Very cold and ultra cold neutron sources for ESS

1.The goals of the HighNESS project

The design of sources of slow neutrons for the European Spallation Source (ESS) has evolved over the years, and among other advances, led to the concept of low-dimensional moderators of high brightness [36]. While the initial configuration [43] placed neutron moderators above and below the spallation target (a 2.5-m diameter rotating wheel made of tungsten), the design optimization resulted in a single moderator system for the initial suite of instruments. For maximum versatility, this moderator system consists of a “bi-spectral” assembly of a water moderator at ambient temperature for thermal neutrons, and a liquid parahydrogen moderator for cold neutrons. Its peculiar shape (the so-called “butterfly”) resulted from the design goal to have thermal and cold neutrons available at all beamports of the facility. This “upper moderator” will be placed above the spallation target and will feed all fifteen instruments of the initial suite dedicated to neutron scattering [3]. Thanks to this optimization, the space below the target remains available for additional neutron sources. This, together with the fact that the initial instruments will occupy less than half of the forty-two beamports that will be available at ESS, offers important upgrade possibilities, which are explored within the HighNESS project [49,50].

HighNESS started in October 2020 as a three-year project, funded by the European Commission, with the goal of performing a design study of a second cold neutron source at ESS. This additional source will be placed below the spallation target, and will possess properties complementary to the upper moderator. The latter was optimised for high brightness, emitting neutrons from a small surface of only ≈6 cm width and 3 cm height. It will serve for applications that need a high flux density, notably the study of small samples of typically a few mm³ size, with moderate beam divergence. In contrast, the lower moderator will be optimised for high intensity (brightness integrated over emission surface), to facilitate applications where the total rate of neutrons emitted from the moderator surface is primordial, even at somewhat reduced brightness. The moderation or conversion of cold neutrons in secondary, colder neutron sources, discussed in later sections of this paper, is such a case where a large and intense primary cold source is crucial. Also, for the neutron instruments, a higher total emitted neutron beam intensity is sometimes preferable over a high neutron flux within only a small area. Some notable examples from particle physics that can profit from large and divergent neutron beams include a search for neutron-antineutron oscillations and measurements of the static electric dipole moment of the neutron [1,11,45]. Nested mirror optics [21,66] seem particularly efficient for delivery of such beams and will be discussed for an “in-beam” ultracold neutron source [8,17,44] in a separate contribution to these workshop proceedings [67].

A neutron spectrum colder than the one delivered by the bi-spectral upper moderator is a second design criterion for the lower source. Besides cold neutrons (CNs), a focus of HighNESS is on Very Cold Neutrons (VCNs, 10–120 Å wavelength range) and Ultra Cold Neutrons (UCNs, λ>500A˚). In most of the configurations discussed in this paper, the central, lower moderator acts as a primary CN source and will directly feed several instruments. The moderator will also serve secondary VCN and UCN sources that, dependent on the concept, can be situated in close proximity or more distant from the primary source. In this context, the important aspects to be considered are the moderation efficiency and resistance of the source material against radiation damage, the heat load on the secondary source, and an efficient cold neutron transport between the primary and the secondary sources.

The discussion of new ideas and possible configurations of intense sources for VCNs and UCNs was one of the main objectives of this workshop. While the first half of the HighNESS project was mostly dedicated to the design of the lower high-intensity CN moderator, these ideas shall be investigated in depth in the remainder of the project. This paper outlines conceptual designs for UCN and VCN sources discussed in the workshop, and their potential implementation using the ESS infrastructure. In the paper, geometries of the ESS source implemented in MCNP version 6.2 [61] are shown. MCNP was used to calculate the CN fluxes, which serve as input for further investigation of the VCN and UCN concepts presented.

2.The high-intensity cold moderator

In a high-power facility, the choice of the moderator material for a cold source is in practice limited to liquid hydrogen and liquid deuterium (LD₂), which are well-established materials for cold moderators exposed to the strong radiation fields near the core of primary neutron production in high-flux research facilities, such as SNS, J-PARC, PSI, and ILL.

Liquid hydrogen is an excellent choice for pulsed sources. For parahydrogen, where the two hydrogen nuclei per molecule couple to zero total spin, (see Ref. [56] for the allowed combinations of nuclear spin, vibrational and rotational states for hydrogen and deuterium, and the corresponding ortho-para designations), a peculiar energy dependence of the neutron scattering cross section leads to a high transparency for CNs. As a result, most of the inelastic scattering events that cool the thermal neutrons entering the liquid parahydrogen occur within about 1–2 cm from the moderator surface. This property is at the basis of the concept of high-brightness low-dimensional moderators [36], which have been adopted for ESS. As discussed in [63], and references therein, it was found that, near the wide thermal premoderator above the ESS spallation target, a thin, flat parahydrogen moderator generates the highest CN brightness (see Fig. 1, left). With increasing moderator height, the brightness steadily decreases, whereas the total rate of CNs escaping from its emission surface increases. However, beyond a few centimeters, this total CN intensity saturates (Fig. 1, right). The choice of thickness made for the upper moderator (3 cm) was the result of a comprehensive analysis work that took into account the brightness and the brightness transfer to samples for the full ESS instrument suite [4].

Compared to liquid hydrogen, LD₂ has a smaller scattering cross section for thermal neutrons and a lower absorption cross section. The slowdown of thermal neutrons is therefore possible in a larger volume, but the larger mean free path also leads to longer tails in the neutron pulses. LD₂ thus offers a complementary option for CN production, where the main advantage is its capability to produce a larger total CN intensity. In fact, a preliminary study performed on the ESS geometry [27], showed that a large LD₂ moderator can give an intensity about 3 times higher than liquid parahydrogen.

Fig. 1.

Illustration, how the thickness of a parahydrogen moderator located above a thermal neutron moderator impacts its CN brightness and intensity. Left: distribution of CN brightness (integrated time average for the neutrons with energy < 20 meV) at the moderator surface for a thin (1.5 cm) parahydrogen moderator (upper figure), and a 10 cm tall moderator (lower figure). Right: Comparison between CN brightness, total number of emitted neutrons, and heat load in a parahydrogen moderator, as a function of the moderator height. Figures adapted from Ref. [63].

The high-intensity moderator based on LD₂ has been designed in the first year of the HighNESS project. The model is shown in Fig. 2. There are two openings for neutron extraction, a large one (40 cm wide, 24 cm high) for the NNBAR experiment [1], designed to improve on an earlier search for neutron-antineutron oscillations [6], and a smaller one (15×15 cm2) for neutron scattering instruments. A comparison of the expected performances of the lower and upper ESS moderators shows their complementarity, see Table 1 and Fig. 3. For the spectra integrated above a short-wavelength cutoff at 2 Å and 4 Å, respectively, the upper moderator is found to have the largest brightness. However, for the long-wavelength part of the spectrum, i.e., with the cutoff set at 10 Å, the 15×15 cm2 opening has almost the same brightness as the upper moderator. Due to the large emission surfaces of both its openings, the lower moderator provides a much higher intensity for all studied spectra.

Of particular interest is also the performance of the lower moderator for 8.9 Å neutrons, which are needed for the CN conversion to UCN in superfluid ⁴He, also called He-II. The brightness of the lower moderator at the NNBAR opening at this wavelength, at 5 MW average power, is about 3.4×1011 n/cm2/s/sr/Å. This number serves as key input to calculating the neutron flux incident on an in-beam UCN source, see [67] for more details.

Fig. 2.

Horizontal cut view through the MCNP model of the high-intensity LD₂ moderator. The moderator, placed below the spallation target, is 45 cm wide (perpendicular to the direction of neutron extraction), 47 cm long, and 24 cm high. There are two windows for neutron extraction, a large one (40 cm wide, 24 cm high) on the NNBAR side, with a cold Be filter to increase the flux above 4 Å (see discussion in Section 4.2.1); on the other side, there is a reentrant hole with dimensions of 15×15 cm2, for neutron scattering experiments. The design is still in progress.

Fig. 3.

Comparison of spectral brightness (left) and intensity (right) of upper and lower moderators, at 5 MW average power. The spectra for the upper moderator are the average over the 42 beamports. For the lower moderator, preliminary spectra are shown for two openings, i.e., for NNBAR and neutron scattering experiments. The calculated intensities refer to the emission windows with sizes specified in the caption of Table 1. Note the effect of the Be filter/reflector on the spectra for the NNBAR opening.

3.Basic concepts and possible locations of UCN sources

The most longstanding user facility for UCN physics, PF2 at the Institut Laue-Langevin, extracts the lowest-energy fraction of neutrons leaving a LD₂ moderator [58]. More recently commissioned UCN sources and the majority of current source projects are based on “superthermal” UCN production [16], using either solid deuterium (SD₂) or He-II as a material for CN conversion to UCNs in direct down-scattering events. The concepts under consideration for ESS are in line with these developments. In contrast to moderation, in a superthermal source, the neutrons stay out of thermal equilibrium with the colder medium, wherein Boltzmann suppression of UCN up-scattering enables buildup of a higher UCN density. To obtain the highest UCN production rates, the energy spectrum of the incident neutrons has to match the phonon spectrum of the medium. For SD₂, existing LD₂ CN sources are close to optimal, providing spectra that correspond to a neutron temperature, Tn, of a few tens of K [20]. For He-II, the optimal CN temperature is significantly lower, as the dominant UCN production mechanism is inelastic scattering with emission of a single phonon of energy near 1 meV, which requires neutrons with wavelengths in a narrow range around 8.9 Å, for which Tn=6 K would be ideal [17]. Multi-phonon processes, which occur at shorter wavelengths, contribute less to UCN production than the single-phonon process. Even at the higher Tn delivered by a standard CN source, at which the neutrons inducing multi-phonon processes have a larger relative weight than at 6 K, the multi-phonon contribution still does not exceed 30% [5,52].

Solid D₂, prepared in the molecular rotational ground state and kept at temperatures below 10 K, is most effectively used “in-pile”, that is, near an intense primary neutron source, where it produces a high rate of UCNs (see Refs. [12,24,26,28–30,32,34,51,53,55] for operating, demonstrated and proposed SD₂ sources). If the UCN source consists of only a small quantity of SD₂, thus acting purely as a neutron converter, the incident neutrons need to be pre-moderated by an external CN source, typically of LD₂. Alternatively, a large block of SD₂ can serve simultaneously for both pre-moderation and final conversion to UCNs. This solution has been adopted at the Paul Scherrer Institute with its 30-L donut-shaped SD₂ source in a dedicated target station at the SINQ neutron source [29]. Due to absorption and up-scattering, UCN storage lifetimes attainable in SD₂ are at best on the order of several tens of milliseconds, requiring suppression of the (rotationally excited) para-deuterium content to a low percentage [31]. The short UCN survival time in SD₂ requires a quick separation of the UCNs from the medium, which can be done using a direct extraction by a guide, or UCN storage in a large buffer volume containing a small converter, thus realizing the concept of a “thin-film” UCN source [14].

In contrast to deuterium, UCN storage lifetimes of several hundreds of seconds can be achieved in He-II, which is due to the unique property of ⁴He to not absorb neutrons at all. A necessary prerequisite is the technically feasible removal of the strongly absorbing isotope ³He from the superfluid [35,62,70], and a temperature of the converter below ≈0.6 K to suppress upscattering. Like SD₂, He-II can be implemented in-pile, resulting in large UCN production rates [30,34,53]. However, the large heat load due to gamma radiation and neutrons near a strong primary neutron source prevents operation of the converter below 1 K, even if a very powerful cryogenic plant is available [53]. The UCN survival times in the warmer He-II are much shorter; however, being in a range of seconds to tens of seconds they are still typically two to three orders of magnitude larger than in SD₂. The corresponding much longer mean free path, together with the outstanding UCN transparency of the perfectly homogeneous He-II [57], enables realization of UCN sources with large active volumes.

Complementary to in-pile locations, as first pointed out in Ref. [17], a He-II converter can also be placed in a CN beam far away from the primary source. Producing UCNs at a smaller rate due to the CN fluxes being lower than in-pile, the heat load at such remote positions can be kept low enough to operate the converter near 0.5 K. This enables the aforementioned long UCN storage lifetimes in the He-II, with an ultimate upper limit of ≈880 s due to neutron beta-decay. It is therefore possible to accumulate the UCNs in the converter, building up a large UCN density prior to extraction to an experiment [15]. A variant of this in-beam type of source employs a thermal beam feeding a He-II converter surrounded by a moderating cold reflector, which produces the required CN spectrum near the source [33].

Possible locations for UCN and VCN sources, identified so far in and around the ESS Target monolith, are shown in Fig. 4 and described in the subsections below. The monolith (shown in red) is a large shielding structure with an outer radius of 5.5 m, which contains the beamports as radial openings. A vertical cut view of a detailed MCNP model of the Target monolith is shown in Fig. 5. The target wheel and the upper and lower moderator assemblies are situated in its center (see Ref. [42] for more information on engineering aspects of the monolith). Any UCN and/or VCN source needs to fit in the existing ESS infrastructure and requires careful modeling and simulation of radiation fields with MCNP to assess UCN production rates but also heat loads and the necessary radiation shielding.

Fig. 4.

A horizontal cut through the target region, at the height of the LD₂ moderator (shown in green) situated below the spallation target (not visible). The cylindrical region of radius 5.5 m around the center represents the shielding monolith (shown in red). The right figure is a zoom of the central part of the left figure. About half of its 42 standard beamports are visible in the cut plane. The possible locations of UCN sources, as studied within the HighNESS project are: (1) inside the “twister”; (2) inside the moderator cooling block; (3) in a standard beamport; (4) in the large beamport (shown as a white segment in the monolith); (5) outside the “bunker”, a heavy concrete shielding structure (shown in orange) placed around the monolith; the minimum distance of this location from the moderator is 15 m. See text for details and explanation of options for the various source positions.

Fig. 5.

Vertical cut of the MCNP geometry of the Target monolith (11 m diameter). Openings in the shielding (in white) serve to extract neutrons from the upper and lower moderators. In this particular view the openings placed at 90° with respect to the proton beam are shown. Notice the shielding between the upper and lower moderator, obstructing the direct view onto the spallation target from the beamports.

The closer the UCN source is placed to the neutron production area, the larger is the UCN production rate, but the required cooling power also becomes larger. Locations beyond a minimal distance of 15 m are situated outside the “neutron bunker” [64]. This additional heavy-concrete shielding structure placed around the monolith contains beamline components such as choppers and shutters. It is accessible only while the spallation source is shut down, for example, for maintenance operations or installation of new instruments. In addition to this limitation, working in this area will be restricted due to activation of components. The more distant locations are therefore best suited if the UCN source requires frequent maintenance or changes of configuration, or if technically advanced equipment is involved, like a magnetic multipole UCN reflector around the converter [23,60]. Since the experiments fed by the source have to be placed outside the neutron bunker as well, provision of UCNs requires only a short UCN guide, which keeps UCN transport losses low. No transport losses at all occur in “in-situ” experiments [2,8,11,69], where the experimental setup is immersed in the bath of the He-II converter exposed to the CN beam.

3.1.Inside the twister (location 1)

The ESS upper and lower moderators are placed in a container called the twister, see Fig. 6. The expected lifetime of the twister plug is about 1 year when the facility operates at full power, corresponding to 5000 hours of operation per year at 5 MW. The lifetime is limited by radiation damage in the moderator vessel. To extract the plug for replacement, it must be rotated about the axis of its shaft, hence the name “twister”. For the first, and possibly second generation of the moderators, only the upper half of the twister contains a moderator and reflector, while the lower half is occupied by a steel structure acting as a shield and fast-neutron reflector. As discussed in Section 2, the high-intensity LD₂ moderator to be implemented in the lower part of the twister will be of large volume. The perhaps only viable option for a UCN source integrated into the limited space left within the twister, seems to be a small-volume SD₂ converter below or beside the moderator (location number 1 in Fig. 4). An alternative option would be to replace the LD₂ moderator by a large block of SD₂ (or another material) capable of producing high rates of both CNs and UCNs. However, this option would present the challenge of cooling a large volume of solid deuterium placed close to the spallation target, which might be not possible.

Fig. 6.

Left: a CAD drawing of the twister and moderator cooling block, with part of the inner shielding for display. Right: horizontal view at the height of the lower moderator (courtesy R. Holmberg). The LD₂ moderator is placed inside the twister. A possible location of a secondary source inside the MCB is indicated by the dashed circle.

3.3.In a standard beamport in the monolith (location 3)

A third possibility is to implement a He-II based UCN source in a standard beamport in the monolith (number 3 in Fig. 4). The beamports start at 2 m from the moderator center, and when they are not used, they are plugged using elongated inserts which are 3.5 m long (Fig. 7, left). The channel limits the available space, but offers some flexibility in the choice of the distance of the source from the LD₂ moderator. The difference with the second possibility, mentioned in Section 3.2, consists in filling the space of a beamport with the UCN converter, rather than using one beamport for the insertion of a converter (He-II or solid D₂) in the MCB close to the target, and, if possible, a second beamport to extract neutrons. In this case, the beamport is filled with a He-II source. It is conceivable that the tip of this source can be placed at a distance from the LD₂ moderator center closer than the 2-m distance at which the guides and beamport inserts begin (cf. Fig. 7, right). It is worth noting that, near the spallation target, a permanent radiation shield at the level of the target (cf. Fig. 5) reduces the height available for the source and its infrastructure, if the source is placed closer than 2 m from the moderator center. At larger distances, the full height of a standard beamport is available and would leave sufficient space to extract UCNs from the back end of the converter via a short section of vertical UCN guide followed by a bend and a horizontal guide towards an experiment. The viability of this technique has previously been demonstrated in He-II converter prototypes at TU Munich and at ILL [46,68,71], however, for much shorter distances of UCN transport than would be needed to deliver UCNs to an experiment out of the neutron bunker.

Fig. 7.

Left: a CAD drawing of a neutron beamport insert (NBPI), courtesy Bengt Jönsson. The dimensions of the tip of the insert are approximately 16 cm×62 cm. The length is 350 cm. The material of the NBPI is stainless steel. The example shown is a NBPI with optics pointing at the upper moderator, with a neutron entrance window in the upper part of the insert. For unused beamports, the NBPIs consist of solid stainless steel blocks with the same shape, but without optics inside. Right: possible location of a He-II UCN source inside a standard neutron beamport. In this concept, He-II would fill a large fraction of the region indicated in gray color, noting that some space is to be reserved for the vacuum housing and thermal screens around the neutron converter vessel.

3.4.In the large beamport (location 4)

A fourth possibility (number 4 in Fig. 4) is to implement a UCN source in the Large Beamport (LBP), of which Fig. 8 shows two photos. This unique location provides free access for neutron extraction over a large angular range corresponding to three standard beamports. A steel frame holds three regular beamports for possible instruments. When both the frame and the three beamport inserts are removed, neutrons from the moderators can be extracted over a much larger solid angle than for a single beamport. The LBP is foreseen to first serve the NNBAR experiment, and could afterwards be exploited for the production of UCNs. A concrete proposal by Serebrov and Lyamkin foresees a vessel of 58 L filled with He-II, surrounded by a LD₂ reflector and a lead shield [54]. Figure 9 shows results of preliminary MCNP simulations performed within HighNESS, providing a map of the unperturbed neutron flux at 8.9 Å, which induces the dominant one-phonon process of UCN production in He-II. The CN reflector is expected to increase this flux. The total heat load on the He-II converter at the indicated source location is about 8 W, which can be removed using existing cooling technology [54].

Fig. 8.

Left: View of the beamport tubes and the monolith vessel during installation (photograph taken in November 2021). The space above the beamport tubes is to be filled with permanent shielding blocks; the 2.7-m radius vessel will eventually host the target wheel, the twister and MCB and other components and shielding blocks. Right: front view of the large beamport frame during installation.

Fig. 9.

Left: MCNP geometry showing the He-II source backed by a LD₂ reflector in the large beamport, concept of Serebrov and Lyamkin. Right: calculated flux map of 8.9 Å neutrons (unperturbed by the UCN source).

4.Basic concepts of VCN sources at ESS

Production of VCNs relies on the same neutron scattering processes as for UCNs, that is, either conversion of thermal or cold neutrons in single-step scattering events in a small-volume source, or by slowdown of neutrons by multiple scatterings in a sufficiently large moderator. The PF2 facility at ILL operates the only beamline presently available for VCN physics and thus sets a benchmark of performance. There, a strongly curved vertical neutron guide extracts the VCN beam from the long-wavelength tail of the spectrum of ILL’s large LD₂ moderator (see Ref. [41] for spectra measured in various sections of the VCN beam. Note, however, that the facility is currently being upgraded, which is expected to change these spectra). It is one of the goals of the HighNESS project to explore both experimentally and theoretically possible improvements on VCN production and their extraction from a source.

In the original proposal of the HighNESS project, two approaches were considered for the design of such a source. One is based on a dedicated VCN source using a suitable material that would provide a high flux of VCNs, while another one, similarly to PF2, extracts VCNs directly from the cold source. At the workshop, an additional concept was suggested, that can be considered a merger of the two approaches. It is only discussed briefly here, as it is outlined in greater detail in Ref. [38].

5.Conclusions

This article has outlined several options for implementation of UCN and VCN sources at ESS. Conceptual and engineering design of UCN sources can build on a broad experience gathered by various research groups around the globe during several decades. The main lines of recent developments, utilizing neutron conversion in SD₂ and He-II, were presented at the workshop. This paper highlighted ideas and concepts for UCN sources discussed for ESS, and gave an overview of the various possible locations. More detailed descriptions can be found in other contributions to the workshop proceedings.

For VCN sources, the state of the art is rather different. At present, the only beamline for VCN exists at the instrument PF2 at ILL, which is fed from a strongly curved vertical neutron guide viewing a LD₂ moderator. Given the potential scientific impact for several fundamental-physics projects, it has been considered as timely to include in the HighNESS project an exploration of possibilities for creation of a dedicated VCN source of high intensity at the ESS. The multifaceted works towards this goal start from basic experimental research, to determine macroscopic scattering cross sections and to study the suitability of advanced materials for neutron reflection and moderation of CNs to the VCN energy range. Materials investigated include nanodiamonds, magnesium hydride and intercalated graphite for neutron reflection, and deuterated clathrate hydrates for neutron moderation at temperatures below 2 K. The experimental data serve to benchmark physical models and also as direct input for Monte Carlo simulations of promising configurations of moderator/reflector assemblies. The project also includes an analysis of the extraction of neutrons in the long-wavelength tail of the spectrum from a LD₂ moderator, using an ND reflector.

All source options discussed in these proceedings will be investigated in depth in the remainder of the HighNESS project. It is the hope of the authors that the community will continue to provide essential input for the design of intense UCN and VCN sources, thus preparing the field for future world-leading experiments at ESS.