J Nanopart Res (2010) 12:1989–1995 DOI 10.1007/s11051-010-9856-y
BRIEF COMMUNICATION
Electrical properties of airborne nanoparticles produced by a commercial spark-discharge generator S. Bau • O. Witschger • F. Gensdarmes D. Thomas • J.-P. Borra
•
Received: 10 August 2009 / Accepted: 11 January 2010 / Published online: 26 January 2010 Ó Springer Science+Business Media B.V. 2010
Abstract A nanoparticle generator based on the principle of electrical discharge (PALAS GFG-1000) was used to produce nanoparticles of different chemical natures. The fractions of electrically neutral particles were then measured by means of a Spectrome`tre de Mobilite´ Electrique Circulaire (SMEC, i.e. radial-flow mobility analyzer) for different operating conditions. The experimental results were compared with the theoretical values calculated from the Fuchs extended charge equilibrium model for spherical particles and agglomerates. For the smallest particles (below 20 nm), the deviations observed
S. Bau (&) O. Witschger Institut National de Recherche et de Se´curite´, INRS, Laboratoire de Me´trologie des Ae´rosols, CS 60027, 54519 Vandoeuvre Cedex, France e-mail:
[email protected] F. Gensdarmes Institut de Radioprotection et de Suˆrete´ Nucle´aire, IRSN, Laboratoire de Physique et de Me´trologie des Ae´rosols, BP 68, 91192 Gif-sur-Yvette Cedex, France D. Thomas Laboratoire des Sciences du Ge´nie Chimique, LSGC/ CNRS, Nancy Universite´, BP 2041, 54001 Nancy Cedex, France J.-P. Borra Laboratoire de Physique des Gaz et des Plasmas, Equipe De´charges Electriques et Proce´de´s Ae´rosols, SUPELEC, Plateau de Moulon, 3 rue Joliot-Curie, 91192 Gif sur Yvette, France
remain below 10%, and tend towards 20% for larger particles (over 35 nm). Keywords Nanoparticle Electrical properties Mobility analysis Bipolar charging model Synthesis Modeling Introduction First introduced by Schwyn et al. (1988), the spark discharge has been identified as versatile in producing carbon, metals, metal oxides (and mix of) nanoparticles in a gas phase (nanoaerosol). The method has already been applied by a number of research groups (Helsper et al. 1993; Roth et al. 1998, 2004; Brown et al. 2000; Evans et al. 2003a, b; Horwath and Gangl 2003; Tabrizi et al. 2009), and amongst the features of interest are the very good temporal stability and reproducibility of the number concentration and nanoparticle size distributions. One property which is rarely evaluated is the electrical charge of the airborne nanoparticles produced. Preliminary rough estimation of the charged fraction using an electrostatic precipitator after the spark generator has been reported (Schwyn et al. 1988; Borra et al. 1998). Recently, Tabrizi et al. (2009) have presented data about charged fraction of nanoparticles produced by a laboratory-built sparkdischarge generator. In their study, the fraction of electrically charged particles was determined by comparing particle size distributions measured with
123
1990
J Nanopart Res (2010) 12:1989–1995
at spark frequency up to 300 Hz with argon pressure supply of 1.5 bars. At the generator exit, the aerosol is diluted by a factor 100 using two dilution stages. Then, the fraction of electrically neutral particles is measured with a set-up using a radial-flow mobility analyser (hereafter referred as SMEC). The SMEC operates with an aerosol sample flow of 0.15 L min-1 and a sheath air flow of 1.35 L min-1. The selected particles are detected with a condensation particle counter (CPC model 3025, TSI) operating at 0.3 L min-1 and an electrometer (model 3068A, TSI) operating at 1.2 L min-1.
a scanning mobility particle sizer (SMPS) with and without use of a neutralizer. However, this set-up does not enable the direct measurement of the fraction of neutral particles. The PALAS GFG 1000 generator is one commercial spark-discharge equipment which has already been used in several studies for different applications like in human or animal inhalation studies (Takenaka et al. 2001) or aerosol instrumentation studies (Bau et al. 2009). For this widely used commercial generator, we have found no data in the literature concerning the electrical charges of the nanoaerosol produced. Therefore, the objective of our study is to measure the fraction of electrically neutral particles produced by the PALAS GFG 1000 generator and to compare experimental data to bipolar charging theory. The results would provide arguments on the necessity of the aerosol neutralisation step in set-ups for inhalation or instrumentation studies, using this spark generator. In this study, we measure the fraction of electrically neutral nanoparticles with a set-up using a radial-flow mobility analyser (hereafter referred as SMEC).
Nanoaerosol generation The nanoaerosols produced by the generator are first directed towards two successive dilution stages fed with nanoparticle-free air coming from a purification unit (model 3074B, TSI). This configuration was used to reduce coagulation phenomenon and to obtain aerosol concentration below 105 cm-3 which is the limit of the CNC 3025. Three different nanoaerosols were generated from three electrode materials: (1) (2)
Materials and methods
(3)
The experimental set-up is presented in Fig. 1. It is composed of the PALAS GFG-1000 generator used
Carbon (graphite, pure), Copper (99.45%—impurities: Al 0.5% and Si 0.04%), Aluminium (93.11%—impurities: Cu 4.76%, Fe 0.66%, Mn 0.6%, Si 0.55% and Mg 0.32%).
Fig. 1 Experimental set-up employed to measure the fraction of neutral particles
clean air HEPA filter
PALAS VKL-10 diluter
dilution volume
Ar excess
excess
excess
aerosol flow clean air flow
PALAS GFG-1000 aerosol generator
1.2 L/min
1.5 L/min
0.15 L/min
20
flow splitter Electrometer TSI # 3068A
0.3 L/min 15000
U SMEC CNC TSI # 3025
123
1.35 L/min
J Nanopart Res (2010) 12:1989–1995
1991
A SMPS, consisting in a differential mobility analyser (DMA model 3085, TSI) and a CNC (model 3025, TSI), was used for measuring the particle size distribution of the aerosols before and after each run to confirm aerosol stability (not shown in Fig. 1). The count median mobility diameters (CMMD) obtained were found in the range 5–50 nm (geometric standard deviation (GSD) between 1.34 and 1.74) and differ with the electrode material, as given in Table 1. As stated by Evans et al. (2003b), the material constituting the electrode plays an important role in the aerosol size. This is mainly due to differences in energy losses within the electrode, depending, in particular, on the thermal and electrical conductivities of the materials. These properties condition the generation efficiency, leading to smaller particles in the aerosol phase with lower number concentration for Cu and Al electrodes, contrary to C. In the case of carbon aerosols, the particles produced are supposed to be agglomerates of primary particles. Schwyn et al. (1988) and later Helsper et al. (1993) observed spherical primary particles of 5 nm in diameter. However, in our work, no samples were taken for TEM observations. Measurement of the fraction of neutral particles The SMEC was characterised in terms of transfer function by Fissan et al. (1998), and has been used both to obtain the particle density (Le Bronec et al. 1999) and to determine the charge distribution of a radioactive aerosol (Gensdarmes et al. 2001). The SMEC configuration used in this study is cumulative, which makes it possible to measure successively the total aerosol concentration, the sum of neutral and positive or negative charged particles (depending on the polarity of the voltage applied to the electrode)
and finally only the neutral particles when a sufficient voltage is applied. The fraction of neutral particles f0 is defined by the following equation: f0 ¼
N0 ; NT
ð1Þ
were NT is the total number concentration (cm-3) and N0 is the number concentration (cm-3) of the particles carrying zero charge. NT is obtained by applying no voltage in the SMEC (U = 0 V, see Fig. 1), while a voltage of 1 kV was applied to obtain N0. The latter voltage produces an electrical field in the SMEC that makes positively and negatively charged particles migrate by electrophoresis and deposit on the upper and lower electrode, respectively. This point was verified by measuring on the aerosol coming out a current equal to zero with the electrometer. It should be noted that the fraction of particles lost by diffusion in the SMEC is neglected. This point is not crucial in our study as the fraction of neutral particles is expressed relatively to the number of particles coming out from the SMEC. Moreover, particle losses by image force near the upper electrode in the absence of voltage are neglected too. Theoretical model to describe the fraction of neutral particles In order to comment our results, theoretical considerations on the possible ion extraction from a sparkdischarge gap as well as post-spark ion densities measurements are presented to justify the choice of the bipolar diffusion aerosol charging model. These considerations suggest that a bipolar charge model could be used to describe the aerosol charging instead
Table 1 Properties of the aerosols generated for the three electrode materials and different spark-discharge frequencies (1.5 bar argon) Electrode material C Al Cu
Air flow (L min-1)
fmin (Hz)
CMMD (nm)
GSD (-)
fmax (Hz)
CMMD (nm)
GSD (-) 1.66
0
15
22
1.54
270
49
30
15
21
1.58
270
37
1.74
0
15
7.3
1.50
270
11.6
1.61
30
15
6.0
1.41
270
8.4
1.60
0
60
5.5
1.34
270
8.0
1.45
30
60
5.1
1.38
270
7.6
1.53
123
1992
of combination of two unipolar charging models. Indeed, positive and negative ions produced by the spark-discharge interact with the particles in the mixing chamber surrounding the electrodes. In electrical discharge at atmospheric pressure, both polarities ions and electrons are produced and then separated in the field towards each electrode depending on their polarities (Borra 2006). In capacitive dc-sparks, the temperature in the order of 103– 104 K leads to reduced electric field (E/N) higher than the threshold-reduced field of ionisation of 40 Td (i.e. 40 V cm-1 Torr-1), related to electric field of 2.3 106 V m-1 at STP. Thus, the spark plasma develops over the entire gap, but for electric fields 10–100 times smaller than the electric field in non-thermal plasmas at STP. Indeed, in the cathode fall region confined close to the cathode, the ion velocity is about 100 m s-1 for a typical ion electrical mobility of 10-4 m2 V-1 s-1 and an electric field of 106 V m-1. Outside of the cathode fall, which concerns around 90% of the gap between electrodes, the electric field is about 104 V m-1, and the ion velocity is about 1 m s-1. The low ion velocity between the electrodes during the discharge lets assume that a significant ion quantity could be flushed out of the interelectrode space by the argon flow and mixed with the aerosol so-produced for post-discharge neutralisation in bipolar ion clouds. Preliminary ion currents were measured only after a capacitive dc-spark (C = 0.12 nF) to evaluate the corresponding post-spark ion densities. A special arrangement of electrostatic precipitator was developed to separate positive and negative ions from the flow for simultaneous current measurements on both electrodes, as described by Bourgeois et al. (2009). The total currents were measured as well to get the net polarity and density of ions in excess. Also, positive and negative ion currents were systematically measured, confirming that the post-spark flow carries ions of both polarities with densities from 106 to 107 elementary charges per cm3. These ion densities lead to charging conditions characterised by the product of the ion concentration by aerosol residence time (Ni 9 t) in the order or higher than the 105 s cm-3 reported by Liu and Pui (1974), to reach the equilibrium state-of-charge in the free molecular regime. These considerations allow us to use a bipolar charging model such as the Fuchs bipolar extended
123
J Nanopart Res (2010) 12:1989–1995
model proposed by Wiedensohler (1988) for ultrafine particles. This model, developed on the basis of experimentations performed on submicronic particles, consists in an empirical formulation of the fraction of particles carrying p elementary charges (fp): i i¼5 X dp ai ð pÞ log ; ð2Þ log fp ¼ d0 i¼0 where dp is the diameter of the spherical particle expressed in nm, d0 = 1 nm is a reference diameter, and ai(p) are adjustment coefficients reported in Flagan (2001). It should be noted that this relationship is valid for p = –1, 0, 1 in the size range 1–1,000 nm, and for p = –2, 2 in the range 20–1,000 nm. Rigorously, the diameter dp taken in Eq. 2 is the particle equivalent charge diameter, notated dqe, which is equal to the equivalent electrical mobility diameter for spherical particles (CMMD). In the case of agglomerates, the equivalent charge diameter taken into account the number (Npp) and the size (dpp) of the primary particles present in the agglomerate (Wen et al. 1984b): dqe ¼
dpp Npp ; ln 2 Npp
ð3Þ
where the number of primary particles can be deduced from the agglomerate mobility diameter according to (Lall and Friedlander 2006): Npp ¼
12 p kg dm : 2 Cuðdm Þ F dpp
ð4Þ
In this relationship, F* = 9.17 is a dimensionless parameter, kg = 66.4 nm is the mean free path of air molecules, and Cu is the Cunningham slip coefficient. For agglomerates, the fraction of electrically neutral particles is then calculated by means of Eq. 2 with dp = dqe.
Experimental results and theoretical calculation Figure 2 shows experimental data obtained for the copper, aluminium and carbon electrodes. It can be observed that for aerosols with CMMD below 10 nm, the fraction of neutral particles is very close to 100%. These aerosols are produced with copper and aluminium electrodes. For the carbon electrodes, the
J Nanopart Res (2010) 12:1989–1995
1993
fraction of neutral particles -f0 (-)
1.2 1.1 20 % deviation for spheres
1.0 0.9 0.8 0.7
C
0.6
Al
0.5
Cu model for spherical particles
0.4
model for agglomerates
0.3 1
10
100
count median mobility diameter -CMMD (nm) Fig. 2 Fraction of neutral particles f0 as a function of the aerosol count median mobility diameter CMMD for three different electrodes (operating conditions: 1.5 bar argon); error bars indicate two standard deviations of measured fraction. The curves represent different calculations using the Fuchs bipolar
extended model for spherical particles (solid line for the mean value and dashed lines corresponding to ±20% deviation) and agglomerates (two dotted lines). For agglomerates, two primary particle sizes were considered: 5 and 10 nm
CMMD goes from 20 to 50 nm and f0 decreases from 90 to 55%. The Fuchs bipolar extended model was used to calculate the theoretical fraction of neutral particles, drawn with a solid line in Fig. 2, assuming that the aerosols are monodisperse (particle size taken equal to the CMMD of the experimental size distribution). Indeed, additional calculation have shown that the aerosol polydispersity induces a very slight difference in terms of neutral fraction, less than 1.8% for lognormal size distribution with such GSD up to 1.7 and CMMD varying between 5 and 50 nm. In the case of carbon aerosols, the equivalent charge diameter was used to calculate the theoretical fraction of neutral particles. Numbers of 5 nmprimary particles from 19 to 110 were found for agglomerates with CMMD between 20 and 50 nm. This case, plotted with dotted line in Fig. 2, highlights that agglomerates are less neutral than spherical particles with identical mobility diameter, which is in concordance with earlier study (e.g. Wen et al. 1984a). It can be observed, from Fig. 2, that for CMMD below 10 nm (case of Al and Cu nanoaerosols), the experimental measurements are in very close agreement with the bipolar charging model assuming
spherical particles. Indeed, the relative deviations vary between -2 and ?7%. It can be observed that the behaviour of aerosols produced from carbon electrodes is a little different. The fraction of neutral particles decreases with increasing particle size, which is in qualitative agreement with the bipolar charging theory. However, considering the model for spherical particles, the relative deviations between experimental and theoretical values range from -18 to ?12%. Compared to the model for agglomerates, the latter deviations vary between ?10 and ?20% considering primary particles of 5 nm in diameter, and between -8 and ?12% for primary particles of 10 nm in diameter. Indeed, the primary particle size of 10 nm was found to lead to the less deviation with experimental measurements. Remaining discrepancies could arise from several mechanisms. One can note that the properties of ions produced in argon are different than those produced in air that may affect particle charge (Wiedensohler et al. 1986; Wiedensohler and Fissan 1991). This influence may be visible only for carbon aerosols due to their diameters upper than 20 nm. Further explanations concerning the influence of ion properties on the fraction of neutral particles would require the
123
1994
measurement of charged fractions (positive and negative) of the aerosols. Moreover, particles may have not reached the state-of-charge equilibrium, due to high particle densities leading to a number of ions per particle smaller than the minimum value of 1000 reported from Borra (2008). This parameter was not measured in this study and reliable information is not available. This hypothesis could only be verified by measuring the ion concentrations within the sparkdischarge generator which is not an easy task, but should be considered as a further experimental study.
Conclusions An experimental method which includes a radial-flow mobility analyser (SMEC) in the order 1 configuration was designed to measure the fraction of neutral particles f0 produced by the PALAS GFG-1000 generator with carbon, aluminium and copper electrodes. For aerosols having a CMMD below 10 nm obtained with aluminium and copper electrodes, the fraction of neutral particles is very close to 100%. For the carbon electrodes, f0 decreases from 90 to 55%, while the CMMD goes from 20 to 50 nm. The experimental data obtained are in close agreement with the neutral fraction calculated with a bipolar charging model for CMMD below 10 nm assuming spherical particles. The relative deviation is less than ±20% for CMMD between 20 and 50 nm. When considering the bipolar charging model for agglomerates, the relative deviation decreases to ±12% for primary particles of 10 nm in diameter.
References Bau S, Witschger O, Gensdarmes F, Thomas D (2009) Experimental study of the response functions of directreading instruments measuring surface-area concentration of airborne nanostructured particles. J Phys Conf Ser 170:012006 Borra J-P (2006) Topical review: Nucleation and aerosol processing in atmospheric pressure electrical discharges: powders production, coatings and filtration. J Phys D 39– 2:R19–R54 Borra J-P (2008) Charging of aerosol and nucleation in atmospheric pressure electrical discharges. Plasma Phys 50:124036 Borra J-P, Goldman A, Goldman M, Boulaud D (1998) Electrical discharge regimes and aerosol production in point-
123
J Nanopart Res (2010) 12:1989–1995 to-plane dc high-pressure cold plasmas: aerosol production by electrical discharges. J Aerosol Sci 29:661–674 Bourgeois E, Jidenko N, Borra J-P (2009) Characterisation of post-DBD ion currents, densities and mobilities. J Phys D Brown JS, Kim CS, Reist PC, Zeman KL, Bennett WD (2000) Generation of radiolabeled ‘‘Soot-like’’ ultrafine aerosols suitable for use in human inhalation studies. Aerosol Sci Technol 32:325–337 Evans DE, Harrison RM, Ayres JG (2003a) The generation and characterisation of elemental aerosols for human challenge studies. J Aerosol Sci 34:1023–1041 Evans DE, Harrison RM, Ayres JG (2003b) The generation and characterization of metallic and mixed element aerosols for human challenge studies. Aerosol Sci Technol 37:975–987 Fissan H, Po¨cher A, Neumann S, Boulaud D, Pourprix M (1998) Analytical and empirical transfer functions of a simplified Spectrome`tre de Mobilite´ Electrique Circulaire (SMEC) for nanoparticules. J Aerosol Sci 29(3):289–293 Flagan RC (2001) Electrical techniques. In: Baron PA, Willeke K (eds) Aerosol measurement: principles, techniques and applications, 2nd edn. Wiley Interscience, New York, pp 537–568 Gensdarmes F, Boulaud D, Renoux A (2001) Electrical charging of radioactive aerosols: comparison of the Clement-Harrison models with new experiments. J Aerosol Sci 32:1437–1458 Helsper C, Molter W, Loffler F, Wadenpohl C, Kaufmann S, Wenninger G (1993) Investigations of a new aerosol generator for the production of carbon aggregate particles. Atmos Environ 27:1271–1275 Horwath H, Gangl M (2003) A low-voltage spark generator for production of carbon particles. J Aerosol Sci 34:1581–1588 Lall AA, Friedlander SK (2006) On-line measurement of ultrafine aggregate surface-area and volume distributions by electrical mobility analysis: I. Theoretical analysis. J Aerosol Sci 37(3):260–271 Le Bronec E, Renoux A, Boulaud D, Pourprix M (1999) Effect of gravity in differential mobility analysers. A new method to determine the density and mass of aerosol particles. J Aerosol Sci 30(1):89–103 Liu BYH, Pui DYH (1974) Electrical neutralization of aerosols. J Aerosol Sci 5:465–472 Roth C, Karg E, Heyder J (1998) Do inhaled ultrafine particles cause acute health effects in rats? I: particle production. J Aerosol Sci 29(Suppl 1):S679–S680 Roth C, Ferron GA, Karg E, Lentner B, Schumann G, Takenaka S, Heyder J (2004) Generation of ultrafine particles by spark discharging. Aerosol Sci Technol 38:228–235 Schwyn S, Garwin E, Schmidt-Ott A (1988) Aerosol generation by spark discharge. J Aerosol Sci 19(5):639–642 Tabrizi NS, Ullmann M, Vons VA, Lafont U, Schmidt-Ott A (2009) Generation of nanoparticles by spark discharge. J Nanopart Res 11(2):315–332 Takenaka S, Karg E, Roth C, Schulz H, Ziesenis A, Heinzmann U, Schramel P, Heyder J (2001) Pulmonary and systemic distribution of inhaled ultrafine silver particles in rats. Environ Health Perspect 109:547–551 Wen YH, Reischl GP, Kasper G (1984a) Bipolar diffusion charging of fibrous aerosol particles. I. Charging theory. J Aerosol Sci 15(2):89–101
J Nanopart Res (2010) 12:1989–1995 Wen YH, Reischl GP, Kasper G (1984b) Bipolar diffusion charging of fibrous aerosol particles. II. Charge and electrical mobility measurements on linear chain aggregates. J Aerosol Sci 15(2):103–122 Wiedensohler A (1988) An approximation of the bipolar charge distribution for particles in the submicron size range. J Aerosol Sci 19:387–389
1995 Wiedensohler A, Fissan HJ (1991) Bipolar charge distributions of aerosol particles in high purity argon and nitrogen. Aerosol Sci Technol 14:358–364 Wiedensohler A, Lu¨tkemeier E, Feldpausch M, Helsper C (1986) Investigation of the bipolar charge distribution at various gas conditions. J Aerosol Sci 17(3):413–416
123
