ACPAtmospheric Chemistry and PhysicsACPAtmos. Chem. Phys.1680-7324Copernicus PublicationsGöttingen, Germany10.5194/acp-17-10733-2017Direct molecular-level characterization of different heterogeneous freezing
modes on mica – Part 1AbdelmonemAhmedahmed.abdelmonem@kit.eduhttps://orcid.org/0000-0002-4348-2439Institute of Meteorology and Climate Research – Atmospheric Aerosol
Research (IMKAAF), Karlsruhe Institute of Technology (KIT), 76344
Eggenstein-Leopoldshafen, GermanyAhmed Abdelmonem (ahmed.abdelmonem@kit.edu)13September2017171710733107416April201720April20173August20177August2017This work is licensed under the Creative Commons Attribution 3.0 Unported License. To view a copy of this licence, visit https://creativecommons.org/licenses/by/3.0/This article is available from https://acp.copernicus.org/articles/17/10733/2017/acp-17-10733-2017.htmlThe full text article is available as a PDF file from https://acp.copernicus.org/articles/17/10733/2017/acp-17-10733-2017.pdf
The mechanisms behind heterogeneous ice nucleation are of
fundamental importance to the prediction of the occurrence and
properties of many cloud types, which influence climate and
precipitation. Aerosol particles act as cloud condensation and
freezing nuclei. The surface–water interaction of an ice
nucleation particle plays a major, not well explored, role in its
ice nucleation ability. This paper presents
a real-time molecular-level comparison of different freezing
modes on the surface of an atmospherically relevant mineral surface
(mica) under varying supersaturation conditions using second-harmonic generation spectroscopy. Two sub-deposition nucleation
modes were identified (one- and two-stage freezing). The nonlinear
signal at the water–mica interface was found to drop following the formation
of a thin film on the surface regardless of (1) the formed phase
(liquid or ice) and (2) the freezing path (one or two step),
indicating similar molecular structuring. The results also revealed
a transient phase of ice at water–mica interfaces during
freezing, which has a lifetime of around 1 min. Such
information will have a significant impact on climate change,
weather modification, and the tracing of water in hydrosphere studies.
Introduction
Clouds influence the energy budget by scattering sunlight and
absorbing heat radiation from the earth and are therefore
considered the major player in the climate system. Formation of
ice changes cloud dynamics and microphysics because of the release
of latent heat and the Bergeron–Findeisen process, respectively
(Pruppacher and Klett, 1997). Ice nucleation in the atmosphere can
be triggered heterogeneously by aerosol particles and ice-nucleating
particles (INPs), or it occurs homogeneously at about
-38 ∘C (Pruppacher and Klett, 1997). Cloud evolution
depends not only on temperature and humidity but also on the
abundance and surface characteristics of atmospheric
aerosols. Understanding the factors that influence ice formation
within clouds is a major unsolved and pressing problem in our
understanding of climate (Slater et al., 2016). Field and
laboratory experiments on cloud formation started decades ago (see
Schaefer, 1949; DeMott et al., 2011; Hoose and Mohler, 2012, and
references therein) and are ongoing. A wide variety of results
and observations has been obtained in cloud microphysics,
especially with respect to the ice nucleation ability of
atmospheric aerosol particles and, hence, the mechanisms of cloud
dynamics, precipitation formation, and interaction with incoming
and outgoing radiation. Aerosol particles act as cloud
condensation nuclei for liquid clouds, immersion or contact
freezing nuclei for mixed-phase clouds, and heterogeneous
deposition nuclei for ice (cirrus) clouds. Depending on whether
water nucleates ice from the vapor or the supercooled liquid
phase, ice nucleation is classified as deposition nucleation or
immersion nucleation, respectively. Despite numerous
investigations aimed at characterizing the effect of particle size
and surface properties of the INP, there is a lack of information
about the restructuring of water molecules on the surface of INPs
around the heterogeneous freezing point.
In this paper, the discrimination between the different modes of the freezing
of water on an ice-nucleating surface using nonlinear optical
spectroscopy is demonstrated. Mica, a widely found layered clay
mineral and one of the most prominent mineral surfaces due to its
atomic flatness and chemical inertness produced by perfect
cleavage parallel to the 001 planes (Poppa and Elliot, 1971), was
selected as a model surface in this study. However, the image of
an inert and atomically smooth surface prepared by cleavage of
muscovite mica in an ambient atmosphere is not quite correct
(Christenson and Thomson, 2016). Surface analytical techniques
found that the surface of muscovite mica cleaved in laboratory
air, as is the case in this work, contains a water-soluble compound,
potassium carbonate crystallites, which may cover a few tenths of
a percent of the surface area (Christenson and Israelachvili,
1987). Nevertheless, definitive chemical analysis showing its
presence is not yet available. However, the mobility of the
potassium ions as potassium carbonate does not necessarily significantly affect any measurement of average surface properties as in
the case of this work, and at high humidities the potassium will
be widely dispersed across the surface. Those readers who are
interested in more details on the nature of the mica surface in
general and the air-cleaved mica surface in particular are referred to
the review paper of Christenson and Thomson (2016) and
papers cited therein.
Mica, as a natural particle, is believed to be among the most
effective ice-nucleating minerals in the deposition mode (Eastwood
et al., 2008; Mason and Maybank, 1958). An early study, which used
a projection microscope and focused on the deposition nucleation of ice on
freshly cleaved synthetic fluorophlogopite mica, which is similar
in structure to muscovite but has the hydroxyl groups replaced by
fluorine, revealed that there is no growth of ice until saturation
with respect to water is reached (Layton and Harris, 1963). The
authors concluded that at temperatures above -40 ∘C,
the growth of ice on mica should be a two-step process: a nucleus
forms as water and then freezes. Experimental evidence of two-step
nucleation was also provided by Campbell et al. (2013) using an
optical microscope. With the help of a scanning optical
microscope, they showed that the nucleation, of various organic
liquids crystallizing from vapor on mica surfaces, favored
specific nucleation sites with surface features such as cleavage
steps, cracks, and pockets. However, they suggested that
a supercooled liquid phase forms first and then freezes after it
has grown to a size which thermodynamically favors the solid
phase. These assumptions were based merely on thermodynamic
observations (temperature and vapor supersaturation). A later
study by the same group has confirmed the role of the surface
features and the two-step process for organic liquids and
strongly suggested a two-step process for water and ice (Campbell
et al., 2017). Recent molecular dynamics (MD) simulations of deposition freezing
revealed that water first deposits in the form of liquid clusters
and then crystallizes isothermally from there (Lupi et al.,
2014). So far, there has been no direct experimental evidence of
two-step freezing based on probing the molecular structuring of
water molecules next to the surface.
In this work, second-harmonic generation (SHG) in total internal
reflection (TIR) geometry was used to probe the change in the
degree of ordering of water on the surface of mica. Compared to sum-frequency generation (SFG), SHG is
a powerful and simple surface-sensitive spectroscopic tool for studying molecules near
surfaces and at interfaces (Shen, 1989b, 1989a). The amplitude and
polarization of the generated field, as a function of the
polarization of the incident fields, carry information on the
abundance and structure of the interfacial molecules between two
isotropic media (Jang et al., 2013; Rao et al., 2003; Zhuang
et al., 1999). More details on SHG and SFG can be found in Sect. 2 and in the Supplement. In the system described here, the
SHG signal originates in the nonresonant electric dipolar
contribution of the interfacial molecules. The signal response
relates to the overall arrangements of the interfacial entities
(Fordyce et al., 2001; Goh et al., 1988; Luca et al., 1995) and is proportional to the incident field and the second-order nonlinear
susceptibility χ(2) of the interface. When the interface
is charged, due to this charge, the static electric field can induce
a third-order nonlinear polarization due to the contribution of
the third-order nonlinear susceptibility
χ(3) of the
solution (Ong et al., 1992; Zhao et al., 1993). In this work, the
contribution of χ(3) to the total SHG signal has been ignored because there was
no significant change in the interface charge with
temperature. The change in pH with temperature is known for
neutral water (e.g., from pH 7 at 25 ∘C to pH 7.47 at
0 ∘C). In addition, this change does not mean that water
becomes more alkaline at lower temperatures because, in the case of
pure water and according to Le Châtelier's principle, there
is always the same concentration of hydrogen and hydroxide ions
and, hence, the water is still neutral (pH = pOH) even if its
pH changes. The pH 7.47 at 0 ∘C is simply the new
way of referencing neutral water pH at 0 ∘C. In addition,
assuming that the surface potential has an influence on the
background signal, this will not change even if the pH changes
with temperature because the surface potential values of the
muscovite basal plane (the surface under study) is pH independent
in the range of pH 5.6 to 10 (Zhao et al., 2008).
The results provide new insight into heterogeneous freezing
processes and show the suitability of the method for studying
current issues relating to ice nucleation. Initially, I found
that the SHG signal drops following the formation of a thin film
regardless of whether the freezing path consists of one or
two steps and with the initially formed phase, liquid or ice,
indicating a similar molecular structuring. In addition, I
observed a transient SHG signal after immersion freezing. The
hygroscopicity of mica is expected to play a role in the described
processes. The hygroscopicity and Langmuir isotherm studies on
mica are available in literature but only for room temperature
where the sample and environment are at equilibrium (Balmer
et al., 2008; Beaglehole et al., 1991; Hu et al., 1995). Such
studies at supercooled surfaces are worth doing and could be
a topic of future work. An atomic force microscopy (AFM) study at 21 ∘C showed no
water absorbed on the surface of mica at RH (relative humidity) =18 % (Hu
et al., 1995). The first uniform water phase, of large
two-dimensional islands with geometrical shapes in epitaxial
relation with the underlaying mica lattice, was observed at
RH =28 %. The growth of this water phase is completed when
the humidity reached between 40 and 50 %. In my experiments on
mica it is not possible to detect sub-monolayers, at least at this
stage, due to technical reasons mentioned later. In the presented
work, only clear steps in the signal were considered.
Experimental setupMaterials and setup
All experiments were carried out using MilliQ water
(18.2 MΩcm). The total organic content in
this water is below 4 ppb. Mica samples were obtained
from Plano GmbH, Wetzlar, Germany. The mica samples were freshly cleaved
parallel to the 001 plane in air right before use. The freshly
cleaved mica exhibits a wetting surface (on which water was
spreading visually). The SHG experiments were conducted using
a femtosecond laser system (Solstice, Spectra Physics) with
a fundamental beam of 800 nm wavelength, 3.5 mJ pulse
energy, ∼80 fs pulse width, 1 kHz repetition rate,
and a beam diameter of ∼2mm at the interface. The
supercooled SHG setup and the measuring cell are similar to those
described in previous publications (Abdelmonem et al., 2015,
2017). Compared to the setup described in Abdelmonem et al. (2015), a single fundamental beam incident on the interface was
used (Fig. 1) and the SM (S-polarized SHG/45∘-polarized
incident) polarization combination was measured. Figure 1 shows
the sample and beam geometry. The polarization direction of the
incident beam was controlled by a half-wave plate followed by
a cube polarizer. The generated signal was collected using
a photomultiplier tube (PMT) placed downstream of an optical
system including band-pass filters for 400 nm and
a polarization analyzer. A sapphire prism was used as an optical
coupler to the surface of a thin mica substrate, the basal plane
of which was exposed to liquid water or water vapor. The
fundamental beam had an incident angle of 15∘ with the
surface normal of the outer side of the prism. Under this
geometry, the reflected fundamental (800 nm) and generated
SHG (400 nm) beams co-propagate to the sapphire–air
interface at the other side of the prism at which both beams are
refracted at two different angles. Only the SHG signal was allowed
to reach the detection path. Before starting the measurements, the
polarization of the SHG signal generated from water at the surface
was analyzed and found to have the expected maxima at S and P
polarizations corresponding to an incident 45∘-polarized
light. The signal was quadratically dependent on the input
power. To study mica in TIR geometry, an index matching gel (IMG)
from Thorlabs (G608N3, RI ∼ 1.45186 at 800 nm) was
used to fix the mica sample on the hypotenuse of the sapphire
prism. The freezing point was not specified by the manufacturer
but tested in the lab. The gel, at least, was not frozen until
-45 ∘C. A detailed description of sample geometry and
the selection of the IMG was published in Abdelmonem et al. (2015). Less than 15 % of the laser output power was coupled to
the setup so as not to destroy the IMG. The fluctuation in the signal
due to reducing the laser power limited the sensitivity of the
system and it was not possible to monitor minor changes which may
arise from pre-adsorption of sub-monolayers at temperatures higher
than the dew point or freezing point.
The sample and beams geometry. See text for details.
A temperature-controlled environmental chamber developed in-house (Fig. 2) was integrated into the setup. A commercially
available cold stage (Linkam model HFS-X350) was used after
modifying the housing to accommodate the SHG setup. The sapphire
prism was place in a copper adaptor which was fixed on the silver
block of the Linkam cold stage. The substrate of interest was
sealed to a circular opening of 8 mm in diameter in
a teflon cell which was purged with the sample gas during the
experiments. The cold stage can perform controlled heating and
cooling ramps, applied to the silver block, at rates between 0.01
and 100 ∘Cmin-1. The temperature stability of the
cold stage is better than 0.1 K. The temperatures of the
air inside the cell and the sapphire prism top and bottom were
measured using four-wire Pt100 elements. The temperature of the
probed spot on the surface was considered to be the average of the
sample top and sample bottom temperatures. However, it should be
emphasized that the exact onset condition of freezing is not the
focus of this work, but the focus is rather the study of the qualitative
behavior of water molecules during freezing on the different
paths. During the experiments, the gas box was filled with
N2 gas to avoid condensation on the outer surfaces of the
prism during cooling. The humid air pumped to the measuring cell
was obtained by mixing dry gas and 100 % humid gas with
different ratios at 21 ∘C using two mass flow controllers
(Tylan 2900). The continuous flow of the gas (either dry or humid)
during the experiment set the temperature inside the cell to 21±0.5∘C. The corresponding fluctuation of the relative
humidity was less than 0.2 %. The corresponding fluctuation in
the dew point, at RH =5±0.2 for instance, was ±0.5∘C. The gas mixing ratio vs. RH was calibrated by setting a mixing ratio, cooling the sample, and recording the condensation/freezing temperature
at which the reflectivity, at an angle equal to the critical angle of TIR for
the air–mica interface, starts to drop due to the violation of the TIR
condition. This
temperature was used to define the corresponding RH using the Arden
Buck equation.
The same method was used to differentiate between liquid film,
liquid bulk, transient ice, and stable ice in this work. The border
between a film at the solid–air interface and a bulk at the
solid–water (or solid–ice) interface was defined experimentally by the point where
the intensity of a TIR reflected light from the solid–air interface
drops due to the violation of the TIR condition when the refractive
index of the contact medium changes drastically from that of air
(na=1) to one of those of water or ice (nw or i>1.3). Whether the contact medium is liquid or ice, this was
defined by the change in the light scattering, observed by a CCD
camera (Guppy F-036 Allied Vision Technology with LINOS Macro-CCD
Lens 0.14 × (1:7) f4) placed close to the detection path (Fig. 1). After immersion freezing, there was a rapid increase in the
signal and then a slow decrease. The maximum after the rapid increase
was defined as the “transient ice” data point. After reaching
a maximum, the signal decreased with time until it stabilized
after a certain time. This stabilized signal was defined as the
“stable ice” data point.
The temperature-controlled environmental chamber configuration and
humidification and temperature sensors. The red dots are the position of the
sample temperature sensors. There is a temperature sensor inside the
measuring cell (not shown).
A control software was developed to adhere to a predefined
temperature profile and to measure the SHG signal and the
temperature of the substrate. Temperature profiles were repeated
several times for each sample to test reproducibility. In each run,
the sample was heated to 110 ∘C while purging with
N2 gas (99.9999 %) to evaporate any residual water; it was then cooled down to 0 ∘C at a rate of
10 ∘Cmin-1 and then down to the heterogeneous
freezing point at a rate of 1 ∘Cmin-1. This
cooling profile was the same for all runs to allow a comparison. An experiment with a mica–N2-gas interface was
carried out to ensure that the change in the refractive indices of
the sapphire prism, IMG, and mica substrate with temperature have
no significant effect on the resulting SHG signal in the range of
freezing temperatures observed in this work (Fig. S5 in the Supplement).
As mentioned above, the incident angle from air of the fundamental
beam was adjusted to 15∘ with respect to the surface normal of
the outer side of the sapphire prism. The corresponding incident
angle on the mica–air or mica–water interface was ∼63.4∘, which is higher than the critical angles of TIR for
mica–air (∼39.7∘) and mica–water (∼59.2∘) interfaces. This guaranteed a TIR condition
regardless of any changes in the Fresnel factors caused by the
change in refractive indices with changing temperature. The
advantage of using an SM polarization combination is its dependence on
only one nonvanishing nonlinear susceptibility tensor element
(χyyz), (Shen, 1989a; Zhuang et al., 1999) at any working
angle, which makes it a direct probe of the degree of order of the
molecules at the interface.
SHG intensity measured at the surfaces of mica (a) and sapphire
(b) in contact with air, ice film, and ice bulk, collected
in an SM polarization combination during the three different cooling cycles.
DF1: sample cooled down first to -38 ∘C under dry N2 gas,
followed by pumping water vapor of RH =5 % to the measuring cell. DF2:
sample cooled down under flow of water vapor of RH =5 %. DF3: sample
cooled down under flow of water vapor of RH =10 %. The signal is
Fresnel-corrected and normalized to the air value. All connection lines
between points serve to guide the eye.
The SHG signal is mainly produced by all polarizable species within
the SHG-active region as long as the inversion symmetry is
broken. The polarizable species at the surface of interest are the
water molecules and the surface hydroxyls (OH). The contribution of water
molecules is limited by the penetration depth inside the second
medium (air, liquid water, or ice). Under the optical geometry
described above, the calculated penetration depths are about 130,
328, and 253 nm for air, liquid water bulk, and
ice bulk as contact media, respectively. Under the thermodynamical
conditions of the presented work, the thickness of the ice (or
water) layer should exceed 1 µm within 1 s after
nucleation, which is far beyond the penetration depth of the
evanescent field. Therefore, although the exact thickness cannot be
determined in this setup it does not affect the probed signal by
the two-interface problem because the second interface (ice–air or
water–air) is not within the focal volume of the pump beam and the
nonresonant signal comes exclusively from the first interface
(water–solid or ice–solid). For the ice layer thickness, the
reader is referred to the calculations of the growth velocity of
a solidification front normal to the ice surface provided in the
Supplement of the recent work of Kiselev et al. (2016). These calculations were taken from Libbrecht (2003, 2005). For calculations of growth due
to condensation, the reader is referred to the Aerosol Calculator
Program (Excel) by Paul Baron which is based on equations from
Willeke and Baron (1993), Hinds (1999), and Baron and Willeke (2001). The signal was not collected until it became stabilized,
and, therefore, it was assumed that the ice layer after deposition was
uniform at the surface covered by the laser spot.
Results and discussion
The RH of the purged gas was set to specific
values in different runs to allow for different freezing modes
on the surface of mica. Figure 3a shows the change in
normalized Fresnel factor-corrected (see Supplement for details) SHG
intensities under an SM polarization combination for deposition
freezing (DF) at -38 ∘C (black solid line and
circles), -21 ∘C (red solid line and empty squares),
and at -12.5 ∘C (green solid line and empty
triangles) labeled DF1, DF2, and DF3, respectively. In DF1,
the cell was filled with N2 gas and the sample was
cooled down to a temperature (-38 ∘C) far below the
dew point at RH =5 % (-21 ∘C). At
-38 ∘C, the cell was purged with humid air of RH
∼5 %. An ice film formed immediately on the surface,
reflected by a drop in the SHG signal compared to that of the air–mica interface. In DF2 and DF3, the cell was purged
continuously with humid air of RH =5 and 11 %,
respectively, and then cooled down until freezing and the growth of
ice were observed on the surface. Deposition freezing and the
formation of an ice film started at -21 and
-12.5 ∘C for RH =5 % (DF2) and 11 % (DF3),
respectively. The results show a drop in the SHG signal with
respect to the signal of the mica–air interface upon freezing
for the three cases DF1, DF2, and DF3. However, the relative
signal drop for DF1 differs from those of DF2 and DF3. DF2 and
DF3 were observed at temperatures equal to the dew points at the
preset RHs, indicating two-step nucleation: first condensation
and then freezing. The coincidence of the SHG signals of the
thin ice film formed in DF2 and DF3 indicates a similar degree of
order of the water on the surface in two-step deposition freezing
regardless of the onset temperature. This means that at
temperatures above -38 ∘C, the growth of ice on mica is apparently a two-step process: water first condenses and then
freezes. This confirms, at the molecular level, the two-stage
nucleation hypothesis which suggests that a nucleus forms as
a liquid cluster and then freezes (Layton and Harris, 1963; Lupi
et al., 2014). Further pumping of the gas mixture to the
measuring cell allows the growth of the ice film by
diffusion. The resulting ice bulk shows a further drop in the
signal for DF2 and DF3. This drop was not observed for DF1, thus
indicating a major difference in the spectroscopic behavior of
ice between one-step and two-step deposition freezing. To ensure
that the lack of change in the SHG signal after the growth of an
ice film to ice bulk in one-stage nucleation (DF1, Fig. 3a) is
not an artifact, DF1 and DF2 were compared using a different
system (sapphire–water interface; Fig. 3b). Figure 3b shows the
change in SHG intensity at the surface of sapphire for DF at
-38 (black solid line and circles) and at -23 ∘C
(red solid line and empty squares), labeled DF1 and DF2,
respectively. As in the case of mica, the drop in SHG intensity
after the formation of an ice film was followed by another
drop in the two-step freezing process (DF2) following further pumping of humid air of RH =5 %, this was not the case in one-step (DF1)
freezing.
Figure 4 shows three different freezing experiments at different
RHs. The gas RH was adjusted to allow liquid condensation (LC)
during cooling at temperatures higher than those of deposition
freezing. Constant pumping of humid air at RH =20, 30, and
40 % and cooling down resulted in the formation of stable
liquid films at -3.5, 3, and 7 ∘C, respectively. At all
RH values, the SHG signal drops down following the formation of
a liquid film by LC. The relative drop in the signal with respect
to the air signal is similar to that observed in the DF
experiments (Fig. 3a). Comparing Figs. 3a and 4, the SHG signals
are in the same range regardless of the film phase (liquid or
ice). By further pumping humid air after LC, liquid bulk forms
at the surface with a signal that is lower than that of the
liquid film. This is mostly due to the contributions from the few
secondary layers of the interfacial water. Further cooling of the
sample in contact with liquid bulk causes water to freeze by
immersion freezing (IF). The observed IF temperatures for IF1,
IF2, and IF3 are similar and center around -11±1∘C, which is within the range of IF temperatures
observed for the freezing of bulk water in contact with the surface of
mica in former studies (Abdelmonem et al., 2015; Anim-Danso
et al., 2016). IF produced a transient ice phase with an SHG
intensity higher than that of the interfacial water of the bulk
liquid. The lifetime of the transient phase is around 1 min (Fig. 5). The values of the SHG intensities plotted in Fig. 4 for
transient ice bulk are the peak values found on the transient
curves shown in Fig. 5 after Fresnel factor corrections and
normalization to the mica–air signal. The transient phase may
have had peak values higher than those obtained from Fig. 5, but
they were not detected due to the fast signal decay right after
nucleation and the limited time resolution of signal detection of
about 2.5 s. A transient phase lasting for several minutes was
reported very recently by Lovering et al. (2107) using SFG at
a water–silica interface. They suggested
a transient existence of stacking-disordered (non-centrosymmetric)
ice during the freezing process at water–mineral
interfaces. Anim-Danso et al. (2016) also observed such transient ice
lasting for a few tens of seconds in SFG experiments at a high-pH
(9.8) solution–sapphire interface. They suggested that charge
transfer and the stitching bilayer are perturbed at high pH, which
leads to a decrease in SFG intensity. The present work shows that
the transient ice occurs at neutral pH on a mica surface and has
a significant nonresonant component, which is observable with the
simple SHG technique. The lifetime of the transient
phase apparently depends on the substrate and probably on the liquid-bulk
size and might play a role in the ice nucleation ability of the
surface. However, this requires a comparative study involving
different substrates, which will be subject of future work.
SHG intensity measured at the surface of mica in contact with air,
liquid film, liquid bulk, transient ice bulk, and stable ice bulk, collected in an SM polarization combination during three different
cooling cycles. LC1 IF1: sample cooled down under flow of water vapor of
RH =20 %. LC2 IF2: sample cooled down under flow of water vapor of
RH =30 %. LC3 IF3: sample cooled down under flow of water vapor of
RH =40 % (see text for details). The signal is Fresnel-corrected and
normalized to the air value. Data before Fresnel correction are shown in
Fig. S4 in the Supplement. All connection lines between points serve to guide the eye.
Typical variation in the SHG intensity around the immersion freezing
points for IF1, IF2, and IF3 during the cooling experiments.
Finally, an explanation is warranted as to why there is a drop, rather
than an increase, in the signal upon adsorption of water (or ice)
on the surface (of either mica or sapphire). The SHG response
reflects the overall arrangements of the polar entities at the
interface between two isotropic media (Fordyce et al., 2001; Goh
et al., 1988; Luca et al., 1995). This signal intensity is
expected to increase when a single (or a few non-centrosymmetric)
layer(s) of water or ice is (are) formed at (or added to) the
surface, while SHG intensity decreases upon deposition freezing,
condensation, and the growth of liquid layers by diffusion as can be
seen in Figs. 3 and 4. One expectation could be phase interference
between two signals originating from two different interfacial
entities of opposite dipole moments. The possible contributors
could be surface hydroxyls, pre-adsorbed water (films), and the
added water. The surface hydroxyls can probably be ruled out,
since mica-basal planes based on structural considerations should
not have such groups, while sapphire-basal planes do. Thus
pre-adsorbed water would be the more reasonable option. Such water films have actually been proposed
on sapphire basal planes based on experimental data (Lützenkirchen et al., 2010) and reported
in MD simulations (Argyris et al., 2011). This scenario, which may
explain the decrease in SHG intensity upon deposition freezing,
condensation, and the growth of liquid layers, involves the
preexistence of a 2-D bilayer of water on the surface. This 2-D
bilayer has the ordered structure of the water monolayer, which
greatly enhances the numbers of H bonds inside the
monolayer. This in turn reduces the likelihood of H-bond formation
between the water molecules of this monolayer with other
molecules, a phenomenon which makes a hydrophilic surface
hydrophobic (Wang et al., 2009; Hu and Michaelides, 2007; Ranea
et al., 2009). Adjacent layers would then be comparable to those
on hydrophobic surfaces and could yield the signal interference
that is observed. Clearly, the drop in signal merits further
investigation, in particular since pre-adsorbed water films are
not necessarily identical for mica and sapphire. In particular the
mica surface exhibits a quite complex interfacial hydration
structure (Cheng et al., 2001).
Conclusions
In summary, a simple SHG setup was used to discriminate and
describe three different freezing paths on the surface of
mica. The results provide decisive evidence for, and confirm, at the molecular level, the previous speculations about the existence of two-stage
deposition ice nucleation at temperatures above
-38 ∘C. One-step and two-step deposition freezing of
a thin film of ice show a water structure similar to that of a thin
film of liquid water. When liquid bulk freezes at the surface
by immersion freezing, there is a transient non-centrosymmetric
ice phase of a highly nonresonant SHG signal. The lifetime of the
transient phase is suggested to be substrate-dependent and
expected to affect ice nucleation efficiency. The presented
results open up new horizons for the role of aerosol surfaces in
promoting and stabilizing heterogeneous ice nucleation. They
provide novel molecular-level insight into different ice
nucleation regimes using a simple spectroscopic
technique. Investigating the structuring of water molecules upon
freezing next to solid surfaces is crucial to many scientific
areas, such as atmospheric physics and chemistry, hydrology, and
environmental and industrial applications.
This work demonstrates the value of investigating different
ice nucleation processes and water structuring upon freezing at the molecular-level using SHG spectroscopy. The paper is intended to serve as the basis of future complementary studies
involving other surfaces and other techniques to precisely
investigate the layer thickness, the surface morphology effect,
the cooling rates, etc. The difficulty of characterizing monolayer
films arose from the use of the IMG, which required reducing the
power. An alternative would be to approach the surface from the
air side which then has the disadvantage of a weak SHG signal and the two-interface problem. These are challenges which will be tackled
in future work.
The data used in this study are available from the
corresponding author upon request
(ahmed.abdelmonem@kit.edu).
Information about the Supplement
The Supplement comprises SHG and SFG theoretical background and data
acquisition and Fresnel factors correction.
The Supplement related to this article is available online at https://doi.org/10.5194/acp-17-10733-2017-supplement.
The author declares that he has no conflict of interest.
Acknowledgements
The work is funded by the German Research Foundation (DFG, AB
604/1-1). The SHG setup was funded by the competence area “Earth
and Environment” of KIT (start-up budget 2012). The author is
grateful to Johannes Lützenkirchen and Alexei Kiselev for the scientific
discussions on surface properties and ice growth, respectively, and
to Thomas Leisner, Claudia Linke, and Maike Schröder from KIT for their
support. The editor thanks Hugo Christenson and two anonymous reviewers
for their assistance in evaluating this paper.The article processing charges for this open-access publication were covered by a Research Centre of the Helmholtz Association.
Edited by: Daniel Knopf Reviewed by: Hugo Christenson and two
anonymous referees
ReferencesAbdelmonem, A., Lützenkirchen, J., and Leisner, T.: Probing
ice-nucleation processes on the molecular level using second harmonic
generation spectroscopy, Atmos. Meas. Tech., 8, 3519–3526,
10.5194/amt-8-3519-2015, 2015.Abdelmonem, A., Backus, E. H. G., Hoffmann, N., Sánchez, M. A., Cyran, J.
D., Kiselev, A., and Bonn, M.: Surface-charge-induced orientation of
interfacial water suppresses heterogeneous ice nucleation on α-alumina
(0001), Atmos. Chem. Phys., 17, 7827–7837, 10.5194/acp-17-7827-2017,
2017.Anim-Danso, E., Zhang, Y., and Dhinojwala, A.: Surface charge affects the
structure of interfacial ice, J. Phys. Chem. C, 120, 3741–3748,
10.1021/acs.jpcc.5b08371, 2016.Argyris, D., Ho, T., Cole, D. R., and Striolo, A.: Molecular dynamics studies
of interfacial water at the alumina surface, J. Phys. Chem. C, 115,
2038–2046, 10.1021/jp109244c, 2011.Balmer, T. E., Christenson, H. K., Spencer, N. D., and Heuberger, M.: The
effect of surface ions on water adsorption to mica, Langmuir, 24, 1566–1569,
10.1021/la702391m, 2008.
Baron, P. A. and Willeke, K.: Aerosol Measurement: Principles, Techniques,
and Applications, 2nd Edn., edited by: Sons, J., Wiley-Interscience, New
York, 2001.
Beaglehole, D., Radlinska, E. Z., Ninham, B. W., and Christenson, H. K.:
Inadequacy of Lifshitz theory for thin liquid films, Phys. Rev. Lett., 66,
2084–2087, 1991.Campbell, J. M., Meldrum, F. C., and Christenson, H. K.: Characterization of
preferred crystal nucleation sites on mica surfaces, Cryst. Growth Des., 13,
1915–1925, 10.1021/cg301715n, 2013.Campbell, J. M., Meldrum, F. C., and Christenson, H. K.: Observing the
formation of ice and organic crystals in active sites, P. Natl. Acad. Sci.
USA, 114, 810–815, 10.1073/pnas.1617717114, 2017.Cheng, L., Fenter, P., Nagy, K. L., Schlegel, M. L., and Sturchio, N. C.:
Molecular-scale density oscillations in water adjacent to a mica surface,
Phys. Rev. Lett., 87, 156103, 10.1103/PhysRevLett.87.156103, 2001.Christenson, H. K. and Israelachvili, J. N.: Growth of ionic crystallites on
exposed surfaces, J. Colloid. Interf. Sci., 117, 576–577,
10.1016/0021-9797(87)90420-6, 1987.Christenson, H. K. and Thomson, N. H.: The nature of the air-cleaved mica
surface, Surf. Sci. Rep., 71, 367–390, 10.1016/j.surfrep.2016.03.001,
2016.DeMott, P. J., Möhler, O., Stetzer, O., Vali, G., Levin, Z.,
Petters, M. D., Murakami, M., Leisner, T., Bundke, U., Klein, H.,
Kanji, Z. A., Cotton, R., Jones, H., Benz, S., Brinkmann, M., Rzesanke, D.,
Saathoff, H., Nicolet, M., Saito, A., Nillius, B., Bingemer, H., Abbatt, J.,
Ardon, K., Ganor, E., Georgakopoulos, D. G., and Saunders, C.: Resurgence in
ice nuclei measurement research, B. Am. Meteorol. Soc., 92, 1623–1635,
10.1175/2011bams3119.1, 2011.Eastwood, M. L., Cremel, S., Gehrke, C., Girard, E., and Bertram, A. K.: Ice
nucleation on mineral dust particles: onset conditions, nucleation rates and
contact angles, J. Geophys. Res.-Atmos., 113, 203-201–203-209,
10.1029/2008jd010639, 2008.Fordyce, A. J., Bullock, W. J., Timson, A. J., Haslam, S.,
Spencer-Smith, R. D., Alexander, A., and Frey, J. G.: The temperature
dependence of surface second-harmonic generation from the air-water
interface, Mol. Phys., 99, 677–687, 10.1080/00268970010030022, 2001.Goh, M. C., Hicks, J. M., Kemnitz, K., Pinto, G. R., Heinz, T. F.,
Eisenthal, K. B., and Bhattacharyya, K.: Absolute orientation of water
molecules at the neat water surface, J. Phys. Chem., 92, 5074–5075,
10.1021/j100329a003, 1988.
Hinds, W. C.: Aerosol Technology: Properties, Behavior, and Measurement of
Airborne Particles, 2nd Edn., Wiley-Interscience, New York, 1999.Hoose, C. and Möhler, O.: Heterogeneous ice nucleation on atmospheric
aerosols: a review of results from laboratory experiments, Atmos. Chem.
Phys., 12, 9817–9854, 10.5194/acp-12-9817-2012, 2012.Hu, J., Xiao, X.-D., Ogletree, D. F., and Salmeron, M.: Imaging the
condensation and evaporation of molecularly thin films of water with
nanometer resolution, Science, 268, 267–269,
10.1126/science.268.5208.267, 1995.Hu, X. L. and Michaelides, A.: Ice formation on kaolinite: lattice match or
amphoterism?, Surf. Sci., 601, 5378–5381, 10.1016/j.susc.2007.09.012,
2007.Jang, J. H., Lydiatt, F., Lindsay, R., and Baldelli, S.: Quantitative
orientation analysis by sum frequency generation in the presence of
near-resonant background signal: acetonitrile on rutile TiO2 (110), J.
Phys. Chem. A, 117, 6288–6302, 10.1021/jp401019p, 2013.Kiselev, A., Bachmann, F., Pedevilla, P., Cox, S. J., Michaelides, A.,
Gerthsen, D., and Leisner, T.: Active sites in heterogeneous ice
nucleation – the example of K-rich feldspars, Science, 355, 367–371,
10.1126/science.aai8034, 2016.Layton, R. G. and Harris, F. S.: Nucleation of ice on mica, J. Atmos. Sci.,
20, 142–148, 10.1175/1520-0469(1963)020<0142:noiom>2.0.co;2, 1963.Libbrecht, K.: Growth rates of the principal facets of ice between
-10 ∘C and -40 ∘C, J. Cryst. Growth, 247, 530–540,
10.1016/S0022-0248(02)01996-6, 2003.Libbrecht, K. G.: The physics of snow crystals, Rep. Prog. Phys., 68, 855,
10.1088/0034-4885/68/4/R03, 2005.Lovering, K. A., Bertram, A. K., and Chou, K. C.: Transient phase of ice
observed by sum frequency generation at the water/mineral interface during
freezing, J. Phys. Chem. Lett., 8, 871–875,
10.1021/acs.jpclett.6b02920, 2017.Luca, A. A. T., Hebert, P., Brevet, P. F., and Girault, H. H.: Surface
second-harmonic generation at air/solvent and solvent/solvent interfaces, J.
Chem. Soc. Faraday T., 91, 1763–1768, 10.1039/ft9959101763, 1995.Lupi, L., Kastelowitz, N., and Molinero, V.: Vapor deposition of water on
graphitic surfaces: formation of amorphous ice, bilayer ice, ice I, and
liquid water, J. Chem. Phys., 141, 18C508, 10.1063/1.4895543, 2014.Lützenkirchen, J., Zimmermann, R., Preočanin, T., Filby, A.,
Kupcik, T., Küttner, D., Abdelmonem, A., Schild, D., Rabung, T.,
Plaschke, M., Brandenstein, F., Werner, C., and Geckeis, H.: An attempt to
explain bimodal behaviour of the sapphire c-plane electrolyte interface, Adv.
Colloid Interfac., 157, 61–74, 10.1016/j.cis.2010.03.003, 2010.Mason, B. J. and Maybank, J.: Ice-nucleating properties of some natural
mineral dusts, Q. J. Roy. Meteor. Soc., 84, 235–241,
10.1002/qj.49708436104, 1958.Ong, S., Zhao, X., and Eisenthal, K. B.: Polarization of water molecules at
a charged interface: second harmonic studies of the silica/water interface,
Chem. Phys. Lett., 191, 327–335, 10.1016/0009-2614(92)85309-X, 1992.Poppa, H. and Elliot, A. G.: The surface composition of Mica substrates,
Surf. Sci., 24, 149–163, 10.1016/0039-6028(71)90225-1, 1971.
Pruppacher, H. R. and Klett, J. D.: Microphysics of Clouds and Precipitation,
2nd edn., Atmospheric and Oceanographic Sciences Library, 18, Kluwer Academic
Publishers, Dordrecht, Boston, 954 pp., 1997.Ranea, V. A., Carmichael, I., and Schneider, W. F.: DFT investigation of
intermediate steps in the hydrolysis of α-Al2O3(0001), J. Phys.
Chem. C, 113, 2149–2158, 10.1021/jp8069892, 2009.Rao, Y., Tao, Y.-S., and Wang, H.-F.: Quantitative analysis of orientational
order in the molecular monolayer by surface second harmonic generation, J.
Chem. Phys., 119, 5226–5236, 10.1063/1.1597195, 2003.Schaefer, V. J.: The formation of ice crystals in the laboratory and the
atmosphere, Chem. Rev., 44, 291–320, 10.1021/cr60138a004, 1949.Shen, Y. R.: Optical second harmonic generation at interfaces, Annu. Rev.
Phys. Chem., 40, 327–350, 10.1146/annurev.pc.40.100189.001551, 1989a.Shen, Y. R.: Surface properties probed by second-harmonic and sum-frequency
generation, Nature, 337, 519–525, 10.1038/337519a0, 1989b.Slater, B., Michaelides, A., Salzmann, C. G., and Lohmann, U.: A blue-sky
approach to understanding cloud formation, B. Am. Meteorol. Soc., 97,
1797–1802, 10.1175/bams-d-15-00131.1, 2016.Wang, C., Lu, H., Wang, Z., Xiu, P., Zhou, B., Zuo, G., Wan, R., Hu, J., and
Fang, H.: Stable liquid water droplet on a water monolayer formed at room
temperature on ionic model substrates, Phys. Rev. Lett., 103, 137801,
10.1103/PhysRevLett.103.137801, 2009.Willeke, K. and Baron, P.: Aerosol Measurement: Principles, Techniques, and
Applications, edited by: Klaus Willeke, P. A. B., Van Nostrand Reinhold, New
York, 1993.
Zhao, X., Ong, S., and Eisenthal, K. B.: Polarization of water molecules at
a charged interface. Second harmonic studies of charged monolayers at the
air/water interface, Chem. Phys. Lett., 202, 513–520,
10.1016/0009-2614(93)90041-X, 1993.Zhao, H., Bhattacharjee, S., Chow, R., Wallace, D., Masliyah, J. H., and Xu,
Z.: Probing Surface Charge Potentials of Clay Basal Planes and Edges by
Direct Force Measurements, Langmuir, 24, 12899–12910,
10.1021/la802112h, 2008.Zhuang, X., Miranda, P. B., Kim, D., and Shen, Y. R.: Mapping molecular
orientation and conformation at interfaces by surface nonlinear optics, Phys.
Rev. B, 59, 12632–12640, 10.1103/PhysRevB.59.12632, 1999.