Melting and Sublimation of Planetary Ices Under Low Pressure Conditions: Laboratory Experiments with a Melting Probe Prototype

Abstract

One possibility to explore the subsurface layers of icy bodies is to use a probe with a “hot tip", which is able to penetrate ice layers by melting. Such probes have been built and used in the past for the exploration of terrestrial polar ice sheets and may also become useful tools to explore other icy layers in the Solar System. Examples for such layers are the polar areas of Mars or the icy crust of Jupiter’s moon Europa. However, while on Earth a heated probe launched into an ice sheet always causes melting with subsequent refreezing, the behaviour of such a probe in a low pressure environment is quite different. We report on the results of some experiments with a simple “melting probe" prototype with two different kinds of hot tips in a vacuum environment. For one of the tips the probe moved into two types of ice samples: (i) compact water ice and (ii) porous water ice with a snow (firn) like texture. It was also found that the penetration behaviour was basically different for the two sample types even when the same kind of tip was used. While in the porous sample the ice was only subliming, the phase changes occurring during the interaction of the tip with the compact ice are much more complex. Here alternating phases of melting and sublimation occur. The absence of the liquid phase has severe consequences on the performance of a “melting probe" under vacuum conditions: In this environment we find a high thermal resistance between the probe surface and the underlying ice. Therefore, only a low percentage of the heat that is generated in the tip is used to melt or sublime the ice, the bulk of the power is transferred towards the rear end of the probe. This is particularly a problem in the initial phases of an ice penetration experiment, when the probe has not yet penetrated the ice over its whole length. In the compact ice sample, phases could be observed, where a high enough gas pressure had built up locally underneath the probe, so that melting becomes possible. Only during these melting periods the thermal contact between the probe and the ice is good and in consequence the melting probe works effectively.

1 Introduction

A lot of information about changes in Earth’s climate was obtained from the analysis of data received from ice drill cores from the Antarctic and Greenland ice sheets (see for example Dahl-Jensen et al. 1998). A detailed analysis of different ice layers on other planets or moons may likewise provide valuable information on the climate history of these extraterrestrial bodies.

To investigate samples from other planets or moons, drilling, coring, or hammering techniques, which are frequently used on Earth, may not be the appropriate tools. They are usually heavy and complex, which makes it difficult to operate them by purely robotic means. However, there is another way to penetrate ice sheets, namely so-called “melting probes". They are robust, can be built with comparatively low weight, and can serve as carriers for various sensors. Different to drills, where the possible depth of the borehole depends on the length of the drill pipe, the penetration depth of a melting probe is only limited by the length of the tether that is providing communication with the surface station and/or the power supply.

Under Earth conditions melting probes have been tested and used for glacier research since the early 1960’s. Pilberth (1962) was the first who built a prototype of a melting probe (see Fig. 1). The penetration behaviour of cylindrically shaped melting probes in temperate glacier ice was studied at the same time by Shreve (1962). Aamot (1967) continued the work of Philberth and developed a theory for predicting the power necessary to make a probe penetrate the ice sheet with a particular speed.

Fig. 1

The Pilberth Probe used for penetrating polar ice sheets. It had a diameter of 10.92 cm and a length of 250 cm. The length of the cable was 3000 m. The power required by the probe to proceed with a speed of 2.5 m/h in compact ice at −28°C was about 5000 W. The figure was adapted from Aamot (1967)

Since then the technical features of the probes were improved. For example, Tüg (2003) worked on a computer controlled melting probe that melts itself automatically through the ice. His main interest was to find biological organisms which live in capillary tubes filled with salt water. For biological and chemical examinations like this the cores have to be melted because the organisms’ ingredients can only be found in the liquid phase of the sample (for details see www.awi.bremerhaven.de; Tüg 2003). Furthermore in this case it is necessary that the melt water is not contaminated by a drilling fluid or metallic abrasions from a drill.

Getting access to planetary subsurface ice layers with the help of melting probes is a topic that has been discussed for several years in connection with proposed Mars and Europa missions (see e.g. Paige 1992; DiPippo et al. 1999; Carsey et al. 1999; Biele et al. 2002). Also theoretical work concerning this topic was performed (Kömle et al. 2002). Furthermore, some first experiments under vacuum conditions with simple spherical prototypes of melting probes were done (Kömle et al. 2004; Treffer 2004; Treffer et al. 2006). Finally, an overview of the theoretical and experimental work with melting probes under vacuum conditions can be found in a recent article by Ulamec et al. (2007).

Since the properties of extraterrestrial ice layers are not known, a generic design for the melting probe has to be found. This design should allow that the probe maintains its direction, even when the ice is contaminated with dust particles, and also that communication with a surface station must be possible at any time.

In case of the application on a planet or moon the quality of wireless communication would heavily depend on the unknown structure and condition of the material to be explored. For this reason the use of a tether for communication purposes is preferable. Also a main point in design requirements is, that the only driving force of the probe is its own weight. The gravity on the extraterrestrial bodies of interest is lower than on Earth (e.g. Mars 3.73 ms⁻², Europa 1.31 ms⁻²).

To maintain a directional stability and a self up-righting ability it is mandatory that the center of gravity of a melting probe has to be as close to the front point of the tip as possible. Another point is that the cross-section of the melting probe has to be small to get a reasonably low power consumption.

At the vacuum chamber of the Space Research Institute in Graz and the Planetary Environments Simulation Chamber of the DLR in Cologne first tests for melting probes with different tip shapes were done. In the following sections the design of these simple prototypes and the results and conclusions from the experimental tests are reported. The experiments done are summarized in Table 1.

Table 1 Summary of the experiments done with the first melting probe prototype

2 The Melting Probe Prototype

2.1 The Ogive-Shaped Tip

For the first design of the prototype an ogive-shaped tip was chosen which is the typical shape of an artillery shell or rifle bullet. Such a tip contains more volume than a cone shaped one and still has a more prominent hot point than a hemispherical one. This form should make it possible for the tip to move through dusty ice layers and helps to put the center of gravity close to the front end. The used tip was made of brass, with a total height of 6.5 cm. Its largest outer diameter is 6.35 cm, it can be fixed to the melting probe envelope, which is a pre-fabricated stainless steel tube with the same maximum diameter. At the inside of the hot tip a cylindrical hole is milled out to a depth of 3 cm with a diameter of 5.4 cm. Thus it was possible to place two heaters inside: a ring-shaped heating foil at the bottom of the cutout, and two rectangular heating foils electrically connected in series, fixed at the inner side wall. At one outer side of the tip a notch is foreseen as outlet for cables and mounting point for external sensors and magnetic markers. Three braces evenly spaced out at an arc distance of 120° can be fixed at the hot tip. At these braces internal compartments of the melting probe can be mounted. Figure 2 shows a sketch of the ogive-shaped tip from the side and a top view where one can see the heating circuits.

Fig. 2

The ogive tip. Left: Sketch of the side view where magnetic markers can be fixed and the outlet for cables and external sensors can be seen. Right: Top view of the tip which shows the two heating circuits

2.2 The Flat Tip

The second tip manufactured for our ice melting tests under vacuum conditions was a flat blunt tip. This shape is of course a limiting case, which is probably not the optimal shape for a real melting probe head, but due to its simple geometry, tests with such a geometry can best be compared with the existing theoretical model of Aamot (1967) for predicting the penetration speed. In this case—in contrast to the ogive-shaped head—the initial penetration speed should not vary with depth, since the cross-section is constant from the beginning. As will be described in the following sections, the experiments with the flat tip probe allowed to gain some valuable insights into the behaviour of a melting probe under vacuum conditions in ices with different texture and porosity. The flat tip design used for the experiments described in the following sections made it possible to mechanically sandwich heating foils between two brass discs and a cylindrically shaped body as illustrated in Fig. 3.

Fig. 3

The flat tip. Top: Sketch of the whole tip including the cylindrical part and the two brass discs. Bottom: Exploded view of the flat tip

The diameter of the two brass discs is 6.35 cm. One of the discs has a height of 0.5 cm and the other disc one of 0.2 cm. The thinner one is sandwiched between the thicker disc and the main part of the tip. The maximum outer diameter of the main part is also 6.35 cm. Its inner diameter is 5.4 cm. The tip has a total height of 3.25 cm (space for heating foils and control thermal sensors included). The three single parts can be fixed together with the help of 6 M3 screws spaced every 60°. Also at this tip three braces evenly spaced at an arc distance of 120° can be fixed. Thus, the same upper section of the melting probe as for the ogive tip can be used. There is only one difference: the probe with the ogive tip had two heating circuits inside the tip and three inside the body of the probe. Some experiments showed that more heating energy concentrated at the tip is necessary (see Sect. 4); therefore, for the design with the flat tip three heating circuits were fixed inside the tip and two inside the body of the probe. This was in a first step done by mounting one ring-shaped commercial Kapton heating foil between the two discs and one ring-shaped heating foil placed between the thinner disc and the cylindrical body. The third heating circuit was an electrical serial connection of two rectangular heating foils fixed at the inner wall of the milled out area of the cylindrical part of the tip. Of course with such a kind of tip it is not possible to penetrate through ice containing dust grains. The flat tip was specifically designed to investigate the necessary power for melting into ice with a compact or a porous texture. In both cases we are talking about samples containing no dust particles.

2.3 Other Components of the Melting Probe and Additional Equipment

In Sects. 2.1 and 2.2 the different hot tips of the melting probe prototype were described. However, the tip is only one of five major components. The other four are: the envelope, the electronics box, the cable magazine and the rear cap. The envelope is made of a pre-fabricated stainless-steel tube and has a diameter of 6.35 cm, a wall thickness of 0.05 cm and a height of 20 cm. It includes three heating circuits each consisting of three commercial Kapton heating foils electrically parallel connected. Although the properties of the heater foils are specified for a wide temperature range, they are operated in a controlled mode to maintain the selected heater set temperature within ± 1 K. This was necessary to ensure that no electrical parts or a potential payload can be overheated. Therefore each heating circuit is equipped with a temperature sensor of type PT100. These sensors monitor the thermal state of the heater foils as housekeeping sensors. In addition, they may serve as reference points for thermal models and in a passive mode of the melting probe they may also be used as scientific sensors for the environment.

The primary purpose of the envelope is to protect the inner compartments of the probe. On the upper side the envelope is capped by a steel lid with a conical opening for the tether. The internal compartments are mounted independent of the envelope on the braces mentioned already above.

The cable magazine is a self-supporting compartment which can be handled independently. For the tests done, different tethers with a maximum length of 4 m were placed inside the magazine in ∞-shaped coils. This is a simple but robust method ensuring that the tether can be pulled out without complications.

The electronics box is another independent part with an outer diameter of 5.715 cm and a height of 6.2 cm. It is mounted in the center of the probe and interfaces with the cable magazine, the heater circuits and—in case of a space mission—the payload. At the top and the bottom the electronics box is covered by 0.1 cm thick steel plates. The payload compartment is defined by the volume below the electronics box and the void area of the tip. The single parts of the probe are shown in Fig. 4. Here the prototype is shown with the ogive-shaped tip and a brass part simulating the weight of a possible payload.

Fig. 4

Picture of the single parts of the melting probe including the ogive tip and an additional weight for the simulation of a possible payload

Another important component necessary for the melting probe experiments is the so-called launch pad. The launch pad is the base frame for the mounting of the melting probe before and at the beginning of the laboratory experiment. It provides directional stability to the probe as long as the rear part is still residing outside the ice. At the launch pad the melting probe is held by two polyethylene filaments. To launch the melting probe these filaments are fused. This is done by heating up two resistances which are placed below the filaments. After the filaments have burnt out, the melting probe drops down and gets in contact with the ice. Then the melting process begins.

3 Aamot’s Melting Penetration Theory

An estimate of the expected penetration speeds in melting probe tests can be obtained by applying the formulae given by Aamot (1967). A cylindrical probe with radius R _probe and length L _probe must be fed with the following power P to penetrate with velocity v into the ice sheet:

$$ P = P_{0}(R_{probe},v) + P_{cond}(R_{probe},L_{probe},v) $$

(1)

This total power P for tip heating consists of two components:

• (1) The power P ₀ needed to heat the ice ahead of the probe from its initial temperature T ₀ up to the melting temperature T _m . In the experiments done T ₀ has a minor variation with depth. The power P ₀ is given by

  $$ P_0 = (\pi R_{probe}^2) v \rho_{ice} [ c_{ice} (T_m-T_0) + H_m ] $$

  (2)

  where ρ _ice = 910 kgm⁻³ is the density of compact water ice, c _ice = 2093 Jkg⁻¹K⁻¹ is the specific heat capacity of water ice, and H _m = 3.35 ×  10⁵ Jkg⁻¹ is the melting heat of water ice. In this form the formula is valid in the terrestrial (normal gas pressure) case, where the ice would melt after having reached the temperature T _m . However, since our tests were performed under low environmental pressure conditions, sublimation of the ice instead of melting must be expected, unless there occurs a significant pressure build-up around the melting head. In this case H _m has to be replaced by H _s , the latent heat of sublimation, which is almost an order of magnitude larger than H _m (H _s  = 2.8 ×  10⁶ Jkg⁻¹).

• (2) The second term in Eq. 1 represents the power losses associated with lateral heat conduction from the cylindrical mantle of the probe into the surrounding ice. In Aamot’s model this contribution is given by the following double integral:

  $$ P_{cond} = {\frac{4 \lambda_{ice} (T_{m}-T_{0})}{R_{probe} \pi^{2}}}(2 \pi R_{probe})\int_{0}^{L_{probe}} \int_{0}^{\infty} {\frac{e^{-\kappa u^{2} s/v}}{u \ [J_{0}^{2}(R_{probe} u) + Y_{0}^{2} (R_{probe} u)]}}\, du\, ds $$

  (3)

  where u is the integration argument for the Bessel functions J ₀ and Y ₀ and s the spatial coordinate along the length of the probe, L _probe . The heat diffusion coefficient κ is related to the thermal conductivity λ _ice , the density ρ _ice and the heat capacity c _ice by

  $$\kappa = {\frac{\lambda_{ice}}{\rho_{ice} c_{ice}}} $$

  (4)

4 Experiment with the Ogive-Shaped Tip

With the tip described in Sect. 2.1 different laboratory experiments were done. One of them (Exp. C1) will now be discussed in more detail. The location chosen for it was the DLR Planetary Environments Simulation Chamber in Cologne. The reason for this decision was that this laboratory is equipped with a vacuum chamber with spacious dimensions and that pressure and temperature of this chamber can be controlled precisely. The chamber itself consists of an outer container that can be evacuated. This thermally isolates the inner experimentation container that can be cooled by liquid nitrogen (LN ₂). Two operation modes are possible:

• Planetary simulation: different gas atmospheres (e.g. Mars)

• Space simulation: vacuum conditions

For the experiment described in the following an ice sample of 120 cm height and a diameter of ∼37 cm was used, which was built in layers. The duration of the test was limited by the time when the chamber could be used for this experiment, and the available amount of LN ₂. Simulations of the duration of the experiment were done prior to the test with the Aamot-model. Calculations with the parameters

• melting probe radius R _probe = 3.175 cm

• melting probe length L _probe = 24.6 cm

• total heating power = 80 W (estimated from the value of the supply voltage without cable losses for the electronics and the resistance of the heater foils given by the producer)

• temperature of the ice = 243 K

predicted a total heating time for the experiment necessary to reach a penetration equal to the length of the probe of 2 days 14 h and 24 min. Previous experiments done with a smaller spherical melting probe showed that the actual measured values of the penetration speed reached only about 20–50% of the predicted value (see Treffer et al. 2006). To keep the duration time of the experiment within a period of less than 4 days, the temperature value of the ice sample was controlled to be 258 K. The mean pressure during the experiment inside the chamber was ∼30 Pa. A supply voltage of 28 V _DC for the electronics was provided which decreased to a value of 23.3V _DC at the heaters due to cable losses. Table 2 summarizes the relevant parameters for the different heaters (designated as H1–H5).

Table 2 Experiment C1: position of the heaters, their set temperature T _set , average measured temperature T _a , average power P and average duty cycles

Before the melting probe was launched it was pre-heated for 1 h. During the early times all heaters were set to a temperature of 303 K. After approximately 19 h it could be visually detected that the melting probe started to slow down and a crater was forming around it. The set temperature of the three heaters fixed in the upper part of the probe was therefore reduced to a value of 283 K. Also the surface of the ice sample was sublimating. The test had to be stopped 47 h and 27 min after its beginning because the LN ₂ reservoir was exhausted and the temperature inside the chamber began to rise.

During the whole test the melting probe was only able to penetrate the ice to a depth of approximately 15 cm. The melting probe slowed down from a speed of ∼1.5 cm/h at the beginning to ∼0.3 cm/h at the end of the test. A change in speed was not unexpected since the penetration speed is inversely proportional to the cross-section of the probe and we were using the ogive tip. But in general the speed was much lower than originally estimated.

With the values for the resistance of the heating foils and the dependence of the resistance on the heater temperature a greater power than the total measured one (P _calculated ∼64 W, P _measured ∼24 W) and therefore a higher velocity (v _calculated = 1.84 cm/h) was expected.

After the end of the experiment the sample was removed from the chamber and cut in halves for a closer examination of the crater that had formed around the probe. Figure 5 shows an image of the melting probe and the ice sample after both were taken out of the chamber.

Fig. 5

Ice sample and probe after the melting test. Left: The probe inside the excavated channel. Right: Image of the crater that builds around the probe together with a cm scale rule

The crater had a maximum diameter of about 10 cm which is approximately one and a half times the diameter of the probe.

This experiment and simulations thereafter showed that with this tip geometry, the effectively necessary heat to melt the ice cannot be sufficiently transported to the front point of the tip. Heating energy was lost via the sides of the melting probe and therefore, instead of a forward motion, the crater surrounding the melting probe increased with time. But the experiment also showed that the melting probe is able to melt into ice under vacuum conditions, at least for ice temperatures ≥ 258 K. Its low speed and the formation of the crater are an indication for a high temperature gradient inside the tip. More energy was dissipated laterally than downwards. The temperatures of the heaters inside the tip decreased rapidly at the moment when the tip got in contact with the ice. The conclusion is that the design of the tip, especially the arrangement or the kind of the heaters, has to be modified in order to allow a more effective ice penetration.

Another reason for the poor performance of the ogive-shaped probe in the Cologne melting test were relatively high losses in the cable through which the individual heaters were supplied with electrical power. This lead to the production of unwanted heat in the probe’s rear part, which was then consequently missing in the tip. This is one reason why the measured power and velocity values differed that much from the predictions of Aamont’s model. Calculations done after the experiment, which included the reduced voltage at the heaters, lead to predicted velocity values approximately twice as high as the measured ones. This result appears acceptable and is comparable with the findings reported from the earlier tests with small spherical probes in Graz (Treffer et al. 2006). In the subsequent experiments described in the following sections heat dissipation in the cables was taken into account when interpreting the results.

5 Experiments with the Flat Tip

5.1 Compact Ice Penetration

With the melting probe design as described in Sect. 2.2 several experiments under vacuum conditions were done in a much smaller vacuum chamber at the Space Research Institute in Graz. This is a cylindrically shaped vacuum chamber of 40 cm diameter. The height can be increased from 40 cm to 120 cm. It can be cooled by LN ₂ and reaches pressures down to 10⁻² Pa. As a first sample for one of these melting probe experiments, further named experiment I, a cylindrical block of clear water ice, typically used as raw material by artists to create ice sculptures, was taken. It had a height of 33 cm and a diameter of 26 cm. In this test the supply voltage for the melting probe was 35 V _DC . At the beginning of the experiment all heaters inside the tip were set to 313 K and the heaters inside the envelope had an initial temperature of 288 K. Each heater was pre-heated for 10 min. Due to losses along the cable the mean supply voltage decreased to ∼28 V _DC at the heaters during the test. The heater properties are given in in Table 3.

Table 3 Experiment I: description of the position of the heaters, their set temperature T _set , average measured temperature T _a , average power consumption P and average duty cycles

A total of 4 h and 20 min after the launch, the LN ₂ supply was stopped and the sample was further cooled by gaseous nitrogen only. The test was carried on under these conditions for 7 h and 10 min, then the experiment was terminated, because the pressure inside the chamber approached the triple point. Similar to the test with the ogive tip, in this case a crater formed around the probe too. However, this time the crater was smaller (d _crater ∼ 8.65 cm) and could only be observed at the upper 3–4 cm of the sample. Below this depth the wall of the melted channel was vertical and no gap between the probe envelope and the wall could be recognized. The probe itself reached a penetration depth of ∼21.5 cm. Notably the inner side wall of the melting channel and also its bottom was plane and smooth without any irregularly structured surface.

During the time when the sample was cooled by LN ₂ the probe reached an effective mean velocity of 0.84 cm/h. After the LN ₂ feeding was stopped and the sample was cooled by gaseous nitrogen only, the temperature of the ice sample and the temperature inside the vacuum chamber started to increase and therefore the probe penetrated faster into the ice (mean velocity ∼2.33 cm/h). This penetration speed was reached by a mean power provided by the foremost heating foil fixed between the two brass discs of ∼36 W, a mean power of ∼ 15W made available by the second heating foil located between the second disc and the cylindrical part of the tip, and a minor contribution of ∼6 W from the last heater fixed at the inner wall of the tip. The heaters located inside the envelope provided no noteworthy contribution to the measured power. A higher temperature than the set one was observed due to the power dissipation of the cable and the electronics box. Figure 6 shows the probe and the sample after they were taken out of the vacuum chamber and the penetration channel which was generated during the test. In this test the duty cycle of the foil closest to the ice surface was approximately 94% of the time at 100% (duration of the experiment: 11 h 30 min). The measured power and the temperature for the first 3 h after the launch are shown in Fig. 7.

Fig. 6

Sample and probe after the experiment. Left: The probe inside the sample. Right: Penetration channel formed during the test

Fig. 7

Measured power and temperature values during the first 3 h after the launch of the probe. In this case the heating foils are mechanically fixed inside the flat tip. Again only the values related to H1 are shown

With the probe as described above and the flat tip further tests were done. On the one hand to verify the influence of a different sample texture, and on the other hand, to clarify if marginal changes in the arrangement of the heaters can increase the penetration speed. As a first step the heating foil sandwiched between the two brass discs of the tip was removed and instead of that a foil with a radius of 6.22 cm was fixed at the front end of the tip. Under these conditions the heating foil could get directly in contact with the ice.

For this experiment, further named experiment II, an ice sample with a height of 33.5 cm and a diameter of 26.5 cm built in layers was used. The main parameters of this test are given in Table 4.

Table 4 Experiment II: position of the heaters, their set temperature T _set , average measured temperature T _a , average power consumption P and average duty cycles

Shortly after the launch the test had to be interrupted because the upper part of the ice sample cracked into rubble and the probe partially lost contact with the ice. This led to a short break during which the heaters were switched off and the surface of the sample was again prepared to get proper test conditions as for the first experiment with the flat tip. After a second launch no penetration could be observed. The probe was temporarily removed from the chamber and H1 was covered with a 0.05 cm thick copper foil. This time the heaters were also pre-heated and after 1 min the probe was released.

Approximately 22 h and 30 min after the start this experiment was stopped. The probe had melted marginally tilted into the ice and reached a penetration depth of about 10 cm, which is much more compared to the test done with the ogive tip, taking into account the shorter duration of the experiment and the fact that the ice was significantly colder than during the Cologne test (T _ice-cologne = 258 K, T _ice-graz ≃216.5 K). The tilt can be explained by the circumstance that the copper tracks of the Kapton heater foil are not equally spaced which lead to a nonuniform temperature distribution in the heater. A lower temperature is reached where the contact wires are fixed, since in this area the space between the copper tracks is wider.

As in the former test a crater formed around the melting channel. This time, it had a smaller diameter of 7.5 cm and a maximum depth of 4 cm. Again the inner side wall of the melting channel was plane and smooth and the diameter of the channel was almost exactly that of the probe itself.

Concerning the covering of H1 with a copper foil, which was necessary to obtain penetration, mathematical simulations done after the experiment showed that the thermal conductivity of Kapton is so small that a bare heater foil can not produce a homogenous temperature field. Thus at some points at the heater the temperature was too low to achieve melting or even sublimation under vacuum conditions. The results of the simulations shown in Fig. 8 indicate that over the short distance of only 107.5μm a temperature drop of about 35 K can be expected. By covering the heater with a layer of higher conducting material the temperature distribution was more even and the heat flow into the ice improved significantly.

Fig. 8

Simulation of the temperature distribution inside the tip above an ice sample. The uppermost layer represents the tip, the second layer the 0.1 mm thin glue film, the third layer the heater with 35μm thick copper tracks isolated by Kapton with a total thickness of 0.25 mm, and finally the bottom part represents the ice sample

In this experiment an interesting phenomenon was observed: H1 worked alternating at full power (duty cycle 100%) and at low power (duty cycle less than 25%). A possible reason for this change in the values of the duty cycle is that closed cavities can form below the probe and therefore the gas pressure increases. Once the triple point is reached locally, melting of the ice instead of sublimation is possible. The liquid phase provides a good thermal contact which increases the thermal flux and the heaters are forced to work at full power. During these times the duty cycle is 100% and a clear heater temperature decrease can be observed. When the cavity collapses and the liquid is lost via crevices or pores, the thermal contact between the probe and the material is again reduced and the thermal flux into the ice returns to the previous low value. As a consequence the duty cycle decreases. Over the whole duration of the test a duty cycle of 100% was observed for 24% of the time. An extract of the data of H1 for a period of 3 h, measured after a heating time of 7 h is shown in Fig. 9. This period was chosen since the change in the duty cycle and the temperature of heater H1 could be very clearly observed during this phase. At this time it also was warranted that the tip had penetrated the ice over its whole length.

Fig. 9

Test with compact ice. Upper panel: The measured power values for 3 h after a heating time of 7 h. Lower panel: The measured temperature of H1. The occurrence of melting is possible for 100% heating power

5.2 Porous Ice Penetration

Next another test (experiment III) with the melting probe in the flat tip configuration was done. In this case the influence of the texture of the sample material was of interest. For that purpose natural snow with a firn-like texture was collected and stored in a freezer. For this test a new heating foil H1 identical to that one described in Sect. 5.1 was used and again H1 was covered by a 0.05 cm thick copper foil. The temperatures of the heaters were set to the values given in Table 5.

Table 5 Experiment III: position of the heaters, their set temperature T _set , average measured temperature T _a , average power P and average duty cycles

The supply voltage for the probe without losses was again set to the maximum possible value of 35 V _DC . The heaters were pre-heated and launched 5 min later. During this test power reduction at the heaters due to cable losses were minor, because a shorter tether was used. The mean value of the voltage was ∼34 V _DC . A total of 6 h and 20 min after the launch the temperature of H1 was increased by about 15 K. Twelve hours after the beginning of the experiment the data acquisition was terminated. The sample and the probe were taken out of the chamber. Again a crater had formed around the probe. It had a larger diameter (d _crater  = 12cm) than in the case of the compact ice sample, a depth of 3 cm and it was asymmetric. A penetration depth of 11.5 cm was reached. This implies also a higher penetration speed compared to the test in Cologne, though the average temperature of the sample was significantly lower (T _ice-cologne = 258 K, T _snow-graz decreases from 250 to 183.5 K). Again the melted channel was tilted. Below the probe a layer of sintered (i.e. harder and more compact material) had formed. The depth of this layer could not be measured exactly, but the estimated average value was in the range of 1–2.5 cm.

In contrast to the test done with compact ice, no phases where H1 worked at the maximum duty cycle were observed. Though all five heaters were active during the whole test, a significant power could only be observed at H1 (mean power during the time when H1 was set to 328 K: ∼11.5 W; mean power during the time when H1 was set to 343 K: ∼14.7 W). The mean powers logged at the other heaters were below 0.1 W. No significant changes in the power distributions could be observed over the whole duration of the experiment. Figure 10 shows the measured power and temperature of heater H1 during the first 3 h of the experiment.

Fig. 10

Power and temperature measured during the first 3 h after the launch of the probe with the flat tip during the experiment with the (firn-like) snow sample for heater H1

At this test no alternating modes of heating (quasi-periodic change between good thermal contact and poor thermal contact) could be observed. We conclude that there was always poor thermal contact between probe and material, with the consequence that the mean duty cycle for H1 was only about 25% throughout; obviously only the sublimation case occurred during this test.

6 Discussion and Conclusions

The experiments done with the different tips in vacuum showed that it is important to have a good thermal connection between the heat source and the ice. While this is not a critical item for melting probes operating under terrestrial (atmospheric) conditions, our experiments showed that it is a very critical point when the probe has to operate in a low pressure environment as the near surface layers of Mars and Europa.

The tests with the ogive tip showed that this shape and the used arrangement of the heaters inside the probe was not sufficient to ensure that losses via the side walls are negligible. Also the reduced supply voltage at the heaters due to cable losses lead to a poor performance.

With the aid of the tests done with the flat tip it could be shown that although the heater—in our case a Kapton foil with enclosed copper conductors—is in contact with the ice, not enough energy can be transferred in a homogeneous way to heat the ice everywhere to the melting point and achieve a reasonable penetration speed. But even a thin metal cover sheet between ice and heater foil can clearly improve this situation. The best way is a tight mechanical fixation of the heater foils sandwiched between good conductors, as it was chosen for experiment I. For such a configuration the best results in terms of heat transfer were obtained.

A comparison between the velocity values observed in experiment I and the velocity obtained from the Aamot-model is given in Fig. 11. Additionally Fig. 11 shows the pressure inside the vacuum chamber and the temperature of the ice sample. For technical reasons both values could not be measured directly at the position of the hot tip. For the values to the right of the vertical line marked in each subplot temperature values were at or above the triple point of ice and melting below the probe occurred globally.

Fig. 11

Results of the test done with the flat tip, where the heating foils were sandwiched between the single parts of the tip. Upper panel: Comparison between the measured and the calculated velocity values. Middle panel: The chamber pressure measured during the test. Lower panel: Temperature of the ice sample during the test. The temperature sensors were placed close to the boundary of the sample. On the left side of the line until t = 255min the sample was cooled by liquid nitrogen (LN ₂) while on the right side only cold N ₂ gas was used for cooling the chamber

To achieve better penetration speed in vacuum another solution for transferring the heating energy from the probe to the ice has to be found. Mathematical simulations showed that the use of Alunit heaters, which have a much better thermal conductivity than Kapton (see www.ceramtec.com) would make the temperature distribution at the outer surface of the tip more homogeneous. This should result in a better heat transfer to the ice and consequently a better performance of the probe. The disadvantage of Alunit is that it is a ceramic material and therefore not as flexible as Kapton. Due to this, the shape of the tip has to be changed from ogive or cylindrical to pyramidal.

From the experimental results of the last two tests one may conclude that in a low pressure planetary environment the type of penetration behaviour strongly depends on the texture of the surrounding ices. On the Martian poles and on the ice crust of Europa the very surface is expected to have a porous texture (similar to the firn-like sample used in our last test). Thus a melting probe without additional mechanical penetration tools like mole or drilling devices would work quite ineffectively, which results in a slow penetration speed, since it would work in the sublimation mode only. The water vapour created below the heated tip can escape through the pores of the surrounding material. Thus, the vapour pressure around the melting head would remain too low to allow temporary local melting. However, once the probe has penetrated over a metre or more, one can expect that the efficiency of the probe improves significantly, because the surrounding ice becomes less porous and a pressure build-up above the triple point of water ice may become possible. This would be comparable to the case of our compact ice experiment.

With further penetration into deeper compact ice layers the penetration mode of the probe would become similar to the behaviour observed on Earth, since the penetration channel would probably close by recondensation and refreezing of the water vapour and/or liquid water above the probe. At which depth this occurs is still a matter of debate and it needs more large scale and long duration experiments to clarify this question.

As mentioned in Sect. 2.2, up to now we only used samples without dust inclusions for our experiments. Although, as stated in the Introduction, melting probes may be a good alternative to drills and hammering devices, it has also to be kept in mind that the application of a pure melting probe without additional capabilities becomes problematic, as soon as the ice contains a certain amount of non-volatile (dusty or rocky) material. In this case dust particles might accumulate ahead of the probe and hinder further penetration or the probe might even get stuck when approaching larger boulders.

On the other hand, the existence of dust inclusions in planetary near surface ices must be considered as quite common. For example, it can be seen from Phoenix lander images that the Martian polar ice is mixed with and to some extent covered by dust. At and below the surface of Jupiter’s moon Europa the situation might be similar.

Taking these considerations into account, the present study should rather be considered as a first step into the development of a more multi-functional probe which would include both the melting capability and mechanical means to penetrate frozen ice/mineral layers. A simple way to overcome dust inclusions might be to include a vibration mechanism, which would allow the probe to divert dust particles and could to some extent avoid the accumulation of dust ahead of the probe. Another concept, namely a melting device including a drill with two counter-rotating blades, has been recently investigated by Weiss et al. (2008). A further idea worth to be considered is to combine the concept of a “mole" (as used for example for the currently developed HP³ experiment aboard ESA’s ExoMars mission, see e.g. Richter et al. 2006) with a melting probe. In the ongoing development, which is presently in progress, the above-mentioned ideas to combine melting and drilling capabilities will be pursued.