- Email: [email protected]
Ice films follow structure zone model morphologies
Thin Solid Films 518 (2010) 3422–3427
Contents lists available at ScienceDirect
Thin Solid Films j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t s f
Ice films follow structure zone model morphologies Julyan H.E. Cartwright, Bruno Escribano ⁎, C. Ignacio Sainz-Díaz Instituto Andaluz de Ciencias de la Tierra, CSIC-Universidad de Granada, E-18071 Granada, Spain
a r t i c l e
i n f o
Article history: Received 13 March 2009 Received in revised form 23 November 2009 Accepted 23 November 2009 Available online 3 December 2009 Keywords: Ice Structure zone model Dendrites Density Scanning electron microscopy
a b s t r a c t Ice films deposited at temperatures of 6–220 K and at low pressures in situ in a cryo-environmental scanning electron microscope show pronounced morphologies at the mesoscale consistent with the structure zone model of film growth. Water vapour was injected directly inside the chamber at ambient pressures ranging from 10− 4 Pa to 102 Pa. Several different substrates were used to exclude the influence of their morphology on the grown films. At the lowest temperatures the ice, which under these conditions is amorphous on the molecular scale, shows the mesoscale morphologies typical of the low-temperature zones of the structure zone model (SZM), including cauliflower, transition, spongelike and matchstick morphologies. Our experiments confirm that the SZM is independent of the chemical nature of the adsorbate, although the intermolecular interactions in water (hydrogen bonds) are different to those in ceramics or metals. At higher temperatures, on the other hand, where the ice is hexagonal crystalline on the molecular scale, it displays a complex palmlike morphology on the mesoscale. © 2009 Elsevier B.V. All rights reserved.
1. Introduction The deposition of films of ceramics, semiconductors, and metals is an extremely active field. An empirical structure zone model has been developed, which predicts the morphology on the mesoscale depending on the film temperature and the ambient pressure at which the gas is being deposited [1,2]. On the other hand, at lower temperatures and pressures than those of Earth's biosphere, water freezes not in a crystalline phase, but as amorphous ice, about which there is a great deal of debate as to whether different structures represent frozen versions of different liquid polymorphs of water with an associated phase transition [3]. These same amorphous ices are of interest to astronomers [4], as dust grains in interstellar space are typically covered with an icy film formed at the very low temperature of the interstellar medium (3–90 K); it has been estimated that most of the ice in the universe is to be found on these interstellar dust grains [5]. Much astrochemistry is presumed to take place on these icy surfaces, and larger icy bodies such as comets and Kuiper-belt objects are supposed to have formed through the accretion of this material. To connect the two bodies of knowledge of film deposition and ice physics we undertook experiments in which we deposited films of ice at 6–220 K. Our aims were to verify the structure zone model for ice, chemically very different to those materials from which it was derived, and thence also to further the understanding of the physics involved in the production of the morphologies.
⁎ Corresponding author. E-mail address: [email protected] (B. Escribano). 0040-6090/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2009.11.076
For many years films of numerous different materials have been deposited onto substrates from the vapour phase. The field is driven by a huge number of technological applications, but also has much scientific interest [2]. One of the key differences between films and bulk materials is in their morphologies, and from the 1960s on efforts have been made to construct a classification of the morphology of a film depending on the conditions of its deposition. This has culminated in the structure zone model. A recent version, reproduced in Fig. 1, shows the distinct film morphologies obtained as a function of the renormalized film temperature T/TM, where TM is the melting point of the material being deposited, and of the ion energy, which is inversely related to the argon gas pressure in the sputtering technique used in that research [1,2]. The different morphologies shown in this model are thus mainly a consequence of the competition between the spatially disordered deposition of particles on the growing film surface and the ordering effect of activated particle mobility processes. This mobility can be driven both by surface temperature and by bombardment of molecules from the vapour, a process often referred as gas-induced mobility [2]. The lowest temperature zone, labelled zone 1, consists of tapered columns separated by voids and is usually referred to as cauliflower morphology. This structure is formed at T/TM b 0.3 and the resulting films are amorphous or nanocrystalline, due to the low adatom mobility and self-shadowing effects. Zone T, transition morphology, occurs at similar low temperatures but at lower gas pressures. It is characterized by a very smooth surface with no long-range structure. These films are also amorphous or nanocrystalline, as the low temperatures do not allow for long-range mobility and molecular ordering. At slightly higher temperatures, 0.4 bT/ TM b 0.5, we find zone M or matchstick morphology, consisting of parallel columns with domed tops. In the case of ice, these films would still be amorphous. Zone 2 is found at 0.5b T/TM b 0.8 and is characterized by
J.H.E. Cartwright et al. / Thin Solid Films 518 (2010) 3422–3427
3423
Fig. 1. A recent version of the structure zone model of film deposition [2].
crystalline columns growing continuously from the substrate to the surface. At this temperature, adatom mobility is enough to compensate the effect of self-shadowing and to produce crystalline films. For even higher temperatures, bulk diffusion comes into place and produces the recrystallized grain structure seen in zone 3. The structure zone model describes well the trends seen in experiments with ceramics, semiconductors, and metals. One question we wished to answer is whether the same morphologies are found in films of ice, despite the very different chemical bonding in ice — which has an extensive network of directional hydrogen bonds — compared to the materials from which the structure zone model was compiled. Our interest has been principally to study ice films at low temperatures, at which there is low admolecule mobility, and at which the ice film on the molecular scale is known to be amorphous [6]. In vapour-deposition experiments hexagonal ice (Ih) is known to be formed above ∼150 K, while between ∼130 and 150 K there appears another crystalline phase; cubic ice (Ic). Below ∼130 K, the ice produced by vapour deposition is amorphous; this ice is frequently termed amorphous solid water (ASW). ASW differs when deposited at lower and higher temperatures. At 30 K, a higher density amorphous form is obtained, while at ∼30–130 K a lower density amorphous form is deposited [5,7]. Others have studied the physical properties of ice films, including density [8], porosity [9], heat transport [10] and crystallization kinetics [11]. Here we investigate their mesoscale morphologies with electron microscopy and compare our findings with the structure zone model. 2. Experimental setup In our experiments we used a FEI Quanta 200 environmental scanning electron microscope (ESEM) equipped with a liquid helium cold stage to grow ice films in situ at low pressures and at temperatures of 6–220 K. As well as a high-vacuum mode as in a conventional scanning electron microscope, an ESEM has further operating modes in which the chamber can function in the presence of water vapour and/or an auxiliary gas at pressures of up to 4000 Pa, and observations can be
made without the need to metal coat the sample. The combination of the ESEM with a liquid helium cold stage makes it possible to grow and image ice films in situ within the instrument. We began by evacuating the chamber in the high-vacuum mode of the microscope (target pressure 6 × 10− 4 Pa) and lowering the substrate to the working temperature. The microscope was set up so that the helium cold finger, together with a thermostat, was directly beneath the substrate on which we grew the ice film. We used several substrates (including brass, carbon, copper, platinum and titanium) in flat samples that were smooth at the micrometer resolution of the microscope and were thermally coupled as well as possible to the cold stage beneath them. We did not detect differences in film morphology between them once the film coverage was complete over the surface. To grow an ice film we switched the microscope to low-vacuum mode, in which we could inject water vapour into the chamber for a given length of time, typically a few seconds to a few minutes, and at target pressures ranging from 10− 4 Pa to 102 Pa. We either used demineralized water alone, or else bubbled helium through the water prior to injection to provide an inert auxiliary gas in the chamber and to reduce the partial pressure of water vapour. The ice film was then deposited on top of the substrate, which was the coldest point within the microscope. We switched again to highvacuum mode to image the results. Images were taken at voltages between 5 and 30 kV, with working distances of 5–16 mm, using a secondary electron detector. The experimental setup had certain limitations. We did not have the means to measure the ice film thickness except when the film peeled away from the substrate, and the temperature was measured at the base of the ice film, not at its surface. The microscope chamber walls were not thermally isolated from the room, so radiative heating of the sample altered the ice surface over a period of tens of minutes to an hour, even though the substrate was maintained at the programmed temperature. If we take into consideration the heat insulating properties of both crystalline and amorphous ice, these technical limitations meant that we could not assume that the film surface temperature was constant during the growth process. It turned out to be easier with this apparatus
3424
J.H.E. Cartwright et al. / Thin Solid Films 518 (2010) 3422–3427
to alter the film surface temperature, not by changing the substrate temperature, but maintaining that fixed and altering instead the deposition time. 3. Results The region of lowest temperature in the structure zone model is occupied by zone 1. Zone 1, or cauliflower morphology, consists of competing void-separated tapered columns whose diameters expand with the film depth according to a power law [2]. The surface resembles a cauliflower, showing self-similarity over a range of scales. We produced this morphology in the example in Fig. 2 using water vapour accompanied by helium as an auxiliary gas in the chamber. The vapour was injected with a target pressure of 10 Pa for 6 s and deposited on a titanium substrate at 6 K. By depositing the ice film at the same substrate temperature as in Fig. 2, but this time at a higher water vapour pressure (133 Pa) for a shorter time (1 s) without an auxiliary gas, we obtained, in the example shown in Fig. 3, zone T, or transition morphology, in which there is no long-range structure above the nanoscale. The general view in Fig. 3(a) shows the featureless surface characteristic of zone T, and also that the film had poor adhesion to the substrate and peeled away in places, enabling us to estimate its depth as 2 μm. A closer view in Fig. 3(b) allows us to see the boundaries of the individual densely-
Fig. 3. Zone T, transition morphology, film produced with water injected with a target pressure of 133 Pa for 1 s deposited on a copper substrate at 6 K.
Fig. 2. Zone 1, cauliflower morphology, film produced with water bubbled through with helium injected with a target pressure of 10 Pa for 6 s deposited on a titanium substrate at 6 K.
packed grains making up the surface, which were made visible as charge accumulated on the grain boundaries of the poorly conducting ice surface during imaging. In qualitative terms, the morphology of zone 1 is clearly driven by a competitive process of growth of clusters at all scales, leading to a fractal geometry, while greater gas-induced mobility of admolecules in zone T than in zone 1 smoothes out the surface. As they occur at the lowest temperatures, both zone 1 and zone T morphologies can be presumed to be composed of high-density amorphous ice. We noted when heating our samples that at close to 30 K the water vapour pressure in the chamber increased sharply, which we associate with sublimation during the structural transition from high- to low-density amorphous ice that occurs around this temperature. We did not however see any mesoscale structural change associated with this molecular-scale transformation. At film temperatures above those of the transition morphology of zone T, a spongelike morphology has been described for metallic films [12]. We were able to reproduce this morphology, which we term zone S, with an ice film, by injecting for a longer time than for zone T. As shown in Fig. 4, this morphology is characterized by a three-dimensional open network of material like a sponge. This film was produced by depositing water vapour with a target pressure of 10 Pa for 83 s on a copper substrate at 6 K. On the other hand at film temperatures above those corresponding to cauliflower morphology (zone 1), and for relatively high gas-induced admolecule mobility there appears the final
J.H.E. Cartwright et al. / Thin Solid Films 518 (2010) 3422–3427
3425
transient growth of whiskers on top of the earlier film at a substrate temperature of 220 K, as we see in Fig. 6(b). These whiskers are almost certainly of hexagonal crystalline ice, whose propensity to form complex dendritic structures is familiar from snowflakes. This knowledge should forewarn us that the structures we should encounter in zones 2 and 3 of the structure zone model, at high substrate temperatures corresponding to the deposition of crystalline ice, would reflect this complexity. Nevertheless we were taken aback by the beauty of the structures we grew in these conditions, seen in Fig. 7, which were grown injecting water at a target pressure of 10 Pa for 15 min over a carbon substrate initially at 6 K. The hexagonal crystalline nature of this ice is reflected in the angles of growth of the branches. This fascinating morphology of branched whiskers is intermediate between the whiskers of Fig. 6(b) and dendritic growth. Although these forms are based on the tendency of ice to dendritic growth, seen, for example, in snowflakes [14], here the environment is significantly different. In snowflakes, dendrites generally terminate in regular hexagonal facets. To obtain these palmlike structures we deposited an ice film for several minutes; the growing ice surface becomes progressively more thermally insulated from the substrate as the film depth increases and the morphology changes correspondingly. The microscope chamber is evacuated but not thermally isolated from its surroundings, so the walls are at room
Fig. 4. Zone S, spongelike morphology, film produced with water injected with a target pressure of 10 Pa deposited on a copper substrate. (a) After 83 s of injection with the substrate at 6 K. (b) Following a further 55 s of injection, during which the substrate temperature rose transiently to 80 K.
zone in the low-temperature region of the structure zone model: zone M, or matchstick morphology. In the example of this morphology we reproduce in Fig. 5 we see very large columns tens of micrometers in diameter resulting from ice deposition over 6 min at a target pressure of 133 Pa, on a platinum substrate at 6 K and using helium as an auxiliary gas. The columns display the domed tips characteristic of matchstick morphology, and also show interesting substructures both on the tip and along their length, presumably from their growth process; their shape is biomimetic, like an icy worm. The transition from zone 1 to zone M, we can view qualitatively in terms of competitive growth leading to a fractal morphology giving way to noncompetitive growth producing a columnar morphology. These large matchsticks may have been the morphology seen by Laufer et al. [13], who described an ice film grown at somewhere between 20 and 100 K as looking like “a shaggy woolen carpet”. With regard to the temperatures involved for zones S and M, they should typically be composed of low-density amorphous ice. The zones in the structure zone model are not disjunct, but form a continuum; in Fig. 6(a) we show an example of a morphology intermediate between zones 1 and M, produced with water injected with a target pressure of 133 Pa for 13 min deposited in a brass substrate at 6 K. Upon heating the substrate in this experiment we induced the
Fig. 5. Zone M, matchstick morphology, film produced with water bubbled through with helium injected with a target pressure of 133 Pa for 6 min deposited on a platinum substrate at 6 K.
3426
J.H.E. Cartwright et al. / Thin Solid Films 518 (2010) 3422–3427
Fig. 7. Film with a dendritic morphology resembling a palm tree produced with water injected with a target pressure of 10 Pa for 15 min deposited on a carbon substrate at 6 K.
Fig. 6. (a) Film with morphology intermediate between zones 1 and M produced with water injected with a target pressure of 133 Pa for 13 min deposited on a brass substrate at 6 K. (b) Whiskers of ice form on top of this morphology during heating of the substrate as the temperature reaches 220 K.
temperature and radiate accordingly, and the film surface temperature reflects this. The temperature of the upper surface can be estimated to be close to 200 K, taking into account the low thermal conductivity of the ice and the radiation. At this temperature there is high admolecule mobility, resulting in surface diffusion (both of molecules and of heat) and sublimation. Although it is clear that this dynamic equilibrium of fluxes along and perpendicular to the surface is the basic physics involved in producing this skeleton of palmlike structures formed by highly ordered branched whiskers, we are far at present from being able to formulate a quantitative theory for their growth [14]. It may be related with the fast ice growth observed at screw dislocations nucleated by surface defects [15]. One interesting open question regarding these higher temperature zones is how the amorphous to crystalline transition at the molecular scale relates to the transition from zones S and T to zones 2 and 3 on the mesoscale. 4. Discussion It is clear that structure zone morphologies do appear in ice films. Can this knowledge contribute to understanding the physics of the structure zone model? The two axes of the structure zone model shown in Fig. 1 are the renormalized film temperature and the ion energy,
which is inversely proportional to the gas pressure in sputtering experiments. On the other hand the deposition technique employed here is simply evaporation, involving thermal energies (3/2kTb 1 eV) for the admolecules. However, the energy required to sputter a water molecule is ∼0.2 eV [16], much lower than the 10–30 eV sputtering threshold for ceramics, semiconductors or metals [2]. So we should renormalize the scale of the second axis by the sputtering threshold, in the same way as the film temperature is renormalized by the melting point for the first axis. It becomes clear then that the physical basis for the two axes is admolecule mobility induced in one case by the temperature of the film itself, and in the other by the impinging admolecules from the vapour. The latter mechanism is related in these experiments to the temperature of the vapour, and to the presence or absence of helium as an auxiliary gas in the chamber that will thermalize with the chamber walls. The different effects of the two sources for admolecule mobility leading to more or less compact structures surely arise because, unlike thermal mobility, admolecule movement induced by bombardment is highly directional. This directionality is quantifiable in terms of the sputtering yield as a function of the angle of incidence of the bombardment, which has a maximum at 45–60° [2]. This is an avenue to explore in moving towards a physical understanding of the structure zone model as a consequence of the competition between the spatially disordered deposition of particles on the growing film surface and the ordering effect of activated particle mobility processes, which will be the subject of future work. Acknowledgments This work was performed by B. E. as part of his Ph.D. under the supervision of J.H.E.C. and C.I.S.D. We thank Daniel Araujo for help with the experimental work and Russ Messier for useful discussions. We acknowledge the financial support of CSIC project Hielocris and MCyT grant CGL2008-06245-CO2-O1. References [1] R. Messier, V.C. Venugopal, P.D. Sunal, J. Vac. Sci. Technol. A 18 (2000) 1538. [2] A. Lakhtakia, R. Messier, Sculptured Thin Films: Nanoengineered Morphology and Optics, SPIE Press, 2005. [3] H.E. Stanley, S.V. Buldyrev, M. Canpolat, M. Meyer, O. Mishima, M.R. Sadr-Lahijany, A. Scala, F.W. Starr, Physica A 257 (1998) 213. [4] P. Ehrenfreund, H.J. Fraser, J. Blum, J.H.E. Cartwright, J.M. García-Ruiz, E. Hadamcik, A.C. Levasseur-Regourd, S. Price, F. Prodi, A. Sarkissian, Planet. Space Sci. 51 (2003) 473. [5] P. Jenniskens, D.F. Blake, M.A. Wilson, A. Pohorille, Astrophys. J. 455 (1995) 389. [6] V.F. Petrenko, R.W. Whitworth, Physics of Ice, Oxford University Press, New York, 1999.
J.H.E. Cartwright et al. / Thin Solid Films 518 (2010) 3422–3427 [7] P. Jenniskens, D.F. Blake, Science 265 (1994) 753. [8] A. Hornekaer, A. Baurichter, V.V. Petrunin, A.C. Luntz, B.D. Kay, A. Al-Halabi, J. Chem. Phys. 124 (2005) 124701. [9] M.E. Palumbo, J. Phys. Conf. Ser. 6 (2005) 211. [10] O. Andersson, A. Inaba, Phys. Chem. Chem. Phys. 7 (2005) 1441. [11] Z. Dohnálek, G.A. Kimmel, P. Ayote, R.S. Smith, B.D. Kay, J. Chem. Phys. 118 (2003) 364.
[12] [13] [14] [15] [16]
3427
A.F. Jankowski, J.P. Hayes, J. Vac. Sci. Technol. A 21 (2003) 422. D. Laufer, E. Kochavi, A. Bar-Nun, Phys. Rev. B 36 (1987) 9219. K.G. Libbrecht, Rep. Prog. Phys. 68 (2005) 855. K. Thürmer, N.C. Bartelt, Phys. Rev. B 77 (2008) 195425. R. Bukowski, K. Szalewicz, G.C. Groenenboom, A. van der Avoird, Science 315 (2007) 1249.
Contents lists available at ScienceDirect
Thin Solid Films j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t s f
Ice films follow structure zone model morphologies Julyan H.E. Cartwright, Bruno Escribano ⁎, C. Ignacio Sainz-Díaz Instituto Andaluz de Ciencias de la Tierra, CSIC-Universidad de Granada, E-18071 Granada, Spain
a r t i c l e
i n f o
Article history: Received 13 March 2009 Received in revised form 23 November 2009 Accepted 23 November 2009 Available online 3 December 2009 Keywords: Ice Structure zone model Dendrites Density Scanning electron microscopy
a b s t r a c t Ice films deposited at temperatures of 6–220 K and at low pressures in situ in a cryo-environmental scanning electron microscope show pronounced morphologies at the mesoscale consistent with the structure zone model of film growth. Water vapour was injected directly inside the chamber at ambient pressures ranging from 10− 4 Pa to 102 Pa. Several different substrates were used to exclude the influence of their morphology on the grown films. At the lowest temperatures the ice, which under these conditions is amorphous on the molecular scale, shows the mesoscale morphologies typical of the low-temperature zones of the structure zone model (SZM), including cauliflower, transition, spongelike and matchstick morphologies. Our experiments confirm that the SZM is independent of the chemical nature of the adsorbate, although the intermolecular interactions in water (hydrogen bonds) are different to those in ceramics or metals. At higher temperatures, on the other hand, where the ice is hexagonal crystalline on the molecular scale, it displays a complex palmlike morphology on the mesoscale. © 2009 Elsevier B.V. All rights reserved.
1. Introduction The deposition of films of ceramics, semiconductors, and metals is an extremely active field. An empirical structure zone model has been developed, which predicts the morphology on the mesoscale depending on the film temperature and the ambient pressure at which the gas is being deposited [1,2]. On the other hand, at lower temperatures and pressures than those of Earth's biosphere, water freezes not in a crystalline phase, but as amorphous ice, about which there is a great deal of debate as to whether different structures represent frozen versions of different liquid polymorphs of water with an associated phase transition [3]. These same amorphous ices are of interest to astronomers [4], as dust grains in interstellar space are typically covered with an icy film formed at the very low temperature of the interstellar medium (3–90 K); it has been estimated that most of the ice in the universe is to be found on these interstellar dust grains [5]. Much astrochemistry is presumed to take place on these icy surfaces, and larger icy bodies such as comets and Kuiper-belt objects are supposed to have formed through the accretion of this material. To connect the two bodies of knowledge of film deposition and ice physics we undertook experiments in which we deposited films of ice at 6–220 K. Our aims were to verify the structure zone model for ice, chemically very different to those materials from which it was derived, and thence also to further the understanding of the physics involved in the production of the morphologies.
⁎ Corresponding author. E-mail address: [email protected] (B. Escribano). 0040-6090/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2009.11.076
For many years films of numerous different materials have been deposited onto substrates from the vapour phase. The field is driven by a huge number of technological applications, but also has much scientific interest [2]. One of the key differences between films and bulk materials is in their morphologies, and from the 1960s on efforts have been made to construct a classification of the morphology of a film depending on the conditions of its deposition. This has culminated in the structure zone model. A recent version, reproduced in Fig. 1, shows the distinct film morphologies obtained as a function of the renormalized film temperature T/TM, where TM is the melting point of the material being deposited, and of the ion energy, which is inversely related to the argon gas pressure in the sputtering technique used in that research [1,2]. The different morphologies shown in this model are thus mainly a consequence of the competition between the spatially disordered deposition of particles on the growing film surface and the ordering effect of activated particle mobility processes. This mobility can be driven both by surface temperature and by bombardment of molecules from the vapour, a process often referred as gas-induced mobility [2]. The lowest temperature zone, labelled zone 1, consists of tapered columns separated by voids and is usually referred to as cauliflower morphology. This structure is formed at T/TM b 0.3 and the resulting films are amorphous or nanocrystalline, due to the low adatom mobility and self-shadowing effects. Zone T, transition morphology, occurs at similar low temperatures but at lower gas pressures. It is characterized by a very smooth surface with no long-range structure. These films are also amorphous or nanocrystalline, as the low temperatures do not allow for long-range mobility and molecular ordering. At slightly higher temperatures, 0.4 bT/ TM b 0.5, we find zone M or matchstick morphology, consisting of parallel columns with domed tops. In the case of ice, these films would still be amorphous. Zone 2 is found at 0.5b T/TM b 0.8 and is characterized by
J.H.E. Cartwright et al. / Thin Solid Films 518 (2010) 3422–3427
3423
Fig. 1. A recent version of the structure zone model of film deposition [2].
crystalline columns growing continuously from the substrate to the surface. At this temperature, adatom mobility is enough to compensate the effect of self-shadowing and to produce crystalline films. For even higher temperatures, bulk diffusion comes into place and produces the recrystallized grain structure seen in zone 3. The structure zone model describes well the trends seen in experiments with ceramics, semiconductors, and metals. One question we wished to answer is whether the same morphologies are found in films of ice, despite the very different chemical bonding in ice — which has an extensive network of directional hydrogen bonds — compared to the materials from which the structure zone model was compiled. Our interest has been principally to study ice films at low temperatures, at which there is low admolecule mobility, and at which the ice film on the molecular scale is known to be amorphous [6]. In vapour-deposition experiments hexagonal ice (Ih) is known to be formed above ∼150 K, while between ∼130 and 150 K there appears another crystalline phase; cubic ice (Ic). Below ∼130 K, the ice produced by vapour deposition is amorphous; this ice is frequently termed amorphous solid water (ASW). ASW differs when deposited at lower and higher temperatures. At 30 K, a higher density amorphous form is obtained, while at ∼30–130 K a lower density amorphous form is deposited [5,7]. Others have studied the physical properties of ice films, including density [8], porosity [9], heat transport [10] and crystallization kinetics [11]. Here we investigate their mesoscale morphologies with electron microscopy and compare our findings with the structure zone model. 2. Experimental setup In our experiments we used a FEI Quanta 200 environmental scanning electron microscope (ESEM) equipped with a liquid helium cold stage to grow ice films in situ at low pressures and at temperatures of 6–220 K. As well as a high-vacuum mode as in a conventional scanning electron microscope, an ESEM has further operating modes in which the chamber can function in the presence of water vapour and/or an auxiliary gas at pressures of up to 4000 Pa, and observations can be
made without the need to metal coat the sample. The combination of the ESEM with a liquid helium cold stage makes it possible to grow and image ice films in situ within the instrument. We began by evacuating the chamber in the high-vacuum mode of the microscope (target pressure 6 × 10− 4 Pa) and lowering the substrate to the working temperature. The microscope was set up so that the helium cold finger, together with a thermostat, was directly beneath the substrate on which we grew the ice film. We used several substrates (including brass, carbon, copper, platinum and titanium) in flat samples that were smooth at the micrometer resolution of the microscope and were thermally coupled as well as possible to the cold stage beneath them. We did not detect differences in film morphology between them once the film coverage was complete over the surface. To grow an ice film we switched the microscope to low-vacuum mode, in which we could inject water vapour into the chamber for a given length of time, typically a few seconds to a few minutes, and at target pressures ranging from 10− 4 Pa to 102 Pa. We either used demineralized water alone, or else bubbled helium through the water prior to injection to provide an inert auxiliary gas in the chamber and to reduce the partial pressure of water vapour. The ice film was then deposited on top of the substrate, which was the coldest point within the microscope. We switched again to highvacuum mode to image the results. Images were taken at voltages between 5 and 30 kV, with working distances of 5–16 mm, using a secondary electron detector. The experimental setup had certain limitations. We did not have the means to measure the ice film thickness except when the film peeled away from the substrate, and the temperature was measured at the base of the ice film, not at its surface. The microscope chamber walls were not thermally isolated from the room, so radiative heating of the sample altered the ice surface over a period of tens of minutes to an hour, even though the substrate was maintained at the programmed temperature. If we take into consideration the heat insulating properties of both crystalline and amorphous ice, these technical limitations meant that we could not assume that the film surface temperature was constant during the growth process. It turned out to be easier with this apparatus
3424
J.H.E. Cartwright et al. / Thin Solid Films 518 (2010) 3422–3427
to alter the film surface temperature, not by changing the substrate temperature, but maintaining that fixed and altering instead the deposition time. 3. Results The region of lowest temperature in the structure zone model is occupied by zone 1. Zone 1, or cauliflower morphology, consists of competing void-separated tapered columns whose diameters expand with the film depth according to a power law [2]. The surface resembles a cauliflower, showing self-similarity over a range of scales. We produced this morphology in the example in Fig. 2 using water vapour accompanied by helium as an auxiliary gas in the chamber. The vapour was injected with a target pressure of 10 Pa for 6 s and deposited on a titanium substrate at 6 K. By depositing the ice film at the same substrate temperature as in Fig. 2, but this time at a higher water vapour pressure (133 Pa) for a shorter time (1 s) without an auxiliary gas, we obtained, in the example shown in Fig. 3, zone T, or transition morphology, in which there is no long-range structure above the nanoscale. The general view in Fig. 3(a) shows the featureless surface characteristic of zone T, and also that the film had poor adhesion to the substrate and peeled away in places, enabling us to estimate its depth as 2 μm. A closer view in Fig. 3(b) allows us to see the boundaries of the individual densely-
Fig. 3. Zone T, transition morphology, film produced with water injected with a target pressure of 133 Pa for 1 s deposited on a copper substrate at 6 K.
Fig. 2. Zone 1, cauliflower morphology, film produced with water bubbled through with helium injected with a target pressure of 10 Pa for 6 s deposited on a titanium substrate at 6 K.
packed grains making up the surface, which were made visible as charge accumulated on the grain boundaries of the poorly conducting ice surface during imaging. In qualitative terms, the morphology of zone 1 is clearly driven by a competitive process of growth of clusters at all scales, leading to a fractal geometry, while greater gas-induced mobility of admolecules in zone T than in zone 1 smoothes out the surface. As they occur at the lowest temperatures, both zone 1 and zone T morphologies can be presumed to be composed of high-density amorphous ice. We noted when heating our samples that at close to 30 K the water vapour pressure in the chamber increased sharply, which we associate with sublimation during the structural transition from high- to low-density amorphous ice that occurs around this temperature. We did not however see any mesoscale structural change associated with this molecular-scale transformation. At film temperatures above those of the transition morphology of zone T, a spongelike morphology has been described for metallic films [12]. We were able to reproduce this morphology, which we term zone S, with an ice film, by injecting for a longer time than for zone T. As shown in Fig. 4, this morphology is characterized by a three-dimensional open network of material like a sponge. This film was produced by depositing water vapour with a target pressure of 10 Pa for 83 s on a copper substrate at 6 K. On the other hand at film temperatures above those corresponding to cauliflower morphology (zone 1), and for relatively high gas-induced admolecule mobility there appears the final
J.H.E. Cartwright et al. / Thin Solid Films 518 (2010) 3422–3427
3425
transient growth of whiskers on top of the earlier film at a substrate temperature of 220 K, as we see in Fig. 6(b). These whiskers are almost certainly of hexagonal crystalline ice, whose propensity to form complex dendritic structures is familiar from snowflakes. This knowledge should forewarn us that the structures we should encounter in zones 2 and 3 of the structure zone model, at high substrate temperatures corresponding to the deposition of crystalline ice, would reflect this complexity. Nevertheless we were taken aback by the beauty of the structures we grew in these conditions, seen in Fig. 7, which were grown injecting water at a target pressure of 10 Pa for 15 min over a carbon substrate initially at 6 K. The hexagonal crystalline nature of this ice is reflected in the angles of growth of the branches. This fascinating morphology of branched whiskers is intermediate between the whiskers of Fig. 6(b) and dendritic growth. Although these forms are based on the tendency of ice to dendritic growth, seen, for example, in snowflakes [14], here the environment is significantly different. In snowflakes, dendrites generally terminate in regular hexagonal facets. To obtain these palmlike structures we deposited an ice film for several minutes; the growing ice surface becomes progressively more thermally insulated from the substrate as the film depth increases and the morphology changes correspondingly. The microscope chamber is evacuated but not thermally isolated from its surroundings, so the walls are at room
Fig. 4. Zone S, spongelike morphology, film produced with water injected with a target pressure of 10 Pa deposited on a copper substrate. (a) After 83 s of injection with the substrate at 6 K. (b) Following a further 55 s of injection, during which the substrate temperature rose transiently to 80 K.
zone in the low-temperature region of the structure zone model: zone M, or matchstick morphology. In the example of this morphology we reproduce in Fig. 5 we see very large columns tens of micrometers in diameter resulting from ice deposition over 6 min at a target pressure of 133 Pa, on a platinum substrate at 6 K and using helium as an auxiliary gas. The columns display the domed tips characteristic of matchstick morphology, and also show interesting substructures both on the tip and along their length, presumably from their growth process; their shape is biomimetic, like an icy worm. The transition from zone 1 to zone M, we can view qualitatively in terms of competitive growth leading to a fractal morphology giving way to noncompetitive growth producing a columnar morphology. These large matchsticks may have been the morphology seen by Laufer et al. [13], who described an ice film grown at somewhere between 20 and 100 K as looking like “a shaggy woolen carpet”. With regard to the temperatures involved for zones S and M, they should typically be composed of low-density amorphous ice. The zones in the structure zone model are not disjunct, but form a continuum; in Fig. 6(a) we show an example of a morphology intermediate between zones 1 and M, produced with water injected with a target pressure of 133 Pa for 13 min deposited in a brass substrate at 6 K. Upon heating the substrate in this experiment we induced the
Fig. 5. Zone M, matchstick morphology, film produced with water bubbled through with helium injected with a target pressure of 133 Pa for 6 min deposited on a platinum substrate at 6 K.
3426
J.H.E. Cartwright et al. / Thin Solid Films 518 (2010) 3422–3427
Fig. 7. Film with a dendritic morphology resembling a palm tree produced with water injected with a target pressure of 10 Pa for 15 min deposited on a carbon substrate at 6 K.
Fig. 6. (a) Film with morphology intermediate between zones 1 and M produced with water injected with a target pressure of 133 Pa for 13 min deposited on a brass substrate at 6 K. (b) Whiskers of ice form on top of this morphology during heating of the substrate as the temperature reaches 220 K.
temperature and radiate accordingly, and the film surface temperature reflects this. The temperature of the upper surface can be estimated to be close to 200 K, taking into account the low thermal conductivity of the ice and the radiation. At this temperature there is high admolecule mobility, resulting in surface diffusion (both of molecules and of heat) and sublimation. Although it is clear that this dynamic equilibrium of fluxes along and perpendicular to the surface is the basic physics involved in producing this skeleton of palmlike structures formed by highly ordered branched whiskers, we are far at present from being able to formulate a quantitative theory for their growth [14]. It may be related with the fast ice growth observed at screw dislocations nucleated by surface defects [15]. One interesting open question regarding these higher temperature zones is how the amorphous to crystalline transition at the molecular scale relates to the transition from zones S and T to zones 2 and 3 on the mesoscale. 4. Discussion It is clear that structure zone morphologies do appear in ice films. Can this knowledge contribute to understanding the physics of the structure zone model? The two axes of the structure zone model shown in Fig. 1 are the renormalized film temperature and the ion energy,
which is inversely proportional to the gas pressure in sputtering experiments. On the other hand the deposition technique employed here is simply evaporation, involving thermal energies (3/2kTb 1 eV) for the admolecules. However, the energy required to sputter a water molecule is ∼0.2 eV [16], much lower than the 10–30 eV sputtering threshold for ceramics, semiconductors or metals [2]. So we should renormalize the scale of the second axis by the sputtering threshold, in the same way as the film temperature is renormalized by the melting point for the first axis. It becomes clear then that the physical basis for the two axes is admolecule mobility induced in one case by the temperature of the film itself, and in the other by the impinging admolecules from the vapour. The latter mechanism is related in these experiments to the temperature of the vapour, and to the presence or absence of helium as an auxiliary gas in the chamber that will thermalize with the chamber walls. The different effects of the two sources for admolecule mobility leading to more or less compact structures surely arise because, unlike thermal mobility, admolecule movement induced by bombardment is highly directional. This directionality is quantifiable in terms of the sputtering yield as a function of the angle of incidence of the bombardment, which has a maximum at 45–60° [2]. This is an avenue to explore in moving towards a physical understanding of the structure zone model as a consequence of the competition between the spatially disordered deposition of particles on the growing film surface and the ordering effect of activated particle mobility processes, which will be the subject of future work. Acknowledgments This work was performed by B. E. as part of his Ph.D. under the supervision of J.H.E.C. and C.I.S.D. We thank Daniel Araujo for help with the experimental work and Russ Messier for useful discussions. We acknowledge the financial support of CSIC project Hielocris and MCyT grant CGL2008-06245-CO2-O1. References [1] R. Messier, V.C. Venugopal, P.D. Sunal, J. Vac. Sci. Technol. A 18 (2000) 1538. [2] A. Lakhtakia, R. Messier, Sculptured Thin Films: Nanoengineered Morphology and Optics, SPIE Press, 2005. [3] H.E. Stanley, S.V. Buldyrev, M. Canpolat, M. Meyer, O. Mishima, M.R. Sadr-Lahijany, A. Scala, F.W. Starr, Physica A 257 (1998) 213. [4] P. Ehrenfreund, H.J. Fraser, J. Blum, J.H.E. Cartwright, J.M. García-Ruiz, E. Hadamcik, A.C. Levasseur-Regourd, S. Price, F. Prodi, A. Sarkissian, Planet. Space Sci. 51 (2003) 473. [5] P. Jenniskens, D.F. Blake, M.A. Wilson, A. Pohorille, Astrophys. J. 455 (1995) 389. [6] V.F. Petrenko, R.W. Whitworth, Physics of Ice, Oxford University Press, New York, 1999.
J.H.E. Cartwright et al. / Thin Solid Films 518 (2010) 3422–3427 [7] P. Jenniskens, D.F. Blake, Science 265 (1994) 753. [8] A. Hornekaer, A. Baurichter, V.V. Petrunin, A.C. Luntz, B.D. Kay, A. Al-Halabi, J. Chem. Phys. 124 (2005) 124701. [9] M.E. Palumbo, J. Phys. Conf. Ser. 6 (2005) 211. [10] O. Andersson, A. Inaba, Phys. Chem. Chem. Phys. 7 (2005) 1441. [11] Z. Dohnálek, G.A. Kimmel, P. Ayote, R.S. Smith, B.D. Kay, J. Chem. Phys. 118 (2003) 364.
[12] [13] [14] [15] [16]
3427
A.F. Jankowski, J.P. Hayes, J. Vac. Sci. Technol. A 21 (2003) 422. D. Laufer, E. Kochavi, A. Bar-Nun, Phys. Rev. B 36 (1987) 9219. K.G. Libbrecht, Rep. Prog. Phys. 68 (2005) 855. K. Thürmer, N.C. Bartelt, Phys. Rev. B 77 (2008) 195425. R. Bukowski, K. Szalewicz, G.C. Groenenboom, A. van der Avoird, Science 315 (2007) 1249.