Content uploaded by Jens Olaf Pepke Pedersen
Author content
All content in this area was uploaded by Jens Olaf Pepke Pedersen on Aug 19, 2014
Content may be subject to copyright.
Aerosol nucleation induced by a high energy particle beam
Martin B. Enghoff,
1
Jens Olaf Pepke Pedersen,
1
Ulrik I. Uggerhøj,
2
Sean M. Paling,
3
and Henrik Svensmark
1
Received 8 February 2011; revised 25 March 2011; accepted 31 March 2011; published 12 May 2011.
[1] We have studied sulfuric acid aerosol nucleation in an
atmosphericpressurereactionchamberusinga580MeV
electron beam to ionize the volume of the reaction chamber.
We find a clear contribution from ion‐ induced nucleation
and consider this to be the first unambiguous observation
of the ion‐effect on aerosol nucleation using a particle beam
under conditions that resemble the Earth’s atmosphere. By
comparison with ionization using a gamma source we fur-
ther show that the nature of the ionizing particles is not
important for the ion‐induced component of the nucleation.
This implies that inexpensive ionization sources ‐ as
opposed to exp ensive accelerator beams ‐ can be used for
investigations of ion‐induced nucleation.
Citation: Enghoff,
M. B., J. O. P. Pedersen, U. I. Uggerhøj, S. M. Paling, and
H. Svensmark (2011), Aerosol nucleation induced by a high
energy particle beam, Geophys. Res. Lett., 38, L09805,
doi:10.1029/2011GL047036.
1. Introduction
[2] Aerosol and cloud research is one of the most critical
frontiers of climate science [Shindell et al., 2009; Bodenschatz
et al., 2010] and the direct radiative forcing and indirect
cloud albedo forcing from aerosols remain the dominant
uncertainty in the radiative forcing of the atmosphere.
Among the uncertainties are the mechanisms behind for-
mation of new aerosols from the gas phase and several
schemes have been suggested [Curtius, 2006] such as ternary
nucleation with amines [Kurtén et al., 2008]. Observations
show that sulfuric acid, together with water and other gas
phase molecules play an important role in atmospheric
nucleation [Lee et al., 2003; Kulmala et al., 2006] and these
systems are also the basis of many theoretical nucleation
works [Wyslouzil et al., 1991; Hanson and Lovejoy, 2006].
[
3] The importance of ion‐induced nucleation for the
Earth’ atmosphere is a subject of intense discussions [Yu,
2010; Sloan and Wolfendale, 2008] which is being studied
both in field measurements [Kulmala et al., 2010] and in
modeling work [Pierce and Adams, 2009; Kazil, 2010].
Previous laboratory studies using gamma rays have found a
positive correlation between ionizing radiation intensity and
aerosol concentration in air [Raes and Janssens, 1985;
Svensmark et al., 2007; Enghoff et al., 2008], but systematic
laboratory experiments on aerosol nucleation with ionizing
particle radiation have so far only been performed with very
high radiation doses far above natural atmospheric levels
using a‐ and b‐sources [Adachi et al., 1996] or high
intensity electron [Hakoda et al., 2003] and proton beams
[Imanaka et al., 2010]. Preliminary measurements with a
more realistic 3.5 GeV/c pion beam have been reported to
show indications of ion‐induced nucleation [Duplissy et al.,
2010], but suffered from unstable conditions in the reaction
chamber. A new experimental facility dedicated to these
studies is under commissioning at the CERN Proton Syn-
chrotron (PS) [CLOUD Collaboration, 2000] and is likely to
resolve these problems.
2. Experimental Methods
[4] The experiments took place in a 1 m long, 50 L
cylindrical, electropolished stainless steel reactor, previously
used by Enghoff et al. [2008]. A mixture of pure humidified
synthetic air is continuously flowed through the vessel at a
rate of 3.1 L/min together with trace amounts of (typically
2 ppb) sulfur dioxide (SO
2
) and approximately 55 ppb
ozone (O
3
) with the relative humidity held at about 50%.
UV light was used to initiate the photochemistry in the
chamber leading to an in situ production of sulfuric acid.
The temperature was kept at 21.2 ± 0.04°C and the pressure
at 1–2 mbar above room pressure.
[
5] A low‐intensity beam of 580 MeV electrons from the
Aarhus Storage Ring ASTRID [Myers et al., 1998] at
the University of Aarhus was used as the ionizing source. The
energy of the electron beam falls within the main part of
the natural cosmic ray spectrum and the induced ion con-
centrations cover the full range of atmospheric values.
[
6] Production of aerosols was measured using a con-
densation particle counter (CPC) with a nominal 50% cutoff
diameter of 4 nm. The chamber volume could be ionized
using the 580 MeV electrons going coaxially through the
chamber covering an elliptical area with semi axes of 4.4
and 6.1 cm (FWHM) at the front end of the chamber. The
ionization level from the beam was adjustable by changing
the amount of beam extracted and, as shown by the esti-
mated energy loss (see auxiliary material), the electrons pass
through the volume of air almost undisturbed.
1
Alternatively
a 33.5 MBq Na‐22 gamma source could be used. For some
of the runs a 5 cm thick scintillator was placed in the beam,
7.5 m upstream of the chamber, dispersing the beam by an
additional 6 cm at the front of the chamber.
[
7] The procedure for each experiment involved turning
the UV lamps on for 10 minutes giving a burst of sulfuric
acid and then recording the resulting particle number. The
time between UV bursts was 60 minutes, in order to allow
1
National Space Institute, Technical University of Denmark,
Copenhagen, Denmark.
2
Department of Physics and Astronomy, University of Aarhus,
Aarhus, Denmark.
3
Department of Physics and Astronomy, University of Sheffield,
Sheffield, UK.
Copyright 2011 by the American Geophysical Union.
0094‐8276/11/2011GL047036
1
Auxiliary materials are available in the HMTL. doi:10.1029/
2011GL047036.
GEOPHYSICAL RESEARCH LETTERS, VOL. 38, L09805, doi:10.1029/2011GL047036, 2011
L09805 1of4
the aerosol concentration (as measured by the CPC) to reach
zero by the end of each burst, with the aerosols being lost to
walls and by dilution.
3. Results and Discussion
[8] The measurement program consisted of varying the
ionization levels by adjusting the electron beam, investi-
gating the effect of dispersing the electron beam, and
comparing the aerosol formation (observed at 4 nm) induced
by, respectively, the gamma source and the electron beam.
A range of ionization levels from background levels to
∼700 cm
−3
s
−1
were tested and the results can be seen in
Figure 1, where the formation rate measured by the CPC is
shown as a function of ion concentration. The runs have
been divided into 3 groups and it is seen that an increase in
ion concentration causes an increase in the aerosol formation
rate for all three groups of runs. Also shown for each group
is the best linear fit to the data. Runs 1–5 were made using
air from the same cylinders. Between run 5 and 6 the air
supply was changed, causing a large shift in the number of
aerosols produced in a burst and between runs 9 and 10 a
shift occurred again. This shift is seen on Figure 1 as a
vertical displacement of the 3 groups of measurements, and
we speculate that the shift is due to varying amounts of trace
impurities in the purified air cylinders. The impurities are
able to pass the filters and influence the neutral component
of the nucleation process. There is also a smaller variation of
the slopes of the linear fits in Figure 1 between the 3 groups
of measurements, but they do not seem to vary in a sys-
tematic way with the vertical shift. While we cannot exclude
that the ion‐induced component of the nucleations are also
affected by the impurities, the effect thus seem to be smaller
on this component, and does not change the main obser-
vation that there is a clear effect of the ion concentration on
the formation rate.
[
9] During run 6 the effect of dispersing the beam was
tested. The formation rate at 4 nm changed from 0.095 ±
0.003 cm
−3
s
−1
to 0.094 ± 0.009 cm
−3
s
−1
, showing that
there was no significant difference between the two settings.
This also shows that the gases in the chamber were well
mixed.
[
10] The sulfuric acid concentrations are found to be
∼7·10
8
cm
−3
(+100/−50)% for runs 1–5 and ∼6·10
8
cm
−3
for runs 6–10, with run 10 having a slightly higher con-
centration than runs 6–9 (see auxiliary material). This could
explain the changes in aerosol production rates observed
during the experiments.
[
11] In all runs the effect of the ions is to produce addi-
tional aerosol formation over a wide range of ion con-
centrations from background levels to 19,000 cm
−3
. Model
results [Kazil and Lovejoy, 2004] indicate that the nucle-
ation rate should peak at high ionization levels, but appar-
ently this range is not reached at the present experimental
conditions, which may be a result of our rather high
[H
2
SO
4
] values. The error bars shown are statistical
uncertainties and the absolute uncertainties are much larger,
but since our purpose is to investigate the relative variation
of the nucleation rate with ion density, we find the statistical
uncertainties to be most relevant. A linear relationship
(which has been observed before in a smaller range of ion
densities by Svensmark et al. [2007]) would show that the
nucleation rate depends on ion concentration, but we cannot
for example exclude a square relationship, which would
indicate that nucleation depends on the ion production rate
and any conclusions would also be complicated by linear
wall losses. An extrapolation of the curves to zero ion
density would give the nucleation rate due to processes not
involving ions, but further experimental work at low ioni-
zation values is needed to resolve this.
[
12] We caution that the combination of a finite residence
time in the reaction chamber and a detector cutoff larger
than the diameter of the critical cluster has the effect that
Figure 1. Formation rate (at 4 nm) versus ion concentration as measured by the CPC (see text for details). Error bars show
statistical uncertainties.
ENGHOFF ET AL.: NUCLEATION INDUCED BY A PARTICLE BEAM L09805L09805
2of4
only a fraction of the nucleated particles will grow to
detectable sizes as recently shown by Sipilä et al. [2010]. In
our experiment the possibility of a faster growth rate of the
ions may artificially enhance the measured particle con-
centrations when using the ionizing sources compared to no
sources. From an estimated residence half‐life of 200 s, and
with [H
2
SO
4
] concentrations in the range 6–7·10
8
cm
−3
we
conclude from Sipilä et al. [2010, Figure 3] that we reach
mean particle diameters above 3 nm, which is close to the
detector cutoff around 4 nm, but may still leave the majority
of the aerosols undetected. Since our measurements were
made under stable conditions where only the ion density
was varied, our observation that ions enhance the formation
rate at 4 nm remains solid in spite of the large absolute
uncertainties; however, the measured relative enhancement
should be considered an upper limit. Sipilä et al. [2010] also
found that the rate of homogenous nucleation measured with
an instrument similar to ours should vary with the 6th power
of [H
2
SO
4
]. If the difference between our runs 1–5 and 6–9
is due to a change in [H
2
SO
4
] by about a factor of 1.2, this
should result in a homogeneous nucleation rate difference
by a factor of 3 which is close to the values found by
extrapolating the rates shown in Figure 1 to zero ion
concentration.
[
13] Using an analytical model (see auxiliary material) we
have calculated the unobserved nucleation rate J1 from the
observed formation rates J4. This has large inherent uncer-
tainties and gives absolute 1 nm nulceation rates in the
order of 1 s
−1
cm
−3
, which is about two orders of magnitude
lower than previous laboratory measurements under similar
conditions and with only background ionization [see Sipilä
et al., 2010, and references therein]. Some explanations
for this discrepancy could be that our sulfuric acid con-
centrations and our counting efficiency are lower than
estimated. A possible remedy would be to normalize our
nucleation rates to the literature values, but we have chosen to
give our estimated absolute values with the added note that
only the relative variations should be considered accurate.
4. Conclusions
[14] An important result from this work is (as seen from
Figure 1) that nucleation induced with the ionization from the
gamma source experiments are indistinguishable from those
using the electron beam. Compared to the 580 MeV elec-
trons the gamma rays have rather low energies and the
electrons emitted through Compton scattering will ionize
very locally, whereas the 580 MeV electrons have a mean
energy loss rate close to the minimum (minimum ionizing)
and will ionize along their path. The gammas may however
ionize more than one atom due to secondary gammas
emitted in the Compton process. Nevertheless our data show
that the fraction of nucleation events due to the electron
beam or the gamma rays is an effect of the produced ions
and that the nature of the ionizing particles is not important.
This demonstrates that while a particle beam can be used for
studies of the ion‐effect on aerosol nucleation future labo-
ratory studies of atmospheric aerosol nucleation can be
greatly facilitated, since they can more easily be done with
gamma radiation sources instead of using complicated and
expensive accelerator beams.
[
15] Further we conclude that under the present experi-
mental conditions (P = 1 atmosphere, T = 21.2°C, [SO
2
]=
2 ppb) there is a clear evidence of ion‐induced nucleation as
a source of aerosol production, which corroborates earlier
measurements using only gamma radiation as the ionizing
source [Svensmark et al., 2007; Enghoff et al., 2008].
However we also stress that this is a qualitative result and
that further work at lower values of P, T,[H
2
SO
4
], and
[SO
2
] is required to perform an extrapolation to real atmo-
spheric conditions, in particular since our [H
2
SO
4
] con-
centrations of 6 – 7·10
8
cm
−3
are at least an order of
magnitude above typical clean‐air concentrations of about
10
7
cm
−3
. Using the lower estimated [H
2
SO
4
] concentration
of 6 · 10
8
cm
−3
in runs 6–9 we find that the effect of
changing the ion density from ∼700 to a typical atmospheric
value of ∼1,500 cm
−3
results in a relative increase in for-
mation rate at 4 nm of approximately 7%, but modelling
results [Kazil and Lovejoy, 2004] show that lowering the
values of the present experimental parameters will influence
both the critical cluster size and the particle production and
may both correlate and anti‐correlate with the ionization
level.
[
16] Acknowledgments. We thank Michel Avngaard for help with
setting up the experiment and the ASTRID staff for delivering the electron
beam. MBE acknowledges support from the Carlsberg Foundation and
SMP thanks the Engineering and Physical Sciences Researc h Counc il
(EPSRC) and the Science and Technology Facilities Council (STFC) for
support.
[17] The Editor thanks two anonymous reviewers for their assistance in
evaluating this paper.
References
Adachi, M., et al. (1996), Exp erimental evaluation of ion‐induced nucle-
ation in nanometer‐aerosol formation by a‐ray radiolysis of SO
2
/H
2
O/
N
2
mixtures, Coll. Surf., 109,39–48.
Bodensch atz, E., et al. (2010), Can we understand clouds witho ut turbu-
lence, Science, 327, 970–971.
CLOUD Collaboration (2000), A study of the link between cosmic rays and
clouds with a cloud chamber at the CERN PS, Rep. CERN/SPSC 2000‐
021, SPSC/P317, Eur. Organ . f or Nucl. Res., Geneva, Switzerland.
[Available at http://cloud.web.cern.ch/cloud.]
Curtius, J. (2006), Nucleation of atmospheric aerosol particles, C. R. Phys. ,
7, 1027–1045.
Duplissy, J., et al. (2010), Results from the CE RN pilot CLOUD experi-
ment, Atmos. Chem. Phys., 10, 1635–1647.
Enghoff, M. B., et al. (2008), Evidence for the role of ions in aerosol nucle-
ation, J. Phys. Chem. A, 112, 10,305–10,309.
Hakoda, T., et al. (2003), Charged nanoparticle formation from humidified
gases with and without dilute benzene under electron beam irradiation,
J. Aerosol Sci., 34, 977–991.
Hanson, D., and E. Lovejoy (2006), Measurement of the thermodynamics
of the hydrated dimer and trimer of sulfuric acid, J. Phys. Chem. A, 110,
9525–9528.
Imanaka, M., et al. (2010), Nanoparticle formation in H
2
O/N
2
and H
2
O/Ar
mixtures under irradiation by 20 MeV protons an dpositive corona dis-
charge, J. Aerosol Sci., 41, 468–474.
Kazil, J. (2010), Aerosol nucleation and its role for clouds and Earth’s radi-
ative forcing in the aerosol‐climate model ECHAM5‐HAM, Atmos.
Chem. Phys., 10, 10,733–10,752.
Kazil, J., and E. R. Lovejoy (2004), Tropospheric ionization and aerosol
production: A model study, J. Geophys. Res., 109, D19206,
doi:10.1029/2004JD004852.
Kulmala, M., K. E. J. Lehtinen, and A. Laaksonen (2006), Cluster activa-
tion theory as an explanation of the linear dependence between formation
rate of 3 nm particles a nd sulphuric acid concentration, Atmos. Chem.
Phys., 6, 787–793.
Kulmala, M., et al. (2010), Atmospheric data over a solar cycle: No connec-
tion between galactic cosmic rays and new particle formation, Atmos.
Chem. Phys., 10, 1885–1898.
Kurtén, T., V. Loukonen, H. Vehkamäki, and M. Kulmala (2008), Amines
are likely to enhance neutral and ion‐induced sulfuric acid‐water nucle-
ation in the atmosphere more effectively than ammonia, Atmos. Chem.
Phys., 8, 4095–4103.
ENGHOFF ET AL.: NUCLEATION INDUCED BY A PARTICLE BEAM L09805L09805
3of4
Lee, S.‐H., et al. (2003), Particle formation by ion nucleation in the upper
troposphere and lower stratosphere, Science, 301, 1886–1889.
Myers, S., L. Liljeby, C. Petit‐Jean‐Genaz, J. Poole, and K. Rensfelt (Eds.)
(1998), Proceedings of the Sixth European Particle Accelerator Confer-
ence, Inst. of Phys., Philadelphia, Pa.
Pierce, J. R., and P. J. Adams (2009), Can cosmic rays affect cloud conden-
sation nuclei by altering new particle formation rates?, Geophys. Res.
Lett., 36, L09820, doi:10.1029/2009GL037946.
Raes, F., and A. Janssens (1985), Ion‐induced aerosol formation in a H
2
O‐
H
2
SO
4
system—1. extension of the classical theory and search for exper-
imental evidence, J. Aerosol Sci., 16, 217–227.
Shindell, D. T., et al. (2009), Improved attribution of climate forcing to
emissions, Science, 326, 716–718.
Sipilä, M., et al. (2010), The role of sulfuric acid in atmospheric nucleation,
Science, 327, 1243–1246.
Sloan, T., and A. W. Wolfendale (2008), Testing the proposed causal link
between cosmic rays and cloud cover, Environ. Res. Lett., 3, 024001.
Svensmark, H., et al. (2007), Experimental evidence for the role of ions in
particle nucleation under atmospheric conditions, Proc. R. Soc. A, 463,
385–396.
Wyslouzil, B., J. Seinfeld, R. Flagan, and K. Okuyama (1991), Binary
nucleation in acid‐water systems. I. Methanesulfonic acid‐water, J. Chem.
Phys., 94, 6827–6841.
Yu, F. (2010), Ion‐mediated nucleation in the atmosphere: Key controlling
parameters, implica tions, and look‐up table, J. Geophys. Res., 155,
D03206, doi:10.1029/2009JD012630.
M. B. Enghoff, J. O. P. Pedersen, a nd H. Svensmark, National Space
Institute, Technical Universi ty of Denmark, Juliane Maries Vej 30,
DK‐2100 Copenhagen, Denmark. (jopp@space.dtu.dk)
S. M. Paling, Department of Physics and Astronomy, Univ ersity of
Sheffield, Sheffield S3 7RH, UK.
U. I. Ug gerhøj, Department of Physics and Astron omy, University of
Aarhus, Ny Munkegade 120, DK‐8000 Aarhus, Denmark.
ENGHOFF ET AL.: NUCLEATION INDUCED BY A PARTICLE BEAM L09805L09805
4of4