
rotation about the fourfold axes, as revealed by NMR
19
and
quasi-elastic neutron scattering studies
20
. Upon cooling to
T=208
K at ambient pressure, adamantane undergoes a
first-order structural phase transition to the tetragonal space
group
P¯
421c
(
a=6.614 ˚
A,
c=8.875 ˚
A,
Z=2
)
16
. In this
phase there is no orientational disorder: the planes of mo-
lecules alternate down the tetragonal
c
axis between the two
high-temperature orientations.
Quantity Value Reference
Tt(ambient p) 208 K 16
pt(ambient T) 4.8 kbar 21–23
∆S16.2 J K−1mol−124
∆H3.38 kJ mol−124
∆VLT→HT/VHT 7.34% 16
Table 1
Thermodynamic data of adamantane’s phase trans-
ition.
∆S
,
∆H
,
∆VLT→HT
are the thermodynamic changes of
the temperature-induced phase transition at ambient pressure.
∆VLT→HTVHT
is the volume change of the unit cell from the low-
temperature (LT) to high-temperature (HT) phase as a percentage
of the high-temperature unit cell volume.
The thermodynamic data of adamantane’s phase transition
are summarised in table 1. As is hinted by the large volume
change, the phase transition can also be induced by pressure:
at ambient temperature, this happens at a pressure of
p=
4.8
kbar
21–23
. Both the large volume change and the phase
transition temperature’s strong sensitivity to pressure are
indicators of potentially large barocaloric effects.
On top of the quasi-elastic neutron studies, inelastic neut-
ron studies have been performed in adamantane’s plastic
phase
25–27
. Here we expand upon the quasi-elastic studies
by measuring the reorientational dynamics under pressure;
the inelastic studies are extended by probing the dynamics
in both phases to reveal the phase transition mechanism.
3 Experiment
3.1 Colossal barocaloric effect
The barocaloric properties of adamantane were measured
using high-pressure differential scanning calorimetry (DSC).
Figure 2(a) and (b) show the heat flow measurements and
the corresponding phase diagram. The heat flow meas-
urements were performed at pressures between 900 and
1000 bar. At lower pressures, the phase transition tem-
perature is below the temperature range of the DSC so it
could not be observed. The phase diagram confirms the
strong pressure-induced shift of the phase transition tem-
perature
Tt
, with
dTt/d p =18.57
K kbar
−1
on cooling and
dTt/d p =19.40
K kbar
−1
on heating. It also reveals an ex-
tremely low hysteresis up to 1.96 K at 1000 bar, which varies
very slightly with pressure. By subtracting the integral of
the heat flow peaks at
p>900
bar from the integral of the
pbase =900
bar peak, we can recover the isothermal entropy
change for releasing pressure down to 900 bar (
p→900
bar)
and adding pressure starting from 900 bar (
900
bar
→p
).
This is shown in figure 2(c). The maximum pressure-induced
entropy change in this experiment is 116.92 J K
−1
kg
−1
. Fi-
nally, the reversible entropy changes are shown in figure
2(d). Starting at a pressure of 900 bar, reversible effects are
already achieved with a pressure of 50 bar to reach 950 bar.
To get an estimate of the barocaloric behaviour at low tem-
peratures and pressures, and to find the pressure at which full
reversibility is reached, the heat flow data were extrapolated
down to 0 bar. First, the phase transition temperature was
extrapolated down to 0 bar using a linear fit to the data shown
in figure 2(a). Next, the 1000 bar heat flow peak was trans-
lated to those phase transition temperatures: as an example,
the resulting predicted heat flow data for 0 bar are shown in
figure 2(b). Using the predicted 0 bar data, the isothermal
entropy change for
p↔0
bar can be estimated; the predic-
tion for
1000 ↔0
bar is shown in figure 2(c), along with
the prediction for
200 ↔0
bar (which uses the two extrapol-
ated heatflow datasets of 0 and 200 bar). Further details of
behaviour at intermediate pressures is available in the sup-
plementary material. Finally, this results in the extrapolated
reversible entropy changes shown in figure 2(d). At
<200
bar full saturation is thus expected, yielding a colossal revers-
ible entropy change of 106 J K
−1
kg
−1
. The small operating
pressures that are necessary for adamantane’s barocaloric
exploitation are very appealing, and are likely a consequence
of the extremely small hysteresis. With the reasonable as-
sumption that the
dTt/dP
relation is linear, the hysteresis at
0 bar is estimated to have a value of only about 1.15 K, far
smaller than that of some adamantane derivatives8.
The barocaloric properties of adamantane compare favour-
ably against other caloric materials. The isothermal entropy
change far surpasses that of electrocaloric and magnetocal-
oric materials (with maximum entropy changes not greater
than
∼50
JK
−1
kg
−1 4,7
). Moreover, the reversible entropy
change is also among some of the largest observed in barocal-
oric plastic crystals (carboranes:
72 −97
JK
−1
kg
−1 10
;
1-X-adamantanes:
∼150
JK
−1
kg
−1
(at 1 kbar)
8
; NPG:
389
JK
−1
kg
−1 6,7
) and crucially, due to the phase transition
temperature’s high sensitivity on pressure, adamantane’s
maximum entropy change normalised by saturation pressure
is larger than any barocaloric plastic crystal known so far,
as shown in figure 3. This means that entropy changes can
be achieved with minimal work. Although adamantane’s
low phase transition temperature makes it unsuitable for
most domestic applications, it may be an excellent candidate
for ultra-low temperature freezers used in vaccine storage
28
3