Phase Transitions of Amorphous Solid Acetone in Confined Geometry

Jun 3, 2014 - and in the atmosphere,1,2 and mineralization of inorganic substances in geology and biology,3 as well as in many chemical applications, ...
0 downloads 11 Views 1MB Size
Article pubs.acs.org/JPCB

Phase Transitions of Amorphous Solid Acetone in Confined Geometry Investigated by Reflection Absorption Infrared Spectroscopy Sunghwan Shin,† Hani Kang,† Jun Soo Kim,*,‡ and Heon Kang*,† †

Department of Chemistry, Seoul National University, 1 Gwanak-ro, Seoul 151-747, South Korea Department of Chemistry and Nano Science, Ewha Womans University, 52 Ewhayeodae-gil, Seoul 120-750, South Korea



ABSTRACT: We investigated the phase transformations of amorphous solid acetone under confined geometry by preparing acetone films trapped in amorphous solid water (ASW) or CCl4. Reflection absorption infrared spectroscopy (RAIRS) and temperatureprogrammed desorption (TPD) were used to monitor the phase changes of the acetone sample with increasing temperature. An acetone film trapped in ASW shows an abrupt change in the RAIRS features of the acetone vibrational bands during heating from 80 to 100 K, which indicates the transformation of amorphous solid acetone to a molecularly aligned crystalline phase. Further heating of the sample to 140 K produces an isotropic solid phase, and eventually a fluid phase near 157 K, at which the acetone sample is probably trapped in a pressurized, superheated condition inside the ASW matrix. Inside a CCl4 matrix, amorphous solid acetone crystallizes into a different, isotropic structure at ca. 90 K. We propose that the molecularly aligned crystalline phase formed in ASW is created by heterogeneous nucleation at the acetone−water interface, with resultant crystal growth, whereas the isotropic crystalline phase in CCl4 is formed by homogeneous crystal growth starting from the bulk region of the acetone sample.

I. INTRODUCTION Phase transitions between liquid and solid phases and between different solid phases are of critical importance in numerous natural processes, such as ice formation on the Earth’s surface and in the atmosphere,1,2 and mineralization of inorganic substances in geology and biology,3 as well as in many chemical applications, including crystallization of active pharmaceutical ingredients4,5 and biological macromolecules.6 Although acetone is one of the most widely used organic compounds in chemical laboratories, the phase transitions of solid acetone remain little understood. These phenomena are of great interest in atmospheric and astrophysical sciences, because carbonyl organic compounds are widely distributed in the Earth’s atmosphere,7 as well as in interstellar clouds as ingredients of ice mantles of dust particles.8 A limited number of studies have been reported on investigations of the phase changes of solid acetone. In 1929, Kelley9 reported an anomalous change of heat capacity near 130 K, which is well below the melting point (178 K) of bulk acetone crystals, indicating a phase transition of the solid. However, the physical origin of the transition remained unclear for many years. In 1995, Ibberson et al.10 proposed that the transition is not of the order−disorder type, based on the observation of a small excess entropy value (2.04 J K−1 mol−1). Later, Allan et al.11 examined acetone crystals prepared either by solidification at room temperature under high pressure or by cooling the liquid below the melting temperature at ambient pressure. They reported stable and metastable structures of the acetone crystals at various temperatures, and attributed the heat capacity anomaly near 130 K to strengthening of the © XXXX American Chemical Society

intermolecular electrostatic contacts of acetone molecules in the crystals. Until now, however, many aspects of the phase transitions of solid acetone have remained unclear. In another line of research, the interactions of acetone with the surfaces of amorphous solid water (ASW) or ice crystals have been actively investigated by several research groups. Schaff and Roberts12−14 reported extensive studies on the adsorption, trapping, and desorption of acetone interacting with ASW and crystalline ice films, using temperature-programmed desorption (TPD) and reflection absorption infrared spectroscopy (RAIRS). Souda15 examined the surface diffusion and hydration of acetone molecules on porous and nonporous ASW films, based on the results of time-of-flight secondary ion mass spectrometry and TPD experiments. More recently, Lasne at al.16,17 reported RAIRS studies of the interactions of acetone and other oxygenated volatile organic compounds with the surfaces of pure or nitric acid-containing ice films. Theoretical studies of this subject include molecular dynamics simulations by Picaud and Hoang,18,19 force-field method calculations by Hammer and Schmidt,20 and quantum chemical calculations by Marinelli and Allouche.21 In addition, flow-tube experiments have been used to investigate the adsorption, trapping, and dissolution behaviors of acetone in ice samples in the temperature range of atmospheric relevance.22−27 Special Issue: Physics and Chemistry of Ice 2014 Received: April 24, 2014 Revised: May 31, 2014

A

dx.doi.org/10.1021/jp503997t | J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

was linearly p-polarized using a wire grid polarizer (Edmund Optics). The beam path between the vacuum chamber and the spectrometer was enclosed within a Plexiglas housing and was purged with dry nitrogen gas. The absorbance spectrum of a clean Ru substrate was used as a background reference. All RAIR spectra were averaged 256 times to reduce the noise to below 5 × 10−5 Abs at a spectral resolution of 4 cm−1.

In the present study, we investigated the phase transitions of amorphous solid acetone samples, which were prepared by acetone vapor deposition on a cold substrate and trapped inside an ASW or CCl4 matrix. Although amorphous-to-crystalline phase transitions have been observed and well established for ASW,28,29 it remains an interesting question whether an analogous process can occur for acetone. We explored this phase transition in diverse environments, i.e., a wide temperature range, spatially confined geometry, and under interfacial conditions with different materials. The occurrence of a phase transition was monitored by measuring the changes in the RAIRS acetone vibrational bands. This paper is organized as follows. The experimental methods for sample preparation and characterization are described in Section II. Section III-1 presents the results of RAIRS and TPD experiments with acetone films trapped in ASW. Section III-2 highlights the anomalous changes in the RAIRS intensities of acetone vibrational bands in these samples, and compares them with those observed for acetone films trapped in other environments. In Section III-3, we analyze the band shapes of ν(C O) and other vibrations to examine the natures of different solid phases of acetone. The experimental findings are discussed in Section IV and summarized in Section V.

III. RESULTS III-1. RAIRS and TPD Studies of Confined Acetone Films in ASW. We prepared an amorphous solid acetone sample confined in ASW as follows. First, a D2O-ASW film of thickness ∼30 ML was prepared by D2O vapor deposition on a Ru(0001) substrate at ∼80 K. An acetone film of thickness ∼14 ML was then grown on top of the ASW film at the same temperature. Finally, a D2O-ASW layer (∼30 ML) was overlaid on the acetone film at the same temperature, completing an ASW-sandwiched acetone structure, in which an amorphous acetone film is trapped between ASW films. The sample was then heated slowly and the resulting phase changes of the acetone film were monitored through RAIRS measurements of the acetone vibrational bands. Figure 1 shows the RAIR spectra

II. EXPERIMENTAL METHODS The experiments were carried out in an ultrahigh-vacuum chamber equipped with instrumentation for RAIRS, TPD, and other surface spectroscopic tools.30,31 The chamber background pressure was maintained below 1 × 10−10 Torr during the experiments. Acetone, water, and CCl4 samples were prepared in the form of thin films grown on a Ru(0001) single-crystal substrate. The substrate surface was cleaned using a standard procedure of Ar+-sputtering at 2 kV (∼5 μA current and ∼20 min duration) and annealing at 1500 K; the surface cleanliness was checked by examining the water TPD shape from a D2O monolayer on Ru(0001).32 The substrate temperature was controlled in the range 60−1500 K using a liquid-helium cryostat and a resistive heater, and the temperature was monitored using an N-type thermocouple wire connected to the substrate. Acetone and CCl4 vapors were deposited on a Ru(0001) substrate surface through a tube doser, at a deposition rate slower than 0.15 monolayers per second (ML s−1). A D2O film was prepared using a backfilling method, at a deposition rate ≤0.1 ML s−1. Acetone, D2O, and CCl4 samples were purified through freeze−vacuum−thaw cycles, and their purities were checked using a quadrupole mass spectrometer (Extrel) installed in the chamber. The thickness of the sample film was estimated based on TPD measurements. For the D2O and CCl4 films, the thickness was calculated by dividing the TPD intensity of each film by that of the corresponding monolayer on Ru(0001). For acetone, the TPD intensity was converted to the total number of acetone molecules in the film, and then to the film thickness using the unit cell volume of an acetone crystal.11 In this paper, we express the film thickness in units of ML, where we use 1 ML = 1.1 × 1015 molecules cm−2 for ASW and crystalline ice films, although this value is derived from an ice Ih(0001) surface, 4.5 × 1014 molecules cm−2 for CCl4, as estimated from the TPD intensity of CCl4 monolayer, and 4.8 × 1014 molecules cm−2 for acetone, as estimated from the crystal density. RAIRS was performed in a grazing angle (84°) reflection geometry,33 using a commercial Fouriertransform IR instrument (PerkinElmer) equipped with a mercury−cadmium telluride detector. The incident IR beam

Figure 1. RAIR spectra of acetone bands measured for ASWsandwiched acetone sample [D2O (∼30 ML)/acetone (∼14 ML)/ D2O (∼30 ML)/Ru(0001)]. Spectrum recorded (a) immediately after sample preparation at 80 K, (b) after heating the sample at 120 K, and (c) after heating to 140 K. (d) Transmission IR spectrum of liquid acetone, shown as a band intensity reference for isotropic molecular orientation.

of an ASW-sandwiched acetone film obtained immediately after preparation of the sample at 80 K (spectrum a), after heating to 120 K (spectrum b), and subsequently heating to 140 K (spectrum c). All these RAIR spectra were obtained at 80 K, i.e., after cooling the sample back to the original temperature (80 K) after each heating cycle; this procedure was used to improve the quality of the difference spectrum by minimizing thermal shifts in the background spectral intensity. In Figure 1, it can be seen that the intensities of the acetone bands, as well as their shapes, change with increasing temperature. The intensity changes are as follows. The CO stretching [ν(C O)] band decreases on heating to 120 K compared with its original intensity at 80 K, but then it regains its original intensity on further heating to 140 K. A similar intensity change occurs for the asymmetric methyl deformation [ν(CH3)asym def] band. In contrast, the asymmetric C−C−C stretching [ν(C− C−C)asym] and symmetric methyl deformation [ν(CH3)sym def] bands show the opposite behaviors, i.e., their intensities increase from 80 to 120 K and then decrease again at 140 K. B

dx.doi.org/10.1021/jp503997t | J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

We interpret that the observed intensity changes arise from reorientation of acetone molecules upon sample heating. RAIRS preferentially detects the vibrations that induce dipole changes in the direction of p-polarization of the light, or perpendicular to the metal substrate surface. Accordingly, if the reorientation of acetone molecules occurs collectively, it will affect the RAIRS intensities of different bands in different ways. The intensity changes are not due to loss of acetone molecules from the sample, as will be shown shortly by TPD experiments. Figure 1d shows the spectrum of liquid acetone, which serves as a reference for isotropic molecular orientation. When compared with this spectrum, the band intensities in spectra a−c indicate that the average molecular orientation of acetone changes from an isotropic distribution in an amorphous solid phase at 80 K (spectrum a) to a molecularly aligned phase at 120 K (spectrum b, called phase I hereafter), and then to an isotropic phase at 140 K (spectrum c, called phase II). In this interpretation, we assume that the acetone phase does not change during sample cooling from 120−140 to 80 K, which was used for the RAIR spectra shown in Figure 1. To check this assumption, we recorded the RAIR spectra directly at 140 K, without cooling the sample. The spectral features were basically identical to those obtained after cooling to 80 K (Figure 1c), which confirmed that sample cooling does not induce a phase change. Figure 2 shows diagrams of the normal modes of the four vibrational bands of acetone shown in Figure 1. If we consider

Figure 3. TPD curves of acetone (lines) and water (shaded area) from ASW-sandwiched acetone samples [D2O (30 ML)/acetone (d ML)/ D2O (30 ML)/Ru(0001)] with different acetone film thicknesses (d ≈ 9, 14, and 64 ML). The TPD curves of acetone are shown after normalization to the same height. The D2O TPD curve was measured from the D2O (30 ML)/acetone (14 ML)/D2O (30 ML) sample. The temperature ramping rate was 1 K s−1.

temperature is substantially below the melting temperature (178 K) of bulk acetone, it is possible that the interfacial region of the acetone film premelts into a fluid acetone layer. Also, it is well-known35,36 that the melting temperature is significantly lowered when the film thickness or particle size is reduced to the nanometer scale. An additional acetone desorption peak appeared at temperatures above the volcano peak, as shown for the 9 ML acetone sample at 168−173 K in Figure 3. The intensity of this desorption closely tracks the D2O desorption curve, and can therefore be ascribed to acetone mixed with the D2O films.14 For a thick acetone film (∼62 ML), acetone desorption starts at a much lower temperature (∼130 K; Figure 3). The desorption rate of acetone increases with increasing temperature, in a similar way to the zeroth-order desorption kinetics of a multilayer, and the volcanic desorption peak does not appear. The TPD curve shows that the majority of trapped acetone molecules desorb below the volcanic desorption temperature. These features indicate that the ASW overlayer does not effectively trap acetone molecules, and they migrate quite freely to the sample surface and desorb from there. As will be shown shortly, an amorphous solid acetone film crystallizes at 80−100 K. When a thick acetone film crystallizes, it may exert a strong pressure against the upper ASW layer and generate cracks there. In this case, acetone molecules can leak through these cracks and cannot be pressurized as in the case of thin sandwiched acetone films. The RAIRS and TPD results presented in this section illustrate that an amorphous acetone film sandwiched between ASW layers changes its phase upon heating from 80 K to the volcanic desorption temperature (157 K). The phase change may occur in multiple stages; for example, amorphous solid acetone with an isotropic distribution at 80 K changes to a molecularly aligned structure (phase I) at 100−130 K and then to an isotropic structure (phase II) at 140 K. Eventually, a fluid phase may be formed before volcanic desorption at 157 K. In the following sections, we investigate these different phases in further detail by examination of differently prepared acetone samples and by performing multiple-peak analysis of the acetone vibrational bands.

Figure 2. Vibrational modes [ν(CO), ν(CH3)asym def, ν(CH3)sym def, and ν(C−C−C)asym] of acetone and their displacement vectors.

the dipole change direction of the vibrations and p-polarization direction of the light, which is perpendicular to the acetone− water interface, we can deduce that phase I involves acetone molecules that are oriented with the CO bond parallel to the interface, and the C−C−C molecular plane perpendicular to the interface. Figure 3 shows the TPD results for ASW-sandwiched acetone films of different thicknesses (9, 14, and 62 ML). The most notable feature in TPD of acetone is a sharp desorption threshold at 156−158 K, which appears for relatively thin (9−14 ML) acetone films. Below this temperature, no desorption of acetone is observed. Schaff et al.14 previously observed this feature, and attributed it to a “molecular volcano” phenomenon. Volcanic desorption can occur as a result of the crystallization of ASW and accompanying crack formation in the upper ice layer, through which acetone molecules burst out.34 This observation indicates that the acetone sample is fluid at ∼157 K. Although this C

dx.doi.org/10.1021/jp503997t | J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

III-2. RAIRS Vibrational Band Intensities: Molecular Orientation of Acetone. We prepared acetone samples in different trapping environments and compared their RAIR spectra. The samples investigated were an acetone film sandwiched between CCl4 films, a homogeneous acetone− D2O mixture sample prepared by codeposition of two vapors, and a pure acetone film grown directly on a Ru substrate, as well as the acetone films of various thicknesses trapped in ASW described in the previous section. These samples exhibit characteristic RAIRS shapes for the acetone vibrational bands, which can be differentiated from one another, and the ways in which the band shapes and relative intensities change with temperature are also different. In this section, we describe how the band intensity changes with temperature; this gives information on the acetone molecular orientations in the samples. Figure 4 shows the relative intensity variations of the ν(C O), ν(CH3)asym def, ν(CH3)sym def, and ν(C−C−C)asym bands for

which the CO bond lies parallel to the plane of the acetone− water interface, and the C−C−C molecular plane is perpendicular to the interface. In the case of a thick (∼60 ML) acetone film in ASW, the decrease in the ν(CO) intensity indicates alignment of the CO bond parallel to the interfacial plane, as for the thin acetone film. However, the ν(CH3)sym def and ν(C−C−C)asym intensities do not increase, which indicate that the direction of the C−C−C molecular plane is not ordered. Interestingly, a thin (∼15 ML) acetone film grown directly on a bare Ru substrate is also ordered at ∼100 K, similar to that trapped between ASW layers. It was not possible to heat the acetone film on Ru above 100 K, because of rapid sublimation of the film. It appears that the only possible interpretation of the formation of an ordered phase in these samples is crystallization of amorphous solid acetone. The other samples, including an acetone film sandwiched between CCl4 layers and an acetone-D2O mixture film, do not show noticeable changes in band intensity during heating. This indicates that these samples do not form an ordered acetone crystal as in phase I. When a thin (∼14 ML) acetone film in ASW was further heated to 140 K, beyond the temperature of crystallization to phase I at 100 K, the intensities of the ν(CO) and ν(CH3)asym def bands increased back to their original intensities at 80 K, and the ν(CH3)sym def and ν(C−C−C)asym band intensities decreased. These changes indicate that the molecularly aligned crystals in phase I change into a new isotropic phase (phase II). This phase appeared when an acetone film was heated to 140 K, a condition that is achievable only for relatively thin acetone films trapped in ASW. III-3. Band Shape Analysis. This section focuses on the vibrational band shapes of acetone samples. Because the ν(C O) band shows more systematic evolution with changing temperature than the other bands, the shape of the ν(CO) band will be examined in detail through multiple-peak analysis. Although the other bands also show noticeable shape changes with changing temperature, they are difficult to interpret systematically. The complex changes of these band shapes are related to intermolecular vibrational couplings in the solid, which are very difficult to analyze without information regarding the packing structures of acetone solids. Figure 5 shows the ν(CO) band of acetone (left frame) and the 1120−1480 cm−1 region (right frame), which encompasses the ν(CH3)asym def, ν(CH3)sym def, and ν(C−C−C)asym bands. The multiple-peak analysis results of the ν(CO) band are also shown. In Figure 5a, an acetone film (∼14 ML) sandwiched between ASW layers shows a broad, structureless ν(CO) band when the sample is prepared at 80 K. The band shape does not change when the sample is prepared at a lower temperature (60 K), as long as it is below the transition temperature to crystalline phase I. These observations support that the morphology of the acetone film is amorphous. When this sample is heated to 120 K, the ν(CO) band shape changes significantly, shown in Figure 5b. The broad band of amorphous solid acetone changes to a shape that consists of three peaks (1717, 1708, and 1699 cm−1), which is indicative of formation of distinct solid structures. The splitting of the ν(CO) band into three components is more clearly seen for a thick acetone film in ASW, shown in Figure 5c. Here, the 1717 cm−1 peak becomes the largest among the three components; this peak seems to be blue-shifted slightly (by 1−2 cm−1) for a thick film, although the shift size is comparable

Figure 4. Relative intensity changes of RAIRS bands for acetone (AC) in different samples during temperature increase: AC (∼14 ML) in D2O (●), AC (∼60 ML) in D2O (○), AC(∼100 ML) in CCl4 (Δ), AC−D2O mixture (▼), and pure AC (∼15 ML) on Ru (★). (a) ν(CO); (b) ν(CH3)asym def; (c) ν(CH3)sym def; (d) ν(C−C−C)asym. The intensities are calculated from the integration of the band area, and shown after normalization against those from the isotropic samples prepared at 60−80 K (see text).

different samples with increasing temperature. The relative band intensities are plotted after normalization with respect to their intensities measured for the initially prepared samples at 60−80 K, which correspond to amorphous solid acetone or homogeneous solid solution phases. In these samples, acetone molecules should have an isotropic orientational distribution. In the case of a thin (∼14 ML) acetone film sandwiched between ASW layers, the ν(CO) intensity decreases upon sample heating from 80 to 100 K (Figure 4a), with a simultaneous decrease in the ν(CH3)asym def intensity (Figure 4b). Conversely, the ν(CH 3 ) sym def and ν(C−C−C) asym intensities (Figures 4c and 4d) both increase simultaneously. As mentioned in the previous section, these changes indicate the formation of a molecularly ordered structure (phase I), in D

dx.doi.org/10.1021/jp503997t | J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

Figure 5. ν(CO) band (left frame) and 1180−1480 cm−1 spectral region (right frame) of various acetone samples. The results of multiple-peak analysis for ν(CO) are shown by dotted curves. (a) Amorphous solid acetone film (∼14 ML) in D2O at T = 80 K, (b) thin (∼14 ML) acetone film in D2O at 120 K (phase I), (c) thick (∼62 ML) acetone film in D2O at 120 K (phase I), (d) homogeneous mixture of acetone and D2O (1:8 ratio) at 120 K, (e) thick (∼100 ML) acetone film in CCl4 at 110 K (phase III), and (f) thin (∼14 ML) acetone film in D2O at 140 K (phase II). In these figures, the vibrational bands are displayed so as to have similar heights for all samples. The actual intensities of these bands can be estimated from the absorbance bars shown in the figure.

to the uncertainty of the fitting procedure. The 1717 and 1708 cm−1 components increase significantly as the thickness of the acetone film increases, whereas the 1699 cm−1 component does not change much with changing film thickness. We therefore interpret that the 1717 and 1708 cm−1 peaks are derived from bulk acetone, whereas the 1699 cm−1 peak is related to acetone molecules in the interfacial region. Previously, the appearance of a red-shifted ν(CO) band (1701−1708 cm−1) was assigned to a H-bonded CO stretching mode in RAIRS studies of acetone adsorption on ASW films.12−14 In IR studies of acetone−water solutions,37 the 1698−1708 cm−1 peaks were assigned to a H-bonded CO stretching mode. However, we observed that acetone films grown on a bare Ru(0001) substrate without water also have these spectral components (not shown), similar to Figure 5c. This casts a doubt on the assignment of the 1699 and 1708 cm−1 peaks uniquely to Hbonded acetone. Our observations indicate that these peaks also appear for a pure acetone film, and the 1708 cm−1 peak may be related to a bulk acetone state. A homogeneous mixture of acetone and D2O (Figure 5d) shows a ν(CO) band shape that is quite different from those

observed for amorphous solid acetone (Figure 5a) or crystalline acetone (Figure 5b,c). The ν(CO) band of the acetone− D2O mixture can be fitted to two broad peaks centered at 1708 and 1699 cm−1. By examining acetone−D2O mixture samples with different acetone/D2O ratios, we verified that the 1708 cm−1 peak intensity increases relative to that of the 1699 cm−1 peak when the acetone/water ratio is increased (data not shown). This again supports our interpretation that the 1699 cm−1 peak observed for ASW-sandwiched acetone samples originates from acetone interacting with water in the acetone− water interfacial region. Heating an acetone−D2O mixture to 120 K did not change the band shapes and intensities. An acetone film sandwiched between CCl4 layers can be used as a reference sample to separate the effects of an ASW surface on acetone crystallization. A CCl4-sandwiched acetone film prepared at 60 K shows the spectral features of amorphous solid acetone, similar to those observed for an ASWsandwiched acetone sample in Figure 5a. However, as shown in Figure 5e, thermal heating of this sample to 110 K changes the band shapes significantly and very differently from the other samples. Specifically, the ν(CO) band in CCl4 consists of a E

dx.doi.org/10.1021/jp503997t | J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

sharp component at 1712 cm−1 and an additional broad component at 1717 cm−1, whereas the ν(CO) band in ASW splits into three components. Note that the ν(CO) band in CCl4 does not have the 1699 or 1708 cm−1 components that appear in ASW. The other vibrations in CCl4 are also different than in ASW; for example, the ν(CH3)sym def band (∼1370 cm−1) develops a new shoulder peak on the lower-frequency side (∼1350 cm−1) on heating the sample in CCl4 to 110 K, whereas this band sharpens in ASW. In addition, the ν(C−C− C)asym band position (∼1237 cm−1) is not changed by heating in CCl4, whereas this peak is slightly blue-shifted in ASW. All these observed distinct differences between the samples in CCl4 and ASW suggest that the crystal phase formed in CCl4 at 110 K is different from the phase I formed in ASW. Regarding the band intensities, the sample in CCl4 shows only marginal variations in the band intensities on heating from 80 to 110 K (Figure 4). This suggests that the crystals formed in CCl4 have nearly isotropic orientational distributions. This new structure is called phase III. The properties of this sample were not examined above 110 K, because of sublimation of the CCl4 film. Figure 5f shows the spectra of an acetone film in phase II, obtained by heating an ASW-sandwiched acetone film (∼14 ML) at 140 K. The spectra were recorded at 80 K, as mentioned above, but they are basically the same as those recorded at 140 K. The ν(CO) band shape changes from the three-component shape of phase I at 120 K to a broad shape centered at 1708 cm−1 at 140 K. The ν(CO) band can be fitted to a single broad Gaussian curve or, alternatively, it can be deconvoluted into the three components (1717, 1708, and 1699 cm−1) of phase I, but with different relative intensities from those in Figure 5b.

Figure 6. Structure of a stable orthorhombic acetone crystal constructed by duplicating the unit cell by 2 × 2 × 1 in the x, y, and z directions, using crystallographic information from ref 11. The unit cell dimensions are a = 9.172(8), b = 7.761(8), and c = 21.66(2), with the Pbca space group, at 110 K. (a) Orthographic and perspective views of the crystal structure are visualized using VMD.38 CO bonds are colored in red, and C−H bonds in white. (b) Two neighboring layers perpendicular to the z-axis, where all CO bonds are aligned perpendicular to the z-axis.

acetone films in ASW. We propose that the crystals in phase I have a structure similar to those reported by Allan et al.11 If, for example, these crystals are formed in a preferential orientation such that the acetone molecules are aligned along the z-axis, as shown in Figure 6a, then the observed spectral features of phase I can be explained. The question of how the crystals grow with such a preferential orientation arises. At the moment, this question lies beyond our experimental capabilities, but some hints are available; for example, phase I crystals are formed in acetone samples trapped in ASW or grown on Ru(0001), but not in those in CCl4. These observations indicate that the interfacial structure is critical for the formation of phase I crystals. Previous molecular dynamics studies of acetone molecules adsorbed on ice surfaces have shown that acetone molecules are aligned with CO bonds slightly tilted from the interface near the acetone−ice interface.18,19 By analogy, we suggest that the ordering of acetone molecules starts from a specific interface such as an acetone−water interface or Ru(0001) surface, and propagates into growth of ordered crystals via heterogeneous nucleation. On the other hand, the porous structure of ASW may not be effective to align acetone molecules collectively to form an ordered crystal phase. B. Isotropic Phase III in CCl4 at 90−110 K. The interactive force between acetone and CCl4 is weaker than that between acetone and water, which involves H bonds. Heterogeneous crystal growth from the interface is therefore less likely to occur for acetone in CCl4 than for acetone in ASW. If homogeneous crystal growth in the bulk prevails over interfacial heterogeneous crystal growth in the case of acetone in CCl4, then the isotropic nature of crystal phase III can be explained by the formation of multiple crystals via homoge-

IV. DISCUSSION The RAIRS study results indicate that amorphous solid acetone trapped in ASW undergoes phase changes twice during temperature increases from 60 to 140 K: first to an aligned crystalline phase (phase I) at 80−100 K and then to an isotropic phase (phase II) at 130−140 K. Also, we observed that amorphous solid acetone in CCl4 crystallizes into an isotropic phase (phase III), which has quite different RAIRS features from those of phases I and II in ASW. In what follows, we further discuss the nature of these different phases produced by the thermal treatment of various acetone samples. A. Molecularly aligned phase I in ASW at 100−130 K. There is clear evidence that an acetone film in ASW forms an aligned crystal phase upon heating to 100 K. The evidence is that the ν(CO) band has a distinct shape (Figure 5c) and the relative intensities of different bands indicate the formation of a molecularly aligned state. The band shapes and relative intensities do not change on further increasing the temperature from 100 to 130 K (Figure 4), which is consistent with the characteristics of a phase transition, indicating that the crystallization has occurred abruptly and is completed somewhere between 80 and 100 K. We therefore conclude that acetone undergoes an amorphous-to-crystalline phase transition at this temperature. Allan et al.11 reported the structures of stable and metastable acetone crystals, where all the CO bonds are aligned along the same plane. Figure 6 shows one of these structures for a stable crystal. In the present samples, the crystallization preferentially aligns CO bonds parallel to the acetone− water interface. The C−C−C molecular plane is also ordered perpendicular to the interface in the cases of thin (∼14 ML) F

dx.doi.org/10.1021/jp503997t | J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

Although current understanding of the underlying principles of solid-phase transitions is insufficient to explain these phenomena clearly, there have been reported examples of experimentally observed crystal morphology changes, and these were interpreted in terms of crystallization theories such as Ostwald’s law of stages, independent nucleation, and cross nucleation.5,39 D. Fluid Phase near 157 K. Volcanic desorption of acetone at ∼157 K indicates that acetone is fluid underneath the crystallizing ASW film near this temperature. The fluid acetone may be formed in the premelting surface region of an acetone film, whose interior remains solid. However, the possibility that the entire acetone film melts cannot be excluded, if thin films have a substantially lower melting temperature than the bulk, as mentioned above. Also, if phase II is acetone−ASW solution rather than pure solid acetone, as we discussed above, then fluid acetone may be produced by freezing out of the acetone−ASW solution to pure ice crystals and acetone liquids near ∼157 K.

neous nucleation, which would result in orientational isotropy on ensemble averaging, regardless of the crystal structure. The RAIRS band shapes are quite different for phases I (Figure 5c) and III (Figure 5e), which leads us to conclude that these phases have different crystal structures. Although their overall band shapes are distinctly different, multiple-peak analysis of the ν(CO) band indicates that the 1717 cm−1 peak may exist commonly in phases I and III, although its intensity in phase III is much weaker than that in phase I. If the 1717 cm−1 peak is uniquely assignable to phase I, then it is possible that the sample in phase III contains a small proportion of phase I crystals, oriented randomly inside the sample. Further discussion of these possibilities will be deferred until further research has been undertaken. C. Isotropic Phase II in ASW at 140 K. The broad ν(C O) band of phase II (Figure 5f) can be fitted either to a single Gaussian curve (1708 cm−1) or deconvoluted into the three components (1717, 1708, and 1699 cm−1) of phase I, but with different relative intensities. The band shapes of phase II and III are well distinguished, which suggests that they are independent structures; for example, the 1708 and 1699 cm−1 components appearing in phase II do not appear in phase III. Although they are not yet definite, we suggest the following interpretations of phase II. When judged in terms of spectral resemblance, an acetone− D2O mixture sample (Figure 5d) is closest to phase II. We therefore propose that a large part of phase II consists of an acetone−water solid solution, which is formed by dissolution of an acetone film into ASW at 140 K. A very thin (