Issue 2 -  2001/02

 ISSN 1311-8978

ATMOSPHERIC DUST AEROSOLS IN SOFIA – ELECTRON MICROSCOPE CHARACTERIZATION

AND ICE NUCLEATION PROPERTIES

Valeria B. Stoyanova, Temenujka N. Kupenova ^(*), Tsenka I. Tsacheva,

Miko V. Marinov, Bogdan S. Ranguelov, Irina S. Georgieva

Institute of Physical Chemistry, Bulgarian Academy of Sciences, Sofia 1113, Bulgaria

^(*) Institute of Nuclear Research and Nuclear Energy,  Bulgarian Academy of Sciences, Sofia 1784, Bulgaria

 Received 10.09.2002; Cited 10.11.2002

Abstract

An elaborate investigation of atmospheric dust aerosols collected from different regions of Sofia city is presented as an attempt to propose a method for their characterizing as complicated systems by introducing of some quantitative physicochemical parameters, more or less formal. Dust particles were collected from smooth surfaces of a lee protected against rainfall where they were accumulated by free sedimentation. The distribution of particle diameters was determined by optical microscope for two samples. The elemental contents of three samples was studied by Electron Probe X-Ray Microanalyzer, and the morphology, shape and sizes of isolated particles of these atmospheric pollutants were additionally observed in Scanning Electron Microscope. The temperature dependence of their steady-state ice nucleation rate was investigated on the basis of experiments on freezing water drops seeded with the same dust aerosols at different concentrations. The ice nucleation activity was investigated in detail in the case of one sample and the assumed three types of active centers were characterized by such parameters as the activity factor F_i , the “wetting” angle q_i^o, and the effective specific surface free energy s_i. This investigation is continuing in the framework of project with the National Fund for Scientific Research under Contract No. F-1008/2001.

Key words: Atmospheric Pollutants, Aerosol Dust, Electron Probe X-ray Microanalyzer, Scanning Electron Microscopy, Ice Nucleation Rate, Nucleation Activity, Morphology, Elemental Content

Introduction

Among the most important aspects in aerosol pollutions study is an understanding of their role as condensation nuclei in the atmospheric processes and their influence on human health. A detail discussion needs knowledge for quantitative description of their sources, transport, composition, size distribution and morphology, as well as laboratory and field tests of their nucleation activity. The great number of such investigations from many years till now gives only a proof how important and notwithstanding insufficient they are [1-7, 27-33].

Systematic electron microscope studies of the characteristics of atmospheric dust in Sofia were carried out by some of the authors (M. Marinov, T. Tsacheva and B. Ranguelov) for many years [8-11]. They aimed at elucidating the peculiarities of dust components in East, Central and West parts of Sofia city characterized as industrial and busy traffic regions. A comparison was made with other regions of Bulgaria – Pirdop and Trojan known as typical representative of metallurgical industry and healthy tourist regions. Their experience was enriched also with detailed analyzing of such exotic airborne objects as the hot particles ejected by exploded Chernobil reactor and examples of dust particles from Sahara desert fallen in Sofia as “yellow rain”.

In the present work, atmospheric dust aerosols collected from different regions of Sofia (Bulgaria) are investigated with an electron microscope. The morphology, size distribution and chemical composition are specified. On the other hand, drops of distilled water were seeded with the same pollutants of different concentrations and were investigated when freezing, as described in detail in our previous papers [12-15]. An attempt was made one to combine the detailed electron microscopic analysis of atmospheric dust particles with investigations and quantitative characterization of their ice nucleation activity.

Microscopic investigations and results

All microscopic investigations in this study of the elemental content of dust samples are performed on JEOL Electron Probe Microanalyzer JXA-733 of the JEOL Superprobe 733 series (EPMA). The morphology of isolated particles of atmospheric aerosols is additionally observed in Scanning Electron Microscope. Individual particles having an average atomic number higher than that of the predominant so-called “dust-soil”(“background”) aerosols are easily outlined and morphologically studied in detail if their minimal size is about 1 mm while 10 mm is the minimal size for their chemical analysis.

Dust aerosols investigated in this work were collected with the help of clean brush from the smooth surfaces of a lee protected against rainfall where these aerosols have been accumulated by free sedimentation for weeks or months by a natural way. The choice of collector places was inspired from the East-West microclimatic division of Sofia valley, as adopted by meteorologists [16] and when taking into account that the Eastern region is formed as an industrial zone of Sofia. All of them are places of busy traffic crossings also. The first two regions are close each to other, the hotel “Pliska” and the district “Slatina”, the corresponding aerosol dusts named “Pl” and “Sl”, respectively. The third dust sample, named “Sim”, was collected from the Southern-East region of Sofia - district Simeonovo, close to cornfields and the road circuit in the foot of the mountain Vitosha. Examination by optical microscope of the first two samples, “Pl” and “Sl”, allowed determination of their particle diameter distribution with a maximum number of particles less than 2 mm (Fig. 1).

 Fig. 1 The size distribution (in percents along Y-axis) of the diameter of dust particles (in microns along X-axis) for the samples “PL” and “Sl”.

 In order to analyze the elemental content of a given sample, a clean double glue scotch tape with pressed on it dust material was stuck on a polished graphite substrate or 1% water suspension of these dust particles was directly dropped and dried onto the graphite substrate. A series of integral spectra (Fig. 2) of each preliminary prepared sample were taken at different sites of each investigated pollution spot under the Electron Microanalyser. As expected, the chemical composition throughout a given sample shows a little variation in the percentage content of elements (see the errors in the table of Fig. 2). The aerosols “Pl” and “Sl” have similar spectra to a great extend, see Fig. 2 and the values in table herein. The big duration of spectrum accumulation (about 300-400s) ensures receiving of statistically correct data for the most typical chemical elements of the samples. The elements Si, Ca, Fe, Al, S, K, Cl, Ti, Mg and Na are present in them, as habitually in more of the typical atmospheric dusts [7]. The metals Mn, Cu and Zn were detected in the aerosol “Sl” only. They usually are known as characteristic elements for the Eastern part of Sofia city associated with the big plants (metallurgical - “Kremikovtci” and others – “Gara Iskar”) in it and the power station (“Iztok”) here [8-11]. We have not any explanation why they were not often met in the region “Pliska” also. The spectrum of aerosol “Sim” shows an absent of Cl, a low content of S and Ca and a little bit rising in the percentage weight of such typical atmospheric dust elements like Al, Si and Fe.

Fig. 2 Elemental contents of dust aerosols (“Pl”, “Sl” and “Sim”), collected at three different places in Sofia,

 as indicated (in percents) in the graph with table below.

In order to clarify these results, the method of analyzing of individual particles was applied. It provides important information about their morphology and composition that otherwise cannot be obtained by “bulk” analysis methods. In morphological aspect, the investigated aerosol samples consist of different individual particles characterized as monolith, porous, agglomerates, etc. Of coarse, all they have either natural origin (the so-called dust-soil-mineral aerosols) or anthropogenic one due to human activities (the so-called technological pollutants). Our analysis shows that from morphological point of view the dust particles are most often agglomerates of varying mineralogical composition rather than pure minerals, confirmed by other authors also [7]. Typical SEM images (in secondary electron emission regime) of samples of the particles under investigation are represented in Fig. 3 (“Pl” - Fig. 3a, “Sl” - Fig. 3b and “Sim” - Fig. 3c), all they done with one and the same magnification (480x). The most of them are probably ordinary soil particles, concluding not only by their shape characteristics but also by their chemical content when one compare the analyzes of many similar single particles. These soil particles can also be characterized as different types, shown in Fig. 4. They also are most often agglomerates, the investigation of which needs separate efforts, and only some of them are monoliths - silicates or minerals containing more heavy elements (with an average atomic number higher than that of the typical soil particles).

Fig. 3 SEM pictures of typical samples of atmospheric dust particles collected in different regions of Sofia city: (a) “Pl”, (b) “Sl” and (c) “Sim”. (All photos are at the same magnification of 480x. For comparison, the rectangular particle close to the right bottom corner of (a) is about 22x35 mm.)

Fig. 4  Typical soil particles with similar (but not the same) chemical content and difference in their shape and sizes, as follows: (a) small 12x16 mm rough particle , (b) large elongated 150x300 mm monolith  particle and (c) middle 45x50 mm oval and more flat particle. (The cases (a) and (c) are from dust “Sl”, and (b) – from “Sim”.) 

Porous vanadium containing particles, as those shown in Fig. 5, occur frequently when one observe the first two objects – “Pl” (Fig. 5a) and “Sl” (Fig. 5b), and rarely are found in the third one - “Sim” (Fig. 5c). As a rule, they possess considerable amounts of S also. Their characteristic shape is spherical or oval one with more or less rough, even strongly eaten away, surfaces. They probably are due to high temperature processes in thermo-electric power plants [11].

Fig. 5  Typical micrographs of vanadium containing particles found in the dust material as follows (a) two oval particles with

  diameters 30 and 40 mm, from “Pl”, (b) spherical 60 mm particle, from “Sl”, and (c) a little bit deformed 20 mm particle, from “Sim”.

Some other dust particles with typical forms are shown in Fig. 6 - namely particles containing preferentially Ca and S (Fig. 6a and b) and a Ca particle (Fig. 6c).

    Fig. 6  Typical micrographs of Ca containing particles found in the dust material as follows: (a) and (b) polygonal,

often rhombic, particles containing Ca and S, which sizes vary  from 20x25 to 25x45 mm, and  (c) 300x300 mm pure Ca particle (all they are from “Sl”).

All particles containing chemical elements with average atomic number higher than that of the predominant dust-soil mineral ones are almost easily detected in the COMPO-type image because of their characteristic “shining”. Microphotographs of a few metal (metal oxide) containing particles, most often met in the sample “Sl”, are shown in Fig. 7 – pure Fe particles (Fig. 7a and 7b), a Zr-containing particle (Fig. 7c), Zn-containing particles (Fig. 7d and 6e). Their origin is either technological (as those in Figs. 7a-c or in some cases they are probably dressed mineral-soil particles (Figs. 7d-e).

 Fig. 7 Pictures of metal containing atmospheric particles from sample “Sl”: (a) pure Fe sphere 5 mm,

(b) almost spherical 50 mm Fe particle, (c) Zn containing garland 120x30 mm, (d) two Zn-containing particles,

almost spherical 25 mm and elongated 30x50 mm, and (e) Zr-containing particle 30x50 mm.

The influence of these complicated chemically content and morphologically pollutants for the atmospheric processes is a difficult problem which solving should pass through simplified laboratory (and field) experiments, for example as the described below.

Experiments with seeded and unseeded water drops

The temperature dependence of a number N_o of water drops was determined experimentally when they are cooling and freezing. In each experiment about 100 water drops or so were placed on a clean polyethylene foil adhering to a thick cooling metal plate. The drops, each one with volume v about 1 mm³, were cooling at constant rate q. The freezing of all drops was observed through an automatically recording computer controlled TV camera. This experimental technique is briefly described in [13, 17-19].

If each drop freezes immediately after the appearance of the first nucleus in it (the case of mononuclear mechanism [20]), the number of nuclei in these N_o drops will be equal to the number N of frozen drops. Then, as shown in detail in [13], at linear decreasing with time t (s) of absolute temperature T(t) = T_m - qt (K) (T_m is the melting temperature of ice) and at independent freezing of any of the drops by the others, the nucleation rate J (m^-3s^-1) can be presented as

J(T) = (q/v) d/dT [ln(1-N(T)/N_o)]                                                                                (1)

i.e. the experimental determination of the nucleation rate J (at known v and q) as a temperature dependent function reduces to finding the temperature derivative of the experimentally obtained quantity ln(1- N/N_o).

The freezing of drops of distilled water and of such drops seeded with some kind of aerosol dust (“Pl”, “Sl” or “Sim”) was studied. Each dust sample was investigated in three series. The weight concentration of aerosol particles in the successive series was C = 0.1, 1.0 and 10.0 mg/cm³. In each of the series, the total number N(T) of frozen drops and the total number N_o of the drops were measured. The experimentally obtained dependence of ln[1 – N(T)/N_o] on T is shown in Fig. 8 by squares, triangles and diamonds for C = 0.1, 1.0 and 10.0 mg/cm³, respectively, and by circles for the unseeded water drops. By the method of smooth approximation [12-15] of experimental data the same dependences were determined and drawn as continuous curves in Fig. 8. The results for the three atmospheric dust samples are shown in Fig. 8a (“Pl”), Fig. 8b (“Sl”), Fig. 8c (“Sim”), and the results for unseeded water are represented for comparison in each of them. The derivatives of the so-obtained smooth curves were used in equation (1) for calculation of the corresponding nucleation rates J(T) and J_o(T) of ice in seeded and unseeded water drops – Figs. 9a - 9c with the same symbols (used in Figs. 8a - 8c).

Fig. 8  Temperature dependence of the relative number of unfrozen water drops unseeded or seeded with different atmospheric dusts, as follows: (a) “Pl”, (b) “Sl” and (c) “Sim”. The seed concentrations are shown in the legend, all experimental data are indicated by symbols and the solid curves represent the smooth approximation of experimental data.

Fig. 9  Dependence of the ice nucleation rate J(cm^-3s^-1) vs T^-3DT^-2 for water drops unseeded or seeded with different atmospheric dusts, as follows: (a) “Pl”, (b) “Sl” and (c) “Sim”. The seed concentrations are shown in the legend, all experimental data are indicated by symbols. The solid curves in the case of dust “Pl” (Fig. 9a) appear the smooth approximation of experimental data, according to eqs. (4) and (6).

The theoretical model

The steady-state rate of isothermal nucleation J(T) can be presented, according to [21-23] and as described in detail in [24], as

J(T) = Aexp(-W^*(T)/kT) = Aexp(-B/T³DT²),                                                               (2)

where T(K) is the absolute temperature, DT=T-T_m is the supercooling, W^* is the nucleation work, Dm is  the chemical potential, and k is the Boltzmann constant.

According to the classical nucleation theory (see [24]), the following relations are valid

W^* = as_eff³v_m²/Dm² ,  Dm = DS_mTDT/T_m ,  s_eff³ = Fs³ , F = (1/4)(2+cosq)(1- cosq)².    (3)

From (2) and (3) it follows that

B = B”s_eff³,  B” = av_m²T_m²/DS_m²k .                                                                          (2a)

Here DS_m is the melting entropy, v_m is the volume of water molecule, s_eff (J/m²) is an effective specific surface free energy and s (J/m²) is the specific surface free energy of the ice/water interface, F is the activity factor of nucleation, q is the angle of “wetting” of the surface of the nucleus, and a is a shape factor.

The kinetic factor A (m^-3s^-1) in equation (2) has weaker temperature dependence than that of the exponential term and is [24-26]

A = A’exp(Dm/kT) = A^”bN_a ,  A” = zn_s^*(kT/hv_m) ,                                                    (2b)

where N_a (m^-3) is the concentration of active centers on which nuclei can be formed and b is a numerical factor accounting for the local viscosity around the nucleus, h is the viscosity of water, z is the Zeldovich factor, and n_s^* is the number of attachment sites of molecules on the nucleus surface (n_s^*»(n_o^*)^2/3, where n_o^* is the number of molecules in ice nucleus, n_o^*=2W^*/Dm).

For unseeded drops of distilled water (where microscopic foreign particles are present uncontrollably), the nucleation rate analogously to eqs. (2), (2a) and (2b) will be

J_o(T) = A_oexp(-B_o/T³DT²), A_o = A_o^”bN_a,o , B_o= B”s_o³                                                           (4)

and the plot of lnJ_o(T) vs 1/T³DT² shows a straight line (see Fig. 9a)

lnJ_o = lnA_o – B_o(1/T³DT²)                                                                                          (5)

From the slope B_o = 8.5x10¹⁰ K⁵ and the intersect lnA_o = 15.2 cm^-3s^-1, both determined by fitting to experimental J_o(T) data (the multiplied to the quantity q/v = 65 Ks^-1cm^-3 temperature derivative of the experimentally obtained quantity ln(1- N/N_o), shown by circles in Fig. 8a), one can obtained s_eff=s_o=11.2 mJ/m² and bN_a,o = 0.00084 cm^-3. All calculations here are done at a=16p/3, s =20 mJ/m², DS_m/kT=2.65, v_m=3.10^-23cm^-3, h=0.005 P, z=0.01, n_s,o^*=20.

The ice nucleation in the presence of atmospheric dust particles in the water drops is additionally stimulated by new active centers existing on the surface of these particles. That is why the resulting nucleation rate J(T) is due to the simultaneous action of the active centers present initially in the unseeded drops and of those introduced in the drops upon seeding them with dust particles. J(T) will be the sum of J_o and the nucleation rates J_i corresponding to the different activities i = 1, 2, 3, …of these centers:

J(T) = A_oexp(-B_o/T³DT²) + SA_iexp(-B_i/T³DT²),         A_i = A_i^”bN_a,i , B_i = B”s_i³ ,                  (6)

Here the i-type active center has a quantitative measure of its activity F_i , a “wetting” angle q_i, and specific surface free energy s_i. As shown explicitly in [13], if the concentration N_a,i (m^-3) of the i-type active center is proportional to the concentration C (g/m³) of the aerosol dust, then N_a,i = p_iC, where p_i (g^-1) shows how many i-type active centers are introduced into the water drops by seeding them with 1g of aerosol dust. It was obtained to a good approximation also that A_i increases linearly with C A_i=(A_o/N_a,o)p_iC], i.e. the nucleation rate J_i of the i-type active center, as well as the overall nucleation rate J, also increase linearly with increasing concentration C of dust particles.

The parameters F_i , q_i and s_i, characterizing the activity of the assumed three types (i = 1,2,3) of aerosol-introduced active centers in one of the dust samples (“Pl”) are given in Table 1 together with that for pure water drops (i = 0). The full analysis and comparison of the obtained data for the activities of the atmospheric aerosols under investigation (“Pl”, “Sl” and “Sim”) is under processing and will be discussed elsewhere.

 Table 1  Activity factor F_i , “wetting” angle q_i^o and effective specific surface energiy s_i for ice nucleation activity on active centers (types i=1,2,3) of dust “Pl” in water, compared to the same parameters of active centers (type i=0) in the same water. 

Conclusion

The present investigation appears an attempt to combine the detailed electron microscopic analysis of atmospheric dust particles with the investigations of their ice nucleation properties studied by freezing of seeded water drops applying some results of the classical nucleation theory. Our aim is to propose a method for characterizing and comparison of complicated systems of atmospheric pollutants with a quantitative even more or less formal description of their physicochemical parameters.

The detailed analysis, demonstrated on the example of the experiments carried out with water drops seeded with atmospheric dust (“Pl”) of region “Pliska” in Sofia, shows at least three types of active centers acting in the investigated temperature interval. The active centers of type i=1 dominate at the lowest investigated temperatures (below about -20^oC), those of type i=2 play a role only in the middle of studied range of temperatures (about -15^oC), while those of type i=3 are active at the highest temperatures (over about –10^oC). That means the aerosol-introduced active centers of type 1 are less active than those of type 2 which in turn are less active than those of type 3.

Unfortunately, the present stage of our limited knowledge of the physicochemical characteristics of the investigated dust samples does not allow the obtained values of the activity parameters immediately to be attributed to specific kind of particles and properties of the dust samples. The troubles with detailed description of such rather complicated system of atmospheric pollutants arise also from the great number of parameters that in some cases makes difficult the comparison with the results of other investigators. However, the present method of analyzing enables the nucleation activity of different aerosols to be characterized quantitatively despite of the fact that the nature of the active centers on them is unknown.

Acknowledgments

The present investigation is partly carried out with the financial support of the Bulgarian National Fund for Scientific Research under Contract No. F-1008/2001. A part of the experimental data is obtained during the stay of one of the authors (V. B. Stoyanova) at the Institute of Atmospherical Physics of the former Czechoslovak Academy of Sciences.

 References

1. Pruppacher H. (1973) in „Chemistry of the lower atmosphere“ (Ed. S. I. Rasol), Plenum Press, NY-London

2. Rosinski J., C. T. Nagamoto (1976) J. Aerosol. Sci. 7, 1; 7, 479

3. Vali G. (1985) J. Rech. Atmos. 19(2-3), 105

4. Havlicek D., R. Pribil, B. Kratochvil (1989) Atmospheric Environment 23(3), 701

5. Figler B., W. Sahle, S. Krantz, U. Ulfvarson (1996) The Science of the Total Environment 193, 77

6. Clausnitzer H., M. J. Singer (1999) J. Environmental Quality 28(5), 1619

7. Falkovich A., Y. Rudich, E. Ganor, Z. Levin (2001) J. Geophys. Res. – Atmosphere 106(D16) 18029

8. Marinov M., T. Tsacheva, R. Stojanchev, N. Pashov (1994) Bulg. Geophys. J., 20(1), 1

9. Marinov M., T. Tsacheva, N. Pashov (1996) JEOL News 32E (1), 24

10. Цачева Ц., М. Маринов, Н. Пашов, М. Йосифова (1996) в Материали от VII Колоквиум “Физиката в опазването на човека и околната му среда” на тема “Измерване и оценка на вредните физически фактори на околната среда”, Гьолечица, стр. 183

11. Ranguelov B., T. Tsacheva, M. Marinov, N. Pashov, M. Iossifova (1998), J. Balkan Ecology 1(1), 35

12. Kupenova T. (1980) Commun. JINR, Dubna P11-80-669

13. Stoyanova V., D. Kashchiev, T. Kupenova (1994) J. Aerosol. Sci. 25, 867

14. Kupenova T., V. Stoyanova (1993) “Physics Computing ’92” (Ed. R. A. de Groot and J. Nadrchal), World Scientific, Singapore, 386

15. Stoyanova V., D.Kashchiev and T.Kupenova (1994) in: “Thin Films and Phase Transitions on Surfaces” (Proc. EWSSW’94, Pamporovo, 14-20.2.1994), Eds. M. Michailov and I. Gutzow, Vanel Ltd., Sofia, , p.363

16. Blaskova D., L. Zlatkova, S. Linchova, Zh. Modeva, L. Sabev (1983) “Climat and Microclimate of Sofia“, Publishing House of the Bulg. Acad. Sci., Sofia, p. 96

17. Dubrovski M., O. Skoloud (1988) in “Atmospheric Aerosols and Nucleation” (Ed. P. E. Wagner and G. Valy), Springer, Berlin

18. Dubrovski M., O. Skoloud (1989) Studia Geophys. Geodes. 33, 287

19. Bilkova E., M. Dubrovski, W. Petera, O. Skoloud (1989) Трудьi Централ. Аеролог. Обсерв. 174, 11

20. Kashchiev D., D. Verdoes, G. M. Rosmalen (1991) J. Crystal Growth 110, 373

21. Volmer M. (1939) “Kinetik der Phazenbildung“, Steinkopff, Dresden

22. Hirth J. P., G. M. Pound (1963) “Condensation and evaporation”, Pergamon, 20, Oxford

23. Walton A. G. (1969) in “Nucleation” (Ed. A.C. Zettlemoyer), Dekker, New York, p. 225

24. Kashchiev D. (2000) “Nucleation. Basic Theory with Applications”, Butterworth-Heinemann, Oxford

25. Kashchiev D. (1984) Cryst. Res. Technol. 19, 1413

26. Gutzow I., D. Kashchiev, I. Avramov (1985) J. Non-Cryst. Solids 73, 477

27. Rosenfeld D., Rudich Y., R. Lahav (2001) Proc. Natl. Acad. Sci. USA 98(11) 5975

28. Clausnitzer, H., M.J. Singer (1996) J. Environ. Qual. 25, 877"

29. Clausnitzer, H, M.J.Singer (2000) Atmospheric Environment 34,1739

30. Ehrman S. H., M. Schwikowski, U. Baltensperger, H. W. Gaeggeler (2001) Atmos. Environment 35, 5371

31. Konarski P., I. Iwanejko, A. Mierzjewska, R. Diduszko (2001) Vacuum 63, 679

32. Eleftheriadis K., I. Colbeck (2001) Atmos. Environment 35, 5321

33. Ro C.-U., K-Y. Oh, H-K. Kim, Y. Chun, J. Osan, J. de Hoog, R. van Grieken (2001) Atmos. Environment 35, 4995