Article pubs.acs.org/JPCA
Probing Methanol Cluster Growth by Vacuum Ultraviolet Ionization Biswajit Bandyopadhyay, Oleg Kostko, Yigang Fang, and Musahid Ahmed* Chemical Sciences Division, Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, California 94720, United States S Supporting Information *
ABSTRACT: The ability to probe the formation and growth of clusters is key to answering fundamental questions in solvation and nucleation phenomena. Here, we present a mass spectrometric study of methanol cluster dynamics to investigate these two major processes. The clusters are produced in a molecular beam and ionized by vacuum ultraviolet (VUV) radiation at intermediate distances between the nozzle and the skimmer sampling different regimes of the supersonic expansion. The resulting cluster distribution is studied by time-offlight mass spectrometry. Experimental conditions are optimized to produce intermediate size protonated methanol and methanol−water clusters and mass spectra and photoionization onsets and obtained. These results demonstrate that intensity distributions vary significantly at various nozzle to ionization distances. Ion−molecule reactions closer to the nozzle tend to dominate leading to the formation of protonated species. The protonated trimer is found to be the most abundant ion at shorter distances because of a closed solvation shell, a larger photoionization cross section compared to the dimer, and an enhanced neutral tetramer precursor. On the other hand, the protonated dimer becomes the most abundant ion at farther distances because of low neutral density and an enhanced charged protonated monomer−neutral methanol interaction. Thomson’s liquid drop model is used to qualitatively explain the observed distributions.
■
INTRODUCTION Understanding the formation and growth of clusters is an active area of research as these processes in principle can capture the dynamical transition between the isolated gas phase and bulk medium. Cluster growth via ionic association pathways are key processes in aerosol physics1 and low temperature interstellar chemistry.2 The reverse process, i.e., dissociation of larger clusters by photoionization, plays a crucial role in atmospheric chemistry.3,4 Mass spectrometry is a powerful technique to study these processes and we used tunable vacuum ultraviolet (VUV) radiation to study small water clusters5,6 and mixed methanol−water hydrogen-bonded systems7 to decipher photoionization dynamics. Measurements of photoionization onsets and mass spectra provided information on fragmentation mechanisms, ion−molecule reactions, and structural rearrangements of neutral and/or ionized clusters. Continuous supersonic expansions from a small nozzle leads to adiabatic cooling, which results in density and temperature drop.8,9 As the expansion proceeds through the nozzle, numerous collisions between gas atoms produces neutral clusters. If a “seed” is added with the buffer gas, coexpansion produces the clusters of interest. For supersonic expansions, cluster formation and growth have been studied by varying the initial temperature and pressure of the expanding gas.10 Various nozzle shapes and sizes have also been used to study the effects on cluster formation.11 Pulsed laser vaporizations,12 electric discharge techniques,13,14 and electron guns15 routinely produce ionic clusters with different intensity distributions and internal temperatures by sampling the various parts of the © XXXX American Chemical Society
supersonic expansion. In those cases, pulsed laser or electron gun timings are adjusted with respect to the gas pulse to sample the regimes of ion−molecule to neutral clustering. For continuous expansions, these regimes in principle can be accessed by varying the photoionization distance from the nozzle. These expansions also provide for a more stable environment compared to pulsed sources by decoupling the fluid dynamics and ionization. In the present work, we report an experimental strategy to systematically study cluster growth and decay processes which are either predominated by ionization induced association and/or dissociation pathways. We produced a continuous molecular beam of methanol and photoionized it at variable distances from the nozzle to sample different regions of the expansion. We chose methanol because of its crucial role in astrochemical processing of hydrocarbons,16−18 atmospheric uptake by ice nanoparticles,19 and local structure of mixed liquids.20 Photoionization properties of alcohol−water clusters are also important in analytical chemistry.21 There have been numerous mass spectrometry based studies of methanol and methanol−water clusters produced by a variety of ionization sources including electron impact, and multi- and single-photon ionization.20,22−38 Photoionization dynamics of these systems are well understood, and hence this experimental approach can Received: January 28, 2015 Revised: April 8, 2015
A
DOI: 10.1021/acs.jpca.5b00912 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
are registered. The photoionization spectrum of xenon has two converging Rydberg series 5p5ns′ and 5p5nd′, the former of which is very narrow. PIE spectra of the 5p58s′ peak are measured using different sizes of the T4 monochromator exit slit and two of the PIE curves are shown in Figure 2a together
examine the effect of ionization distance on the mass spectrometry.
■
EXPERIMENTAL SECTION The experiments are carried out in a continuous supersonic expansion cluster machine coupled to a three meter VUV monochromator of a newly commissioned terminal (T4) on the Chemical Dynamics Beamline (9.0.2), located at the Advanced Light Source, Berkeley, California. Figure 1 shows a
Figure 1. Cross-sectional view of the molecular beam machine.
schematic of the experiment. One bar of argon is passed through a bubbler containing 99.9% pure methanol and expanded through a 100 μm nozzle to a differentially pumped chamber which is kept at a pressure of 2 × 10−4 Torr during the expansion. The molecular beam is intersected with the VUV radiation at various axial distances from the nozzle (x = 2−25 mm) and the resulting cluster ions are sampled into a time-offlight mass spectrometer. The ionization distance is varied by changing the nozzle position with respect to the point of intersection of molecular and VUV beams. The VUV beam is rectangular with a dimension of 1000 × 330 μm2. A set of four electrodes are used to guide the ions from the ionization region to the mass spectrometer through a skimmer. The lenses are kept at small potentials (+5, 0, −3, and 0 V, respectively) and the skimmer is grounded. The mass spectrometer is kept at 2 × 10−6 Torr in a second differentially pumped chamber. A start pulse for the TOF is provided by pulsing the repeller and accelerator plates because of the quasi-continuous (500 MHz) nature of the synchrotron light. The ions are pulse-extracted by fast switching of repeller and accelerator plates to 1100 V using a pulse width of 7.0 μs. Ions are accelerated perpendicularly to their initial flight path through the field free region and detected by a microchannel plate (MCP) detector which is installed at the end of the flight tube. Mass spectrometer settings are kept fixed while ionization distances are varied. The time dependent electrical signal from the detector is amplified by a fast preamplifier, collected by a multichannel scalar card, and then integrated with a computer. TOF spectra are measured at different positions in the photon energy range between 9.5 and 13.0 eV. The photoionization efficiency (PIE) curves are obtained by integrating peak intensities at each photon energy and normalized by the photon flux. The molecular beam end-station is used to measure the resolution of terminal T4’s three meter normal incidence monochromator (McPherson, Inc., Chelmsford, MA). PIE measurements are performed on a mixture of 2% xenon in helium and intensities of Xe peaks in time-of-flight mass spectra
Figure 2. (a) Xenon 5p58s′ resonance shown after subtraction of underlying broad 5p56d′ peak for two sizes of T4 monochromator exit slit: 100 and 600 μm. Red lines represent a Gaussian function fit to experimental data, the dependence of the full width at half-maximum (fwhm) on width of T4 monochromator exit slit is shown in panel (b). Error bars in panel (b) correspond to standard deviation of 2σ.
with a Gaussian fit. The dependence of full width at halfmaximum (fwhm) of the Gaussian on the exit slit size is shown in Figure 2b. Most of the experimental data points have a linear dependence. Only for the slit size below 100 μm do the experimental points start to deviate from linear dependence, converging to a maximum resolving power (E/ΔE) of 1000. The T4 photon flux is measured by a silicon photodiode SXUV-100 (Opto Diode Corp.) whose quantum efficiency is known. The flux under typical operational conditions is about 2 × 1013 photons/s.
■
RESULTS A. Mass Spectrometry of Methanol Clusters. Figure 3 displays a representative mass spectrum measured at a distance of 10 mm from the nozzle at 11 eV photon energy. The main ions observed in this experiment are similar to those in which VUV ionization was performed inside the mass-spectrometer chamber (MS-ionization).7 Protonated methanol clusters [H+(CH3OH)n] show the most intense distribution followed b y a m u ch w e ak e r p r o t o n a t e d m e t h a n o l− wa te r [H+(CH3OH)nH2O]. Unprotonated species apart from methanol monomer are not observed. Intermediate sizes (n ∼ 30) are observed for both protonated methanol and methanol− water clusters. Protonated methanol dimethyl ether progression B
DOI: 10.1021/acs.jpca.5b00912 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
Figure 3. Mass spectrum showing various cluster distributions measured at an ionization distance of 10 mm from the nozzle at 11 eV photon energy. Expanded view of protonated methanol and methanol dimethyl ether distributions are shown in insets. Figure 4. Intensity distributions of protonated methanol and methanol−water clusters measured at x = 2, 15, 20, and 25 mm. Ion intensities are plotted in logarithmic scale to highlight the prominent distributions observed for these two species. Inset in the upper trace shows the expanded view for n = 1−5.
[H+(CH3OH)n(CH3)2O] is also detected, albeit the signal is relatively small compared to the other cluster distributions. The appearance of protonated methanol upon ionization has been observed previously using electron impact and single and multiphoton ionizations.7,23,25,26,32,35,37−39 The mechanisms of methanol and methanol−water22,34 cluster formation were discussed in these earlier studies and here we will briefly describe some of the key features. It has been suggested that the ionization of neutral hydrogen bonded clusters leads to the formation of protonated cluster ions via rapid proton transfer and fragmentation:
distances, while the dimer is the most abundant ion at x = 20 mm. For n = 4−30, the overall ion intensities decrease with increasing ionization distances. There are some interesting features observed for these distances. At x = 2 mm, a gradual slow decrease is observed for n = 4−30. Mass spectra at all other distances show a general trend in the n = 4−30 range: a rapid decay at smaller clusters followed by a slow rise to intermediate sizes, and a moderate fall at larger species. As the distance increases, the intensities of smaller clusters decay more rapidly and an inverted-well shaped distribution is observed for the intermediate sizes. The maximum of this well increases with distance. For H+(CH3OH)nH2O, clusters up to n = 4 are not detected and n = 5−6 are very weak; a plateau is observed from n = 8−30 with a maximum around n = 15. As the ionization distance increases, the intensity of a specific size decreases, similar to those observed for protonated methanol clusters. C. Photoionization Efficiency (PIE) Curves and Appearance Energies (AE) at Various Distances. Figures 5 and 6 display the PIE curves of methanol and protonated methanol clusters measured in the 9.5−13.0 eV region. Figure 5 depicts PIE curves for each species with distance variation while Figure 6 shows PIE curves for different species while keeping the distance fixed to highlight trends in the mass spectra. The appearance energies (AE) for monomer cation and protonated clusters evaluated from the PIE curves are tabulated in Table S1 in the Supporting Information file (SI). The AE of methanol was found to be the same (10.8 eV) for MSand in-source ionization. AE values obtained for protonated methanol monomer, dimer, and trimer are in the range of 10.1−10.3, 9.8−10.1, and 9.6−10.1, respectively. Previous work from our group has shown that the ionization energy of methanol decreases upon clustering, reaching an asymptotic limit of around 9.8 eV for a cluster of size ≥4.7 PIE curves of the unprotonated monomer cation at various distances show sharp rise at 10.8 eV and this value is exactly the same as obtained for MS ionization. PIE curves of the protonated monomer shows a
(CH3OH)n + hν → (CH3OH)+n + e− → H+(CH3OH)n − 1 + CH3O + e−
(1)
Protonated methanol−water are formed either from neutral methanol−water cluster (CH3OH)n H 2O + hν → H+(CH3OH)n − 1H 2O + CH3O + e−
(2)
or from protonated methanol clusters H+(CH3OH)n → H+(CH3OH)n − 2 H 2O + (CH3)2 O
(3)
Castleman and Garvey have suggested that reaction 3 occurs for size n ≥ 9 since the lowest cluster size observed was H+(CH3OH)7.23−25,34 Morgan and Castleman23 have also suggested that this reaction does not occur for the smaller clusters, since the formation of a methyl bound complex intermediate is not facile. In our experimental conditions, n = 8 and 9 are more intense than n = 7 in the mixed methanol− water cluster series indicating that the observed distribution arises from reaction 4. B. Intensity Distribution of Cluster Ions at Various Distances. Figure 4 depicts the integrated intensities of protonated methanol [H+(CH3OH)n] and methanol−water [H+(CH3OH)nH2O] clusters at four various ionization distances (x = 2, 15, 20, 25 mm). Intensities are shown in logarithmic scale to highlight the interesting size distributions. For H+(CH3OH)n, relative intensities of smaller sizes (n = 1− 3) do not change significantly at all distances. The protonated methanol trimer is found to be the most intense ion at shorter C
DOI: 10.1021/acs.jpca.5b00912 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
Figure 5. PIE curves for methanol (M+) and protonated methanol clusters (MnH+) at various distances. Insets show the expanded view showing the onsets of ionization.
Figure 6. PIE curves for methanol (M+) and protonated methanol clusters (MnH+) at x = 2, 15, 20 and 25 mm.
small shoulder around 10.2 eV and sharp rise at 10.8 eV. As distance increases, AE values change slightly (on the order of ∼100−150 meV) for monomer. The change in AE values is most prominent for trimer, where the difference in AE between x = 2 and 25 mm is about 500 meV. PIE curves for these species also show some interesting features as the distance and photon energies are varied. For all distances, the methanol cation shows a sharp rise at 10.8 eV and then a gradual decrease up to 13 eV (Figure 5). Here, it should be noted that all the PIE curves have an absorption due to argon (used in the gas filter of the beamline to remove higher harmonics) at 11.8 eV.
For protonated methanol at 2 mm, after a sharp rise at 10.8 eV, a rapid decay (around 11 eV) followed by a sharp rise (at around 12.5 eV) is observed. At farther distances, a gradual increase in ionization efficiency is seen (ignoring the dip at around 11.8 eV). For dimer and trimer at 2 mm, PIE curves show a slow rise from 9.5 to 12.5, followed by a sharp increase. For the farthest distances, the curves show a slow rise up to 10.8 eV and then a rapid increase due to methanol absorption. The PIE curve for the methanol cation crosses that of the protonated methanol around 12.25 eV at all distances (Figure 6). At x = 2 mm, the dimer and trimer show a slow rise at lower D
DOI: 10.1021/acs.jpca.5b00912 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
Figure 7. (a) Schematic of a supersonic expansion. (b) A COMSOL simulation of a nozzle of diameter 100 μm. Distances are in mm.
clusters to get an idea if a liquid-like property is indicated by the intensity distributions. A general theory of cluster growth processes (Thomson’s liquid drop model) will be discussed for data analysis to explain the observed distributions at various distances. A. Characterization of the Supersonic Expansion. Supersonic expansion from a small nozzle leads to adiabatic cooling, resulting in rapid temperature and density drop as the expansion proceeds farther from the nozzle. Clusters are formed as a result of sufficient collisions close to the exit before the expansion becomes rarefied and cooling stops. Figure 7a shows a schematic of a supersonic jet expansion into a region of background pressure Pb. The gas accelerates in the nozzle throat where Mach number (M) becomes close to 1 at the exit and continues to accelerate (M ≫ 1) up to the Mach disc.
energies and crosses the monomer curve twice: once at 11.1 eV and again around 12.3−12.4 eV. For all other distances after the initial rise, dimer and trimer show steady ionization yields in the 11.0−13.0 eV range.
■
DISCUSSION Mass spectrometry and photoionization dynamics of methanol clusters have been well studied, and here we will mostly discuss the effect of variable ionization distance. In doing so, we will first briefly describe the supersonic expansion from a small nozzle and how cluster generation can be influenced by changing the distance. In the following section, the intensity distributions, PIE curves, and AEs for small clusters will be discussed to explain the effect of distance on these features. Finally, we will discuss the transition from small to intermediate E
DOI: 10.1021/acs.jpca.5b00912 J. Phys. Chem. A XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry A
and at 25 mm, the dimer is clearly the most abundant ion. The other interesting feature is the relative populations of the intermediate sized clusters. At 2 mm, after the initial rise from n = 1−3, the relative population decreases gradually showing an almost smooth distribution. On the other hand, as the distance increases, intermediate size clusters go through a maximum. In our previous study (MS-ionization), we found that the protonated methanol trimer is the most stable ion up to 14 eV photon energy. This observation is consistent with several other studies, including an electron impact26,32 one, where a closed solvation shell is proposed around a protonated methanol ion. In our case, the protonated trimer is the most abundant ion at 2 mm probably because of three factors: (a) closed solvation shell around a protonated methanol; (b) photoionization cross section is greater than the dimer up to 14 eV energy; and (c) enhancement of the neutral tetramer precursor at closer distances since the neutral density is higher. The protonated dimer becomes the most stable ion above 14 eV energy (in MS ionization) and also at farther ionization distances (in-source ionization). The dimer is most stable ion beyond 14 eV, as discussed in our previous work, because clusters with higher internal energies tend to fragment to lower sizes via evaporative cooling. In the in-source ionization scheme, the neutral density decreases as the expansion is moved farther from the nozzle. Up to the point of ionization, neutral clusters which are produced during the supersonic expansion remain unperturbed. As the distance increases, neutral density decreases and condensation around a charged species would be more probable for monomer (making a protonated dimer cation) rather than a dimer which would make a protonated trimer ion. Ion−molecule reactions involving the protonated monomer cation is unlikely since unprotonated clusters other than monomer cation are not observed. The AE of methanol monomer is found to be 10.8 eV at all distances (Figures 5 and 6). The AEs of protonated clusters are shifted around 100−300 meV as a function of distance. The AE of protonated monomer is in the range of 10.1−10.3 eV, similar to that found in the MSionization study. Since at all distances, the AEs do not shift closer to or are greater than the IE of methanol, it is clear that the protonated monomer is formed from fragmentation of larger clusters. This situation is also true for the protonated dimer and trimer, where for all distances, the AE of the protonated dimer is less than that of the protonated monomer; similarly, the AE of the trimer is less than that of the dimer. Another interesting feature is that at higher energies, the ionization yield of the monomer cation goes down at all distances, whereas in the case of the dimer and trimer, the yield is steady at all energies except at 2 mm ionization distance. At 2 mm, for dimer and trimer, the initial rise in the PIE curve is slow and gradual but shows a sharp rise at higher energies indicating that evaporative cooling from the higher cluster leads to more dimers and trimers. The evaporative cooling effect is not observed when the ionization is carried out at farther distances. Having described the nature of mass spectrometry and PIE curves of small clusters, we now discuss the effect of ion− molecule reactions to the cluster distributions. The thermochemistry of methanol and methanol−water clustering has been studied by Kebarle and Meot-Ner.26−28 They have computed enthalpy (ΔH), entropy (ΔS), and free energies (ΔG) of formation of reaction 4
Mach disc location (xM) is related to the nozzle diameter (d), stagnation pressure (P0), and background pressure (Pb) in a very simple but accurate empirical form
P xM = 0.67 b d P0 We use the nozzle diameter (100 μm) and pressure values from our experiment to the above expression and obtain a Mach disc position of ∼67 mm. The “zone of silence” is the core of the jet bounded by the barrel shock and the Mach disc. This is a region of very low pressure and molecular beams are extracted from it. Figure 7b shows a representation of the expansion from a 100 μm nozzle by solving the Navier−Stokes (N−S) equations using the COMSOL multiphysics software package. We assumed that the gas flow in the interfacial region remains at near thermal equilibrium continuum. This treatment is similar to continuum models applied in previous studies which used different nozzles and mass spectrometers.40,41 The complete description and parameters used for solving this model are provided in the SI. We have used the high Mach number flow module in the COMSOL multiphysics, which is valid for Knudsen number
