Aerosol Science 39 (2008) 609 – 617

www.elsevier.com/locate/jaerosci

A study of ionization and collection efficiencies in

electrospray-based electrostatic precipitators
Gary Teppera, b,∗ , Royal Kessickc
a Sentor Technologies Inc., 11551Q Nuckols Rd., Glen Allen VA 23059, USA
b Department of Mechanical Engineering, Virginia Commonwealth University, Box 843015, Richmond VA 23284, USA
c Sentor Technologies Inc., 11551Q Nuckols Rd., Glen Allen VA 23059, USA

Received 2 November 2007; received in revised form 25 March 2008; accepted 26 March 2008

Abstract
Ozone-free electrostatic precipitators incorporating an array of electrospray wick charged aerosol ionization sources were built
and tested. The particle ionization and collection efficiencies were independently measured for airborne particles as a function of air
flow rate, electric field magnitude, and particle size. The results indicate that, while particle ionization efficiencies near 100% can be
achieved using the electrospray ionization method, the number of elemental charges deposited onto a given particle depends strongly
on the particle diameter and for submicron particles is very low. High overall particle collection efficiency, therefore, requires a
carefully designed precipitation region where the maximum particle collection time, associated with the particles with the lowest
mobility, is less than the residence time within the collector.
䉷 2008 Elsevier Ltd. All rights reserved.

Keywords: Electrospray; Air purification; Electrostatic precipitation

1. Introduction

Air purification systems based on electrostatic precipitation (ESP) have been used for years in both industrial and
commercial applications and typically employ a corona discharge to ionize air contaminants and a series of metal plates
to collect the ionized species (Jaworek, Krupa, & Czech, 2007). The popularity of ESP systems is due, in part, to a
combination of quiet operation and low maintenance (e.g. no filter replacement). However, the corona ionization process
produces ozone and the usefulness of these systems has come into question because of health concerns (Hazucha &
Lefohn, 2007).
Recently, we reported on the development of ozone-free electrostatic precipitation systems employing an array
of electrospray wick charged aerosol sources for particle ionization (Tepper, Kessick, & Pestov, 2007). Electrospray
aerosols are used in applications such as analytical chemistry (Fenn, Mann, Meng, Wong, & Whitehouse, 1989), particle
generation (Rulison & Flagan, 1994), coatings (Seaver & Eckhardt, 1988), and space propulsion (Bartoli, von Rohden,
Thompson, & Blommers, 1984). Typically, in what is known as the cone-jet mode of electrospraying, a jet of charged

∗ Corresponding author at: Department of Mechanical Engineering, Virginia Commonwealth University, Box 843015, Richmond, VA 23284,
USA. Tel.: +1 804 827 4079; fax: +1 804 827 7030.
E-mail addresses: gtepper@vcu.edu, gary.tepper@sentorinc.com (G. Tepper).

0021-8502/$ - see front matter 䉷 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jaerosci.2008.03.005
610 G. Tepper, R. Kessick / Aerosol Science 39 (2008) 609 – 617

liquid emerges from the tip of a single conical protrusion known as a Taylor cone, which is formed at a liquid surface
through a competition between an applied electric force and the liquid surface tension (DelaMora, 2007). The charged
liquid jet is unstable and eventually breaks up into a plume of charged liquid droplets. The initial droplet diameter is
determined by various parameters including the liquid conductivity, surface tension, flow rate and the diameter of the
spraying orifice and is typically just a few microns (Ganan-Calvo, Davila, & Barrero, 1997; Gomez & Tang, 1994). As
the droplets shrink due to solvent evaporation, they undergo a series of Coulombic explosions when the electrostatic
repulsive force on the surface of the charged droplet exceeds the liquid surface tension. This process repeats and
produces a charged aerosol of nanoscale droplets, which can be used for the ionization and collection of gas-phase
contaminants (Tepper, Fenn, Kessick, Pestov, & Anderson, 2006).
In addition to conventional electrospray sources, nanoelectrospray (NanoES) sources (Wilm & Mann, 1996) and
polymer monolith sources (Koerner, Turck, Brown, & Oleschuk, 2004) have been reported for applications in mass
spectrometry and microfluidics. In both NanoES and polymer monolith sources, the effective diameter of the spraying
orifice is greatly reduced thereby reducing the initial droplet diameter and liquid flow rate. The liquid flow in these
sources is driven primarily by capillary forces thus eliminating the need for solvent pumps. However, fabrication
of NanoES or polymer monolith sources is relatively cumbersome, requiring either drawing and tapering the orifice
diameter to about 1 m or a polymerization reaction within a microcapillary. Therefore, we used porous polymer wicking
materials from commercial vendors such as the Porex or Filtrona Corporations as electrospray aerosol sources. Porous
polymer wicking materials such as bonded synthetic fibers are currently used in applications such as air fresheners,
ink jet printers, filters, and medical testing kits. These low-cost, wettable, porous materials are available in a variety
of shapes, sizes, and porosities and we found them to be ideal electrospray aerosol sources for applications in air
purification.
In our previous electrospray-based ESP systems, the particle ionization and collection steps were integrated making
it difficult to separately quantify particle ionization and collection efficiencies. However, the overall particle removal
efficiencies were quite promising and clean air delivery rates (CADR) greater than 100 for both dust and cigarette smoke
were achieved (Tepper et al., 2007). The CADR number is used by the Association of Home Appliance Manufacturers
(AHAM) to quantify and compare the performance of commercial air purification systems and, as a rule of thumb,
should be at least 23 the square footage of the room being purified. Here we report on a quantitative evaluation of the
particle ionization and collection efficiencies in electrospray-based ESP systems with separate particle ionization and
collection regions.

2. Experimental

Two prototype air purification systems, one small and one large, were constructed and tested. In each system, particle
ionization was accomplished by charge transfer from an aerosol of charged droplets dispersed into an air stream from an
array of electrospray polymer wick sources. Wick electrospray sources were employed instead of conventional hollow
needles in order to eliminate the need for mechanical components such as syringe pumps and valves and to uniformly
distribute the charged aerosols throughout an air flow channel. A wick is self-balancing and, through capillary action,
will automatically replenish the solvent dispersed by the electrospray aerosol. The wicks consist of a microporous,
bonded synthetic polymer material from two commercial vendors; Filtrona and Porex Corporations. The wicks from
Filtrona were 2 mm in diameter and the wicks from Porex were 1 mm in diameter. The wicks were cut to a length of about
20 mm and wetted by placing one end into a liquid reservoir containing a 90/10 water/ethanol solution. Electrospray
was initiated by applying a high voltage between the liquid reservoir and a counter electrode located opposite the wick
tip, which could be either blunt or cut to a 45◦ bevel to enhance the local electric field magnitude.
Fig. 1 is a plot of the electrospray current versus the applied potential difference between a blunt wick tip and a
grounded counter electrode located at a distance of 1 cm for a 1 mm Porex and 2 mm Filtrona wick. For both wicks, the
electrospray current initially increases with applied voltage as expected, but then levels off and stays relatively constant
until reaching the point of corona discharge where the current increases dramatically. Due to the higher electric field
concentration in the smaller diameter Porex wick, the electrospray initiates at a lower voltage and the onset of corona
discharge also occurs at a lower voltage (7.2 kV versus 8.4 kV).
The current–voltage characteristics of the wick sources depends on several factors including the wick diameter and
porosity and the solvent viscosity and surface tension and two primary spray modes were observed. In the first mode,
when the maximum wicking rate of the porous polymer as determined by the Washburn (1921) equation exceeds the
 G. Tepper, R. Kessick / Aerosol Science 39 (2008) 609 – 617 611

3.5 Corona

Electrospray Current (µAmps)

3
Corona
2.5

1.5 Electrospray
region
1

0.5 Filtrona
Porex
0
4 5 6 7 8 9
Voltage (kV)

Fig. 1. Electrospray current versus applied voltage for the 1 mm Porex and 2 mm Filtrona wicks (90/10 water/ethanol electrospray solution).

liquid flow rate from the electrospray jet, the wick will replenish and maintain the flow from a single liquid droplet at
the wick tip with a single electrospray cone jet. In the second mode, the maximum wicking rate of the porous polymer
is below the liquid flow rate established by the electrospray jet. In this mode, the single droplet at the tip of the wick
quickly becomes depleted and disappears and the electrospray continues from a plurality of microscopic Taylor cones
emerging from the surface of the wick. This second mode is very similar to that of a nanoES or polymer monolith
electrospray system and is responsible for the I–V curves shown in Fig. 1.
Microscope imaging of the wick tip during electrospray revealed that, while the macroscopic current is constant, the
flow from any single microscopic Taylor cone is periodically interrupted as the fluid is depleted and then replenishes
through capillary action. The net current from the wick appears constant because at any given time, a certain fraction
of the microsprays are on, while the others are replenishing. The level region in the curves of Fig. 1 before the onset
of corona and where the current is relatively independent of the applied voltage can be explained by the fact that the
electrospray current is being limited by the maximum capillary flow rate of the wick.
Due to its very high surface tension (approximately 72 dyn/cm in air at standard temperature and pressure), pure
water has a relatively high wicking rate in both the Porex and Filtrona materials and is more likely to electrospray
from a single large Taylor cone (mode 1). However, the voltage difference between the onset of electrospray for pure
water and the onset of corona can be small and we, therefore, used a 90/10 water/ethanol solution to reduce the surface
tension to a value near 45 dyn/cm. The change in the solution viscosity and density due to the addition of ethanol is
small in comparison to the change in surface tension and the net result is a significant decrease in the liquid wicking
rate which is why the water/ethanol solution is more likely to electrospray in mode 2 from these wick sources.
Because the initial droplet diameter is much smaller in mode 2, the ratio of the electrospray current to the liquid flow
rate is significantly (about 50×) higher and this is of great advantage in air purification systems where liquid collection
and recirculation is undesirable. All of the data reported here was obtained from electrospray wick sources using 90/10
water/ethanol as the electrospray solvent and operating in mode 2 with a typical current–voltage characteristic shown
in Fig. 1. We plan to report a more comprehensive analysis of polymer wick-based electrospray sources as a function
of solvent properties and wick porosity in a separate paper.
In the smaller system, prototype A, illustrated schematically in Fig. 2a, the electrospray ionization region consists
of 44 individual Filtrona wicks cut to a 45◦ bevel at the tip and arranged in four parallel rows. A photograph of
the wick array used in prototype A is shown in Fig. 3a. A positive high voltage of 6.8 kV was applied between the
liquid reservoir and a grounded counter electrode located at a distance of 2.6 cm from the wick tips. At this voltage
the maximum electrospray current from the 44 component array was 18 A or about 400 nA per wick. In the larger
system, prototype B, illustrated schematically in Fig. 2b, the electrospray ionization region consists of 440 individual
Porex wicks arranged in two square arrays and a grounded counter electrode located at a distance of about 5 cm from
the wick tips. A positive high voltage of up to 20 kV was applied between the wick array and the grounded counter
electrode producing a maximum electrospray current of about 80 A or an average of about 180 nA per wick. Fig. 3b
is a photograph of one half of the electrospray ionization array of prototype B.
612 G. Tepper, R. Kessick / Aerosol Science 39 (2008) 609 – 617

Wick array
Collector plates

Ionization region
+HV
Counter electrode
-HV

Wick array
Collector plates

Ionization region
+HV
Counter electrode
- HV

Fig. 2. Schematic diagram illustrating the main components in the electrospray-based ESP systems: (a) small prototype and (b) large prototype.

In both prototypes the electrospray aerosol emerges from the wick tips and, therefore, the particles in the air flowing
around the body of the wick and below the electrospray tips might avoid the charged aerosol and not become ionized.
We used air fins at the entrance to the ionization region in order to direct the incoming air flow into the electrospray
aerosols, but did not see any measurable difference in the resulting air purification efficiency. Smoke visualization of the
flow revealed that the presence of the sharp wick tips in the channel produces turbulent mixing in the ionization region
such that all of the entering streamlines have a high probability of interacting with the charged aerosols. Therefore, the
fins were not used in any subsequent testing.
The particle collection region in prototype A consists of two parallel square metal plates measuring approximately
30 cm on a side and separated by 2 cm. A negative potential was applied to one of the plates and the second plate was
grounded as illustrated in Fig. 2a. The particle collection region in the larger prototype B consists of an array of 19
parallel metal plates 23 cm long and separated by about 7 mm. A negative potential was applied to alternating collection
plates with respect to ground as illustrated in Fig. 2b.

3. Results and discussion

Fans were used to direct ambient air with a measured particle concentration through the ionization region of each
prototype and into a separate particle collection region and a particle counter was used to measure the single pass
particle removal efficiency, defined as the entering minus exiting particle count divided by the entering count, in each
prototype as a function of air flow rate, particle size, and collector plate bias. Fig. 4 is a plot of the particle removal
efficiency in prototype A for 0.3, 0.5, and 5 m particles as a function of collection plate differential voltage at a
measured air flow rate of 21.6 l/min. The particle removal efficiency increased monotonically at all particle sizes and
approached 100% at a collection plate voltage of −8 kV for 5 m particles and −13 kV for 0.3 and 0.5 m particles.
 G. Tepper, R. Kessick / Aerosol Science 39 (2008) 609 – 617 613

Solvent reservoir

Wicks

Fig. 3. Photograph of wick ionization regions in (a) prototype A and (b) prototype B.

100

80
Purification Efficiency %

60

5.0 micron
40 0.5 micron
0.3 micron

20

0
0 2 4 6 8 10 12 14
Collector Voltage (- kV)

Fig. 4. Collection efficiency versus collector plate voltage at an air flow rate of 21.6 l/min (prototype A).

From the data of Fig. 4 we conclude that the particle ionization efficiency, defined as the fraction of particles becoming
charged, is very close to 100% for all three particle sizes at this electrospray current and air flow rate. If the ionization
614 G. Tepper, R. Kessick / Aerosol Science 39 (2008) 609 – 617

efficiency was less than 100%, the particle collection efficiency could not approach 100% even at very high collector
plate voltages.
The trajectories of charged particles in devices such as differential mobility analyzers (DMAs) have been investigated
for years (Kim, Woo, Liu, & Zachariah, 2005) and can be used to analyze the particle collection efficiencies in these air
purification systems. The collection efficiency for the ionized particles depends on the particle diameter, the amount
of charge deposited onto a particle and the magnitude of the electric field in the collection region. At an air flow
rate of 21.6 l/min, the average air velocity between the collection plates is about 6 cm/s. Assuming that the average
particle velocity is the same as the average air velocity, it corresponds to an average particle residence time between the
collection plates of about 5 s. Therefore, for complete particle collection, the transverse particle velocity (perpendicular
to the stream lines) due to the electric field must be equal to 0.4 cm/s. By extrapolating the linear regions of the curves
of Fig. 3, we determine that the particle collection efficiency approaches 100% at an electric field magnitude of 6.5,
6, and 4 kV/cm for the 0.3, 0.5, and 5 m particles, respectively. Therefore, the minimum particle electrical mobility
required to traverse the 2 cm distance between the plates is simply the transverse velocity divided by the electric field
magnitude and is 6 × 10−5 , 7 × 10−5 , and 1 × 10−4 cm2 /V s for 0.3, 0.5, and 5 m particles, respectively.
The actual particle mobility (Z) depends on the amount of charge on the particle and the particle diameter and for
particles with diameters significantly larger than the mean free path of the air molecules is given by (Knutson & Whitby,
1975)

Z = nQ/3k D (1)

where n is the number of elemental charges on the particle, Q is the magnitude of an elemental unit of charge, k is
the kinematic viscosity of air (1.8 × 10−5 kg/m s) and D is the particle diameter. The actual amount of charge on a
particle of diameter D emerging from the ionization region will follow a probability distribution centered around a most
probable value of n, given by n. The average amount of charge n on the particles of diameter D can, therefore, be
determined by setting the experimentally determined mobility equal to the theoretical mobility from Eq. (1) and solving
for n. From this calculation we find that there is on average 2, 4, and 53 units of charge on the 0.3, 0.5, and 5 m
particles, respectively. For all three particle sizes the slopes of the curves shown in Fig. 4 decrease as the collection
probability approaches 100%. We believe that this decrease is due to the statistical variation in the amount of charge
on the particles. That is, the particles of a given diameter with an amount of charge less than n will require a larger
electric field in order to be collected. The change in slope is more apparent for the larger 5 m diameter particles with
n = 53 because there is a much broader charge distribution in comparison to the smaller 0.5 and 0.3 m particles with
n = 4 and 2, respectively.
The maximum possible or saturation charge of a spherical particle in an electric field increases as the square of the
radius as given by (Pauthenier & Moreau-Hanot, 1932)

Qmax = 40 r 2 pE (2)

where r is the particle radius, E is the electric field magnitude, and p is a parameter that depends on the dielectric
constant of the particle and is about 2 for a typical dust particle. From Eq. (2) we estimate that the saturation charge in
prototype A is about 30, 90, and 9000 for the 0.3, 0.5, and 5 m particles, respectively. For this calculation we estimated
that the electric field magnitude in the ionization region of prototype A is about 2.6 kV/cm, which is qualitative since
the electric field in the vicinity of the wick tips will be much higher due to the sharp radius of curvature while the
electric field in the electrospray plume could be decreased by space charge effects. Nevertheless, the results confirms
that the amount of charge deposited onto a given particle in the electrospray ionization region depends strongly on the
particle diameter and, for all diameters, is significantly less than the saturation limit.
The collection efficiency results for prototype A suggest the following simple design rule for maximizing particle
collection in systems with parallel collection plates,

d/L < {Zmin E}/vmax (3)

where d/L is the ratio of the collection plate spacing to the collection plate length, Zmin is the electrical mobility of the
slowest-moving particles, E is the electric field magnitude, and vmax is the maximum air velocity between the plates.
Of the three particle sizes tested in our systems, the slowest moving particles in the applied electric field were the small,
0.3 m particles because they had the least amount of charge. According to the above calculations, the 0.3 m diameter
 G. Tepper, R. Kessick / Aerosol Science 39 (2008) 609 – 617 615

1.2

Normalized Mobility
0.8

0.6

0.4

0.2

0
0 0.2 0.4 0.6 0.8 1
Particle Size (microns)

Fig. 5. Normalized particle electrical mobility versus particle diameter.

particles picked up, on average, 2 units of fundamental charge from the electrospray aerosols. Larger particles picked
up more charge and, presumably, smaller particles will pick up less. However, 1 unit of charge is the minimum amount
and the 0.3 m particles are already very close to this limit. Therefore, we expect the particle collection efficiency as
a function of particle size to go through a minimum corresponding to the largest particle that picks up a single unit
of charge from the electrospray droplets. Smaller, singly ionized particles will have a higher mobility as will larger
particles with a higher degree of ionization. Therefore, assuming that 0.2 m particles are the largest singly ionized
particles, the average amount of particle ionization n produced in prototype A as a function of particle diameter D
can be approximated as
n = 1 for D < 0.2 m
= (11 × D − 1.2) for D > 0.2 m (4)

Fig. 5 is a plot of Eq. (1) with the particle ionization n from Eq. (4) and normalized such that the maximum particle
mobility is equal to 1. The curve of Fig. 5 predicts that the collection efficiency in electrospray-based electrostatic
precipitators will be lowest for submicron particles with a diameter near 0.2 m. This prediction is in general agreement
with the experimental air purification efficiencies reported for our earlier prototypes (Tepper et al., 2007).
In all cases, in order to maximize the collection efficiency, the electric field should be as large as possible, but below
the air breakdown field and, for a given air flow rate, the air velocity between the plates can be reduced by using multiple
parallel plates with alternating electrical polarity as was used in the collection region of prototype B. Using a Zmin value
of 5 × 10−5 cm2 /V s, an electric field magnitude of 28 kV/cm and a d/L = 0.03 we find the maximum allowable air
velocity through the collection region of prototype B from Eq. (3) to be 31 cm/s, which corresponds to a volumetric air
flow rate of about 36 cfm. That is, we expect the collection region of prototype B to operate at close to 100% collection
efficiency as long as the air flow rate is below 36 cfm and assuming that the particle mobility distribution is similar to
that of prototype A.
Fig. 6 is a plot of percent particle reduction for 0.3, 0.5, and 5 m diameter particles in prototype B as a function of
electrospray current and at an air flow rate of 24.3 cfm, which is well below the maximum predicted value for prototype
B. The particle collection efficiency increases linearly with electrospray current and approaches a value of 100% at an
electrospray current of 65 A. Therefore, we conclude that there is a minimum amount of electrospray current required
to produce a certain degree of particle ionization at a given air flow rate. From the data of Fig. 6 for prototype B we
find that the electrospray current must be greater than 65 m at an air flow rate of 24.3 cfm in order for the total air
purification efficiency to approach a value of 100%.
Even though the particle ionization and collection regions were physically separated in both prototypes there will
be some particle collection on the grounded counter electrode in the ionization region and this could affect the above
616 G. Tepper, R. Kessick / Aerosol Science 39 (2008) 609 – 617

100

% Particle Reduction

90

80

5 micron
0.5 micron
0.3 micron
70
25 35 45 55 65
Electrospray Current (µAmps)

Fig. 6. Collection efficiency versus electrospray current in prototype B at an air flow rate of 24.3 cfm.

estimates on the degree of particle ionization. We estimate, using Eq. (3), that the maximum amount of particle collection
in the ionization region as a percentage of the amount of particle collection in the collection region is less than 7% in
prototype A and less than 3% in prototype B when the typical collector voltage is applied. These values, while small
are not insignificant and could affect the above mobility and particle ionization estimates.
The typical electrospray liquid flow rate in both prototypes is on the order of 100 nl/min per wick. The flow rate
depends on a number of parameters including the solvent properties, wick porosity, the applied electric field, and the
air flow rate, which can affect the solvent evaporation rate from the sides of the wick. Nevertheless, using 100 nl/min
per wick as a typical liquid flow rate it is possible to estimate factors such as the weekly liquid consumption rate and
the concentration of ethanol introduced into an air stream by a large array of wick sources. For example, 1000 wick
aerosol sources injecting a 90/10 water/ethanol solution into an air stream with a volumetric flow rate of 100 cfm would
consume about 1 l of solvent per week and, assuming no additional dilution, would result in an ethanol concentration
in air of about 1 ppm.

4. Conclusions

Arrays of electrospray aerosols from wick sources were used for particle ionization in ozone-free electrostatic
precipitation systems. The results indicate that the electrospray aerosols are highly efficient at transferring charge
onto particles in air, but the amount of deposited charge depends strongly on the particle diameter and is very low
for submicron particles. The particle collection efficiency depends on the collector plate design and particle electrical
mobility. The minimum mobility will be for the largest diameter particles that pick up a single unit of charge from the
electrospray aerosols.

Acknowledgments

This work was supported in part by the Defense Threat Reduction Agency under ARO contract no. W911NF-06-C-
0164. Initial charged droplet air sampling work was supported by the US Army under STTR contract no. W9132V-04-
C-0023.
The authors would like to thank Professor John Fenn for many illuminating discussions and suggestions.
 G. Tepper, R. Kessick / Aerosol Science 39 (2008) 609 – 617 617

References

Bartoli, C., von Rohden, H., Thompson, S. P., & Blommers, J. (1984). A liquid cesium field ion source for space propulsion. Journal of Physics D,
17, 2473–2483.
DelaMora, J. F. (2007). The fluid dynamics of taylor cones. Annual Review of Fluid Mechanics, 39, 217–243.
Fenn, J. B., Mann, M., Meng, C. K., Wong, S. F., & Whitehouse, C. M. (1989). Electrospray ionization for mass-spectrometry of large biomolecules.
Science, 246(4926), 64–71.
Ganan-Calvo, A. M., Davila, J., & Barrero, A. (1997). Current and droplet size in the electrospraying of liquids. Scaling laws. Journal of Aerosol
Science, 28(2), 249–275.
Gomez, A., & Tang, K. Q. (1994). On the structure of an electrostatic spray of monodisperse droplets. Physics of Fluids, 6(1), 404–414.
Hazucha, M. J., & Lefohn, A. S. (2007). Nonlinearity in human health response to ozone: Experimental laboratory considerations. Atmospheric
Environment, 41(22), 4559–4570.
Jaworek, A., Krupa, A., & Czech, T. (2007). Modern electrostatic devices and methods for exhaust gas cleaning: A brief review. Journal of
Electrostatics, 65, 133–155.
Kim, S. H., Woo, K. S., Liu, B. Y. H., & Zachariah, M. R. (2005). Method of measuring charge distribution of nanosized aerosols. Journal of Colloid
and Interface Science, 282, 46–57.
Knutson, E. O., & Whitby, K. T. (1975). Journal of Aerosol Science, 6, 443–451.
Koerner, T., Turck, K., Brown, L., & Oleschuk, R. D. (2004). Porous polymer monolith assisted electrospray. Analytical Chemistry, 76, 6456–6460.
Pauthenier, M. M., & Moreau-Hanot, M. (1932). Journal de Physique at le Radium, 3, 59–613.
Rulison, A. J., & Flagan, R. C. (1994). Synthesis of yttria powders by electrostatic pyrolysis. Journal of the American Ceramic Society, 77,
3244–3250.
Seaver, A. E., & Eckhardt, C. J. (1988). US Patent No. 4,748,043.
Tepper, G., Fenn, J., Kessick, R., Pestov, D., & Anderson, J. (2006). A new air sampler based on electrically charged liquid nanodroplets. In
Proceedings of the sixth IEEE conference on nanotechnology (vol. 2(17–20), pp. 781–782).
Tepper, G., Kessick, R., & Pestov, D. (2007). An electrospray-based, ozone-free air purification technology. Journal of Applied Physics, 102, 113305.
Washburn, E. W. (1921). Physical Review, 17, 273.
Wilm, M., & Mann, M. (1996). Analytical properties of the nanoelectrospray ion source. Analytical Chemistry, 68, 1–8.