Available via license: CC BY 3.0
Content may be subject to copyright.
Journal of Physics: Conference Series
PAPER • OPEN ACCESS
A high precision calorimeter for hunting the sterile neutrino in the SOX
experiment
To cite this article: L.Di Noto et al 2020 J. Phys.: Conf. Ser. 1342 012015
View the article online for updates and enhancements.
This content was downloaded from IP address 179.61.181.168 on 18/01/2020 at 00:32
Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd
XV International Conference on Topics in Astroparticle and Underground Physics
Journal of Physics: Conference Series 1342 (2020) 012015
IOP Publishing
doi:10.1088/1742-6596/1342/1/012015
1
A high precision calorimeter for hunting the sterile
neutrino in the SOX experiment
L.Di Noto1, M.Agostini2, K.Altenm¨uller2, S.Appel2, A.Caminata1,
L.Cappelli3, R.Cesereto1, S.Farinon1, M.Gschwender4, H.Hess2,
J.Martyn5, R.Musenich1, B.Neumair2, M.Nieslony5, T.Lachenmaier4,
L.Oberauer2, M.Pallavicini1, L.Papp2, C.Rossi1, S.Rottenanger4,
S.Sch¨onert2, G.Testera1, A.Trantel4, S.Weinz5, M.Wurm5,
S.Zavatarelli1
1Dipartimento di Fisica, Universit´a degli studi e INFN, 16146 Genova, Italy
2Physik-Department and Excellence Cluster Universe, Technische Universit¨at M¨unchen,
85748 Garching, Germany
3INFN Laboratori Nazionali del Gran Sasso, 67010 Assergi (AQ), Italy
4Eberhard Karls Universit¨at T¨ubingen, 72072 T¨ubingen, Germany
4Johannes Gutenberg-Universit¨at Mainz, 55099 Mainz, Germany
E-mail: lea.dinoto@ge.infn.it
Abstract. A thermal calorimetric apparatus was designed, built and calibrated for measuring
the activity of the artificial 144C e −
144 P r antineutrino source. This measurement will be
performed at the Laboratori Nazionali del Gran Sasso in Italy, just before the source insertion
in the tunnel under the Borexino detector and a precision better than 1% is required for a
disappearance technique measurement in the SOX (Short distance neutrino Oscillation with
BoreXino) project. In this work the apparatus is described and the most important results
from the calibration measurements are shown, where the final precision of few per thousand is
demonstrated.
1. Introduction
Since the neutrino anomalies can be explained by the existence of a sterile neutrino and they
hint to the short distance neutrino oscillations ( δm2= 1 eV2)[1], the SOX experiment [2] was
proposed with the goal of using a 144 Ce−144 P r antineutrino artificial source for a possible direct
observation of the short distance oscillations and at the same time, for a precise disappearance
experiment, if the source activity is measured better than 1% precision. The 144C e-144 P r
artificial antineutrino source of ∼150 kCi activity (1200 W) is being produced at the Mayak
Production Association in Russia after a chemical extraction from spent nuclear fuel. After
the transportation to the Laboratori Nazionali del Gran Sasso, before the insertion in the
tunnel under Borexino detector, the calorimetric measurement will occur thanks to two different
calorimeters (the INFN-TUM calorimeter and the CEA calorimeter). In this work the INFN-
TUM apparatus is presented.
XV International Conference on Topics in Astroparticle and Underground Physics
Journal of Physics: Conference Series 1342 (2020) 012015
IOP Publishing
doi:10.1088/1742-6596/1342/1/012015
2
Figure 1. At left: the section view of the calorimetric apparatus; the main components are
shown. At right: picture of the apparatus before the tungsten alloy shield is lowered inside the
copper heat exchanger. In the copper heat exchanger the water pipe is clearly visible.
2. Apparatus description
The radioactive source is made of few kilograms of CeO2powder pressed into a two double
wall stainless steel container, closed inside a tungsten alloy biological shield, with a minimum
thickness of 19 cm (see Fig. 1). Since the shield debars a precise and direct βor γspectroscopic
activity measurement, a calorimetric apparatus was designed in order to estimate the heat
released from the source that is directly linked to the total activity. In the INFN-TUM
calorimeter the thermal heat is absorbed by the water flowing inside a pipe, embedded in a
copper heat exchanger, directly in contact with the tungsten shield and by measuring the mass
flow ˙mand by knowing the temperature of the entering Tin and the outgoing Tout water respect
with the copper exchanger, the power Pemitted by the source can be obtained by the relation:
P= ˙m[h(Tout, p)−h(Tin , p)] (1)
where h(T, p) is the water enthalpy function, which depends on the temperature and on water
the pressure pin the pipe line, and tabulated with about 0.1% accuracy by the International
Association for the Properties of Water and Steam in [5].
The apparatus was designed in order to minimize all the heat losses and in particular the
convection was reduced thanks to a big vacuum chamber, where a residual pressure of the order
of p∼10−5mbar can be achieved by a scroll and a turbo molecular pump. In addition, the
source and the copper heat exchanger were placed on a platform kept suspended by three Kevlar
ropes (whose heat capacitance is low) in order to reduce any conduction between the copper
exchanger and the external vacuum chamber. Finally in order to minimize the radiation, two
stages of super insulators (10 layers each) were built and located between the copper exchanger
and the vacuum chamber and a water pipe was soldered on the chamber surface for allowing the
chamber thermalization to a temperature closer to the average copper exchanger temperature.
The apparatus here described is shown in Fig.1, where all the main components are visible in
the section view.
The water circulating inside the copper exchanger is pumped and cooled by a chiller, the
massflow is measured by two Coriolis flowmeters (0.05% precision) and the flow is regulated and
actively controlled by a pneumatic valve in a feedback loop with the flowmeter. Just before the
inlet and the outlet of the copper exchanger the water temperature is measured by two Semi-
Standard Platinum Resistance Thermometers immersed in the water line and read out by a high
precision thermometer, ensuring a final precision of 3 mK on the water temperature. Thanks
to an active Proportional, Integrative and Derivative control based on a 150 W heater in a
XV International Conference on Topics in Astroparticle and Underground Physics
Journal of Physics: Conference Series 1342 (2020) 012015
IOP Publishing
doi:10.1088/1742-6596/1342/1/012015
3
feedback loop with the thermometer, the temperature of the entering water can be stabilized up
to 0.04◦C of maximum oscillation around the average value, compensating any environmental
fluctuation or chiller instability during the measurements, and preventing any instability also
on the Tout temperature.
In addition a water loop was installed in order to heat the vacuum chamber to the desired
temperature; in this case a centrifugal pump and a 1000 W heater allows the hot water
to circulate with an high value of massflow (around 30 g/s for reducing to few degrees the
temperature gradient on the chamber) inside the stainless steel pipe soldered on the external
surface of the chamber. Thanks to a dedicated feedback loop based on the temperature sensor
on the chamber, a temperature of the order of 20−35◦C can be set, with a maximum fluctuation
around the average value of about 0.4 C.
During the calibration phase, the INFN-TUM calorimeter was tested with a mock-up source
made by six electrical heaters inserted in a copper cylinder, closed inside the tungsten alloy
shield. A power ranging from 700 W up to 1200 W was set with 0.04% uncertainty and the set
value was obtained by measuring the current flowing in the circuit and the voltage as applied
just close to the stainless steal flange on the top of the tungsten shield. The electrical power was
continuously controlled by regulating the applied voltage in order to compensate any heating
effect on the heaters resistance during the measurement or to reproduce the source decay during
the time.
3. Calibration measurements results
The calibration measurements were divided in two groups: in the first phase the optimization
of the system parameters was performed in order to evaluate the losses in different conditions,
whilst in the second phase the behavior to an exponential decaying power simulating the source
was investigated. For each measurements the temperatures in many points inside the apparatus
were acquired continuously and the losses were investigated as a function of the temperature
distribution. The parameters that can be set are:
•the massflow value ˙m, that ranges between 5 and 13 g/s: in fact values lower than 5 g/s
generate too high Tout value, inducing the formation of bubble in the water, whilst values
higher than 13 g/s produce a positive (bigger than 0.4 W) spurious power due to the friction;
•the temperature of the vacuum chamber Tch that can be increased of about 10◦C above
the environmental temperature;
•the temperature Tin , ranging between 10 and 20◦C and it was chosen around 16◦C, just
few degrees lower than the environmental temperature for allowing the temperature control.
3.1. Calibration with a constant set power
Many measurements were performed in order to estimate the losses and in order to optimize the
massflow value and the chamber temperature. For each measurement a constant power was set
in the range 700-1200 W and all the temperature sensors were acquired during the time. Since
the big mass of the shield, the heat capacity of the tungsten and the thermal contacts between
all the components, from the inner copper mockup source through the tungsten alloy shield and
to the external copper heat exchanger, a transient period of about 3 days was found, during
which all temperature sensors increase their values up to reach a stable value in the equilibrium
condition. The stability was evaluated by calculating the incremental ratio of Tout during a time
period of 3 hours and only the phase in which the ratio was lower than 20 mK/h was selected
for the analysis. For each measurement the power was calculated by the equation 1, where
Tin,Tout and ˙mare the average values as extracted during the equilibrium condition, of more
than 4 hours duration. Since the Tin and ˙mare controlled by the feedback loop, they fluctuate
around the average value and the accuracy of the instruments (3 mK for Tin and 0.05% for ˙m)
XV International Conference on Topics in Astroparticle and Underground Physics
Journal of Physics: Conference Series 1342 (2020) 012015
IOP Publishing
doi:10.1088/1742-6596/1342/1/012015
4
was taken as error, while for Tout the real fluctuation during the equilibrium condition (around
10 mK and different for each point) was considered. In addition the accuracy on the enthalpy
h(T, p) knowledge (∼0.1%) as tabulated in [5], was considered and it contributes significantly
to the final error that is of the order of 0.2%.
The losses were studied by varying the parameters of the system as the mass flow or the
temperature of the vacuum chamber Tch with respect to the copper heat exchanger Texc, for
different set powers. As it results from the plot shown in Fig.2a) the losses are influenced
mainly by the difference Texc −Tch and they are always lower than 0.4% even in the case when
Texc −Tch ∼16◦C, but it can be close to zero if Texc −Tch <8◦C. It is worth noting that the
massflow value directly influences the Tout and Texc temperature and thus it has to be chosen
accordingly to the set power in order to obtain a Tout value around 40◦C, in such a way that
Tch can be set around 30◦C. Since the losses are induced by the temperature distribution, the
massflow can be increased for higher power without increasing the total losses and by comparing
measurement at different powers and different massflows it can be concluded that a massflow
value close to 12 g/s has to be chosen if the power released by the source is expected to be 1200
W.
The losses were estimated also with a pressure of the order of 10−3mbar (instead of 10−5mbar)
achieved by switching off the turbo molecular pump. In that case they resulted of the order of
0.6% but they were lowered to 0.4% when the difference Texc −Tch was reduced as well.
From these measurements the massflow and the chamber temperature were optimized in such
a way that if the residual pressure is of the order of 10−5mbar, the losses are always lower
than 0.2% and, since the applied value was always found inside the statistical error of 0.2%, no
additional systematic uncertainty has to be introduced.
3.2. Calibration with an exponential decaying power
In order to simulate the radioactive 144Ce −144 P r source a power function
P(t) = P0e−t
τCe (2)
where τCe = 411 days is 144 Ce lifetime was applied. In this case the equilibrium condition is
never reached since the generated heat is measured with a delay dependent on the global thermal
diffusivity of system and on the decay time τCe. In practice, due to the massive biological shield
and to the several interfaces, the delay cannot be a priori neglected and it was estimated to
be about 0.2 days, by means of many finite elements simulations performed during the design
phase [4]. The goal of these tests was to evaluate its influence on the final power estimation.
In Fig.2b) the measured power is shown as a function of the time and the transient phase
of 3 days is however visible, after which the system reaches the final phase where the measured
power behavior follows the negative exponential function as well. In principle, assuming that the
losses are really close to zero, the measured power is expected higher respect with the applied
power, due to the delay that shifts the measured power in the time scale, but in the real case
either this effect or the vertical shift due to losses, that decreases the measured power value, can
happen and both can contribute to the final measured power value as described in the equation
3.
Pm(t) = P0e−t−δt
τCe −Ploss (3)
For these measurements the optimized parameters for massflow and Tch, as resulted from the
last section, were set and the measured power was calculated during the time, following the
equation 1, where in this case Tin and ˙mare the average values, and Tout(t) is the value as
acquired at each time.
A plot of Pm(t) is shown in Fig.2 and the difference between the set power and the measured
power during the time is shown in the inset. The exponential function of eq. 2 was fitted both
XV International Conference on Topics in Astroparticle and Underground Physics
Journal of Physics: Conference Series 1342 (2020) 012015
IOP Publishing
doi:10.1088/1742-6596/1342/1/012015
5
Figure 2. a): the percentage losses as a function of the temperature difference between Texc and
Tch. b): The measured power (in red) and the power applied to the electrical heater (in black)
are shown as a function of the time. The yellow line corresponds to the fit function done by using
the eq.2, performed both on the set and on the measured power. The green line corresponds to
the error band of 0.2% on the measured power. In the inset the difference between the applied
and the measured power during the time is shown.
on the applied power and the measured power and the P0value as resulted from the fit on the
last part of the measured power was taken as average value for the final power estimation. As a
confirmation that the losses are really small, it is worth noting that the measured power value
is higher than the set value due to the delay effect that dominates over the losses contribute.
Assuming that the losses are close to zero, the delay time can be estimated of the order of 0.1-0.2
days in agreement with the simulations results and since the difference between the set and the
measured power is of the order of 0.3-0.4 W (0.03% of the measured power), the effect is well
below the statistical uncertainty and it can be neglected.
Many measurements were performed in the optimized conditions and with different power
values and in all the cases the set value was found well inside the error band of 0.2% precision,
evidencing that no additional systematic error has to be added.
4. Conclusions
A calorimetric apparatus here described was designed and built for measuring the 144C e −144 P r
activity with 1% accuracy. The calibrations performed with an electrical source, replacing the
radioactive source, demonstrated that the heat losses in the optimized condition are close to zero
and can be neglected. In particular, when the residual pressure is of the order of 10−5mbar and
when the massflow value and the chamber temperature are properly set, the exponential decaying
behavior can be recognized in the measured power and the final precision of the measurement
was found to be 0.2%.
References
[1] C. Giunti et al, Phys. Rev. D 88, (2013) 073008
[2] G. Bellini et al (Borexino collaboration), JHEP 08 (2013) 038
[3] J. Gaffiot et al, Phys. Rev. D 91 (2015) 072005
[4] Farinon et al., Thermal analysis of the antineutrino 144Ce source calorimeter for the SOX experiment,INFN-
16-08/GE (2016)
[5] The International Association for the Properties of Water and Steam, Revised Release on the IAPWS
Industrial Formulation 1997 for the Thermodynamic Properties of Water and Steam, R7-97 (2012),
(http://www.iapws.org),
