Article pubs.acs.org/Langmuir
X‑ray Fluorescence Imaging of Frozen Aqueous NaCl Solutions Kouki Tokumasu, Makoto Harada,* and Tetsuo Okada* Department of Chemistry, Tokyo Institute of Technology, Meguro-ku, Tokyo 152-8551, Japan S Supporting Information *
ABSTRACT: In frozen aqueous NaCl, the liquid phase (LP) should coexist with ice at temperatures between the melting point and eutectic point of the system. The LP forms grooves on the surface of ice. In the present study, the morphology of the surface LP is examined by synchrotron radiation X-ray fluorescence (XRF). These measurements afford 2D distribution maps for Cl− as a major component, in addition to coexistent metal ions (Mn2+, Co2+, Cu2+, and Zn2+). The 2D images obtained for Cl− indicate that a Y-shaped surface groove is formed in an observation area. Mn2+ and Co2+ are simply enriched in the LP together with Cl−. In contrast, Cu2+ and Zn2+ exhibit behavior that is different from that of Mn2+ and Co2+, and they are concentrated at particular locations that are obviously off of the LP; this tendency is more apparent for Zn2+. The 2D images are converted into 3D images by taking into consideration the freeze concentration in the LP and the attenuation of X-rays. The depth is largest in the middle of the groove, particularly at the intersection of surface grooves.
N
aturally occurring ice, e.g., glaciers, snow, clouds, and sea ice, usually contains impurities.1 NaCl is among the most common impurities found therein.2 When the temperature of NaCl-containing ice is lower than the eutectic point of the system (i.e., −21.3 °C), NaCl forms a hydrated salt.3 When the temperature of this solid mixture increases past the eutectic point, the liquid phase (LP), which is highly concentrated aqueous NaCl in this system, is formed within the frozen matrix. The volume fraction of the LP increases as the temperature is further increased. Thus, the LP coexists with ice in the temperature range between the eutectic point and the melting point, though the existence of a small amount of the LP has been demonstrated at temperatures below the eutectic point.4 The LP can act as an important medium for various environmentally important reactions and may accelerate reactions by various mechanisms, including the freeze concentration5,6 and pH shifts.7 In addition, the LP coexistent with ice has been recently employed to design analytical systems such as chromatography,8−10 microreactors,11−13 and preconcentration systems.14,15 The morphology of the LP, which provides a basis for the physical characteristics of a frozen aqueous solution, has received much attention because the LP coexistent with ice is of fundamental and environmental interest. The dihedral angle at the ice/LP interface has been discussed on the basis of the balance between the interfacial tension of the ice/LP interface and that of the ice/ice interface.16−18 The dihedral angle of the triple junction, where three ice grains meet, is expected to be smaller than 60° when the boundary is filled with the LP. This suggests that the cross section of the LP has a concave triangular shape, as schematically shown in Figure 1. This predicts that an LP-filled groove on the ice surface is wedgeshaped (Figure 1B). The LP growing on the ice surface at a temperature lower than the melting point is known as a quasi-liquid layer © 2015 American Chemical Society
Figure 1. Schematic representations of the cross section of (A) a grain boundary at a triple junction and (B) a surface groove filled with the LP.
(QLL).19−22 Although both experimental and theoretical approaches, such as vibrational spectroscopy,20 chromatography,21 and molecular dynamics,22 have been applied to study this structurally perturbed layer, there is no consensus even on its thickness.19 Sazaki et al.23−25 have developed laser confocal microscopy combined with differential interference contrast microscopy, which has a depth resolution greater than 0.37 nm, and have successfully evaluated the growth of the quasi-liquid layer (QLL) on the ice surface. They have revealed that there are two types of QLLsdropletlike and thin layer QLLand their appearance is independent of the ice crystal facet. However, the application of this method has been confined in the pure ice surface thus far. Cryo-SEM has been often employed for this purpose and has led to the detection of the formation of veins with a triangular cross section.26 McCarthy et al.27 recently succeeded in direct measurements of the dihedral angle of the LP in frozen aqueous Received: December 1, 2015 Published: December 29, 2015 527
DOI: 10.1021/acs.langmuir.5b04411 Langmuir 2016, 32, 527−533
Article
Langmuir sulfuric acid with cryo-SEM and showed that the angle between ice and sulfuric acid-containing liquid is approximately 26 ± 2° at −35 °C. However, this powerful method usually requires low pressure, which induces the sublimation of surface water from an ice sample. Neither the real ice surface nor the LP is visualized in a direct manner. Recently, environmental cryoSEM has been applied to the visualization of the ice surface.28 The sublimation of ice is substantially reduced because it can be operated at moderately high pressures of around 1 kPa. For the selective visualization of the LP, the incorporation of uranyl ions was studied. Although the high spatial resolution and surface selectivity of SEM-based methods are attractive for fine imaging of the ice surface, there remain difficulties in their application to the visualization of the LP. Magnetic resonance imaging is another choice for selective visualization of the LP that coexists with ice.29 However, this method has limited spatial resolution and does not show the fine morphologies of the LP. X-ray is also an efficient probe for imaging materials.30 X-ray computed tomography (CT) is a powerful tool for visualizing the interior structure of various materials and microphases. This method has been employed to probe ice crystal growth in frozen food,31 He-hydrate formation at the interface between ice and air,32 void formation in stones by freeze−thaw cycles,33 and so forth. An advantage of X-ray CT is the possibility of 3D mapping with a resolution of some tens of micrometers. However, this method gives poor contrast for light elements and no clear information on the identification of materials. It is thus difficult to distinguish ice from the LP when these two phases coexist. X-ray reflection fluorescence (XRF) specifies the element based on the energy of an emitted fluorescence Xray.34 Because this method cannot resolve ice from the LP, selective imaging is impossible unless at least one phase contains elements with high atomic weight. As stated earlier, when ice contains salts as impurities, the salts are expelled and dissolved in the LP, which is selectively visualized on the basis of the XRF signals of the elements contained therein. This article reports the selective imaging by XRF of the grooves on the surface of NaCl-doped ice. When ice contains other minor impurities, these elements also allow XRF imaging. In this study, we discuss the behavior of some transition-metal ions as minor components and that of the chloride ion as a major component upon freezing an aqueous NaCl solution.
■
Figure 2. Schematic illustration of experimental setup for XRF measurements. 60 XRF spectra. The energy-dispersion spectral data were acquired at each spot for 10 s. After one line scan was completed, the subsequent scan was carried out along the next line shifted by 2.5 μm. Thus, full scanning of the 150 μm × 150 μm area produced 3600 XRF spectra. These spectra were processed by principal component analysis (PCA) with a program written in Matlab R2009b (Mathworks) to remove noise components (script given in the Supporting Information).35,36 Figure S1 shows a change in the diagonal variation obtained by the PCA calculation. Although the 4 largest components recovered basic features of the original spectra, the first 10 components were used to keep minor spectral characteristics and regenerate noise-filtered spectra. The X-ray fluorescence peaks on the spectra were fitted with the Gaussian function to calculate peak areas. The XRF intensity was calibrated with an aqueous solution in the same cell as that used for measuring frozen samples. Figure S2 shows a solution-phase XRF spectrum for 25 mM Mn, 15 mM Co, 5.0 mM Cu, 5.0 mM Zn, and 0.20 M NaCl. Similar measurements were performed for different concentrations to construct calibration graphs.
■
RESULTS AND DISCUSSION Frozen samples can be regarded as a NaCl/water binary system because the NaCl concentration is around 1000 times higher than those of other components. The phase behavior of frozen samples can therefore be described by the NaCl/water phase diagram depicted in Figure S3. From this diagram, the concentration of NaCl in the LP is, for example, estimated to be 1.33 and 2.98 M at −4.6 and −11.5 °C, respectively. Upon freezing, all of the minor water-soluble components are enriched in the LP by the ratio of the NaCl concentration in the LP (cNaClLP) to that in an original unfrozen solution (cNaClori), unless specific removals such as evaporation, deposition, and entrapment by ice occur. The concentration ratios (cNaClLP/cNaClori) are thus estimated to be 78.5 and 176 for freezing at −4.6 and −11.5 °C, respectively. These ratios are assumed to be applicable to other minor solutes, i.e., transitionmetal ions in the present case. Figure 3 shows selected XRF spectra measured for a frozen sample at −4.6 °C after PCA noise reduction. For comparison, the corresponding spectra before PCA filtration are shown in Figure S4. The baseline noise is substantially eliminated by this procedure, and peak areas can thereby be reliably estimated. Because the frozen sample is heterogeneous on the micrometer scale, the XRF spectra measured at different locations have different features. The concentrations of the transition-metal ions in an original solution ranged from 5 to 25 μM. These
EXPERIMENTAL SECTION
Figure 2 shows a schematic representation of an XRF cell used in this study. The temperature of a sample was controlled on a Peltier array, which was driven by a model TDC-2020R Peltier controller (Cell System, Yokohama, Japan). A solution sample containing Mn2+, Co2+, Cu2+, Zn2+, and NaCl was put in a glass vessel (5 mm in diameter, 5 mm in depth) that was fitted into an Al holder with silver paste to maintain high thermal conductivity. The solution was frozen at −4.6 °C. The atmosphere over the frozen sample was purged with He to prevent frost adhesion. After complete freezing, XRF spectra were recorded at −4.6 and −11.5 °C. The temperature was measured on the surface of the sample using a thermocouple. The standard solutions of transition metals were prepared from the corresponding nitrate salts. XRF was measured on beamline BL-4A at KEK (Tsukuba, Japan). The excitation X-ray was monochromated at 10 keV by a doublecrystal Si(111) monochromator. The X-ray was focused on a sample area of 5.0 μm × 5.0 μm with a Kirkpatrick-Baez mirror. The excitation X-ray intensity was measured with an air-filled ion chamber. Fluorescence X-rays were detected by a silicon drift detector. Line scanning was performed over a distance of 150 μm on the sample, with an XRF spectrum measured every 2.5 μm. A single line scan thus gave 528
DOI: 10.1021/acs.langmuir.5b04411 Langmuir 2016, 32, 527−533
Article
Langmuir
purposes because the freezing process is contamination-free and does not require organic solvents and a concentration ratio can be managed by setting an appropriate temperature. The XRF measurements produce 2D images representing the distribution of individual elements in frozen samples. The images measured for Cl−, Mn2+, Co2+, Cu2+, and Zn2+ at −4.6 and −11.5 °C are depicted in Figure 4. A Y-shaped grain boundary can be seen in the images. The 2D images for Cl−, Mn2+, and Co2+ were similar, implying that these ions are simply concentrated in the LP. An environmental cryo-SEM study has shown that rapid freezing possibly causes the inhomogeneous distribution of the solutes contained in the LP.28 Because the present sample was slowly frozen at −4.6 °C, the concentrations of the components are expected to be uniform within the entire LP because the grain boundaries should be connected and the LP should have a homogeneous composition. Clear surface grooves can be seen in the images obtained for Cl−, Mn2+, and Co2+ at −11.5 °C, whereas the shape of the grooves is partially deformed and the signal intensities become weaker at −4.6 °C. The signals are treated as the projection on the surface plane. Although grain boundaries in the ice interior can contribute to XRF signals, this effect is not significant in the present case because of the short attenuation lengths of these elements. Intense signals are evident in the center of the surface grooves and also at their intersection. The signal intensity is related not only to the concentration of the element but also to the depth of the LP. This aspect is discussed in the following section. The simple freeze concentration of Cl−, Mn2+, and Co2+ in the LP is confirmed by close correlations among the XRF signals from these elements. Figure 5 illustrates the relationships between the XRF signal intensity from Cl and that from Mn or Co. The intensities measured at −4.6 and −11.5 °C are plotted in the figure in different colors. Data obtained at different temperatures fall on single lines in both Cl−Mn and Cl−Co correlations. This indicates that Mn2+/Cl− and Co2+/
Figure 3. Selected XRF spectra after PCA filtration measured on a frozen sample at −4.6 °C. The original solution (before freezing) contained Mn2+, Co2+, Cu2+, Zn2+, and NaCl (25 μM, 15 μM, 5 μM, 5 μM, and 17 mM, respectively).
concentrations give negligible XRF signals if measured in an unfrozen state. The metal ions are enriched in the LP by a factor of 78.5 at −4.6 °C and consequently give clear XRF peaks. The freeze concentration thus allows the XRF detection of transition-metal ions even at concentrations in the micromolar range. In the present study, the LP embedded in ice was irradiated by excitation X-rays to generate XRF signals. The depth of the LP is not uniform and depends on a measurement point on the surface of a frozen sample. Therefore, although the direct application of freezing to trace analyses is difficult with the present experimental setup, the freeze concentration is potentially efficient for quantification
Figure 4. Two-dimensional XRF images measured for Cl, Mn, Co, Cu, and Zn at −4.6 and −11.5 °C. The original solution (before freezing) contained Mn2+, Co2+, Cu2+, Zn2+, and NaCl (25 μM, 15 μM, 5 μM, 5 μM, and 17 mM, respectively). 529
DOI: 10.1021/acs.langmuir.5b04411 Langmuir 2016, 32, 527−533
Article
Langmuir
Figure 5. Correlations of the XRF intensity for Cl with those for Mn and Co in frozen aqueous NaCl at −4.6 °C (blue) and −11.5 °C (red).
Cl− concentration ratios are constant in the LP and not dependent on the locations on the frozen sample surface; i.e. Cl−, Mn2+, and Co2+ are concentrated in the LP at the same ratio along a scenario of freeze concentration. The 2D images of Cu2+ and Zn2+ show different features from those obtained with Cl−, Mn2+, and Co2+. Although the grain boundary can be seen in the image of Cu2+, the contrast is poorer than the corresponding images of Cl−, Mn2+, and Co2+. For Zn2+, the grain boundary cannot be seen; instead, intense spots are found at locations off the grain boundary. These observations suggest that Cu2+ and Zn2+ are not necessarily enriched in the grain boundary by a simple freeze concentration mechanism. Figure S5 shows 2D images obtained with a different sample. These images were measured over a larger area (600 μm × 600 μm) to ensure the specific enrichments of Cu2+ and Zn2+. The strong XRF signals from Cl− are detected mainly along diagonal surface grooves, where Mn2+ and Co2+ also provide intense XRF signals. In contrast, Cu2+ and Zn2+ exhibit entirely different features; as in Figure 4, intense XRF signals can be seen at the locations that are obviously off of the grain boundaries. A different mechanism from the freeze concentration should thus be responsible for the enrichment of Cu2+ and Zn2+. The freezing potential is induced by the imbalance between anionic and cationic partitions of the ice phase upon freezing an aqueous electrolyte.37 When aqueous NaCl is frozen, both Na+ and Cl− are partially entrapped in the ice crystal, but the latter is more preferably distributed in ice than the former. The resulting excess negative charges in ice and excess positive charges in the LP are relaxed by the transfers of H+ and OH− to the individual phases. This pH change in the LP in frozen alkali chloride has been precisely evaluated by fluorescence ratiometry using pyranine as a fluorescent pH indicater.38 The results indicated that the distribution of ions in the ice phase increases in the order of Li+ < Na+ < K+ < Cl−. The excess concentration of Cl− over Na+ in ice was determined to be on the order of 10−4 M. This possibly makes the LP basic compared to the original solution before freezing. Thus, freezing possibly facilitates the formation of transition-metal hydroxides. The reported values of pKSP for Mn(OH)2, Co(OH)2, Cu(OH)2, and Zn(OH)2 are 12.8, 14.2, 19.9, and 16.5−17.4, respectively,39,40 suggesting that Zn2+ and Cu2+ are more liable to form the hydroxides in moderately basified solutions than are Mn2+ and Co2+. Additional effects due to the
ice wall are expected. The ice surface is negatively charged by the facilitated dissociation of water molecules.37 More acidic metal ions should be more preferably entrapped by the surface negative charges, to be deposited as the corresponding metal hydroxides. Sea ice is suggested as a carrier of trace nutrients that control oceanic biological activity and production. Solid particulate matter, supplied mainly from the atmosphere to sea ice, is considered to be the main source of metallic elements.41,42 Iron, which is one of the most important elements governing biological activity in the oceans, was not studied in this work because of the low solubility of its hydroxide. The XRF spectra have shown that Cu2+ and Zn2+, whose hydroxides are much more soluble in water than Fe(OH)3, accumulate upon freezing. The present work implies that the accumulation of some metal ions in oceanic ice possibly contributes to the circulation of trace elements because they travel with sea ice, which is carried on ocean currents. When a simple freeze concentration is responsible for the enrichment of ions in the LP, the 2D elemental mapping can be converted to a 3D image by the detailed procedure given in the Supporting Information. This holds true for Cl−, Mn2+, and Co2+ as shown in Figure 5. In such cases, solutes are concentrated in the LP by a factor of cNaClLP/cNaClori upon freezing, unless a specific mechanism removes them from the LP. Thus, the concentrations of such components in the LP can be estimated from the phase diagram of NaCl/water. The depth, xi, of the LP determined from the XRF of an element, i, is given by xi = −
⎡ Ifi (d0 + di) ⎤ d0di ⎥ ln⎢1 − d0 + di ⎢⎣ Ifiεici d0di ⎥⎦
(1)
where d0 and di are the attenuation lengths of the incident X-ray and the XRF coming from i, respectively, Ifi is the XRF intensity from i, and εi is the excitation efficiency of i. For a homogeneous medium, the attenuation length, d, of an X-ray is given by43−45
d = ρ− 1 μ M − 1
(2)
where ρ is the density of the medium and μM is a mass absorption coefficient, which is given by μM =
∑ wiμM i
530
i
(3) DOI: 10.1021/acs.langmuir.5b04411 Langmuir 2016, 32, 527−533
Article
Langmuir
Table 1. Attenuation Length of Excitation X-ray and XRF from Cl, Mn, Co, Cu, and Zn in Ice, Water, and the LP at −4.6 and −11.5 °C attenuation length/μm energy/keV ice water LP at −4.6 °C LP at −11.5 °C
excitation
Cl Kα
Mn Kα
Co Kα
Cu Kα
Zn Kα
10.0 2071 1906 1229 844
2.621 40.3 37.1 35.7 34.1
5.894 432 398 264 184
6.924 696 640 420 292
8.040 1084 997 650 449
8.630 1338 1231 799 551
where wi and μMi denote the mass fraction and the mass absorption coefficient of element i. The mass absorption coefficient can be estimated by μ M = Ciλ 3 − Diλ 4 i
(4)
where Ci and Di are the Victoreen coefficients at a wavelength of λ.43 The d0 and di values calculated in this way are summarized in Table 1. The 3D images for the surface grooves can be determined by calculation using eq 1 with the parameters listed in Table 1 and a factor of cNaClLP/cNaClori. However, self-absorption should be taken into consideration to estimate the depth of the LP from the XRF intensity because the LP is a highly concentrated solution. The calibration curves (Figure S6), which assumed that all of the components are enriched in the LP at the same ratio, were used to correct the self-absorption in the LP. As stated above, this assumption is applicable to Cl−, Mn2+, and Co2+, which are simply enriched in the LP by the freeze concentration mechanism. In addition, the actual attenuation length of a given element cannot be explicitly estimated because the present sample has a heterogeneous composition. The XRF intensity is critically affected by the density of an intervening layer. The attenuation length is larger in ice than in the LP, suggesting that the estimated depth of the groove depends on the assumption of the phases lying along the path of the X-ray. Two extreme cases are assumed for the calculation of the depth of the groove; i.e., the path is composed of ice or the LP. The 3D images obtained with Cl−, Mn2+, and Co2+ as probes at −4.6 and −11.5 °C are shown in Figure 6, where the LP was assumed to be the XRF path. For comparison, the corresponding images obtained assuming ice as the XRF path are given in Figure S7. The largest depths estimated at the intersection of surface grooves are also shown in the corresponding figures. The LP assumption gives slightly greater depths than the ice assumption, but the difference in the depth is no larger than 10%. Thus, an uncertainty in the depth arises from the assumption of the intervening material but is not critical in the discussion of the order of the value. Therefore, the images obtained with the LP assumption (Figure 6) are discussed in the following text. The images depicted from the Cl− distributions show that wedge-shaped grooves with a depth of 10−30 μm are formed on the surface. The grooves are deeper at −4.6 °C than that at −11.5 °C, and the shape on the surface is deformed at the higher temperature because of the increased volume of the LP. The deepest LP appears at the intersection of surface grooves; the depth at this point is estimated to be 61.1 μm at −4.6 °C and 24.9 μm at −11.5 °C. The 2D images (Figure 4) indicate that the signal intensities for Cl− are higher at −11.5 °C than at −4.6 °C. We cannot determine from the 2D images whether the concentration of Cl− in the LP or the depth causes this
Figure 6. Three-dimensional images converted from the 2D XRF images given in Figure 4 using eq 1. The depth is given in micrometers. The LP was assumed to be the intervening layer for XRF.
difference. However, the conversion of images from 2D to 3D clearly indicates that high signal intensities at −11.5 °C arise from high concentrations of the analytes in the LP at the lower temperature and that the shallow LP emerges at a lower temperature, as expected. The 3D images determined with Mn2+ and Co2+ give almost the same morphologies as that obtained with Cl−, though the depth is slightly different depending on the probe ions. In the present experimental setup, the X-ray was incident on the sample at an angle of 45° because of the restriction of the optical alignment. This beam angle possibly deforms the shape of a groove. The groove width shown in Figure 6 may be larger than the actual value. The dependence of the shape, depth, and width of grooves on the incident angle is important in discussing their physical nature in more detail and is in progress in our laboratory.
■
CONCLUSIONS XRF imaging of the morphology of the LP on the surface of NaCl-doped ice has been presented. The 2D images can be 531
DOI: 10.1021/acs.langmuir.5b04411 Langmuir 2016, 32, 527−533
Article
Langmuir
(4) Cho, H.; Shepson, P. B.; Barrie, L. A.; Cowin, J. P.; Zaveri, R. NMR Investigation of the Quasi-Brine Layer in Ice/Brine Mixtures. J. Phys. Chem. B 2002, 106, 11226−11232. (5) Takenaka, N.; Bandow, H. Chemical Kinetics of Reactions in the Unfrozen Solution of Ice. J. Phys. Chem. A 2007, 111, 8780−8786. (6) Pincock, R. E. Reactions in frozen systems. Acc. Chem. Res. 1969, 2, 97−103. (7) Takenaka, N.; Tanaka, M.; Okitsu, K.; Bandow, H. Rise in the pH of an Unfrozen Solution in Ice due to the Presence of NaCl and Promotion of Decomposition of Gallic Acids owing to a Change in the pH. J. Phys. Chem. A 2006, 110, 10628−10632. (8) Shamoto, T.; Tasaki, Y.; Okada, T. Chiral Ice Chromatography. J. Am. Chem. Soc. 2010, 132, 13135−13137. (9) Tasaki, Y.; Okada, T. Ice Chromatography. Characterization of Water-Ice as a Chromatographic Stationary Phase. Anal. Chem. 2006, 78, 4155−4160. (10) Tasaki, Y.; Okada, T. Control of Ice Chromatographic Retention Mechanism by Changing Temperature and Dopant Concentration. Anal. Chem. 2011, 83, 9593−9599. (11) Anzo, K.; Harada, M.; Okada, T. Enhanced Kinetics of Pseudo First-Order Hydrolysis in Liquid Phase Coexistent with Ice. J. Phys. Chem. A 2013, 117, 10619−10625. (12) Hashimoto, T.; Tasaki, Y.; Harada, M.; Okada, T. ElectrolyteDoped Ice as a Platform for Atto- to Femtoliter Reactor Enabling Zeptomol Detection. Anal. Chem. 2011, 83, 3950−3956. (13) Tasaki, Y.; Okada, T. Up to 4 Orders of Magnitude Enhancement of Crown Ether Complexation in an Aqueous Phase Coexistent with Ice. J. Am. Chem. Soc. 2012, 134, 6128−6131. (14) Ito, K.; Okada, T. Freeze Sample Enrichment Highly Adaptable to Capillary Electrophoresis. Anal. Methods 2013, 5, 5912−5917. (15) Qu, H.; Arai, Y.; Harada, M.; Okada, T. Freeze Enrichment Protocol based on Voltammetric Probing of Liquid-Phase Growth in Frozen Aqueous Electrolyte Solutions. Anal. Chem. 2015, 87, 4314− 4320. (16) Mader, H. M. Observations of the Water-Vein System in Polycrystalline Ice. J. Glaciol. 1992, 38, 333−347. (17) Nye, J. F. The Geometry of Water Veins and Nodes in Polycrystalline Ice. J. Glaciol. 1989, 119, 17−22. (18) Walford, M. E. R.; Nye, J. F. Measuring the Dihedral Angle of Water at a Grain Boundary in Ice by an Optical Diffraction Method. J. Glaciol 1991, 125, 107−112. (19) Ewing, G. E. Thin Film Water. J. Phys. Chem. B 2004, 108, 15953−15961. (20) Sadtchenko, V.; Ewing, G. E. Interfacial melting of thin ice films: An infrared study. J. Chem. Phys. 2002, 116, 4686. (21) Tasaki, Y.; Okada, T. Ice Chromatographic Characterization of Thin Liquid Layer at the Interface between Water-Ice and Organic Solvent. J. Phys. Chem. C 2008, 112, 2618−2623. (22) Carignano, M. A.; Shepson, P. B.; Szleifer, I. Ions at the ice/ vapor interface. Chem. Phys. Lett. 2007, 436, 99−103. (23) Asakawa, H.; Sazaki, G.; Nagashima, K.; Nakatsubo, S.; Furukawa, Y. Prism and Other High-Index Faces of Ice Crystals Exhibit Two Types of Quasi-Liquid Layers. Cryst. Growth Des. 2015, 15, 3339−3344. (24) Sazaki, G.; Zepeda, S.; Nakatsubo, S.; Yokomine, M.; Furukawa, Y. Quasi-liquid layers on ice crystal surfaces are made up of two different phases. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 1052−5. (25) Sazaki, G.; Zepeda, S.; Nakatsubo, S.; Yokoyama, E.; Furukawa, Y. Elementary steps at the surface of ice crystals visualized by advanced optical microscopy. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 19702−7. (26) Blackford, J. R. Sintering and Microstructure of Ice: A Review. J. Phys. D: Appl. Phys. 2007, 40, R355−R385. (27) McCarthy, C.; Blackford, J. R.; Jeffree, C. E. Low-temperatureSEM study of dihedral angles in the ice-I/sulfuric acid partially molten system. J. Microsc. 2013, 249, 150−7. (28) Krausko, J.; Runstuk, J.; Nedela, V.; Klan, P.; Heger, D. Observation of a Brine Layer on an Ice Surface with an Environmental Scanning Electron Microscope at Higher Pressures and Temperatures. Langmuir 2014, 30, 5441−7.
converted to 3D images by considering the freeze concentration. Thus, the nature of eutectic mixtures renders 3D imaging possible. A similar approach is possible for other systems forming eutectic mixtures. In addition, the present work has revealed that the specific mechanism allows the enrichment of Cu2+ and Zn2+ upon freezing. They are possibly deposited as the corresponding hydroxides. The identification of the deposited species is now being studied in our laboratory. The specific deposition of metal ions on the ice surface should be related to their circulation in the northern oceans. The present study proposes a possible mechanism that controls the oceanic elemental circulation. Extending this work to other elements of biological importance should contribute to the understanding of this environmentally important feature.
■
ASSOCIATED CONTENT
* Supporting Information S
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.5b04411. Matlab script used for PCA filtration, typical change in the variance calculated in PCA, XRF spectrum of a standard solution, phase diagram of the water/NaCl system, raw XRF spectra obtained for a frozen sample, 2D images measured for Cl−, Mn2+, Co2+, Cu2+, and Zn2+ at −8.0 °C, calibration graphs for the XRF determination of Cl−, Mn2+, Co2+, Cu2+, and Zn2+ in the LP, 3D images converted from 2D XRF images given in Figure 4 with eq 1, and determination of the depth of an LP-filled groove (PDF)
■
AUTHOR INFORMATION
Corresponding Authors
*Phone and Fax: +81-3-5734-2612. E-mail: tokada@chem. titech.ac.jp (T.O). *
[email protected] (M.H.). Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank Professor Takeshi Hasegawa, Kyoto University, for helping us write Matlab programs for PCA. This work has been supported by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science. This work was performed under the approval of the Photon Factory Advisory Committee (proposal nos. 2013P105 and 014G705).
■
REFERENCES
(1) Bartels-Rausch, T.; Jacobi, H.-W.; Kahan, T. F.; Thomas, J. L.; Thomson, E. S.; Abbatt, J. P. D.; Ammann, M.; Blackford, J. R.; Bluhm, H.; Boxe, C.; Domine, F.; Frey, M. M.; Gladich, I.; Guzmán, M. I.; Heger, D.; Huthwelker, T.; Klán, P.; Kuhs, W. F.; Kuo, M. H.; Maus, S.; Moussa, S. G.; McNeil, V. F.; Newberg, J. T.; Pettersson, J. B. C.; Roeselo, M.; Sodeau, J. R. Relationship between Snow Microstructure and Physical and Chemical Processes. Atmos. Chem. Phys. Discuss. 2012, 12, 30409−30541. (2) Cullen, D.; Baker, I. Observation of Impurities in Ice. Microsc. Res. Tech. 2001, 55, 198−207. (3) Cohen-Adad, R., Vallee, P., Lorimer, J. W., Eds.; Alkali Metal and Ammonium Chlorides in Water and Heavy Water (Binary Systems); Pergamon Press: Oxford, England; Solubility Data Series, 1991; Vol. 47, pp 2255−2243. 532
DOI: 10.1021/acs.langmuir.5b04411 Langmuir 2016, 32, 527−533
Article
Langmuir (29) Galley, R. J.; Else, B. G. T.; Geilfus, N. X.; Hare, A. A.; Isleifson, D.; Ryner, L.; Barber, D. G.; Rysgaard, S. Morphology and Distribution of Liquid Inclusions in Young Sea Ice as Imaged by Magnetic Resonance. Cryosphere Discussions 2013, 7, 4977−5006. (30) McRae, R.; Bagchi, P.; Sumalekshmy, S.; Fahrni, C. J. In Situ Imaging of Metals in Cells and Tissues. Chem. Rev. 2009, 109, 4780− 4827. (31) Mousavi, R.; Miri, T.; Cox, P. W.; Fryer, P. J. A Novel Technique for Ice Crystal Visualization in Frozen Solids Using X-Ray Micro-Computed Tomography. J. Food Sci. 2005, 70, E437−E442. (32) Jin, Y.; Nagao, H.; Hayashi, J.; Shimada, W.; Ebinuma, T.; Narita, H. Observation of Xe Hydrate Growth at Gas-Ice Interface by Microfocus X-ray Computed Tomography. J. Phys. Chem. C 2008, 112, 17253−17256. (33) De Kock, T.; Boone, M. A.; De Schryver, T.; Van Stappen, J.; Derluyn, H.; Masschaele, B.; De Schutter, G.; Cnudde, V. A Pore-Scale Study of Fracture Dynamics in Rock Using X-ray Micro-CT under Ambient freeze-Thaw Cycling. Environ. Sci. Technol. 2015, 49, 2867− 2874. (34) Nakano, K.; Nishi, C.; Otsuki, K.; Nishiwaki, Y.; Tsuji, K. Depth Elemental Imaging of Forensic Samples by Confocal Micro-XRF Method. Anal. Chem. 2011, 83, 3477−3483. (35) Hubert, M.; Verboven, S. A Robust PCR Method for HighDimensional Regressors. J. Chemom. 2003, 17, 438−452. (36) Vogt, S.; Maser, J.; Jacobsen, C. J. Data Analysis for X-ray Fluorescence Imaging. J. J. Phys. IV 2003, 104, 617−622. (37) Drzymala, J.; Sadowski, Z.; Holysz, L.; Chibowski, E. Ice/Water Interface: Zeta Potential, Point of Zero Charge, and Hydrophobicity. J. Colloid Interface Sci. 1999, 220, 229−234. (38) Watanabe, H.; Otsuka, T.; Harada, M.; Okada, T. Imbalance between Anion and Cation Distribution at Ice Interface with Liquid Phase in Frozen Electrolyte As Evaluated by Fluorometric Measurements of pH. J. Phys. Chem. C 2014, 118, 15723−15731. (39) Charlot, G. Les Reations Chimique en Solutions; Masson et CIE: Paris, 1969. (40) Lange’s Handbook of Chemistry; 13th ed.; McGraw-Hill: New York, 1985. (41) Noble, A. E.; Moran, D. M.; Allen, A. E.; Saito, M. A. Dissolved and Particulate Trace Metal Micronutrients under the McMurdo Sound Seasonal Sea Ice: Basal Sea Ice Communities as a Capacitor for Iron. Front. Chem. 2013, 1, 1−18. (42) Nomura, D.; Nishioka, J.; Granskog, M. A.; Krell, A.; Matoba, S.; Toyota, T.; Hattori, H.; Shirasawa, K. Nutrient Distributions Associated with Snow and Sediment-laden Layers in Sea Ice of the Southern Sea of Okhotsk. Mar. Chem. 2010, 119, 1−8. (43) Teo, B. K. EXAFS: Basic Principles and Data Analysis. SpringerVerlag: Berlin ; Tokyo, 1986. (44) Thomsen, V. Basic Fundamental Parameters in X-ray Fluorescence. Spectroscopy 2007, 22, 46−50. (45) Malzer, W.; Kanngießer, B. Calculation of Attenuation and Xray Fluorescence Intensities for Non-Parallel X-ray Beams. X-Ray Spectrom. 2003, 32, 106−112.
533
DOI: 10.1021/acs.langmuir.5b04411 Langmuir 2016, 32, 527−533
