Criticality in the slowed-down boiling crisis at zero gravity

Boiling crisis is a transition between nucleate and film boiling. It occurs at a threshold value of the heat flux from the heater called CHF (critical heat flux). Usually, boiling crisis studies are hindered by the high CHF and short transition duration (below 1 ms). Here we report on experiments in hydrogen near its liquid-vapor critical point, in which the CHF is low and the dynamics slow enough to be resolved. As under such conditions the surface tension is very small, the experiments are carried out in the reduced gravity to preserve the conventional bubble geometry. Weightlessness is created artificially in two-phase hydrogen by compensating gravity with magnetic forces. We were able to reveal the fractal structure of the contour of the percolating cluster of the dry areas at the heater that precedes the boiling crisis. We provide a direct statistical analysis of dry spot areas that confirms the boiling crisis at zero gravity as a scale-free phenomenon. It was observed that, in agreement with theoretical predictions, saturated boiling CHF tends to zero (within the precision of our thermal control system) in zero gravity, which suggests that the boiling crisis may be observed at any heat flux provided the experiment lasts long enough.

Figure 1

Residual gravity map calculated for HYLDE containing a ferromagnetic insert for $I_h=65.55$ A of (a) axial, (b) radial components of the force $-{\vec{g}}^{*}/g$, and (c) its modulus $g^{*}/g$.

Figure 2

Experimental cell: (a) photo, (b) drawing showing two way optical cell observation paths, and (c) more detailed drawing of the cell alone.

Figure 3

Photos of the phase distribution in the lower part of the cell at magnetic gravity compensation [(a)–(c), side view] and the corresponding sketches. The exact compensation point is shown in the sketches with a dot. The gray color corresponds to the liquid, while white indicates the gas. (a) Compensation point in the middle of the cell at zero heating; the residual gravity at the heater level is directed downward. The liquid column does not touch the heater. (b),(c) Compensation point is at the heater level during boiling $(0g$ situation): (b) before and (c) right after the boiling crisis. Figure (d) sketches a situation similar to (a) but with the column touching the heater.

Figure 4

Left column: boiling of ${\mathrm{H}}_2$ at $T_c-T=10$ mK and $0g$ seen through the transparent heater. Images (a)–(c) correspond to the nucleate boiling far from BC at heater current $I_h=0$, 2, and 5 mA, respectively, whereas images (d)–(f) show the BC dynamics under $I_h=7$ mA, $290/30\mathrm{th},20/30\mathrm{th}$, and $0/30\mathrm{th}$ s before the dryout, respectively (see Supplemental Material [47]). Right column: dry spot size distributions of the corresponding left-hand images. Dashed lines correspond to $h\left(s\right)\propto s^{-1.55}$.

Figure 5

Boiling crisis dynamics in ${\mathrm{H}}_2$ at $T_c-T=78$ mK, $I_h=19.5$ mA, and $0g$ seen through the transparent heater. The images correspond to (a) 37/30th s, (b) 14/30th s, (c) 6/30th s, and (d) 1/30th s before the heater dryout.

Figure 6

Boiling crisis dynamics in ${\mathrm{H}}_2$ seen through the transparent heater at $T_c-T=17$ mK, $I_h=7$ mA, and $0.02g$. The images correspond to (a) 489/30th s, (b) 384/30th s, (c) 213/30th s, (d) 27/30th s, (e) 23/30th s, (f) 9/30th s, (g) 6/30th s, and (h) 4/30th s before total heater dryout. The arrow shows the diameter of the portion of heater area in contact with the liquid that decreases with time.

Figure 7

Image (a) and its corresponding processed snapshot (b) where white denotes vapor phase. In this case $I_h=7$ mA. Panels (c)–(f): box counting method. For more details, see the text.

Figure 8

Top panels: density of the bubbles with size $s>0.011{\mathrm{mm}}^2$ (left axes, black solid curves) and $s>1.1{\mathrm{mm}}^2$ (right axes, red dotted curves) as functions of time (a) and boiling regime (b). Bottom panels: vapor fraction (left axes, black solid curves) and mean dry spot size (right axes, red dotted curves) as functions of time (a) and boiling regime (b). Panels (a) correspond to the case $I_h=7$ mA, whereas panels (b) also include the data for the nucleate boiling at $I_h=0$, 2, and 5 mA [corresponding to Figs. 4–4].

Figure 9

Box counting dimension (BCD) of dry spot perimeters for the experiment at $I_h=7$ mA corresponding to Figs. 4–4. (a) BCD averaged over the whole evolution as a function of the dry spot area. (b) BCD averaged over dry spots area $s>0.011{\mathrm{mm}}^2$ as a function of time. Data for the experiments at $I_h=2$ mA and 5 mA [i.e., for the snapshots shown in Figs. 4 and 4] are also included.