibit this specific behavior.

where qr IS the amount of SO2 adsorbed at the equilibrium per g of particulate matter and q, the amount of SO* adsorbed at time 1. Constant K, is almost the same for atmospheric dusts, being in the range 45 x lo-’ min-’ as is shown in Fig. 5. The reaction seems to be of the first order and it appears that adsorption is a quite fast event. To this ptocess, which is believed to be of environmental significance, a second rather slow process occurs which is rendered evident by the additional uptake of SO2 by a sample dust previously treated with this gas. It has been observed that after equilibration with SO2 a dust sample adsorbs a further amount of SO2 after a certain time (l-4 days). This process is described by the relationship:

The adsorption experiments described in Fig. 2 allow the interaction of SO2 with particulate matter to be evaluated; this seems to occur through two processes : (I) adsorption; (2) oxidation of adsorbed SO2 to sulfate. Measurements carried out with atmospheric particulate matter, which does not adsorb SO2 quantitatively, indicate that adsorption occurs in two steps. The first step, which is very fast, can be described by the equation:

where qrds is the amount of SO, adsorbed at time t. K2 = 6 x 10-5min-‘. Given the dynamics of the atmosphere, this process may be of no relevance for the environment. In all dusts where the adsorbed SO2 is quantitatively desorbed by heating at 175°C. no chemisorption occurs and no sulfate formation was observed. On the other hand, in dust samples where SO2 adsorbed is not released at 175°C such as the alkaline dust sampled in a cement factory stack, no free SOz was detected and the sulfate weight increase corresponds to the amount of sulfur dioxide adsorbed. The information obtained from the adsorptiondesorption process ate supported by the application of t.g.a. and d.t.a. and by X-ray photo-electron spectroscopy. Thermal analysis of atmospheric particulate matter shows an endothermic process due to desorp-

b C4./(% - dl = K,t

tion of SO*, which occurs at about 125°C. The energy

Interaction of SO2 with particulate matter


Adsorption and oxidation of sulfur dioxide



(c) 2-



* 3-

0 E






dY-&Adi Temperature,


Fig. 6. (a) thermal analysis of an atmospheric dust (sample 1 of Table 1). (b) thermal analysis of the same sample after SO2 adsorption (120 pg g- ‘). (c) differential thermal analysis (DTA) of (b) vs (a).

involved in this process corresponds to a physical adsorption. By exposing the sample to sulfur dioxide the desorption peak increases, as is shown in Fig. 6. The desorption peak is not observed in samples taken from stacks of various emissions, whereas it is found in soot and combustion ashes. By keeping urban atmospheric dusts or fly ashes either in a desiccator or in the open atmosphere for long periods of time (three months) thermal analysis shows exactly the same desorption peak, indicating that the SO2 content does not change appreciably. This indicates that in most dusts oxidation of sulfur dioxide does not occur to an appreciable extent. If, however, a dust is alkaline, the conversion of adsorbed SO2 to sulfate is very fast. Additional information was obtained by using X-ray photoelectron spectroscopy (XPS). The study of the kinetic energy of photoelectrons expelled from a dust sample irradiated with mono-energetic X-ray provides a direct measurement of the electron binding energy of one element. Since the binding energies are modified by the valence electron distribution, it is possible to obtain a general picture of the various

sulfur containing species (Allegrini and Mattogno, 1977). The XPS spectrum of particulate material from an atmospheric dust (sample 3) shows a peak corresponding to a binding energy of 154 eV, due to Si2, a small band at 162 eV due to S( -2A a barely visible break assigned to S( +4) and a large peak at 168 eV characteristic of S( + 6) (Fig. 7). The XPS spectrum of another urban dust has a similar shape (sample 1) which shows only variations of the peak size due to dilferences in surface concentration. On the other hand, the spectrum of a ‘fresh dust collected from a cement factory (sample 5) does not show the shoulder due to S(+4), in agreement with our findings (Fig 8(a)). In agreement with this, other stack emissions have a similar behavior, as appears from sample 6 (dust from a manganese plant) where the S(+4) peak is not detected (Fig. 8(b)). ROLE OF ORGANIC3 IN



Adsorption experiments carried out with dusts, which were first extracted with benzene to bring most

BIndIng energy ,


Fig. 7. XPS spectrum on an atmosphere dust (sample 3).



Stack emissions with an alkalme reactlon adsorb SO1 but, since sulfites have never been detected in these particles. presumably chemlsorptlon occurs wnh formation of sulfate. AdsorptIon of SO2 b) this material leads to quantitative converslon to sulfate. provided the dust has 3 baste reaction. In certain cases. dusts with an acid reaction are observed which have a high concentratton of sulfuric acid. As the high acidity 1s usually found m ashes, this means that H,SO, formatlon takes place to a large extent m the combustion process. The gaseous SO, reacts with the chemisorbed oxygen at a high temperature and sulfate

ions are formed. The technical literature emphasizes the catalytic actlon of atmospheric particulate matter, in connectlon with carbon particles (Novakov and Chang, 1974). vanadium pentoxide (Barbaray er al., 1977) and Fig. 8. XPS spectra: (a) dust from a cement factory stack (sample 5); (b) dust from a manganese plant stack (sample 6). organ& into solution, show that in most cases a larger amount of sulfur dioxide is adsorbed. Experiments on dusts l-4 show that more than twice the amount of SO2 is adsorbed by the same dust after extraction. The larger amount of SO2 adsorbed means that a larger number of active sites occupied by organic compounds becomes available for SO1 adsorption. Since several organic compounds have been detected in atmospheric particulate matter (Brocco et al., 1975). the problem arises whether, besides oxidation to sulfate, SO2 might also react with other species. However, the concentration of various sulfur-containing compounds is small and it appears that they might come from exhaust gases and stack emissions, rather than through combination of SO2 with organic compounds from various sources. Though it is also reported (Conte et al.. 1976) that by equilibrating particulate matter with SOZ. sulfonic acid derivatives are formed, the formation of these compounds seems to be of minor importance in the conversion


of S02.


The variety of atmospheric dusts which have been examined and of the experimental conditions tested allows some general conclusions on the interaction between atmospheric particulate matter and SO1 at ambient temperature to be drawn. The main process SOI undergoes is adsorption, which takes place to the extent of several pgg-’ of dust, through a quite rapid first order reaction, which is strongly influenced by humidity. SO2 remains on the particulate matter as such. or in a hydrated form, so that it is introduced

into the living organism not only in the gas phase but also adsorbed on particles. These results may be of some importance in assessing the biological action of dust, as it always carries a layer of free SO1. Conversion to sulfate at room temperature does

not take place to a considerable extent on atmospheric dusts unless aerosols due to industrial emissions of a specific nature affect their composition.

ferric oxide (Chung

and Quon,

1973). In most cases,

however, experiments have been carried out at fairly high temperature and it is likely that these conditions occur only at the outlet of a chimney where the high reactants concentration and the high temperature may favor the oxidation rate of SOZ. At ambient temperature, however, the heterogeneous non-photochemical sulfate formation reaction strongly depends upon the reaction of the aerosol surface. An acid-base reaction is in most cases the determining factor, whereas on carbon it seems that sulfur dmxide adsorbs at actlvr sites of oxygen complex on carbon.







--C-OH and -COOH (Smith. 1959). An oxidation mechanism, which involves surface radicals (especially the hydroxyl group) might be therefore introduced (Yue ef al., 1976). It can be concluded that whereas in an open atmosphere SO2 adsorption is by far the most important process, a competitive reaction occurs at the chimney outlet. Sulfate-bearing particles are emitted into the atmosphere. together with SOZ, as a primary pollutant. The ratio SO :-SO, depends upon the combustion regime. :he type of the emissions, the size and especially

the nature

of the surface

area of particles.

These conclusions are in accord with the results of Forrest and Newman (1977) who pointed out that most oxidation might occur in a coal-fired power plant durmg the early history of the plume with virtually no further conversion taking place down wind. REFERENCES

Allegrmi I. and Mattogno G. (1977) Analysis of environmental particulate matter by means of ESCA. Paper preoared for the Seminar on Fine Particulates (ECE-ONU) iillach (Austria) 17-21 October. Barbaray B.. Contour J. P. and Mouvier G. (1977) Sulfur dioxide oxidation over atmospheric aerosols; X-ray photoelectron spectra of sulfur dioxide adsorbed on V,O, and carbon. Atmospheric Environment II, 351-356. Brocco D.. Liberti A. and Ponsanzini M. (1976) Determinazlone dell’acidlt8 del materiale particolato in atmosfere urbane e mdustriali Ann. 1.~1 Super. Sanirti 12. 49-55

Adsorption and oxidation of sulfur dioxide Brocco D, Liberti A. and Possanzini M. (1975) AdsorptIon desorption of sulfur dioxide by air-borne particulate matter. Presented at the 3th Technical Symposium COST Project 61a Ispra 18-20 November. Cheng R. T., Corn M. and Frohliger J. 0. (1971) Contribution to the reaction kinetics of water-soluble aerosols and SO1 in the air at ppm concentrations. Atmospheric Environment

5, 987-1008.

Chun K. C. and Quon J. E. (1973) Capacity of ferric oxide particles to oxidize sulfur dioxide in air. Encir. Sci. Technol. I, 532-538.

Conte C., Devitofrancesco G. and Starace G. (1976) Behaviour of SO, in the atmosphere-interactions wth the particulate matter. Atmospheric Pollution, pp. 243-253. Elsevier. Amsterdam. Corn M. and Cheng R. T. (1972) Interaction of sulfur dioxide with insoluble suspended particulate matter. J. Air Pollut.


Ass. 11, 870-875.


Devitofrancesco G. and Liberti A. (1966) Determining the dust concentration by surface measurement. Stauh 26. 13-15. Forrest J. and Newman L. (1977) Further studies on the oxidation of sulfur dioxide in coal-fired power plant plumes. Atmospheric Environment 11, 465-474. Judeikis H. S. and Siegel S. (1973) Particle-catalyzed oxidation of atmospheric pollutants. Atmospheric Enoironmenf 7, 619-631.

Novakov T. and Chang S. G. (1974) Catalytic oxidation of SO, on carbon particles. Presented at the 76th National AICHE Meeting, Tulsa. Oklahoma IO-13 March. Smith R. M. (1959) The chemistry of carbon-oxygen surface compounds. Q. Rev. 13, 287-395. Yue G. K.. Mohnen A. V. and Kiang C. S. (1976) Sulfur dioxide to sulfate conversion in the atmosphere. 12th Int. Colloquium Poll. Atm, Paris.