ArticlePDF Available

Abstract and Figures

The sorption of vapor molecules onto pre-existing nanometer sized clusters is of importance in understanding particle formation and growth in gas phase environments and devising gas phase separation schemes. Here, we apply a differential mobility analyzer-mass spectrometer based approach to observe directly the sorption of vapor molecules onto iodide cluster ions of the form (MI)xM+ (x = 1-13, M = Na, K, Rb, or Cs) in air at 300 K and with water saturation ratios in the 0.01-0.64 range. The extent of vapor sorption is quantified in measurements by the shift in collision cross section (CCS) for each ion. We find that CCS measurements are sensitive enough to detect the transient binding of several vapor molecules to clusters, which shift CCSs by only several percent. At the same time, for the highest saturation ratios examined, we observed CCS shifts of up to 45%. For x < 4, cesium, rubidium, and potassium iodide cluster ions are found to uptake water to a similar extent, while sodium iodide clusters uptake less water. For x ≥ 4, sodium iodide cluster ions uptake proportionally more water vapor than rubidium and potassium iodide cluster ions, while cesium iodide ions exhibit less uptake. Measured CCS shifts are compared to predictions based upon a Kelvin-Thomson-Raoult (KTR) model as well as a Langmuir adsorption model. We find that the Langmuir adsorption model can be fit well to measurements. Meanwhile, KTR predictions deviate from measurements, which suggests that the earliest stages of vapor uptake by nanometer scale species are not well described by the KTR model.
Content may be subject to copyright.
Analysis of heterogeneous water vapor uptake by metal iodide cluster ions via
differential mobility analysis-mass spectrometry
Derek Oberreit, Vivek K. Rawat, Carlos Larriba-Andaluz, Hui Ouyang, Peter H. McMurry, and Christopher J.
Hogan Jr.
Citation: The Journal of Chemical Physics 143, 104204 (2015); doi: 10.1063/1.4930278
View online:
View Table of Contents:
Published by the AIP Publishing
Articles you may be interested in
Element specificity of ortho-positronium annihilation for alkali-metal loaded SiO2 glasses
J. Chem. Phys. 142, 094307 (2015); 10.1063/1.4913976
Superconductivity and spectroscopy of homo- and heterofullerides of alkali metals and thallium
Low Temp. Phys. 37, 245 (2011); 10.1063/1.3580512
Structure prediction of high-pressure phases for alkali metal sulfides
J. Chem. Phys. 121, 2289 (2004); 10.1063/1.1766013
Escape length of ultraviolet induced photoelectrons in alkali iodide and CsBr evaporated films: Measurements
and modeling
J. Appl. Phys. 84, 2890 (1998); 10.1063/1.368395
Ultraviolet photoabsorption measurements in alkali iodide and caesium bromide evaporated films
J. Appl. Phys. 83, 7896 (1998); 10.1063/1.367967
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded to IP: On: Mon, 14 Sep 2015 15:33:42
Analysis of heterogeneous water vapor uptake by metal iodide cluster
ions via differential mobility analysis-mass spectrometry
Derek Oberreit,1,2 Vivek K. Rawat,1Carlos Larriba-Andaluz,1,a) Hui Ouyang,1
Peter H. McMurry,1and Christopher J. Hogan, Jr.1,b)
1Department of Mechanical Engineering, University of Minnesota, Minneapolis, Minnesota 55455, USA
2Fluid Measurement Technologies, Inc., Saint Paul, Minnesota 55110, USA
(Received 28 June 2015; accepted 26 August 2015; published online 14 September 2015)
The sorption of vapor molecules onto pre-existing nanometer sized clusters is of importance in
understanding particle formation and growth in gas phase environments and devising gas phase sepa-
ration schemes. Here, we apply a dierential mobility analyzer-mass spectrometer based approach to
observe directly the sorption of vapor molecules onto iodide cluster ions of the form (MI)xM+(x =1-
13, M =Na, K, Rb, or Cs) in air at 300 K and with water saturation ratios in the 0.01-0.64 range. The
extent of vapor sorption is quantified in measurements by the shift in collision cross section (CCS)
for each ion. We find that CCS measurements are sensitive enough to detect the transient binding
of several vapor molecules to clusters, which shift CCSs by only several percent. At the same time,
for the highest saturation ratios examined, we observed CCS shifts of up to 45%. For x <4, cesium,
rubidium, and potassium iodide cluster ions are found to uptake water to a similar extent, while
sodium iodide clusters uptake less water. For x 4, sodium iodide cluster ions uptake proportionally
more water vapor than rubidium and potassium iodide cluster ions, while cesium iodide ions exhibit
less uptake. Measured CCS shifts are compared to predictions based upon a Kelvin-Thomson-Raoult
(KTR) model as well as a Langmuir adsorption model. We find that the Langmuir adsorption model
can be fit well to measurements. Meanwhile, KTR predictions deviate from measurements, which
suggests that the earliest stages of vapor uptake by nanometer scale species are not well described by
the KTR model. C2015 AIP Publishing LLC. []
Vapor molecule sorption (heterogeneous uptake, defined
as the sorption of one species onto a chemically distinct spe-
cies) onto nanometer scale ions is of interest for several rea-
sons. In many gas phase environments, heterogeneous uptake
can control the rates of formation and growth of condensed
phase entities (e.g., molecular clusters and aerosol parti-
cles).16Measurement systems can be developed in which
heterogeneous uptake alters the size and structure of chemi-
cally distinct ions to varying degrees; this enables instruments
which separate ions based upon structure (e.g., low field and
high field ion mobility spectrometries) to discriminate be-
tween ions which are similar in structure in the absence of
vapor dopants but exhibit varying degrees of uptake.712 In
modeling heterogeneous uptake, it is commonplace to apply
classical models, namely, the Kelvin-Thomson model1315
and the Köhler model.1618 These models enable prediction
of equilibrium sorption coecients, which are ratios of the
number concentrations of ions with gvapor molecules sorbed
to the number concentrations with g1 sorbed at equilib-
rium and which govern the extent of uptake in controlled
vapor concentration environments. While classical calcula-
tions agree qualitatively with experimental measurements of
a)Current address: Department of Mechanical Engineering, Indiana
University-Purdue University, Indianapolis, Indiana 46202-5132, USA.
b)Author to whom correspondence should be addressed. Electronic mail: Tel.: 1-612-626-8312. FAX: 1-612-625-6069.
sorption in several instances,19 as well as with measurements
of condensed phase entity growth,14,20,21 there are a series
of experimental observations of vapor molecule uptake that
are not explained by these models, such as the influence of
ion chemical composition and polarity on the extent of up-
take14,22 and unanticipated uptake rate functional dependencies
on temperature.23,24 Further, there are quantitative dierences
in classically predicted and measured equilibrium sorption
coecients.25 As an alternative, computational approaches
can now be used to theoretically study sorption2631 without
invoking classical assumptions (i.e., that the sorbed species
has identical properties to the bulk condensed phase); however,
experiments remain necessary to better test both classical and
computational predictions.
Experiments to-date have not clearly established the link
between structure/size shifts and equilibrium sorption coef-
ficients for ions in the nanometer size range. Heterogeneous
uptake has been examined with tandem mobility analysis3235
as well as with electrodynamic balances36 and optical trapp-
ing.37,38 These methods are usually limited to ions in the >5 nm
size range (with supermicrometer particles needed for electro-
dynamic balances and optical trapping) and are further insen-
sitive to the addition or loss of a single vapor molecule from
the surface of an ion. Conversely, single vapor molecule sorp-
tion events are detectable in high pressure mass spectrometry
(HPMS) systems,19,25 yet HPMS is limited to vapor concen-
trations well below saturation, thereby limiting the number
of attached vapor molecules that can be measured.19 Field
0021-9606/2015/143(10)/104204/11/$30.00 143, 104204-1 ©2015 AIP Publishing LLC
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded to IP: On: Mon, 14 Sep 2015 15:33:42
104204-2 Oberreit et al. J. Chem. Phys. 143, 104204 (2015)
asymmetric waveform ion mobility spectrometry-mass spec-
trometry (FAIMS-MS) systems often exploit dierential
amounts of heterogeneous uptake by ions to distinguish iso-
mers from one another.7,8As operated, existing FAIMS-MS
technology only provides separation capability; it does not
provide quantitative information on the extent of heteroge-
neous uptake. Heterogeneous uptake may also be studied in
expansion chamber based “nucleation probability” experi-
ments;14,15,39 though enabling very precise measurement of
the vapor saturation ratio, such experiments only facilitate
observation of ions after they are grown into micrometer sized
droplets, making it dicult to link measurements to the early
stages of uptake. Finally, infrared spectroscopy can be used
to quantitatively explore the structures of sorbed vapor-ion
complexes.4044 With this technique as well, there are limita-
tions to what can be probed in terms of the sizes and chemical
complexity of the ions.
There exists a “window” in the sub 2 nm size range
with vapor saturation ratios in the 0.10–1.0 range, in which
vapor sorption onto ions is of interest as both a naturally
occurring phenomenon and in the design of gas phase sepa-
ration schemes, but wherein quantification of the extent of
sorption has been dicult. In this work, we develop a method
to link the extent of vapor molecule sorption by chemically
identified ions to changes in ion size and structure. This
method involves measurements with a low field Dierential
Mobility Analyzer-Mass Spectrometer (DMA-MS a form of
ion mobility spectrometry-mass spectrometry), wherein the
DMA is used to separate and select ions based on their collision
cross sections (CCSs). The DMA-MS method is applicable
to ions in the 1 nm size range and can be used to examine
vapor molecule uptake at any vapor molecule concentration
up to saturation. In the sections titled “Experimental Methods”
and “Results and Discussion,” the DMA-MS measurement
method is described and an analysis approach linking the shift
in CCSs inferred from measurements to equilibrium sorption
coecients for successive vapor molecules is provided. The
method is applied to measurements of water vapor molecule
uptake by positively charged metal salt cluster ions of the
form (MI)xM+, where M =Cs, Rb, K, or Na and x=1-13.
These ions are chosen mainly for ease of formation in the gas
phase and the past precedent in study ions of this type.4548
The observed extents of heterogeneous uptake are compared
to modified classical theory predictions as well as to Langmuir
adsorption isotherm based models as a demonstration of how
measurements can be used to test predictions of equilibrium
sorption coecients.
Differential mobility analysis-mass spectrometry
A schematic of the DMA-MS system as operated in this
study is shown in Figure 1. DMA model P5 (SEADM, Boe-
cillo, Spain) was interfaced with a QSTAR XL mass spec-
trometer (Applied Biosystems); the DMA was operated as
described previously.46,49,50 Positively charged cluster ions (the
test ions for this study) of sodium, potassium, rubidium, and
cesium iodide were produced via positive mode electrospray
FIG. 1. A schematic diagram of the dierential mobility analyzer-mass
spectrometer (DMA-MS) system used to examine vapor molecule sorption
by electrospray generated cluster ions.
ionization (ESI) of 10 mM salt solutions in high performance
liquid chromatography grade methanol and were directed into
the DMA electrostatically against a counterflow of air. Ultra-
high purity (UHP) air (Airgas) was used for the DMA sheath
flow (and correspondingly for the counterflow). Distinct from
prior studies utilizing DMA-MS measurements, to humid-
ify the sheath flow, a custom-made nebulizer was used to
introduce controlled amounts of water vapor into the sheath
flow. Details of the nebulizer design and a schematic of it
are provided in Oberreit et al.33 A chilled-mirror dewpoint
hygrometer (General Eastern, Hygro M4) was attached to the
ESI chamber and was used to determine the total water content
of the sheath and counterflow air. The sheath flow temperature
was controlled at 299-300 K using a fan cooled heat exchanger
attached to the sheath flow recirculation tubing. The combined
water vapor content and temperature control system facilitated
mobility measurements in the water saturation ratio (S)range
of 0.01–0.64.
The DMA was stepped in 10 V increments from 900
to 3600 V applied across its electrodes, with the ESI source
voltage floating above the upper electrode. Mass spectra were
collected using the time-of-flight tube of the QSTAR XL sys-
tem at each applied potential dierence in the DMA. The DMA
was calibrated through measurement of a known-mobility
standard ion;51 because the DMA is a linear-mobility spec-
trometer, the ratio V*ZS, where Vis the voltage of maximum
transmission and ZSis the ion’s mobility measured at satura-
tion ratio S, is a constant. The ion selected for calibration in
prior work has been the tetraheptylammonium+ion. However,
at the higher saturation ratios examined, we found that the
mobility of the tetraheptylammonium+ion shifted noticeably
(but only by several percent), which was indicated by an
increase in the DMA voltage required to transmit the ion.
Also at higher saturation ratios, the mass spectrometer detected
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded to IP: On: Mon, 14 Sep 2015 15:33:42
104204-3 Oberreit et al. J. Chem. Phys. 143, 104204 (2015)
ions not only at the expected m/z (410 Da) but also at the
m/zcorresponding to tetraheptylammonium+-H2O (428 Da).
Heterogeneous uptake of water molecules may hence shift
the mobility of tetraheptylammonium+ions, rendering them
unsuitable for instrument calibration under humidified condi-
tions. We instead used the tetradodocylammonium+ion for
calibration; for this ion, a shift in the DMA voltage required
for maximal transmission was not observed and water adduct
ions were never observed. The inverse mobility (1/ZS)of the
tetradodecylammonium+ion at 293 K and near 101 kPa was
measured to be 1.401 V s cm2by Ude and Fernandez de la
Mora.51 For an ion of this inverse mobility, the influence of
gas molecule polarization (the ion induced dipole potential
between gas molecules and the cluster ion) is expected to
be minimal.46,5255 The published mobility was thus adjusted
to the measurement temperature by multiplying by the fac-
tor (293 K/T)1/2, which is based on the assumption that
tetradodecylammonium+undergoes hard sphere interactions
with the background gas molecules.
Observations of heterogeneous uptake
As discussed by Ouyang et al.,46 positive mode ESI of
iodide salt solutions leads to the formation of ions of the type
(MI)x(M+)z, where M =Na, K, I, or Cs. Here, we elect to
focus on water uptake by selected ions with x =1-13, and
z=1 (the x =0 ions were also detected, but appear to bind
transiently to low concentration gas phase impurities, compli-
cating assessment of heterogeneous uptake). Uptake of water
vapor by cluster ions leads to not only a shift in ion mobility
but also a shift in mass, and ideally, it would be possible to
assess the extent of uptake by examining changes to mass
distributions. Although not shown, in our experiments, for the
smallest ions examined (x =1-2), we commonly observed ions
shifted in mass by +18 Da units, which is indeed indicative
of water sorption. However, upon being transmitted through
the DMA, ions enter a high pressure drop, high electric field
system, in which they can either (1) undergo high energy gas
molecule collisions, leading to ion heating and dissociation
of sorbed vapor molecules and loss of a cation-anion pairs,56
or (2) travel along trajectories wherein the vapor saturation
ratio increases, leading to additional condensation of water
onto ions.57 Subtle changes to inlet operating conditions have
been shown to induce appreciable amounts of either water
condensation or evaporation from hydrated ions,58 and for this
reason, the mass distributions of water molecules bound are
not reliable measures of the extent of heterogeneous uptake at
the prescribed vapor saturation ratio.
Conversely, mobilities were measured in a temperature
and pressure controlled region, with ions migrating at low
speed relative to their mean thermal speeds. Therefore, mobil-
ity shifts can be directly correlated with the extent of heteroge-
neous uptake occurring under well-defined conditions. Repre-
sentative inverse mobility spectra at 4 water saturation
ratios are shown in Figure 2for the (CsI)2Cs+,(CsI)4Cs+, and
(CsI)6Cs+ions, with inverse mobility derived from DMA cali-
bration. Plotted signal intensities (arbitrary units) are normal-
FIG. 2. Normalized inverse mobility spectra for mass selected cesium iodide
cluster ions at four discrete water saturation ratios. Shifts to larger inverse
mobilities are indicative of water vapor uptake by cluster ions.
ized for each ion; higher signal intensities were observed
for less massive ions, and with increasing S, we observed a
decrease in absolute signal intensity, particularly for the most
massive ions examined. We attribute the decreasing signal
intensity at large saturation ratios to water condensation onto
ions in the mass spectrometer inlet (after DMA measurement),
growing ions to masses larger than can be detected/transmitted
in the mass spectrometer. For all detected ions, a shift to
larger inverse mobilities is observed with increasing water
saturation ratio. To further analyze measurements, we convert
each measured inverse mobility to a mean ion-neutral CCS
(denoted as Sat saturation ratio S), via the Mason-Schamp
S=π µ
in which zis the ion charge state, eis the unit electron charge, k
is Boltzmann’s constant, Tis the gas temperature, µis the gas
molecule-ion reduced mass (approximated in the absence of
sorbed vapor), and ρgas is the gas mass density. The Mason-
Schamp equation is itself an approximation, which is only
applicable in suciently low electric field strength systems60
as well as in instances where the ion size relative to the gas
molecule mean free path (near 66 nm here) is small.61,62 While
these criteria are met in the presented experiments, even in
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded to IP: On: Mon, 14 Sep 2015 15:33:42
104204-4 Oberreit et al. J. Chem. Phys. 143, 104204 (2015)
the low-field limit, the CCS is a complex parameter that is
dependent not only the physical cross section of the ion but also
the measurement temperature and the interactions between gas
molecules and ions during impingement. A change in any of
these parameters will undoubtedly lead to a change in the
CCS. However, the focus of this work is to understand the
changes in CCSs due to water sorption onto ions under constant
temperature and pressure conditions; under these conditions,
the CCS is primarily a measure of ion size and structure and
heterogeneous uptake will typically lead to an increase in CCS.
For all measurements, the ratio S/0, i.e., the ratio of the
CCS measured at saturation ratio Sto the dry condition CCS,
is provided in Table S1 of the supplementary material,63 and
for ions of specific nvalues, the value of S/01 is plotted
as a function of Sin Figure 3. Several features are apparent
in these plots. First, under nearly all circumstances, we find
that S/01 versus Scurves are slightly concave down-
ward, yet CCSs appear to increase continuously with increas-
ing saturation ratio. This behavior is similar to that observed by
Rawat et al.64 for isopropanol uptake by peptide ions, though
they observed that uptake ceased beyond a critical saturation
ratio. We compare and contrast the shapes of the S/01
versus Scurves with model calculations in the sections titled
“Modeling heterogeneous uptake” and “Measurement-model
comparison.” Second, for lower values of x (1-2), we find
that ions composed of potassium, rubidium, and cesium iodide
exhibit similar uptake behavior, with noticeably less uptake
by sodium iodide cluster ions. Conversely for x 3, we find
that sodium iodide cluster ions exhibit the largest extents of
uptake, followed by potassium iodide, rubidium iodide, and
finally cesium iodide cluster ions. Third, we find that the extent
of uptake does not correlate strongly (positively or negatively)
with x for potassium and rubidium iodide clusters, while the
extent of uptake appears to increase and decrease with x, for
sodium iodide and cesium iodide cluster ions, respectively.
However, for no ion do we observe a monotonic increase or
decrease with x in the extent of uptake (which is examined in
further detail subsequently). In total, we find that the extent of
uptake depends on both cluster size and chemical composition,
and hence ion structure.
Modeling heterogeneous uptake
The goal of this study is not only to demonstrate that
DMA-MS measurements can be used to probe heterogeneous
uptake at modest saturation ratios but also to show that the
FIG. 3. The measured value of
S/01 as a function of the water
saturation ratio in the dierential
mobility analyzer for mass identified
cluster ions.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded to IP: On: Mon, 14 Sep 2015 15:33:42
104204-5 Oberreit et al. J. Chem. Phys. 143, 104204 (2015)
shifts in CCS can be linked directly to equilibrium sorption
coecients. To establish this link, first, we note that the number
of water molecules bound to an ion is not a constant; rather,
water molecules continuously sorb and desorb with each ion
probing the equilibrium distribution for the number of bound
vapor molecules. We define the probability of an individual ion
having gvapor molecules bound at any given instant as Pg. We
also define eective unimolecular equilibrium sorption coe-
cients (K
eq,g , dimensionless equilibrium constants) through the
eq,g =ng
where ngand ng1are the number concentrations of ions with g
and g1 water vapor molecules bound. Noting the ergodicity
of systems in equilibrium, the probability that a single ion has
gmolecules bound to it is equivalent to the fraction of ions with
gvapor molecules bound, leading to Pgdefined as
Combining Equations (2a) and (2b),Pgcan subsequently be
expressed as
for g1 (2c)
for g=0.(2d)
Each ion traverses the DMA in a time ttot, and with a linear
electric field in a parallel-plate DMA of magnitude E, the
electrode to electrode distance (LDMA, traversed by the ions) is
equal to the product of ttot,E, and the ion’s measured mobility,
LDMA =ZSttotE=κ0t0E+κ1t1E+κ2t1E
+· · · κgtgE+· · · +κtE,(2e)
where tgdenotes the time an ion spends within the DMA with
gvapor molecules bound, and κgis the cluster ion’s mobility
specifically with g vapor molecules sorbed (in constant to ZS,
the mobility measured at saturation ratio S). With the equilib-
rium relation Pg=tg/ttot, the ratio ZS/Z0is expressed as
Using Equation (1) to link mobility and CCS (both for Zand
κ) enables Equation (2f) to be rewritten as
where gis the CCS of the cluster ion specifically with gvapor
molecules bound. Equation (2g) therefore facilitates compar-
ison between DMA-MS observed structural modifications to
ions and predictions of equilibrium sorption coecients (from
any theoretical model), provided the ratio 0/gcan be esti-
mated for all gand provided that the introduction of vapor
molecules into the DMA does not substantially alter ion mobil-
ities due to the change in gas composition (an eect which is
anticipated to be negligible for water and is addressed in the
work of Rawat et al.64).
Comparison of measurements to models of heterogeneous
uptake requires both models of the CCSs of ions with a specific
number of vapor molecules bound and methods to calculate
eq,g . The calculation of CCSs is non-trivial. Here, we leverage
recent developments in gas molecule scattering calculations
for air made by Larriba-Andaluz and coworkers52,53,6567 as
well as the measurements of Ouyang et al.46 of the CCSs of
bare (dehydrated) iodide salt cluster ions. Both calculations
and measurements suggest that the CCSs of iodide salt ions
of the form (MI)nM+can be approximated by the relationship,
where PAgis the orientationally averaged projected area of a
cluster-gas molecule complex with g vapor molecules bound, ξ
is the momentum scattering coecient, found by Ouyang et al.
to be 1.36 for NaI, 1.27 for KI, 1.23 for RbI, and 1.19 for CsI
(dependent upon the manner in which gas molecules impinge
and are reemitted from cluster structure surfaces52,53,68,69 and
assumed independent of the extent of water vapor sorption
here). Λis a factor which accounts for the increase in the CCS
due to attractive forces between the ion and polarizable gas
molecules, defined as53
Λ1+Ψpol 0.322 +1
ξ0.0625 +0.1212ΨpolΨpol <1,
Ψpol =παpolz2e2
where αpol is the gas molecule polarizability and ε0is the
permittivity of free space. With the approximation that ξis
independent of the extent of heterogeneous uptake, Equa-
tions (3a)(3c) lead to
Prediction of 0/ghence amounts to prediction of projected
area ratios and implementation of a function accounting for
the ion-induced dipole potential. In related studies of hetero-
geneous uptake, performed with a DMA coupled to an atmo-
spheric pressure drift tube ion mobility spectrometer,33 we
quantified the extent of vapor molecules sorption onto 2–7 nm
nanoclusters through a relationship similar to Equation (2g).
We approximated all nanoclusters as spheres, linking the pro-
jected areas to nanocluster diameters, and thus linking their
size and structures to their mobilities in the manner utilized
by Ku and Fernandez de la Mora70 and Larriba et al.71 While
the spherical approximation for nanoclusters composed of 102-
103cation-anion pairs is reasonable, it is not valid for clusters
with n 13.46 We instead modeled cluster ion structures using
density functional theory (DFT) and calculated the projected
areas of DFT inferred structures. Structures for clusters of the
type (MI)xM+*(H2O)g(x =1-3, g =0-30) were generated using
the Gaussian 09 software package (Gaussian, Inc., Walling-
ford, CT), as described by Ouyang et al.46 The B3LYP den-
sity functional72 was employed with the basis set LANL2DZ,
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded to IP: On: Mon, 14 Sep 2015 15:33:42
104204-6 Oberreit et al. J. Chem. Phys. 143, 104204 (2015)
which applies Los Alamos ECP (eective core potential) plus
DZ (double zeta).7375 Symmetry restrictions were not applied,
and vibration frequencies were calculated. All structures evalu-
ated had positive frequencies, indicating they are local minima
rather than transition states. Complete characterization of clus-
ter structures requires the determination of the number of local
minimum structures and their energy dierences.76 However,
except in rare circumstances where a cluster has both a linear
and compact stable isomer (only found at x <5), the projected
areas calculated (using the procedure noted subsequently) for
dierent isomers dier by only several percent. For computa-
tional simplicity, we based projected areas for implementation
in Equation (3d) on the lowest energy structure obtained.
Depictions of DFT calculated structures for clusters with
x=6 and with varying numbers of bound water molecules
are shown in Figure 4. All atoms are depicted as spheres with
relative radii proportional to the radii used in projected area
calculations: Na (blue): 1.16 Å, K(gold): 1.52 Å, Rb (purple):
1.66 Å, Cs (yellow): 1.81 Å, I (green): 2.06 Å, H (white):
1.20 Å, and O (red): 1.52 Å, which are in line with the ionic
radii for the charged species (cations and anions) and the van
der Waals radii for hydrogen and oxygen. Qualitatively, clus-
ters with a limited number of water molecules sorbed (g<6)
retain their core structures observed in the absence of water,
and the water molecules themselves appear to bind to clusters
FIG. 4. Depictions of (MI)6M+structures predicted via density functional
theory with selected numbers of water molecules (g)bound.
at specific sites. As the number of sorbed water molecules
increases, the cluster structures appear to open, leading to
water present in cluster interstices. However, even with 30
water molecules bound, total dissolution of most salt clusters
is not observed. Similar results were obtained for clusters with
higher and lower x, except that dissolution was possible for
the smallest xclusters. The projected areas of all structures
were calculated using the projected area calculator of the IMoS
software package (available freely from Dr. Carlos Larriba-
Andaluz and described in detail previously52) with an added
“probe radius” to account for the size of the impinging mole-
cules (1.5 Å for air). Values for PAgwhere a structure was not
predicted were found by linear interpolation of the calculated
PAgdata up to the largest calculated cluster gmax. Beyond gmax,
PAgwas approximated using the equation,
where vwis the volume of the condensed phase vapor mole-
cule (based upon its bulk density in the liquid phase33). A
summary of the calculated PA values is provided in Table
S2 of the supplementary material,63 and corresponding plots
of g/01 for select clusters are shown in Figure 5. For
all clusters examined, the sorption of fewer than 15 water
molecules shifts the CCS by 30% or more, and we therefore
anticipate it is 100-101(time based average) water molecules
which are sorbing to clusters at the saturation ratios examined.
However, we reiterate that as each cluster migrates through
the DMA, the number of water molecules bound is transient,
leading to Equation (2g) describing the shift in mobility.
The sorption and desorption of water molecules can
be described by models of K
eq,g . We elect to compare
two such models to experimental results. First, similar to
Oberreit et al.33 and Rawat et al.,64 based upon the classical
Kelvin-Thomson model with incorporation of Raoult’s law
(the Kelvin-Thomson-Raoult (KTR) model), K
eq,g can be
expressed as
eq,g =S
exp Eg
kT   µv, g
µv, g 11/2PAg1
×ηψD,g 1(KTR),(4a)
where Sis the saturation ratio, axis the activity coecient
of water (the sorbing vapor) over the cluster surface (with x
cation-anion pairs), µv, g is the reduced mass of the sorbing
vapor-cluster ion pair, η(ψD,g 1)is an enhancement factor
in vapor-cluster ion collision rate considering the ion-dipole
potential (significant because vapor dopants have non–
negligible dipole moments, µD), and ψD, g is the ion-dipole
energy to thermal energy ratio,
ψD,k=ze µD
4εokT PAg
Based upon the analysis of Su and Bowers,77 we calculate
η(ψD,k)with the equation,
with the approximation C=0.6 (with C=1 corresponding
to complete dipole alignment and C=0 corresponds to no
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded to IP: On: Mon, 14 Sep 2015 15:33:42
104204-7 Oberreit et al. J. Chem. Phys. 143, 104204 (2015)
FIG. 5. The values of g/01 for selected cluster ions, as predicted for model cluster structures.
dipole alignment). Egis the change in cluster ion enthalpy
associated with the sorption of a single vapor molecule (from
g1 to gmolecules sorbed), described considering the Kelvin
and Thomson influences,
where σis the eective surface tension of the sorbed liquid and
εris the dielectric constant (of water). Equation (4d) is written
approximating each cluster as a sphere, which is an assumption
invoked in derivation of the Kelvin and Thomson enthalpy
changes. Second, considering Langmuir-like adsorption, K
can be described by the equation,64
eq,g =S
g1  µv,g
µv, g 11/2PAg1
×ηψD,g 1(Langmuir),(5)
where ζxis the maximum (integer) number of water molecules
which be sorbed to a cluster. In implementing Equation (5)
with Equation (2g), it is necessary to truncate all sums and
multiplicative sums at ζx, as in the Langmuir model, no ion can
uptake more than ζxvapor molecules. In addition, unlike the
KTR model, the Langmuir model does not contain an enthalpy
term; uptake is entropically driven.
Measurement-model comparison
While we elect to use Equations (4) and (5) to compare
to experimental results, we remark that both equations are
derived under the assumption that cluster ion structure
TABLE I. A summary of the fit parameters used in Figure 6plots with the
KTR and Langmuir models.
KTR Langmuir
Cluster type x axσ(N m1)axζx
Sodium 1 0.62 0.108 0.95 5
Iodide 4 0.98 0.057 11.0 34
9 0.95 0.020 2.00 33
13 0.45 0.020 2.50 40
Potassium 1 1.00 0.090 1.20 15
Iodide 4 0.75 0.055 1.00 10
9 0.95 0.020 1.50 18
13 0.70 0.010 0.50 7
Rubidium 1 0.95 0.085 4.30 17
Iodide 4 0.80 0.040 1.50 13
9 0.50 0.010 0.85 12
13 0.60 0.015 0.50 6
Cesium 1 0.8 0.09 1.20 6
Iodide 4 0.9 0.045 4.00 10
9 0.98 0.015 2.20 10
13 0.55 0.018 1.00 4
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded to IP: On: Mon, 14 Sep 2015 15:33:42
104204-8 Oberreit et al. J. Chem. Phys. 143, 104204 (2015)
FIG. 6. Comparison of the measured S/01 ratios for selected cluster ions to predictions based upon the KTR model with the fit parameters provided in
Table I.
negligibly aects vapor sorption, which is not consistent with
experimental results. We thus perform comparison qualita-
tively; for selected clusters with the KTR model, we attempt
to fit values of σand ax, and with the Langmuir model, we
fit values of ζxand ax. For clusters with x=1,4,9, and 13,
the fit KTR and Langmuir model parameters are provided
in Table I. Focusing first on the KTR model, comparison of
measurements and model calculations is plotted in Figure 6.
FIG. 7. Comparison of the measured S/01 ratios for selected cluster ions to predictions based upon the Langmuir model with the fit parameters provided
in Table I.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded to IP: On: Mon, 14 Sep 2015 15:33:42
104204-9 Oberreit et al. J. Chem. Phys. 143, 104204 (2015)
FIG. 8. The CCS weighted “Average Sorbed Water Molecules” for selected cluster ions, as a function of saturation ratio, determined by direct comparison of
measured Svalues and calculated gvalues.
For all cluster ions, predictions can be brought to be of the
same order of magnitude as measurements using fitted sur-
face tensions of similar order to the bulk surface tension of
water (0.073 N m1) and activity coecients below unity (ex-
pected for salt cluster ions). However, in all circumstances, the
shapes of the calculated curves are qualitatively dierent from
the curves derived from experimental measurements; unlike
experimental measurements, the KTR model predicted curves
are consistently concave upward (with semi-log axes). For all
clusters, better agreement is found between with Langmuir
model predictions (Figure 7). Fitted activity coecients above
unity presumably arise because of the lack of an enthalpy
barrier to uptake in the Langmuir model; because the fitting
procedure is qualitative, we elect not to modify the Langmuir
model to include an enthalpy term. That the Langmuir model
can be fit to results is further evidence that the KTR model
insuciently describes the earliest stages of heterogeneous va-
por uptake and suggests the KTR model should not be invoked
in predicting either the extent of uptake or uptake rates in this
As a final note, the values of ζxused in fits range from 4
to 40. Additional support that the number of water molecules
bound to clusters in experiments falls within this range (but
below ζxfor all clusters) is provided in Figure 8, which is
a plot of the CCS weighted average number of sorbed water
molecules, determined by comparison between the shift in
CCSs observed in experiments and the calculated shifts in Fig-
ure 4(with linear interpolation used when observed shifts fall
between integer numbers of water molecules bound). Though
not the true average number of water molecules bound during
transit through the DMA, these values do enable us to estimate
how hydrated each particular cluster becomes. With the excep-
tion of cesium iodide, the least hydrated cluster is found to be
the (MI)2M+cluster, with the average number of water mole-
cules below 2 for most examined saturation ratios. Conversely,
for all cluster types, the (MI)9M+is the most hydrated. Figure 8
results additionally confirm that larger sodium iodide clusters
sorb water to a larger extent than potassium iodide or rubidium
iodide, and cesium iodide clusters of all sizes sorb relatively
few water molecules. We did not observe sucient hydration
for complete dissolution of any of the examined clusters.
We describe an IMS-MS based approach, with a dif-
ferential mobility analyzer-mass spectrometer to examine
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded to IP: On: Mon, 14 Sep 2015 15:33:42
104204-10 Oberreit et al. J. Chem. Phys. 143, 104204 (2015)
heterogeneous uptake of vapor molecules by cluster ions. We
also develop methods to link experimentally observed CCS
shifts to model predictions of equilibrium sorption coecients.
The measurement approach is applied in examining water
uptake by singly charged alkali metal iodide cluster ions, with
comparison made to KTR and Langmuir models, each with
two fit parameters. Based on these measurements, we make
the following remarks.
1. If cluster ions rapidly equilibrate with their surroundings
during mobility measurement, then the mobility measured
for a given ion at a given temperature, pressure, and vapor
concentration depends upon the distribution of vapor mole-
cules bound at equilibrium.
2. Precise mobility measurements enable detection of shifts
in CCSs as small as 1% of the baseline CCS. Such shifts
for smaller ions correspond to the transient binding of few
vapor molecules, and IMS-MS measurements provide a
means to probe vapor uptake on an individual vapor mole-
cule level. We note that the ability to detect such small
shifts in CCS are facilitated by exact mass identification
subsequent to mobility analysis. They are further made
possible because shifted peaks in mobility spectra need not
be resolved from bare ion peaks (i.e., since all ions probe
the equilibrium distribution of bound vapor molecules, all
ion CCSs shift by equal amounts).
3. Comparison between measurements and models requires
both a method to predict the CCSs of ions with specific
numbers of vapor molecules bound, as well as a model
for equilibrium sorption coecients. Here, CCSs were pre-
dicted using cluster ion structural models with an approx-
imation (Equation (3a)) based upon prior measurements
of iodide salt cluster, and equilibrium sorption coecients
were based on functional forms from the KTR model and
Langmuir adsorption model. For the cluster ions and wa-
ter saturation ratios examined, we find that the Langmuir
adsorption model can be fit reasonably well to measure-
ments at 300 K, while the KTR model shows significant
deviation from measurements. This finding is of relevance
in understanding not only vapor sorption at equilibrium
but also new particle formation in gas phase, and it sug-
gests that the earliest stages of vapor uptake are inade-
quately described by Kelvin based models (which include
the KTR model, as well as the Köhler model18). Deriva-
tions of the KTR and Köhler models invoke assumptions
of a bulk liquid layer on the surface of cluster as well as
dissolution of the cluster constituents within the sorbing
vapor (i.e., the sorbed vapor is a solvent and other cluster
components are a solute); these assumptions are not consis-
tent with the structures of clusters with a limited number
of vapor molecules bound. Future research eorts should
be devoted to developing improved approaches to equilib-
rium sorption coecient determination, including compu-
tational approaches,30,78 multilayer sorption models,79 as
well as heuristic models based on experimental measure-
ments.80 Conveniently, IMS-MS measurements (using a
dierential mobility analyzer, drift tube, or other linear
mobility spectrometer at a fixed, prescribed temperature)
of CCS shifts can be compared to any model for the equi-
librium sorption coecients, independent of the model’s
4. Throughout this work, we have used the terms heteroge-
neous uptake and vapor sorption to refer to the binding
of water vapor molecules to cluster ions, without making
the distinction between adsorption (binding of vapor to the
cluster ion surface) and absorption (binding of vapor in the
cluster interstitial regions). We note that measurement of
CCS shifts alone cannot be used to distinguish between
adsorption and absorption; the CCS is global structural
parameter and its determination provides little informa-
tion about internal structure. At the same time, we note
that for cluster ions composed of a limited number of
molecules/atoms, the demarcation between adsorption and
absorption becomes blurred, as most atoms are exposed on
the cluster surface.
5. Though the reported measurements were made at 300 K,
of interest are measurements across a wide temperature
range to better understand how equilibrium binding coe-
cients vary with temperature (i.e., to examine separately en-
thalpic and entropic eects). Coupled with measurements at
temperatures appreciably higher or lower than 300 K, it will
be necessary to develop appropriate methods to predict ion
mobilities under these conditions, as the method invoked
here has only been tested near room temperature.
This work was supported by National Science Foundation
(NSF) Grant No. CHE-1011810. D.R.O. acknowledges sup-
port from a NSF Graduate Research Fellowship (NSF GRFP),
H.O. acknowledges support from a University of Minnesota
Doctoral Dissertation Fellowship, and C.L.A. acknowledges
support from a Ramon-Areces Fellowship.
1I. Kusaka, Z. G. Wang, and J. H. Seinfeld, J. Chem. Phys. 102, 913 (1995).
2I. Kusaka, Z. G. Wang, and J. H. Seinfeld, J. Chem. Phys. 103, 8993 (1995).
3F. Q. Yu, J. Chem. Phys. 122, 084503 (2005).
4S. P. Fisenko, D. B. Kane, and M. S. El-Shall, J. Chem. Phys. 123, 104704
5A. B. Nadykto and F. Q. Yu, J. Geophys. Res., [Atmos.] 108, 4717,
doi:10.1029/2003JD003664 (2003).
6U. Horrak, J. Salm, and H. Tammet, J. Geophys. Res. 103, 13909,
doi:10.1029/97jd01570 (1998).
7L. C. Rorrer and R. A. Yost, Int. J. Mass Spectrom. 300, 173 (2011).
8C. W. Tsai, R. A. Yost, and T. J. Garrett, Bioanalysis 4, 1363 (2012).
9G. A. Eiceman, E. V. Krylov, N. S. Krylova, E. G. Nazarov, and R. A. Miller,
Anal. Chem. 76, 4937 (2004).
10W. E. Steiner, S. J. Klopsch, W. A. English, B. H. Clowers, and H. H. Hill,
Anal. Chem. 77, 4792 (2005).
11M. Zhu, B. Bendiak, B. H. Clowers, and H. H. Hill, Anal. Bioanal. Chem.
394, 1853 (2009).
12G. V. de Miguel, M. Macia, C. Barrios, and J. Cuevas, Anal. Chem. 87, 1925
13J. Fernandez de la Mora, Aerosol Sci. Technol. 45, 543 (2011).
14P. M. Winkler, G. Steiner, A. Vrtala, H. Vehkamaki, M. Noppel, K. E. J.
Lehtinen, G. P. Reischl, P. E. Wagner, and M. Kulmala, Science 319, 1374
15P. M. Winkler, A. Vrtala, G. Steiner, D. Wimmer, H. Vehkamaki, K. E. J.
Lehtinen, G. P. Reischl, M. Kulmala, and P. E. Wagner, Phys. Rev. Lett. 108,
085701 (2012).
16H. Kohler, Trans. Faraday Soc. 32, 1152 (1936).
17A. K. Shchekin and A. I. Rusanov, J. Chem. Phys. 129, 154116 (2008).
18M. Kulmala, V. M. Kerminen, T. Anttila, A. Laaksonen, and C. D. O’Dowd,
J. Geophys. Res.: Atmos. 109, D04205, doi:10.1029/2003JD003961 (2004).
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded to IP: On: Mon, 14 Sep 2015 15:33:42
104204-11 Oberreit et al. J. Chem. Phys. 143, 104204 (2015)
19I. Dzidic and P. Kebarle, J. Phys. Chem. 74, 1466 (1970).
20T. Seto, K. Okuyama, L. de Juan, and J. Fernandez de la Mora, J. Chem.
Phys. 107, 1576 (1997).
21A. G. Nasibulin, J. Fernandez de la Mora, and E. I. Kauppinen, J. Phys.
Chem. A 112, 1133 (2008).
22S. V. Hering and M. R. Stolzenburg, Aerosol Sci. Technol. 39, 428
23A. Kupc, P. M. Winkler, A. Vrtala, and P. Wagner, Aerosol Sci. Technol. 47,
i (2013).
24S. Schobesberger, P. M. Winkler, T. Pinterich, A. Vrtala, M. Kulmala, and
P. E. Wagner, ChemPhysChem 11, 3874 (2010).
25A. W. Castleman, P. M. Holland, and R. G. Keesee, J. Chem. Phys. 68, 1760
26R. Bahadur and L. M. Russell, J. Chem. Phys. 129, 094508 (2008).
27V. Talanquer and D. W. Oxtoby, J. Chem. Phys. 119, 9121 (2003).
28D. Suh and K. Yasuoka, J. Phys. Chem. B 115, 10631 (2011).
29D. Suh and K. Yasuoka, J. Heat Transfer 135, 101002 (2013).
30I. K. Ortega, O. Kupiainen, T. Kurten, T. Olenius, O. Wilkman, M. J.
McGrath, V. Loukonen, and H. Vehkamaki, Atmos. Chem. Phys. 12, 225
31H. Y. Tang and I. J. Ford, Phys. Rev. E 91, 023308 (2015).
32D. J. Rader and P. H. McMurry, J. Aerosol Sci. 17, 771 (1986).
33D. R. Oberreit, P. H. McMurry, and C. J. Hogan, Phys. Chem. Chem. Phys.
16, 6968 (2014).
34G. Biskos, A. Malinowski, L. M. Russell, P. R. Buseck, and S. T. Martin,
Aerosol Sci. Technol. 40, 97 (2006).
35H. Ouyang, S. He, C. Larriba-Andaluz, and C. J. Hogan, J. Phys. Chem. A
119, 2026 (2015).
36I. N. Tang and H. R. Munkelwitz, J. Colloid Interface Sci. 98, 430 (1984).
37G. Hargreaves, N.-O. Kwamena, Y. H. Zhang, J. R. Butler, S. Rushworth,
S. L. Clegg, and J. P. Reid, J. Phys. Chem. A 114, 1806 (2010).
38L. Mitchem, J. Buajarern, R. J. Hopkins, A. D. Ward, R. J. J. Gilham, R. L.
Johnston, and J. P. Reid, J. Phys. Chem. A 110, 8116 (2006).
39P. E. Wagner, D. Kaller, A. Vrtala, A. Lauri, M. Kulmala, and A. Laaksonen,
Phys. Rev. E 67, 021605 (2003).
40A. Kamariotis, O. V. Boyarkin, S. R. Mercier, R. D. Beck, M. F. Bush, E. R.
Williams, and T. R. Rizzo, J. Am. Chem. Soc. 128, 905 (2006).
41R. A. Jockusch, A. S. Lemo, and E. R. Williams, J. Am. Chem. Soc. 123,
12255 (2001).
42R. J. Cooper, S. Heiles, M. J. DiTucci, and E. R. Williams, J. Phys. Chem.
A118, 5657 (2014).
43M. J. DiTucci, S. Heiles, and E. R. Williams, J. Am. Chem. Soc. 137, 1650
44T. Jin, B. B. Zhang, J. Song, L. Jiang, Y. Qiu, and W. Zhuang, J. Phys. Chem.
A118, 9157 (2014).
45A. Aguado, A. Ayuela, J. M. Lopez, andJ. A. Alonso, Phys. Rev. B 58, 9972
46H. Ouyang, C. Larriba-Andaluz, D. R. Oberreit, and C. J. Hogan, J. Am. Soc.
Mass Spectrom. 24, 1833 (2013).
47Q. Zhang, C. J. Carpenter, P. R. Kemper, and M. T. Bowers, J. Am. Chem.
Soc. 125, 3341 (2003).
48S. Kruckeberg, D. Schooss, M. Maier-Borst, and J. H. Parks, Phys. Rev. Lett.
85, 4494 (2000).
49C. J. Hogan and J. Fernandez de la Mora, Phys. Chem. Chem. Phys. 11, 8079
50J. Rus, D. Moro, J. A. Sillero, J. Royuela, A. Casado, F. Estevez-Molinero,
and J. Fernandez de la Mora, Int. J. Mass Spectrom. 298, 30 (2010).
51S. Ude and J. Fernandez de la Mora, J. Aerosol Sci. 36, 1224 (2005).
52C. Larriba and C. J. Hogan, J. Comput. Phys. 251, 344 (2013).
53C. Larriba and C. J. Hogan, J. Phys. Chem. A 117, 3887 (2013).
54J. Fernandez-Garcia and J. Fernandez de la Mora, J. Am. Soc. Mass Spec-
trom. 24, 1872 (2013).
55H. Tammet, J. Aerosol Sci. 26, 459 (1995).
56C. J. Hogan and J. Fernandez de la Mora, J. Am. Soc. Mass Spectrom. 21,
1382 (2010).
57J. B. Fenn, Int. J. Mass Spectrom. 200, 459 (2000).
58M. F. Bush, R. J. Saykally, and E. R. Williams, Int. J. Mass Spectrom. 253,
256 (2006).
59E. A. Mason and E. W. McDaniel, Transport Properties of Ions in Gases
(Wiley, New York, 1988).
60L. A. Viehland and E. A. Mason, Ann. Phys. 91, 499 (1975).
61B. K. Annis, A. P. Malinauskas, and E. A. Mason, J. Aerosol Sci. 3, 55
62C. Zhang, T. Thajudeen, C. Larriba, T. E. Schwartzentruber, and C. J. Hogan,
Aerosol Sci. Technol. 46, 1065 (2012).
63See supplementary material at for
tables displaying the ratio of the CCS measured at saturation ratio S to
the dry condition CCS and a summary of the calculated PA values for all
cluster ions.
64V. K. Rawat, G. V. de Miguel, and C. J. Hogan, “Modeling vapor uptake
induced mobility shifts in peptide ions observed with transversal modulation
ion mobility spectrometry-mass spectrometry,Analyst (in press)
65A. Kumar, S. Kang, C. Larriba-Andaluz, H. Ouyang, C. J. Hogan, and R. M.
Sankaran, Nanotechnology 25, 385601 (2014).
66C. Larriba-Andaluz, J. Fernandez-Garcia, M. A. Ewing, C. J. Hogan, and
D. E. Clemmer, Phys. Chem. Chem. Phys. 17, 15019 (2015).
67C. Larriba-Andaluz and C. J. Hogan, J. Chem. Phys. 141, 194107 (2014).
68H. Ambaye and J. R. Manson, Phys. Rev. E 73, 031202 (2006).
69H. Ambaye and J. R. Manson, J. Chem. Phys. 125, 084717 (2006).
70B. K. Ku and J. Fernandez de la Mora, Aerosol Sci. Technol. 43, 241
71C. Larriba, C. J. Hogan, M. Attoui, R. Borrajo, J. Fernandez-Garcia, and
J. Fernandez de la Mora, Aerosol Sci. Technol. 45, 453 (2011).
72A. D. Becke, J. Chem. Phys. 98, 5648 (1993).
73P. J. Hay and W. R. Wadt, J. Chem. Phys. 82, 270 (1985).
74P. J. Hay and W. R. Wadt, J. Chem. Phys. 82, 299 (1985).
75W. R. Wadt and P. J. Hay, J. Chem. Phys. 82, 284 (1985).
76Z. H. Li and D. G. Truhlar, J. Am. Chem. Soc. 130, 12698 (2008).
77T. Su and M. T. Bowers, J. Am. Chem. Soc. 95, 7609 (1973).
78S. L. Girshick, P. Agarwal, and D. G. Truhlar, J. Chem. Phys. 131, 134305
79C. S. Dutcher, X. L. Ge, A. S. Wexler, and S. L. Clegg, J. Phys. Chem. A
117, 3198 (2013).
80M. Chen, M. Titcombe, J. Jiang, C. Jen, C. Kuang, M. L. Fischer,F. L. Eisele,
J. I. Siepmann, D. R. Hanson, J. Zhao, and P. H. McMurry, Proc. Natl. Acad.
Sci. U. S. A. 109, 18713 (2012).
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: Downloaded to IP: On: Mon, 14 Sep 2015 15:33:42
... Currently, DMA-MS was applied to mobility and mass analysis of nanoparticles Fernandez de la Mora, 2009, 2010;Hogan et al., 2011;Criado-Hidalgo et al., 2013) and established their mobility-volume relationship based on measurements (Larriba et al., 2011). Atmospheric relevant clusters such as iodide salt (Oberreit et al., 2014;Oberreit et al., 2015), dimethylamine-sulfuric acid Thomas et al., 2016) and hybrid iodine pentoxide-iodic acid (Ahonen et al., 2019) were investigated by DMA-MS and their physicochemical properties such as density, thermostability, extent of hydration and organic vapors uptake ability were illustrated. Compared with previous techniques such as API-TOF/CI-API-TOF, DMA-MS could obtain the mobility and chemical composition at the same time. ...
Full-text available
The chemical composition of aerosol particles is crucial to understanding their formation and evolution in the atmosphere. However, very limited information is available for the chemical composition of ultrafine particles, particularly for nanoclusters and newly formed particles, due to the lack of valid analytical methods. This work reviews the online measurement techniques for characterizing the chemical composition of atmospheric clusters and sub-20 nm particles. The commonly used instrumentations are divided into two categories: direct techniques based on mass spectrometry and indirect measurement techniques mainly relying on the physical properties (e.g., hygroscopicity and volatility). The advantages of these techniques are compared and their limitations in the lab and field application are summarized. The combination of direct and indirect techniques is discussed, and this may provide more comprehensive understanding of chemical information of atmospheric clusters and particles. We propose that the newly developing instrumentations are needed to improve the collection efficiency for direct techniques, or the chemical resolution for indirect techniques. Future development should focus on obtaining simultaneous measurements of particle physical and chemical properties, which can be helpful in improving the accuracy of modeling and the understanding of particle formation and evolution.
... Proton affinities, intramolecular hydrogen bonds, and steric and inductive effects, are considered when computing the interaction energies of analyte-SR and, consequently, the most used factors to explain mobility changes induced by SR are the number of adduction sites, interaction energies, and SR size (Fernandez-Maestre 2018). Oberreit et al (2015) applied differential mobility-mass spectrometry to observe the sorption of 1-13 vapor molecules onto iodide cluster ions in air at 300 K. Measured CCS shifts were compared to predictions based on the Kelvin-Thomson-Raoult (KTR) and Langmuir adsorption models. They found that the models fit very well to measurements, but the earliest stages of vapor uptake were not well described by the KTR model. ...
Full-text available
Ion mobility spectrometry is widely used for the detection of illegal substances and explosives in airports, ports, custom, some stations and many other important places. This task is usually complicated by false positives caused by overlapping the target peaks with that of interferents, commonly associated with samples of interest. Shift reagents (SR) are species that selectively change ion mobilities through adduction with analyte ions when they are introduced in IMS instruments. This characteristic can be used to discriminate false positives because the interferents and illegal substances respond differently to SR depending on the structure and size of analytes and their interaction energy with SR. This study demonstrates that ion mobility shifts upon introduction of SR depend, not only on the ion masses, but on the interaction energies of the ion:SR adducts. In this study, we introduced five different SRs using ESI-IMS-MS to study the effect of the interaction energy and size on mobility shifts. The mobility shifts showed a decreasing trend as the molecular weight increased for the series of compounds ethanolamine, valinol, serine, threonine, phenylalanine, tyrosine, tributylamine, tryptophan, desipramine, and tribenzylamine. It was proved that the decreasing trend was partially due to the inverse relation between the mobility and drift time and hence, the shift in drift time better reflects the pure effect of SR on the mobility of analytes. Yet the drift time shift exhibited a mild decrease with the mass of ions. Valinol pulled out from this trend because it had a low binding energy interaction with all the SR and, consequently, its clusters were short-lived. This short lifetime produced fewer collisions against the buffer gas and a drift time shorter compared to those of ions of similar molecular weight. Analyte ion:SR interactions were calculated using Gaussian. IMS with the introduction of SR could be the choice for the free-interferents detection of illegal drugs, explosives, and biological and warfare agents. The suppression of false positives could facilitate the transit of passengers and cargos, rise the confiscation of illicit substances, and save money and distresses due to needless delays. Keywords: Adduction, ion mobility spectrometry, mass spectrometry, shift reagent, valinol, buffer gas modifier
... Ion mobility is sensitive to molecular geometry, enabling it to separate isomers and other species that are similar in mass, which is one of the challenges of mass spectrometry, especially as ion mass-to-charge ratios are increased (Krechmer et al., 2016). For AN measurements in particular, DMA-based separation may provide an extra advantage over typical IMS drift tubes since they operate at atmospheric pressure, near room temperature, and with a variety of carrier gases (Oberreit et al., 2015), which can result in higher cluster stability under atmospherically-relevant conditions. IMS-MS and DMA-MS have been shown to be powerful tools for measuring AN properties in laboratory experiments and an IMS-MS instrument has also been successfully deployed in the field (Krechmer et al., 2016). ...
Aerosol nanoclusters (AN), defined here as molecular aggregates suspended in a gas with dimensions between 2 and 10 nm, are the link between substances that we think of as molecules, or the “gas phase,” and those that we consider as particles, or the “condensed phase.” The ability to measure and model the physical and chemical properties of size-resolved AN, which at present is rudimentary at best, is crucial for understanding how particles form and evolve in a number of environments that are natural or influenced by human activities. This review describes the current state-of-the-art for measuring and modeling the size-resolved composition of atmospheric AN. We focus specifically on instruments, many relying on mass spectrometry, that show promise for closing the measurement gap under atmospherically relevant conditions by increasing the size of measurable gas-phase clusters (bottom-up approaches) and by decreasing the size of measurable nanoparticles (top-down approaches). Theoretical methods for predicting AN composition have similarly relied on bottom-up approaches that extend the accuracy of quantum chemistry calculations to larger molecular systems, as well as top-down approaches that correct bulk composition aerosol models for size-dependent properties such as viscosity and volatility. Current measurement and modeling challenges that must be overcome in order to close the gap are discussed.
Full-text available
Condensation and evaporation of vapor species on nanoparticle surfaces drive the aerosol evolution in various industrial/atmospheric systems, but probing these transient processes is challenging due to related time and length scales. Herein, we present a novel methodology for deducing nanoparticle evaporation kinetics using electrical mobility as a natural size indicator. Monodispersed nanoparticles are fed to a differential mobility analyzer which serves simultaneously as an evaporation flowtube and an instrument for measuring the electrical mobility, realizing measurements of evaporation processes with time scales comparable to the instrument response time. A theoretical framework is derived for deducing the evaporation kinetics from instrument responses through analyzing the nanoparticle trajectory and size-mobility relationship, which considers the coupled mass and heat transfer effect and is applicable to the whole Knudsen number range. The methodology is demonstrated against evaporation but can potentially be extended to condensation and other industrial/atmospheric processes involving rapid size change of nanoparticles.
Measurement of the gas-phase ion mobility of proteins provides a means to quantitatively assess the relative sizes of charged proteins. However, protein ion mobility measurements are typically singular values. Here, we apply tandem mobility analysis to low charge state protein ions (+1 and +2 ions) introduced into the gas phase by nanodroplet nebulization. We first determine protein ion mobilities in dry air and subsequently examine shifts in mobilities brought about by the clustering of vapor molecules. Tandem mobility analysis yields mobility-vapor concentration curves for each protein ion, expanding the information obtained from mobility analysis. This experimental procedure and analysis is extended to bovine serum albumin, transferrin, immunoglobulin G, and apoferritin with water, 1-butanol, and nonane. All protein ions appear to adsorb vapor molecules, with mobility "diameter" shifts of up to 6-7% at conditions just below vapor saturation. We parametrize results using κ-Köhler theory, where the term κ quantifies the extent of uptake beyond Köhler model expectations. For 1-butanol and nonane, κ decreases with increasing protein ion size, while it increases with increasing protein ion size for water. For the systems probed, the extent of mobility shift for the organic vapors is unaffected by the nebulized solution pH, while shifts with water are sensitive to pH.
Mass spectrometry is uniquely suited to identify and quantify environmentally relevant molecules and molecular clusters. Mass spectrometry alone is, however, not able to distinguish between isomers. In this study, we demonstrate the use of both an experimental set-up using a differential mobility analyser, and computational ion mobility calculations for identification of isomers. In the experimental set-up, we combined electrospray ionisation with a differential mobility analyser time-of-flight mass spectrometer to separate environmentally relevant constitutional isomers, such as catechol, resorcinol and hydroquinone, and configurational isomers, such as cyclohexanediols and fatty acids (i.e., oleic and elaidic acids). Computational ion mobility predictions were obtained using the Ion Mobility Software (IMoS) program. We find that isomer separation can be achieved with the differential mobility analyser, while for catechol, resorcinol and hydroquinone, the computational predictions can reproduce the experimental order of the ion mobilities between the isomers, confirming the isomers identification. Our experimental set-up allows analysis both in the gas and liquid phase. The differential mobility analyser can, moreover, be combined with any mass spectrometry set-up, making it a versatile tool for the separation of isomer.
Full-text available
In ion mobility spectrometry (IMS), reduced mobility (K 0 ) is an identification parameter of gas-phase ions but, frequently, these values are different whether there is contamination with moisture and other volatile...
Aerosol plays a vital role in atmosphere pollution, climate change, and health hazard. The mass transfer process between the aerosol particles and their ambient gas critically affects the evolution of aerosol particles and their phase states. A novel kinetic model is proposed to describe the gas-particle mass transfer among a particle bulk, a gas-particle interface and the ambient sources, based on Maxwell-Stefan relations and Langmuir adsorption theory. Two kinds of typical aerosol, which are respectively consisted of oxalic acid solution (aqueous solutions) and sucrose solution (easily to form a glassy state in a special condition), are chosen to demonstrate the feasibility of this proposed kinetic model, along with the available experimental data. The results showed that aerosol particles, which could kinetically form amorphous states during evaporation, are limited by the bulk viscosity. In contrast, the bulks consisting of aqueous solutions are controlled by the surface adsorption and desorption of molecules during mass transport.
Vapor assisted mobility shift measurements were made with atmospheric pressure drift tube ion mobility-mass spectrometry (IM-MS) to determine the thermodynamic properties of weakly bound ion-molecule clusters formed from protonated phenylalanine and neutral vapor molecules with hydroxyl functional groups. Relative binding energies and gas-phase association energies of amino acid ions clustered with small organic molecules have been established previously using high pressure mass spectrometry. However, the issue of volatility largely prohibits the use of high pressure mass spectrometry for the determination of gas-phase associations of amino acid ions clustered with neutral vapor molecules in many instances. In contrast, ion mobility measurements can be made at atmospheric pressure with volatile vapor additives near and above their boiling points providing access to clustering equilibria not possible using high vacuum techniques. In this study we report the gas-phase association energies, enthalpies, and entropies for protonated phenylalanine ion clustered with three neutral vapor molecules: 2-propanol, 1-butanol, and 2-pentanol based upon measurements at temperatures ranging from 120 oC to 180 oC. The gas-phase enthalpy and entropy changes ranged between -4 to -7 kcal/mol and -3 to 6 cal/(mol K), respectively. We found enthalpically favored ion-neutral cluster reactions for phenylalanine with entropic barriers for the formation of phenylalanine-1-butanol and phenylalanine-2-pentanol cluster ions, while phenylalanine- 2-propanol cluster ion formation is both enthalpically and (weakly) entropically favorable. Under the measurement conditions examined, phenylalanine-vapor modifier cluster ion formation is clearly observed via shifts in the drift time for the three test vapor molecules. In comparison, negligible shifts in mobility are observed for protonated arginine exposed to the same vapor modifiers.
Full-text available
Interest in understanding gas-to-particle phase transformation in several disciplines such as atmospheric sciences, material synthesis, and combustion has led to the development of several distinct instruments that can measure the particle size distributions down to the sizes of large molecules and molecular clusters, at which the initial particle formation and growth takes place. These instruments, which include the condensation particle counter battery, a variety of electrical mobility spectrometers and the particle size magnifier, have been usually characterized in laboratory experiments using carefully prepared calibration aerosols. They are then applied, alone or in combination, to study the gas-to-particle transition in experiments that produce particles with a wide range of compositions and other properties. Only a few instrument intercomparisons in either laboratory or field conditions have been reported, raising the question: how accurately can the sub-10 nm particle number size distributions be measured with the currently available instrumentation? Here, we review previous studies in which sub-10 nm particle size distributions have been measured with at least two independent instruments. We present recent data from three sites that deploy the current state-of-the-art instrumentation: Hyytiälä, Beijing, and the CLOUD chamber. After discussing the status of the sub-10 nm size distribution measurements, we present a comprehensive uncertainty analysis for these methods that suggests that our present understanding on the sources of uncertainties quite well captures the observed deviations between different instruments in the size distribution measurements. Finally, based on present understanding of the characteristics of a number of systems in which gas-to-particle conversion takes place, and of the instrumental limitations, we suggest guidelines for selecting suitable instruments for various applications.
Full-text available
Low field ion mobility spectrometry-mass spectrometry (IMS-MS) techniques exhibit low orthogonality, as inverse mobility often scales with mass to charge ratio. This inadequacy can be mitigated by adding vapor dopants, which may cluster with analyte ions and shift their mobilities by amounts independent of both mass and mobility of the ion. It is therefore important to understand the interactions of vapor dopants with ions, to better quantify the extent of dopant facilitated mobility shifts. Here, we develop predictive models of vapor dopant facilitated mobility shifts, and compare model calculations to measurements of mobility shifts for peptide ions exposed to variable gas phase concentrations of isopropanol. Mobility measurements were made at atmospheric pressure and room temperature using a recently developed transversal modulation ion mobility spectrometer (TMIMS). Results are compared to three separate models, wherein mobility shifts due to vapor dopants are attributed to changes in gas composition and (I) no vapor dopant uptake is assumed, (II) site-specific dopant uptake by the ion is assumed (approximated via a Langmuir adsorption model), and (III) site-unspecific dopant uptake by the ion is assumed (approximated via a classical nucleation model). We find that mobility shifts in peptide ions are in excellent agreement with model II, site-specific binding predictions. Conversely, mobility shifts of tetraalkylammonium ions from previous measurements were compared with these models and best agreement was found with model III predictions, i.e. site-unspecific dopant uptake.
Full-text available
A pending issue in linking ion mobility measurements to ion structures is that the collision cross section (CCS, the measured structural parameter in ion mobility spectrometry) of an ion is strongly dependent upon the manner which gas molecules effectively impinge and are reemitted from ion surfaces (when modeling ions as fixed structures). To directly examine gas molecule impingement and reemission processes and their influence, we measured the CCSs of positively charged ions of room temperature ionic liquids 1-ethyl-3-methylimidazolium dicyanamide (EMIM-N(CN)2) and 1-ethyl-3-methylimidazolium tetrafluoroborate (EMIM-BF4) in N2 using a differential mobility analyzer-mass spectrometer (DMA-MS) and in He using a drift tube mobility spectrometer-mass spectrometer (DT-MS). Cluster ions, generated via electrosprays, took the form (AB)N(A)z, spanning up to z = 20 and with masses greater than 100kDa. As confirmed by molecular dynamics simulations, at the measurement temperature (~300 K), such cluster ions took on globular conformations in the gas phase. Based upon their attained charge levels, in neither He nor N2 did the ion-induced dipole potential significantly influence gas molecule-ion collisions. Therefore, differences in the CCSs measured for ions in the two different gases could be primarily attributed to differences in gas molecule behavior upon collision with ions. Overwhelmingly, through comparison of predicted CCSs with selected input impingement-reemission laws to measurements, we find that in N2, gas molecules collide with ions diffusely- they are reemitted at random angles relative to the gas molecule incoming angle- and inelastically. Meanwhile, in He, gas molecules collide specularly and elastically and are emitted from ion surfaces at deterministic angles. Results can be rationalized on the basis of the momentum transferred per collision; in the case of He, individual gas molecule collisions minimally perturb the atoms within a cluster ion (internal motion), while in the case of N2, individual gas molecules have sufficiently large momentum to alter internal motion in organic ions.
Full-text available
Structural characterization of ions in the gas phase is facilitated by measurement of ion collision cross sections (CCS) using techniques such as ion mobility spectrometry. Further information is gained from CCS measurement when comparison is made between measurements and accurately predicted CCSs for model ion structures and the gas in which measurements are made. While diatomic gases, namely molecular nitrogen and air, are being used in CCS measurement with increasingly prevalency, the majority of studies in which measurements are compared to predictions use models in which gas molecules are spherical or non-rotating, which is not necessarily appropriate for diatomic gases. Here, we adapt a momentum transfer based CCS calculation approach to consider rotating, diatomic gas molecule collisions with polyatomic ions, and compare CCS predictions with a diatomic gas molecule to those made with a spherical gas molecular for model spherical ions, tetra-alkylammonium ions, and multiply charged polyethylene glycol ions. CCS calculations are performed using both specular-elastic and diffuse-inelastic collisions rules, which mimic negligible internal energy exchange and complete thermal accommodation, respectively, between gas molecule and ion. The influence of the long range ion-induced dipole potential on calculations is also examined with both gas molecule models. In large part we find that CCSs calculated with specular-elastic collision rules decrease, while they increase with diffuse-inelastic collision rules when using diatomic gas molecules. Results clearly show the structural model of both the ion and gas molecule, the potential energy field between ion and gas molecule, and finally the modeled degree of kinetic energy exchange between ion and gas molecule internal energy are coupled to one another in CCS calculations, and must be considered carefully to obtain results which agree with measurements.
Full-text available
The production of metal nanoclusters composed of less than 10(3) atoms is important for applications in energy conversion and medicine, and for fundamental studies of nanomaterial nucleation and growth. Unfortunately, existing synthesis methods do not enable adequate control of cluster formation, particularly at atmospheric pressure wherein formation typically occurs on sub-millisecond timescales. Here, we demonstrate that ligand-free, unagglomerated nickel nanoclusters can be continuously synthesized at atmospheric pressure via the decomposition of bis(cyclopentadienyl)nickel(II) (nickelocene) in a spatially-confined microplasma process that rapidly quenches particle growth and agglomeration. The clusters were measured on line by ion mobility spectrometry (IMS) and further analyzed by atomic force microscopy (AFM). Our results reveal that stable clusters with spherical equivalent mean diameters below 10 [Formula: see text] are produced, and by controlling the nickelocene concentration, the mean diameter can be tuned up to ∼50 [Formula: see text]. Although diameter is often the sole metric used in nanocluster and nanoparticle characterization, to infer the number of atoms in AFM and IMS detected clusters, we compare measured AFM heights and IMS inferred collision cross sections to theoretical predictions based on both bulk matter approximations and density functional theory and Hartree-Fock calculated Ni nanocluster structures (composed of 2-15 atoms for the latter). The calculations suggest that Ni nanoclusters composed of less than 10(2) atoms can be produced repeatably with simple microplasma reactors.
Recent studies of new particle formation events in the atmosphere suggest that nanoclusters, i.e the species formed during the early stages of particle growth which are composed of 10^1-10^3 molecules, may consist of amines and sulfuric acid. The physicochemical properties of sub-10 nm amine-sulfuric acid clusters are hence of interest. In this work we measure the density, thermostability, and extent of water uptake of < 8.5 nm effective diameter dimethylamine-sulfuric (DMAS) nanoclusters in the gas phase, produced via positive electrospray ionization. Specifically, we employ three systems to investigate DMAS properties; ion mobility spectrometry (IMS, with a parallel-plate differential mobility analyzer) is coupled with mass spectrometry to measure masses and collision cross sections for < 100 kDa positively charged nanoclusters, two differential mobility analyzers in series (IMS-IMS) are used to examine thermostability, and finally a differential mobility analyzer coupled to an atmospheric pressure drift tube ion mobility spectrometer (also IMS-IMS) is used for water uptake measurements. IMS-MS measurements reveal that dry DMAS nanoclusters have densities of ~1567kg/m3 near 300 K, independent of the ratio of dimethylamine to sulfuric acid originally present in the electrospray solution. IMS-IMS thermostability studies reveal that partial pressures of DMAS nanoclusters are dependent upon the electrospray solution concentration ratio, R= [H2SO4]/[(CH3)2NH]. Extrapolating measurements, we estimate that dry DMAS nanoclusters have surface vapor pressures of order 10-4 Pa near 300 K, with the surface vapor pressure increasing with increasing values of R through most of the probed concentration range. This suggests that nanocluster surface vapor pressures are substantially enhanced by capillarity effects (the Kelvin effect). Meanwhile, IMS-IMS water uptake measurements show clearly that DMAS nanoclusters uptake water at relative humidities beyond 10% near 300 K, and that larger clusters uptake water to a larger extent. In total our results suggest that dry DMAS nanoclusters (in the 5-8.5 nm size range in diameter) would not be stable under ambient conditions; however, DMAS nanoclusters would likely be hydrated in the ambient (in some cases above 20% water by mass), which could serve to reduce surface vapor pressures and stabilize them from dissociation.
The excess free energy of a molecular cluster is a key quantity in models of the nucleation of droplets from a metastable vapour phase; it is often viewed as the free energy arising from the presence of an interface between the two phases. We show how this quantity can be extracted from simulations of the mechanical disassembly of a cluster using guide particles in molecular dynamics. We disassemble clusters ranging in size from 5 to 27 argon-like Lennard-Jones atoms, thermalised at 60 K, and obtain excess free energies, by means of the Jarzynski equality, that are consistent with previous studies. We only simulate the cluster of interest, in contrast to approaches that require a series of comparisons to be made between clusters differing in size by one molecule. We discuss the advantages and disadvantages of the scheme and how it might be applied to more complex systems.
Structures and reactivities of gaseous Fe(CN)6(3-)(H2O)n were investigated using infrared photodissociation (IRPD) kinetics, spectroscopy, and computational chemistry in order to gain insights into how water stabilizes highly charged anions. Fe(CN)6(3-)(H2O)8 is the smallest hydrated cluster produced by electrospray ionization and blackbody infrared dissociation of this ion results in loss of an electron and smaller dianion clusters. Fe(CN)6(3-)(H2O)7 is produced by the higher activation conditions of IRPD and this ion dissociates both by loss of an electron and by loss of a water molecule. Comparisons of IRPD spectra to those of computed low-energy structures for Fe(CN)6(3-)(H2O)8 indicate that water molecules either form two hydrogen bonds to the trianion or form one hydrogen bond to the ion and one to another water molecule. Magic numbers are observed for Fe(CN)6(3-)(H2O)n for n between 58 - 60, and the IRPD spectrum of the n = 60 cluster shows stronger water molecule hydrogen bonding than that of the n = 61 cluster, consistent with the significantly higher stability of the former. Remarkably, neither cluster has a band corresponding to a free O-H stretch, and this band is not observed for clusters until n ≥ 70, indicating that this trianion significantly affects the hydrogen-bonding network of water molecules well beyond the second and even third solvation shells. These results provide new insights into the role of water in stabilizing high valency anions and how these ions can pattern the structure of water even at long distances.
A prototype is introduced based on the Transversal Modulation Ion Mobility Spectrometry (TMIMS) technique, which provides a continuous output of mobility selected ions, which greatly eases the synchronization between the different analyzing stages. In the new architecture, two stages of filtration are used to drastically reduce the background produced by one stage alone. The two-stages TMIMS was coupled with two different Atmospheric Pressure Interface Mass Spectrometer MS. The new system enables IMS-IMS-MS analysis and other modes of operation: IMS pre-filtration, IMS-IMS, and full transmission mode. It provides a resolving power R>60 in IMS mode, and R>40 in each stage in IMS-IMS mode. 2-propanol vapors were introduced in one of the stages to enhance the mobility variations, and its effect was studied on a set of tetra-alkyl ammonium ions. We found that concentrations as low as 1% (in partial pressure) produce mobility variations as high as 20%, which suggest that IMS-IMS separation using dried N2 (in one stage) and a dopant (in the other stage), could be a very powerful way to enhance the separation capacity of the IMS-IMS pre-filtration approach.
Condensation on a cubic seed particle was simulated by classical molecular dynamics. Seed size and supersaturation ratio of the system were the factors that were examined in order to observe the effects of the dimension of seeds and thermodynamic conditions. Two stages of nucleation were observed in the phenomenon, where the first stage is from the seed growth and the second from homogeneous nucleation. Therefore, the nucleation rate and growth rate were each calculated by the Yasuoka-Matsumoto method. As seed size increased the growth rate decreased, but there was no clear seed influence on the homogeneous nucleation characteristics. Besides the classical nucleation theory, cluster formation free energy and kinetic analysis were conducted. The free energy in the exponential term of the classical nucleation theory and that obtained from the cluster formation free energy showed different characteristics.
The structures of hydrated guanidinium, Gdm+(H2O)n, where n = 1 - 5, were investigated with infrared photodissociation spectroscopy and with theory. The spectral bands in the free O-H (~3600 - 3800 cm-1) and free N-H (~3500 - 3600 cm-1) regions indicate that for n between 1 and 3, water molecules bind between the NH2 groups in the plane of the ion forming one hydrogen bond with each amino group. This hydration structure differs from Gdm+ in solution, where molecular dynamics simulations suggest that water molecules form linear H-bonds with the amino groups, likely a result of additional water-water interactions in solution that compete with the water-guanidinium interactions. At n = 4, changes in the free O-H and bonded O-H (~3000 - 3500 cm-1) regions indicate water-water H-bonding and thus the onset of a second hydration shell. An inner shell coordination number of n = 3 is remarkably small for a monovalent cation. For Gdm+(H2O)5, the additional water molecule forms hydrogen bonds to other water molecules and not to the ion. These results indicate that Gdm+ is weakly hydrated, and interactions with water molecules occur in the plane of the ion. This study offers the first experimental assignment of structures for small hydrates of Gdm+, which provide insights into the unusual physical properties of this ion.