Heat transfer analysis study for the design of the New Polychromatic Beam Neutron Reflectometer CANDOR (Chromatic Analysis Neutron Diffractometer Or Reflectometer)

Abstract

A newly developed polychromatic beam neutron reflectometer CANDOR (Chromatic Analysis Neutron Diffractometer Or Reflectometer) on NG-1 at the NIST Center for Neutron research (NCNR) utilizes a wavelength-sensitive neutron detector consisting of 324 analyzing highly-oriented pyrolytic graphite (HOPG) crystals positioned sequentially in rows. Known for having a small thermal diffuse scattering cross section, HOPG crystals can lead to low signal-to-noise ratios in wavelength-sensitive detectors such as CANDOR. Even though it is possible to mathematically separate the desired signal from thermal diffuse scattering; by cooling the detector array of HOPG crystals in order to minimize the Debye Waller effect generates a better solution to this problem. In this heat transfer analysis study we show, within the instrument design constrains and thermodynamic considerations, technical feasibility and test results for the development of the New Polychromatic Beam Neutron Reflectometer CANDOR (Chromatic Analysis Neutron Diffractometer Or Reflectometer) at the NIST Center for Neutron Research.

Keywords

CANDOR thermodynamics simulation COMSOL

1. Introduction

The newly developed polychromatic beam neutron reflectometer CANDOR (Chromatic Analysis Neutron Diffractometer Or Reflectometer) located at the NG-1 guide at the NIST Center for Neutron research (NCNR) [2,13] utilizes a wavelength-sensitive neutron detector consisting of 324 analyzing highly-oriented pyrolytic graphite (HOPG) crystals positioned sequentially in rows within the detector bank (Fig. 1). This new detector allows researchers to utilize a much larger portion of the available incident neutron spectrum for scattering studies of the structure of condensed matter on a microscopic level. Each crystal observes a narrow band of wavelengths successively from 4 to 6 Angstroms in constant increments of 0.5° between adjacent crystals. The incoming neutron beam is reflected with almost 100 % efficiency into an adjacent scintillation plate detector as shown in Fig. 1. The neutrons which are captured in a ⁶Li matrix initiate a cascade of events which ultimately result in an electrical current signaling the capture of a neutron.

Fig. 1.

Schematic of the CHRNS CANDOR – white-beam reflectometer.

Although HOPG analyzer crystals are widely used in neutron scattering research because of their favorable neutron characteristics, e.g. high reflectivity and low absorption, HOPG also has a small thermal diffuse scattering cross section. Though this occurrence is negligible in most neutron scattering instruments, it can lead to low signal-to-noise ratios in wavelength-sensitive detectors like the one described here for reflectometry under certain circumstances. Despite the possibility of mathematically separating the desired signal from thermal diffuse scattering, it is preferable to eliminate or significantly suppress the scattering instead. One way of achieving this, is to cool down the HOPG crystals in order to minimize the Debye Waller effect [7,14]. This approach also has its own difficulties due to the number of HOPG crystals, the physical constrains of the detector bank, as well as the thermodynamic considerations involved in cooling the detector bank without directly affecting the adjacent scintillation plate detectors. The process below describes the empirically confirmed heat transfer analysis study of the technical feasibility for the development of the New Polychromatic Beam Neutron Reflectometer CANDOR (Chromatic Analysis Neutron Diffractometer Or Reflectometer) at the NIST Center for Neutron Research [13].

2. Experimental

A detector bank testing platform was designed with a series of HOPG sample crystals to ascertain the feasibility of construction for the novel detector bank required for the success of CANDOR. A cross-sectional view of the complete detector prototype is shown in Fig. 2. The design, for purpose of the heat transfer study, comprises of HOPG crystals mounted inside a mock-up neutron guide (Fig. 2). The upper, lower, crystals support structure was designed using aluminum 6061 and coated with $m = 4$ supermirror [2]. Whereas spring-loaded set screws in the slotted bottom support keep the crystals secured in place and allow for thermal contraction, the upper support contains slots where the HOPG crystals are free to move. Sensors were attached to the extremities of the copper bottom support as seen on Fig. 2. The sides where designed using 0.5 mm thick aluminum sheets and primarily ensure a homogenous temperature gradient of the pyrolytic graphite at cryogenic temperatures. A thin sheet of neutron absorbing borated aluminum is mounted between adjacent rows of HOPG to prevent signal cross-talk.

Fig. 2.

CANDOR detector bank testing platform. (a) Double stage closed-cycle refrigerator. (b) HOPG crystals array. (c) Scintillation plate detectors and electronics. The first stage of the CCR is thermally connected to a heat shield enveloping the HOPG crystals array. Both sensors were attached to the extremities of the Cu plate holding each HOPG crystal.

For the CANDOR detector a two stage commercially available closed-cycle refrigerator (CCR) was chosen to cool the HOPG crystals to cryogenic temperatures. Both stages share a common vacuum with the detector prototype (Fig. 2). The first stage cools the system down to 40 K. It also serves as a gradient heat shield between the exterior compartment, which is at room temperature (300 K) and the second stage. It was considered that, after the first stage reaches its target temperature the second stage lowers the temperature of the cold head down to 4 K. The interface between cold head and HOPG holder was fabricated from oxygen-free highly conductive (OFHC) copper which has exceptional thermal properties. Whereas the thermal conductivity of most materials decreases with decreasing temperature, it increases by orders of magnitude(s) below 100 K for OFHC copper (Fig. 3).

Fig. 3.

Thermal conductivity of 6061 aluminum alloy, oxygen free highly conductive (OFHC) copper condition rrr50 and HOPG.

Heat transfer analysis was done to evaluate the performance of the cooling process and the ability to produce the required extremely low temperatures. The analysis performed by using FEA software COMSOL Multiphysics 5.3 [6] and the results used to answer of major questions (radiation heat transfer mechanism effect, thermal design optimization, material effects, etc.) that are crucial to cool down the HOPG to the goal temperature. The geometry of the detector prototype (Fig. 4a) was performed by PTC-Parametric Cryo CAD [12] and imported to COMSOL as a CAD file. In order to reduce computing resources, the geometry was separated into four distinct parts (Fig. 4b) and used symmetry boundary conditions to take the missing sections into account as a part of the simulation results.

Fig. 4.

(a) CANDOR geometry of the detector prototype performed by PTC-Parametric Cryo CAD and imported to COMSOL as a CAD file. (b) Candor detector quarter geometry for boundary conditions determination.

A time dependent heat transfer analysis in solids was applied and surface to surface radiation was included. Moreover, a cluster server VM with 22 CPUs running at 2.50 GHz per CPU and 300 GB of memory ran for about 50 hours during each simulation.

The thermal conductivity profile of the materials implements onto the model as a function of the temperature was considered for Aluminum 6061-T6, Stainless steel 303, Graphite, and OFHC Copper. The ability to implement the real thermal conductivity profile of the materials into the model was critical for this simulation of the heat transfer at cryogenic temperature where the materials properties have strong temperature dependence. The heat load transfer into the internal system, and thus the HOPG crystals, caused by the outside walls temperature sink to ambient (300 K) that introduce energy to the system by radiation mechanism was also considered.

The calculation structural mesh used is a combination of free tetrahedral elements and bricks. The heat transfer model with surface to surface radiation is an extremely high computational resource demand, so we increased the mesh size as one of the options to minimize the use of physical memory required to solve the problem. In order to do this, we distributed the mesh size according to the temperature distribution. We modeled the area with the constant temperature with extra-large elements and refined the elements in the rest of the area with relation to the geometry limitations or temperature gradient. The total number of elements allowed to vary were approximately 100 thousand. Figure 5 shows the representation of the calculation structural mesh in the geometry.

Fig. 5.

Representation of the calculation structural mesh in the geometry of candor detector prototype.

It was determined, in extensive discussion with the neutron instrument designing committee that cooling the HOPG crystals to a temperature at most of 10 K was ideal to minimize the Debye Waller effect. Therefore, it was important that all of the internal detector graphite elements reach temperatures below 10 K without considerable temperature gradients between crystals. For that reason, it was necessary to evaluate the minimum distance between the graphite columns to maximize the number of HOPG at the same time minimizing the overall footprint of the detector bank without detriment to the cooling process.

In order to meet the first requirement, we decided to create the heat removing model using a commercially available Closed-cycle refrigerator (CCR) to cool down the graphite to a temperature below 10 K. The CCR uses helium gas as the refrigerant employing the Gifford-McMahon Refrigeration Scheme [4] where the gas compressor (operating near room temperature) and the cryogenic expansion cylinder (cold head) are thermally linked by a regenerator or thermal storage device. The CCR includes two cooling stages, the first stage is responsible for removing the major heat from the system, acting as a thermal anchor for radiation shields and reducing temperature to the range of 80 K–40 K, and a second stage that removes the remaining heat from the system to the minimum possible temperature, that in some cases below 4 K. The capability of the CCR cooling power used for modeling and subsequent test was 45 Watts at 40 K for the first stage and 1.5 Watts at 2.5 K for the second stage.

To achieve the low temperature in the detector, it is necessary to ensure high thermal efficiency of the design; thus, to obtain less energy loss and maximize the capability of the CCR. In terms of heat removal, our goal in modeling the system was to reduce the heat flux which is subsequently removed by the CCR from the system. Only after the system modeled is capable to reach temperatures below 10 K that creating a test prototype becomes viable.

Heat transfer in our modeled system occurs via the three main mechanisms that can cause heat from outside to “creep” into the system: Conduction, convection, and thermal radiation.

In the case of conduction, heat transfer by means of molecular agitation within a material without any motion of the material as a whole and may be described by the expression: $\begin{matrix} (1) & ρ C_{p} \frac{\partial T}{\partial t} = \nabla \cdot (k \nabla T) + Q \end{matrix}$

That gives the relationship between the variation of the temperature (T) over time to the spatial variation of the temperature. Q describes the source of energy that can be add or remove from the system, and the density of the material is given by ρ, $C_{p}$ describes the specific heat and k represents the thermal conductivity of the material.

For the case of convection, that involves the transfer of heat by mixing one parcel of fluid with another, the motion of the fluid may be entirely the result of differences of density resulting from the temperature gradients as in natural convection or the motion may be produced by mechanical means as in forced convection. Here energy is also transferred simultaneously by molecular conduction. The heat transfer mechanism describes by the expression. $\begin{matrix} (2) & ρ C_{p} \frac{\partial T}{\partial t} + ρ C_{p} u \cdot \nabla T = α_{p} T (\frac{\partial p_{A}}{\partial t} + u \cdot \nabla p_{A}) + r : S + \nabla \cdot (k \nabla T) + Q \end{matrix}$

And the Stefan-Boltzmann law where thermal radiation, the energy transferred by the emission of electromagnetic waves, is given by the relationship governing the net radiation: $\begin{matrix} (3) & P = e σ A (T^{4} - T_{C}^{4}) \end{matrix}$ Where T is the source of heat in the system, $T_{c}$ is the heat sink to the system, A is the surface area, e is the emissivity of the heat sink surface, and σ is the Boltzmann constant (in units of [ $\frac{Watt}{m^{2} k^{4}}$ ]).

Any thermodynamic model is highly dependent on the heat dissipation (leakage) and the ability of the thermal sink to absorb the heat leaking into the system. In order to identify the major mechanism driving this energy into the system we analyzed each mechanism according to system conditions. Figure 6 shows a cross section of the internal system structure used in the calculations, the location of the graphite elements and boundary condition. The HOPG crystal plates are located inside an enclosed box under vacuum condition to the order of $10^{- 5}$ mbar or better.

Fig. 6.

Cross section of the internal system structure, the location of the graphite elements and initial boundary conditions. The 1st and 2nd stages of the CCR are represented by the 1st and 2nd heat sinks respectively. The 1st stage creates a thermal heat shield enveloping the detector bank containing the HOPG crystals thermally sunk to the 2nd stage.

Under vacuum, the motion and the pressure terms of equation Eq. (2) (the velocity term u and variation of the pressure over time $\frac{\partial p_{A}}{\partial t}$ ) that describe the movement of the fluid or its driving force to move become negligible. Therefore, the two distinct mechanisms driving the thermodynamic process are conduction and radiation. Nevertheless, the vacuum regime inside the system creates a very good thermal isolation between the internal components (red color in Fig. 6) to the peripheral walls (gray color in Fig. 6).

When the internal parts of the system begin to cool down from room temperature (300 K) the system holds an initial amount of stored energy related to heat capacity ( $C_{p}$ ) of the materials in the system. The energy is removed through conduction and over time the temperature begins decreasing. The rate of the energy removal from the system is a function of: design (interfaces, insulation, mass, etc.), materials, and the heat removal capability (cooling power) of the CCR. Only after consideration of these parameters one is able to evaluate the cooling power needed to cool down the system to the working temperature (∼ 10 K or below) in a reasonable period of time (here this period of time is also a function of the working characteristics of the neutron scattering instrument this detector bank will serve). The gradient between the temperatures of the first and second stages of the CCR (cold fingers) and the peripheral walls (that remain at their initial temperature) establishes the heat conduction paths and generate yet another heat gradient between the non-contacting surfaces.

Radiation being a function of surface area, absorptivity, and emissivity; is the only mechanism that can allow energy transfer into the system from the peripheral walls and limit the heat removal capacity of the CCR to cool the HOPG crystal elements down to the required temperature. In this work we used Finite Element Analysis (FEA) software to perform a complete physical model of the system to evaluate the cooling process with a focus on the radiation mechanism and its impact on the cooling process. Additionally, we used the model to examine the thermal design and the material benefits to support and optimized the detector thermal mechanical design.

Minimizing the distance between the wall and the 1st stage heat shield (represented by the 1st heat sink in Fig. 6) is necessary to optimize the number of graphite column in the final design and, as a result, increase the number of graphite elements inside the detector bank. Nevertheless, reducing the distance between the wall and the 1st stage heat shield raises questions in reference to the impact of radiation in the heat transfer model. Due to the increase in surface area, very short distance between the 1st stage shield and the detector wall (Fig. 6); the cooling process may encounter a thermodynamic “short-circuit” of sorts. Therefore, there is a “minimal distance” between the 1st stage shield and the detector wall where the heat transfer is a minimum.

In order to evaluate the minimum distance between the first stage shield (b) and the detector wall (c), we considered all three heat transfer mechanisms and its individual impact on closing this gap, physical distortions, and the view factor [11].

Under vacuum regime (<1Pa) the number of the particles in the system is significantly low compare to their number in atmospheric pressure. Despite this very small number of particles, under very low pressures we can consider the remaining air a path for conduction heat transfer purposes. This type of conduction heat transfer mechanism between two very close surfaces surrounded by air in very low pressure is governed by the molecular theory [9], where the length scale of the system and the mean free path of the particles becomes a major player to determine the conductivity of air at low pressure. In our case we consider the air that fills the gap between the detector walls and the CCR first stage shield as the vehicle for this transfer of heat between the plates. The thermal conductivity in this case given by empiric correlation for very low air pressure between two plates: $\begin{matrix} (4) & \frac{K_{e}}{K_{0}} = \frac{1}{1 + \frac{C}{P P}} \end{matrix}$ Where $K_{e} / K_{0}$ is the thermal conductivity ratio between $K_{e}$ thermal conductivity at reduced pressures in [W/mK] and $K_{0}$ the thermal conductivity at 1 bar in [W/mK], C is a constant ( $7.6 * 10^{- 5}$ [mK/N]), the pressure parameter $P * d / T$ is given by PP in [N/mK] and T in Kelvin, P is the pressure in Pascals, and the plate distance d is given in meters. This correlation provides the tool to determine and understand the impact of the distance between the detector walls to the first stage shield on the thermal conductivity due to gas particles at low pressure [10].

The second criterion to evaluate with the minimum gap between the wall and the first stage shield is the physical distortions (buckling, twisting, shrinkage etc.) that can appear due to metal movement as result of temperature variation and difference between the characteristics of the materials.

Most metals undergo volume changes while cooling/warming. The dimensions change in accordance to the thermal expansion coefficient for each material. Developing a system that is subjected to temperature variation using different materials must take into consideration the different thermal expansion coefficients. Those differences may cause physical distortion in some of the elements in the system, and may be exacerbated by material production process. Distortions due to production process and dissimilar coefficients of thermal expansion are difficult to predict but easily circumvented during the system design by carefully considering materials and tolerances.

Finally, the view factor depends only on geometry when under vacuum, and at short distances the intensity of the emitted and absorbed radiation remains a constant aside from surface area considerations. The total energy absorbed by a surface is a function of the view factor – “fraction of the energy leaving and diffuse from a surface 1 (by emission or reflection), that directly impinges on another surface 2, where it is absorbed, reflected, or transmitted” [11]. For the case of parallel plates at a closed distance, we can assume that the view factor between two parallel equal rectangular plates is the dominant factor. For plates with high ratio between length or width dimension to the distance between the plates (W/H ≫ 1) we can assume the view factor close to 1; that is, most of the emitted energy from the hot surface impinges on the cool surface, so we can neglect radiation as a condition that limits the minimum distance between the 1st stage heat shield and the detector wall.

Once the factor affecting design considerations are accounted; another necessary criterion to enter our model is the ability of the closed-cycle refrigerator to remove heat from the system. To simulate the cooling process using FEA software, we needed to calibrate the heatsink source of the software with experimental data from the CCR cooling power. There is limited information available for such an end due to the industrial “secrecy” surrounding each different brand of CCRs [1,3]. Reliable heatsink calibration across the entire temperature range of interest ( $300 K ⩾ T ⩾ 4 K$ ) allows us to model and predict an accurate temperature distribution inside the internal parts of the detector over time. The closed-cycle refrigerator (CCR), also known as Gifford-McMahon Cryocoolers, use the Gifford-McMahon refrigeration process where helium gas is used as the refrigerant. The gas compressor (operating near room temperature) and the cryogenic expansion cylinder (cold head) are thermally linked by a regenerator or thermal storage device. The gas is driven through a 4-step cycle [5]. The advantage of the use of the CCR is that this cryocooler can have many distinct temperatures multistage with each stage having distinct temperature and cooling power profiles. For the current project a two-stage CCR was used. In a two-stage CCR, refrigeration is achieved at both distinct temperature levels by varying the displacer volumes. For this reason, we conducted an experiment using a two-stage CCR to cool down a well define mass of metal with a known heat capacity. We used a commercially available CCR [1] with a 1st stage heat removal capacity of 50 Watts @ 43 K, and a 2nd stage heat removal capacity of 1 Watt @ 4.2 Kelvin to cool down a 3 kg cylinder of oxygen free copper.

3. Results

The CANDOR detector bank testing platform was designed with a series of HOPG sample crystals to ascertain the feasibility of construction for the novel detector bank to be constructed. PTC-Creo Parametric [12] was used to model the geometry of the CCR and the 3 kg cylinder of oxygen free copper. This model was then imported it into COMSOL FEA software to model the heat transfer physics in compliance to the real experiment system boundary conditions and physical conditions. The empirical results of cooling the copper block were used to calibrate the heatsink source of the software for our model. However, In order to tune the heat source, we used analytical functions that describe the power capability of each CCR stage versus temperature and the manufacturer limited power map to extrapolate the CCR power for higher temperatures and used an empirical process to attain the correct match between the simulation and the experimental results. Figure 7 shows the graphs of temperature over time comparison between the experimental result and the simulation. There is a good match between the simulation and the experimental result. The simulation of the 1st stage cooling head closely track the experimental results, to good degree of agreement, especially over the high temperature regime (high cooling rate). The 2nd stage cooling head results display the same behavior as the experiment 2nd stage result with a small difference between the curves that apparently is caused by the complicated coupling mechanism between the two stages of the CCR which is hard to simulate accurately by the software. The good match between the simulation and the experiment give us confidence to rely on the CCR model as a cooling source to work in other cooling systems.

Fig. 7.

Graph of temperature over time for 3 kg of copper under cooling process. Comparison between simulation and experiment.

Convergence of the temperature distribution and cooling time simulation was possible due to the good agreement between the experimental data and empirical results of cooling the copper block used to calibrate the heatsink source in the model. The simulation run time was approximately 50 hours on NCNR’s computational server. The temperature of the HOPG crystals (Figs 2 and 6) thermally sunk to the CCR’s 2nd stage reached approximately 6 K; while the temperature of the 1st stage heat shield reached 31 K, indicating that the 1st stage heat shield absorbed some of the emitted radiation from the peripheral walls. The energy transferred to the surfaces of the 1st stage heat shield by radiation was successfully removed by the 1st stage cooling head of the CCR to enable the graphite to cool down to the working temperature by the 2nd stage.

The simulation was performed as a time dependent model which provides the opportunity to monitor the cooling process over time, as well as temperature gradients that could indicate hot spots throughout the structure. Figure 8 displays the cooling process graph (temperature over time) of three main simulated elements: 1st and 2nd cooling stages and the HOPG crystals. The temperature variation over these three elements appears reasonable and matches the characteristic temperature that was observed in the calibration experiment. In the simulation, the time to cool down the HOPG crystals to 6 K was approximately 6 hours, with a minimum equilibrium temperature of 5.8 K reached within 7 hours. The temperature gradient between the 2nd stage and HOPG crystals was small in the range below 77 K. The results indicate the efficiency of the thermal design and its use for the proposed cooling process application.

Achieving a minimum distance between the wall and the 1st stage heat shield in order to optimize the number of graphite column is a result much needed in this simulation in order to increase the number of graphite elements inside the detector bank. Analyzing the impact on each of the three heat transfer mechanisms of the distance between the walls showed that only thermal conductivity has an impact on the minimum distance between the walls [5]. To emphasize the variation of the thermal conductivity of air at low pressures as a function of distance and pressure we arbitrarily chose 3 difference distances and assume an average temperature ( $T_{av}$ ) of 200 K between the detector wall and the 1st stage heat shield, parameters chosen to comply to the system condition and the proposed design. The results observe molecular flow theory, based on the statistical probability model of collision between particles. Therefore, reducing the distance between the detector wall to the 1st stage heat shield will reduce the probability of particle interaction, and therefore results in lower thermal conductivity. In fact, as the distance decreases the thermal conductivity decreases, and thus theoretically there is no minimum limit to the physical gap; however, systems that are subjected to cryogenic temperature tend to suffer from distortion effects. Such effects are mainly due to residual stress in the metals that compose the detector because of fabrication process.

Fig. 8.

Simulation results of temperature vs. time at three locations inside the detector: 1st and 2nd cooling stages and HOPG crystals. Solid lines are a guide to the eye. Also, the result of the HOPG detector bank testing platform cooling is shown. Within 6 hours the 1st stage and 2nd stage temperatures reach a minimum in the simulation; while the HOPG detector bank testing platform achieves a minimum after 4.5 h. And with excellent agreement with the simulation.

The detector bank testing platform was constructed using the thermodynamic parameters outline in this simulation and tested. Temperature tests of the detector bank testing platform (Fig. 8) show that the knowledge base acquired through this simulation yielded results that superseded expectations, even though we noticed a deviation from the simulation results below 100 K. Within 6 hours the 1st and 2nd stages temperature reaches a minimum in the simulation; while the first cooldown test of the HOPG detector bank testing platform achieves a minimum after only 4.5 hours proving the design concept for the further design and manufacturing of the spectrometer’s final detector array. Subsequent long duration tests displayed slight increase of base temperature as well as in cooling time that were resolved by the use of cryogenic good work practices [8]; Nevertheless, this deviation from simulation addresses the many facets of actual manufacturing effects mentioned previously coupled with thermodynamic effects such as the different metals heat capacity interactions.

4. Conclusion

Thermal analysis for the construction of the detector bank testing platform for the newly developed polychromatic beam neutron reflectometer CANDOR (Chromatic Analysis Neutron Diffractometer Or Reflectometer) on NG-1 at the NIST Center for Neutron research (NCNR) under cryogenic temperature has been completed using COMSOL FEA software (Geometry files and COMSOL simulation files are available upon request). It was demonstrated in the simulation of the detector bank testing platform that the cooling process of the graphite from room temperature to below 10 K would require in the order of 4 to 5 hours. The simulation relies on empirical data collected from a cooling source of one commercially available closed-cycle refrigerator while cooling a 3 kg copper block. The CCR cooling power was also modeled and used on the cooling power calibration test. Radiation heat transfer was modeled using an emissivity coefficient value of 0.1 for both aluminum and copper. The recommend minimum distance between the detector walls (cassette) and the 1st stage heat shields was determined to be 2 mm to account for effects due to residual stress in the metals that compose the detector bank. A detector bank testing platform prototype was designed using the data acquired from these simulations, built, and tested. Although it was observed a divergence of cooling time observed between simulation and detector bank testing platform, the prototype testing proved to exceed the minimum temperature required ( $T < 10 K$ ) for proof of concept. The first cooldown results from the detector bank testing platform exceed expectations, and demonstrated the concept viability of using commercially available CCR as a means to cool the final detector bank for the polychromatic beam neutron reflectometer CANDOR (Chromatic Analysis Neutron Diffractometer Or Reflectometer) located at the NG-1 guide at the NIST Center for Neutron research (NCNR). A final simulation for the design of the full detector bank containing the entire array of 324 analyzing highly oriented pyrolytic graphite (HOPG) crystals was also conducted and showed a time to base temperature to be in the order of 55 hours.

Any mention of commercial products within this article is for information only; it does not imply recommendation or endorsement by NIST.

References

http://www.shicryogenics.com/products/4k-cryocoolers/rdk-408d2-4k-cryocooler-series/.

https://www.nist.gov/ncnr/chrns-candor-white-beam-reflectometer.

https://www.arscryo.com/de-215-specifications.

R.W.

Ackermann, Cryogenic Regenerative Heat Exchangers, Plenum Press, New York, 1997.

R.F.

Barron, Cryogenics Systems, 2nd edn, 1985.

“COMSOL Multiphysics Reference Manual, version 5.3”, COMSOL, Inc, www.comsol.com.

Debye, Annalen der Physik.348(1) (1913), 49–92. doi:10.1002/andp.19133480105.

J.W.

Ekin, Experimental Techniques for Low-Temperature Measurements, Oxford University Press, New York, 2007.

Fluid Flow Databook, General Electric, Genium Publishing, Section, 410.2, May 1982.

10.

C.J.M.

Lasance, The Thermal Conductivity of Air at Reduced Pressures and Length Scales, November 1, 2002.

11.

Martinez, Radiative View Factors, http://webserver.dmt.upm.es/~isidoro/tc3/Radiation%20View%20factors.pdf.

12.

Parametric Technology Corporation, https://www.ptc.com/en/products/cad/creo/.

13.

J.M.

Simmons

et al., J. Appl. Cryst.46 (2003), 1347–1352.

14.

Waller, Zeitschrift für Physik A.17(1) (1923), 398–408. doi:10.1007/BF01328696.