First demonstration of an all-solid-state optical cryocooler

First demonstration of an all-solid-state optical cryocooler Solid-state optical refrigeration uses anti-Stokes fluorescence to cool macroscopic objects to cryogenic temperatures 3+ without vibrations. Crystals such as Yb -doped YLiF (YLF:Yb) have previously been laser-cooled to 91 K. In this study, we show for the first time laser cooling of a payload connected to a cooling crystal. A YLF:Yb crystal was placed inside a Herriott cell and pumped with a 1020-nm laser (47 W) to cool a HgCdTe sensor that is part of a working Fourier Transform Infrared (FTIR) spectrometer to 135 K. This first demonstration of an all-solid-state optical cryocooler was enabled by careful control of the various desired and undesired heat flows. Fluorescence heating of the payload was minimized by using a single-kink YLF thermal link between the YLF:Yb cooling crystal and the copper coldfinger that held the HgCdTe sensor. The adhesive-free bond between YLF and YLF:Yb showed excellent thermal reliability. This laser-cooled assembly was then supported by silica aerogel cylinders inside a vacuum clamshell to minimize undesired conductive and radiative heat loads from the warm surroundings. Our structure can serve as a baseline for future optical cryocooler devices. Introduction that can provide truly vibration-free cooling to cryogenic temperatures All-optical cooling of a solid was first observed in 1995 . by Epstein et al. , and extensive developments over the Solid-state laser cooling is achieved using anti-Stokes past two decades in materials, characterization techni- fluorescence, a process in which the average wavelength ques, and optical designs have laid the groundwork for of the fluorescence (λ ) emitted by a material is shorter practical applications. Early work focused on the cooling than the wavelength (λ) of the laser used for excitation . 3+ 3+ of Yb -doped fluoride glasses (e.g., ZBLANP:Yb ), the The cooling efficiency η is given by the ratio of the cooling efficiency of which was later found to be limited cooling power (P ) and absorbed power (P ): cool abs 3+ by the substantial inhomogeneous broadening of the Yb P λ 2 cool crystal-field transition in the amorphous glass host . The η ¼ ¼ η η  1 ð1Þ c ext abs 3+ P abs much smaller inhomogeneous broadening in Yb -doped 3+ fluoride crystals (e.g., YLiF :Yb ) allowed for higher where η = η W /(η W +W ) is the external quantum cooling efficiencies, which helped enable the break- ext e r e r nr efficiency and η = α (λ)/[α (λ)+α (λ)] is the absorption through into the cryogenic regime in 2010 . This break- abs r r b efficiency. Here, W and W are the radiative and non- through has fueled further research into solid-state optical r nr 4, 5 radiative decay rates of the emitting state, respectively; η refrigeration , which is currently the only technology is the fluorescence escape efficiency; α (λ) is the resonant absorption coefficient of the cooling transition; and α (λ) is the parasitic background absorption coefficient. The Correspondence: Mansoor Sheik-Bahae (msb@unm.edu) radiative process of cooling by anti-Stokes fluorescence Los Alamos National Laboratory, P.O. Box 1663, Los Alamos, NM 87545, USA competes with a variety of non-radiative processes that Department of Physics & Astronomy, University of New Mexico, Albuquerque, NM 87131, USA reduce η and η by multiphonon relaxation of the ext abs © The Author(s) 2018 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to theCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; Hehlen et al. Light: Science & Applications (2018) 7:15 Page 2 of 10 emitting state (W ) and absorption by impurities (α ), cryocooler capable of refrigerating a HgCdTe infrared nr b 3+ respectively. Yb -doped solids have received particular sensor to 135 K. This device is the only solid-state cooling 3+ attention because Yb can be excited with high-power device that works in the cryogenic regime, i.e., in the lasers of ~1 μm wavelength and can emit with high temperature range of liquid cryogens, such as Xe, with a quantum efficiency due to its simple energy-level struc- boiling point of 166 K. The most common solid-state 3+ ture. However, laser cooling of Yb -doped solids to refrigerators, thermoelectric coolers, cannot achieve temperatures nearly as low as 135 K. cryogenic temperatures is only practical if η η exceeds ext abs 0.98–0.99, a challenging requirement that necessitates a Figure 1 shows a block diagram of the solid-state optical cooling crystal with exquisite purity . High-purity YLiF cryocooler architecture used in this study. The payload is 3+ crystals doped with 10 mol% Yb (YLF:10%Yb) have attached via a coldfinger, mirror, and thermal link to the recently been cooled to 91 K by exciting them with a laser laser-cooling material. This assembly (refrigerator cold 3+ tuned to the E4→E5 crystal-field transition of Yb at side) is mounted with a support element and surrounded 1020 nm . by a closely fitting clamshell (refrigerator hot side), which Most solid-state laser cooling studies to date have is lined with a low-emissivity coating. The breakthrough focused on the optical refrigeration of the laser-cooling performance of the present solid-state optical cryocooler material itself or a load that is transparent to fluores- is enabled by the materials and geometries of both the cence . For practical applications, however, it is neces- thermal link and support element combined with a high- 3+ sary for the laser-cooling material to refrigerate an performance YLF:10%Yb crystal. The design of various attached arbitrary external payload, such as a sensor or components and the performance of the optical cryo- electronic component. The primary challenges for cooler are discussed in detail in the following sections. advancing from a basic laser-cooling setup to a practical optical cryocooler device involve: (1) managing the Materials and methods numerous heat and radiation flows in the system and (2) Optical components providing a sturdy, thermally insulating support structure We used the YLF:10%Yb laser-cooling crystal that was for the laser-cooled assembly. In this article, we address cooled to 91 K in a previous study . This crystal was cut these challenges and describe the design and first so that the linearly polarized pump laser has E k ~ c because experimental demonstration of a solid-state optical the absorption cross section of the E4→E5 crystal-field Herriott cell TiNOX low-ε coating input mirror YLF:10%Yb crystal 1020 nm Herriott cell laser YLF thermal Focusing back reflector link lens Ag/Au mirror HgCdTe Cu cold sensor finger Silica aerogel support Cu clamshell Vacuum Loquid-cooled base chamber Fig. 1 Block diagram (not to scale) showing the components of the solid-state optical cryocooler. The laser-cooling material (blue) is placed inside an astigmatic Herriott cell to enable multi-pass excitation by the pump laser. The sensor payload (red) is connected by a coldfinger (black), mirror (gray), and thermal link (green) to the laser-cooling material. A support element (purple) provides the mounting of this assembly within a closely fitting clamshell (orange), which is lined with a low-emissivity coating (olive) and mounted onto a liquid-cooled base (dark red). The cryocooler is contained within a vacuum chamber (gray) 4.10 3.30 4.10 Hehlen et al. Light: Science & Applications (2018) 7:15 Page 3 of 10 transition in a YLF:Yb crystal at 1020 nm is ~1.7× greater yielded hydrophobic aerogels that reversibly adsorbed <2 12 15 in this geometry than that with E?~ c polarization . wt% water during long-term storage in ambient air . This The thermal link was fabricated from a Czochralski- process also allowed custom aerogel shapes to be directly grown undoped YLF crystal boule (AC Materials, molded during synthesis. We fabricated hydrophobic Tarpon Springs, FL), which was found to have low silica aerogels with a 0.1 g/cm density in the shape of 5 parasitic background absorption coefficients of mm diameter × 5 mm tall cylinders as the basic support −4 −1 5 5 element. The samples were found to absorb <2 × 10 of ~ ~ α E k ~ c ¼ 9:6 ´ 10 cm and α E?~ c ¼ 4:4 ´ 10 b b the incident power of a 1015-nm laser and were therefore −1 cm , as determined by exposing the sample to a 1020- expected to produce a negligible heat load when exposed nm laser (43 W) and measuring the resulting temperature to YLF:Yb fluorescence. increase relative to a reference heat load using a thermal The solid-state optical cryocooler was placed in a camera (Nanocore 640, L3 Communications Corporation, −7 vacuum chamber (5 × 10 Torr), and the YLF:10%Yb Garland, TX, USA). The YLF thermal link was fabricated crystal was excited by a linearly-polarized continuous- such that its ~ c axis was aligned parallel to the ~ c axis of the wave fiber laser (IPG Photonics, Inc., custom-made) YLF:10%Yb crystal at the respective mating interface that provided 47 W at 1020 nm. The temperature of (Fig. 1). This crystallographic alignment minimized the the coldfinger was monitored with a calibrated silicon thermal stresses induced by the anisotropic coefficients of diode (LakeShore Cryotronics, DT-670-SD) located thermal expansion (CTE) of YLF. The thermal link sur- in the coldfinger base. In addition, non-contact differ- face that was attached to the coldfinger was subsequently 3 ential luminescence thermometry (DLT) was used to coated with a metallic mirror by first etching the YLF in monitor the temperature of the cooling crystal by col- an argon plasma followed by sputter deposition of 200 nm lecting fluorescence with a multimode optical fiber of silver, followed by 200 nm of gold. This produced a inserted flush with the clamshell wall near the YLF:10% silver mirror with high reflectance on the inside of the Yb crystal. thermal link that was protected from oxidation by the gold layer. Figure 2a, b show the mirrored YLF thermal Thermal model 13, 14 ® link before and after Adhesive-Free Bonding (AFB The dynamics of the observed temperature changes of by Onyx Optics, Inc., Dublin, CA) to the YLF:10%Yb the cooling crystal and coldfinger follow an intricate crystal, respectively. The elevated temperatures during the interplay between the absorbed laser power, cooling effi- AFB process caused interdiffusion of the Ag and Au lay- ciency, parasitic losses (due to fluorescence absorption) at ers, as shown by the discoloration of the external gold interfaces, radiative loads, conductive loads, and useful layer in Figure 2b. No discoloration of the optically loads (e.g., heat lift from the sensor). A rigorous model important inside Ag mirror was observed. Furthermore, that numerically solves the 3-dimensional heat equation no mechanical failure was observed when temperature including the exact structure of the optical cooler could, cycling an equivalent assembly 22 times from 300 to 75 K in principle, reproduce the observed thermal dynamics. (at −6.9 K/min) and back to 300 K (at +21.6 K/min), However, we were able to determine the essential features demonstrating that a thermo-mechanically reliable AFB and underlying parameters by considering a simple 3- between undoped YLF and YLF:10%Yb was achieved. element model, as depicted in Figure 3, that consisted of Silica aerogels were synthesized by a sol–gel process, the YLF:Yb cooling crystal (at temperature T ) connected followed by drying in supercritical methanol, which to the coldfinger (at temperature T ) by a thermal link (at ab c 105° 3.97 0 10 20 010 20 75° 10.20 mm mm Fig. 2 Geometry, fabrication, and bonding of the YLF thermal link. a Fabricated YLF thermal link with a sputtered Ag/Au mirror on the cold- finger interface. b YLF thermal link AFB-bonded to the Brewster-cut YLF:10%Yb cooling crystal. c Cross sectional view and dimensions (in mm) of the δ= 15° geometry with a 90° inside angle. The dimension in the orthogonal direction is 4.10 mm. The orientation of the ~ c axis is indicated, and it is parallel to the ~ c axis of the YLF:10%Yb crystal 8.76 Hehlen et al. Light: Science & Applications (2018) 7:15 Page 4 of 10 area of the cooling crystal, link, and cold-finger, Cold respectively. ϵ  ϵ are the effective emissivities, ε/(1 + χ), x l finger Link of YLF in a tight clamshell coated with a low emissivity (ε ) TiNOX solar absorber (Almeco GmbH, Bernburg, Fluorescence abs T T T Germany) held at constant temperature T , and ϵ is the c f x l f Payload Heat conduction effective emissivity of the copper coldfinger. Here, χ= (1−ε )ε A /ε A (Ref. [11]), where ε and ε are the c l l c c l c Yb:YLF emissivities and σ is the Stefan-Boltzmann constant. β represents the fraction of the total fluorescence power that effectively leaks though the link and is consequently Fig. 3 Three-body thermal model of the optical cryocooler. The absorbed by the coldfinger. P is the useful thermal load lift model consists of a cooling crystal (YLF:Yb) at temperature T and a (payload heat lift) and small parasitic load from the sensor thermal link (YLF) with a mid-length temperature of T connected to a wires and aerogel supports. copper coldfinger at temperature T . These elements are placed in a vacuum clamshell whose walls are coated with low-emissivity materials at a constant (ground) temperature of T Results and discussion Cooling crystal and Herriott cell As showninEq. (1), the cooling power of a solid-state temperature T ), which in this case, was also constructed optical refrigerator linearly scales with the laser power from a YLF (undoped) crystal. absorbed by the cooling crystal (YLF:Yb), P = P η , abs in cpl Assuming that the adhesive-free bond between the which in turn depends on N , α ,and L .The tem- rt r x cooling crystal and thermal link introduces negligible perature dependence of α scales as α (T )∝[1+exp(δE / r r x g −1 −1 2 thermal resistance and parasitic loss, the equations gov- k T )] ,where δE ≈ 460 cm is the width of the F B x g 7/2 erning the temperature evolution of the three elements ground-state multiplet in YLF:Yb, and k is the Boltz- are approximated as follows: mann constant . The decrease of α at low tempera- tures requires N to be ~13× greater at 135 K than at rt dT 4 4 C ¼ ηðÞ T P ðÞ T þ ϵ σ T  T A x x abs x x x c c x 300 K to realize the same P .Torealizealarge P ,the abs abs dt ð2Þ Brewster-cut YLF:10%Yb crystal was placed inside an þKðÞ TðÞ T  T l x l x astigmatic Herriott cell formed by a concave spherical dT 4 4 C ¼KðÞ TðÞ T  T þ ϵ σ T  T A l l x l x l l c l input mirror (R = 50 cm) and a cylindrical back reflector dt ð3Þ (R = ∞, R = 50 cm) with the laser focused into the cell x y þKðÞ T T  T f l f l with a spherical lens (f = 5 cm) through a 400 μmdia- dT 4 4 C ¼ KðÞ T T  T þ ϵ σ T  T A meter center hole in the input mirror. We estimate f f l f l f f c f dt N ≈ 40 in the final configuration, which yields η > rt cpl þβ η ðÞ T P ðÞ T þ P x abs x lift −1 l ext 0.9999 with L = 1.11 cm and α = 0.12 cm at 1020 nm x r ð4Þ ~ and 135 K for E k ~ c. v v where C ¼ V C ðÞ T , with V and C denoting x;l;f x;l;f x,l,f x;l;f x;l;f Thermal link the known volume and heat capacity (J/K/m ), respec- Attaching the payload directly to the YLF:Yb crystal is tively, of the cooling crystal (x), link (l), and coldfinger (f). impractical. Although this would provide an excellent The temperature dependence of the YLF heat capacity is thermal contact, the payload would be directly exposed to taken from Ref. 16. P (T) = P η is the absorbed abs in cpl YLF:Yb fluorescence and absorb radiation that could power drawn from the incident power P (at λ = 1020 in exceed P . Therefore, a transparent thermal link was cool nm). Here, η = 1−exp(−2N α (λ,T)L ) is the laser cpl rt r x inserted between the YLF:Yb crystal and payload (Fig. 1) coupling efficiency, where N is the number of roundtrips rt to reduce the amount of fluorescence reaching the pay- the laser makes through the crystal of length L placed load while maintaining an efficient thermal path between inside the astigmatic Herriot cell and α (λ,T ) is the r x 18, 19 the payload and YLF:Yb crystal . The thermal, 3+ resonant absorption coefficient of the Yb crystal-field mechanical, and optical properties of the resulting inter- transition . K (T ) = 2κA/L is the effective thermal l l l face between the YLF:Yb crystal and thermal link require conductance of the link with a cross-sectional area of A careful consideration. The use of adhesives to bond the and an effective length L . The factor of 2 arises from the two elements can introduce optical absorption and cor- fact that T is the temperature of the link halfway between responding heating. We avoided this by bonding the the crystal and the coldfinger. κ(T ) is the temperature- cooling crystal and thermal link by van-der-Waals forces dependent thermal conductivity of the thermal link 13, 14 in an Adhesive-Free Bond (AFB®) . Although sapphire (undoped YLF). A , A , and A denote the total surface x l f is an attractive thermal-link material due to its low optical Radiative load Laser 140 mW 120 mW 100 mW 80 mW 60 mW 40 mW P = 20 mW Hehlen et al. Light: Science & Applications (2018) 7:15 Page 5 of 10 ab 0.10 10.0 Mirror 9.8 δ Inside angle 9.6 0.08 9.4 YLF thermal link YLF:Yb 9.2 35° 9.0 30° 8.8 0.06 20° 8.6 8.4 8.2 0.04 15° 8.0 10° 7.8 15° 7.6 0.02 5° 0° 7.4 7.2 0.00 7.0 0 5 10 15 20 25 30 35 020 40 60 80 100 120 140 P [mW] Tilt angle δ [deg] fluor Fig. 4 Optimization of the YLF thermal link geometry. a The fraction of fluorescence power reaching the link/mirror interface η for different tilt angles δ as obtained from a raytracing simulation (red circles) and a fit of the data to a single exponential function (black line). The inset shows a cross-sectional view of the kinked thermal link (green) attached to the YLF:Yb laser-cooling crystal (blue) and coated with a mirror on the interface that attaches to the coldfinger (see Fig. 1). The tilt angle δ is indicated. b Calculated heat load on the thermal link from thermal radiation P vs. the rad optical heat load due to fluorescence absorption at the link mirror P for links with different δ values. The calculation assumed link and clamshell fluor temperatures of 135 and 300 K, respectively, P =47 W, η = 0.999, η = 0.996, and R = 0.97. Link geometries with an inside angle of (90 −δ) and in cpl ext m 90 are shown as filled and open circles, respectively. Lines of constant total heat load P = P + P are indicated L fluor rad absorption and high thermal conductivity (κ ≈ 400 W/m K Given these constraints, we optimized the YLF thermal at 100 K ), our preliminary experiments exploring the link in the geometry of a short kinked YLF waveguide AFB of YLF:10%Yb and sapphire found frequent bond (Fig. 4a, inset). The goal was to minimize the total heat failures upon thermal cycling from 300 to 77 K. The stress load introduced by the thermal link P = P +P , L fluor rad at the AFB interface induced by a temperature change ΔT where P is the fraction of fluorescence power absorbed fluor is proportional to ΔTΔα, where Δα is the difference in the at the thermal link mirror surface (see Fig. 1) and P is rad coefficients of thermal expansion (CTE) of the two bon- the heat load due to radiative heat transport from the 21 −6 ded materials . The CTE of sapphire (α = 5.22 × 10 coated clamshell surface (warm) to the thermal link −1 −6 −1 20 K and α = 5.92 × 10 K at 296 K ) and YLF (α = (cold). Our measurements have shown that negligible heat b a −6 −1 −6 13.0…14.3 × 10 K and α = 8.0…10.1 × 10 at 300 is produced by the absorption of the fluorescence in the 17, 22 −6 −1 K ) indicate a Δα ≈ (8−9) × 10 K and a resulting bulk of the YLF thermal link. We performed an optical stress of 180–200 MPa for ΔT = 223 K, which exceeds the raytracing simulation of the light transport in the thermal tolerable stress of mechanically reliable AFB composites link using the FRED Optical Engineering software (Pho- of dissimilar materials by ~10×. We therefore chose to ton Engineering, LLC, Tucson, AZ, USA). A 4.1 × 4.1 use undoped YLF instead of sapphire as the thermal link mm link cross section was chosen to match the width of material, nominally yielding Δα≈0 and thus a stress-free the existing YLF:10%Yb crystal. The model used λ = adhesive-free bond between the YLF:10%Yb cooling 1004.7 nm as measured for YLF:Yb at 135 K, and a cor- crystal and the undoped YLF thermal link. However, this responding averaged refractive index of n = (2n + n )/3 o e results in a ~14× lower thermal conductivity of YLF (κ = = 1.456, where n = 1.448 and n = 1.471 , was used for a o e 24 W/m K and κ = 34 W/m K at 100 K ) than that of both undoped YLF and YLF:10%Yb. The fluorescence was sapphire. The undoped YLF thermal link, therefore, must represented by 10 rays that were uniformly and iso- be as short as possible to minimize the temperature dif- tropically emitted within the attached YLF:Yb crystal. As ference between the payload and YLF:Yb crystal, while shown in Figure 4a, an exponential decrease of the frac- still providing good fluorescence light rejection. tion of fluorescence power reaching the link/mirror P [mW] rad Hehlen et al. Light: Science & Applications (2018) 7:15 Page 6 of 10 interface η occurs when increasing the tilt angle δ; η P = 31.3 mW, of which 75% is due to residual mirror L L L does not significantly change for δ > 15°. This simulation absorption. Compared with a typical P = 470 mW (for cool yields P = P η η η (1 − R ), where R is the P = 47 W, η ≈ 1, and η = 0.01), the heat load intro- fluor in cpl ext L m m in cpl c reflectance of the mirror, which is set as 0.97 for the duced by the YLF thermal link is relatively small. sputtered silver layer used in this study, and P = 47 W, in η = 0.999, and η = 0.996 are assumed. The radiative Silica aerogel supports cpl ext heat load on the thermal link is approximately The laser-cooled assembly consisting of the YLF:Yb 4 4 P ¼ ε A σ T  T =ðÞ 1 þ χ . A 1-mm gap is assumed crystal, YLF thermal link, mirror, coldfinger, and payload rad l l c l between the thermal link and the TiNOX-coated clam- (Fig. 1) must be mounted within the closely fitting shell surfaces as well as ε =0.9 (from CaF ), ε = 0.05 clamshell structure in a manner that introduces minimal l 2 c (TiNOX solar absorber datasheet), T = 135 K, and T = thermal conduction from the clamshell to the cooled l c 300 K. Figure 4b plots P vs. P for thermal links with assembly, is mechanically reliable, and can be easily rad fluor different δ. The δ = 15° design achieves the lowest P , assembled. Earlier experiments using silica optical fibers whereas links with a smaller δ have a significantly larger with sharpened tips as support elements were difficult to P due to the greater η , and links with greater δ have a assemble and mechanically unreliable. The new approach fluor L slightly larger P due to their larger surface areas A and implemented in this study uses several silica aerogel rad l A . The link performance is slightly improved by provid- cylinders to support the cooled assembly under the ing a 90° inside angle (see inset Fig. 4a), a geometry that is coldfinger. Silica aerogels are open-celled mesoporous also more favorable for fabrication than geometries with SiO networks with low mass densities of typically smaller inside angles. The final thermal link design 0.01–0.2 g/cm as well as low optical absorption in the (Fig. 2c) is estimated to produce a total heat load of visible and near-infrared spectral range. Pure SiO ab YLF:Yb YLF:Yb crystal YLF crystal link YLF link HgCdTe sensor Cu coldfinger Cu coldfinger (1) (2) (3) 1′′ Cooled assembly Aerogel cylinder Clamshell #1 (4) (5) (6) Herriott cell mirrors Payload Clamshell #2 Clamshell #3 viewport Fig. 5 Solid-state optical cryocooler design and assembly. a Cooled assembly consisting of a YLF:Yb crystal, YLF thermal link, and copper coldfinger. b Image of the cooled assembly showing the HgCdTe sensor payload attached to the coldfinger. c Cryocooler assembly sequence consisting of the insertion of seven aerogel cylinders (1), placement of the cooled assembly (2), installation of side and top clamshell layers (3,4,5), and positioning and alignment of Herriott cell mirrors (6) Hehlen et al. Light: Science & Applications (2018) 7:15 Page 7 of 10 ab Coldfinger 280 YLF:Yb Initial slope: 0.31 K/mW 134.9 K 01 2 34 0 50 100 150 200 Time [h] Applied heat load [mW] Fig. 6 Performance of the solid-state optical cryocooler. a Temperature of the coldfinger (solid trace) and YLF:10%Yb crystal (dotted trace) as a function of time after turning on the 1020-nm laser in steps from 0, 12, 25, to 47 W. A steady-state coldfinger temperature of 134.9 K was reached after 4 h. b Heat load curve of the cryocooler pumped at 47 W, showing an initial slope of 0.31 K/mW aerogels have low thermal conductivities of ~0.004 W/m optical baffle that prevented residual fluorescence from K (at 300 K) and ~0.001 W/m K (at 100 K) in a vacuum , reaching the payload. making them attractive for use as support structures in The clamshell serves as the heat sink for the fluores- optical cryocoolers. The laser-cooled assembly (19.1 g cence emitted by the YLF:Yb crystal and must be designed total weight) rested on three 5 mm diameter × 5 mm tall to minimize the radiative heat loads [Eqs. (2)–(4)] it aerogel cylinders (58.9 mm total area), thus producing a imparts on the cooled assembly, which requires a geo- pressure of 3.18 kPa. Four additional aerogel cylinders metry that closely envelops the cooled assembly to were used around the outside perimeter of the coldfinger minimize the surface area and maximize χ. Figure 5c base to laterally secure the assembly. No mechanical shows the assembly sequence of the layered structure, degradation of the aerogel cylinders was observed under which allows for a stepwise installation of the clamshell at these conditions, which is consistent with the compressive a separation of 1 mm around the cooled assembly. We strength of silica aerogels, for which values of 1 MPa (0.25 used a TiNOX Energy solar absorber (Almeco GmbH, 3 26 3 27 g/cm ) and 0.047–0.11 MPa (0.1 g/cm ) have been Bernburg, Germany) glued to the inside surfaces of the reported. clamshell with silver epoxy to provide high absorbance of the YLF:Yb fluorescence and low thermal emissivity. The Laser-cooled assembly and clamshell clamshell also included a viewport that provided optical Earlier solid-state optical cryocooler designs envisioned access to the HgCdTe sensor payload. attaching the payload directly to the coated side of the 19, 28 thermal link , an approach that creates significant size Laser cooling of a HgCdTe sensor constraints for the placement of the support elements and Figure 6a shows the coldfinger temperature (measured potentially larger payloads. We therefore introduced a with a silicon diode) and the cooling crystal temperature copper coldfinger in the present cryocooler, allowing for (measured by DLT) as a function of time after turning on more space to accommodate larger payloads, aerogel the 1020-nm laser. The 1020 nm laser power was supports, as well as other instrumentation, such as a increased in steps from 0, 12, 25, to 47 W as a precau- temperature sensor and heating element. Figure 5a, b tionary measure to reduce the potentially large tempera- show the cooled assembly consisting of the YLF:Yb ture gradient and associated thermally induced stresses crystal/YLF thermal link unit and the HgCdTe infrared between the YLF:Yb crystal and coldfinger that develop at photo-sensor chip (InfraRed Associates, Inc.) attached to early times. This is evident in the stepwise decrease of the the copper coldfinger. The necked shape of the coldfinger YLF:Yb crystal temperature as measured by DLT. The near the thermal link interface was created to provide an coldfinger temperature reached 134.9 K after 4 h, Temperature [K] Cold finger temperature [K] Hehlen et al. Light: Science & Applications (2018) 7:15 Page 8 of 10 a b 300 296.0 Measurement Measurement Model Model 295.9 295.8 295.7 295.6 295.5 250 295.4 0 100 200 300 400 500 600 0 20 40 60 80 100 Time [s] Time [s] Fig. 7 a YLF:Yb cooling crystal temperature (T ) and b copper coldfinger temperature (T ) at early times. The measured (solid lines) and calculated x f (dotted lines) temperatures are shown as a function of time after turning on the 1020-nm pump laser at t = 0 with N = 1 rt representing the first ever demonstration of cooling a small temperature increase due to the small fluorescence payload to the cryogenic regime using solid-state optical leakage β at a rate given by ≈β η P (T )/C (T ) before l l ext abs c f c refrigeration. Furthermore, these data experimentally cooling from the crystal reverses this process in a time validate the various geometry and material choices for the scale approximated by τ ≈ K (T )/C (T ), as shown in l c l c thermal link, coldfinger, aerogel supports, and clamshell Figure 7b. Beyond this point, both the crystal and discussed above. cold-finger continue to cool and reach the steady- The results shown in Figure 6a were obtained without state condition with a temperature difference link link an electrical current flowing through the HgCdTe sensor, ðT  T Þ  2P =K ðT Þ, where P denotes the f x l l final load load i.e., without an external heat load introduced by the total heat load power (parasitic and useful) carried payload. A separate resistor installed in the coldfinger through the link. The data used in Figure 7 and the cor- base allowed for the application of well-defined heat loads responding fits were obtained for N = 1 (one laser rt and measurement of the corresponding increase in the roundtrip). The parameters obtained from these fits are coldfinger temperature. Figure 6b presents the heat load β ≈0.2%, η (T )≈1.2% at λ=1020 nm (corresponding to l c c −4 −1 curve obtained with the YLF:10%Yb crystal pumped at 47 η ≈0.992), α ≈2.1×10 cm , and ϵ  ϵ  0:38. The ext b x l W. The initial slope is 0.31 K/mW, a value that is char- values for η and α are in close agreement with pre- ext b 3+ acteristic for this particular cryocooler. Powering up the viously measured values for this 10% Yb -doped YLF HgCdTe sensor through a Fourier-Transform Infrared crystal. β also agrees with the calculated value from the (FTIR) spectrometer (Midac M4401) at the 134.9 K base raytracing model. T was fixed at 283.15 K, which is the temperature resulted in a temperature increase of ~2.5 K, final measured temperature of the clamshell. With the which corresponds to an estimated 8 mW heat load above parameters fixed, the model predicts a final cold- introduced by the active HgCdTe sensor. This value is in finger temperature T = 135.0 K with a temperature drop good agreement with the 8.8 mW heat load calculated of ΔT = T −T = 6.8 K across the link when the number f x from the 22.5 mA bias current and 17.3 Ω resistance (at of roundtrips is increased to N ≥30, as expected in the rt 134.9 K) of the HgCdTe sensor. Herriott cell arrangement. Considering the simplicity of The thermal model presented above [Eqs. (2)–(4)] our model, these calculated values are in excellent allows us to gain insight into the roles of the various terms agreement with our observed values of T = 134.9 K and in the observed temperature dynamics. The cooling ΔT = 6.1 K (Fig. 6a). The above calculations assumed a crystal temperature (T ) initially drops rapidly with the negligible heat lift (P = 0 mW) in comparison to the x lift slope given by ≈ −η (T )P (T )/C (T ) before the radia- existing radiative and fluorescence loads. Furthermore, c c abs c x x tive load (from the chamber) and conductive load the measured initial slope of 0.31 K/mW (Fig. 6b) is in (through the link) slow down this process, as shown in good agreement with ∂T /∂P ≈ 0.31 K/mW obtained f lift Figure 7a. The coldfinger, however, initially experiences a from the model calculations. Temperature [K] Temperature [K] Hehlen et al. Light: Science & Applications (2018) 7:15 Page 9 of 10 Laser-cooled HgCdTe sensor (135 K) Liquid-nitrogen cooled HgCdTe sensor (77 K) 2.0 1.5 1.0 0.5 0.0 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 –1 Wavenumbers [cm ] Fig. 8 FTIR spectroscopy using a laser-cooled HgCdTe infrared sensor. FTIR spectrum of a sheet of low-density poly-ethylene (LDPE) measured with the laser-cooled HgCdTe sensor (135 K, red trace) and, for comparison, an equivalent liquid-nitrogen cooled HgCdTe sensor (77 K, black trace). The larger relative noise observed at the extremes of the spectra is simply a result of the smaller signal (HgCdTe sensitivity and/or source brightness) in these regions; the noise amplitude is wavelength-independent FTIR spectroscopy with the laser-cooled HgCdTe sensor represents a breakthrough in this field and opens the door In a final experiment, the laser-cooled HgCdTe sensor to using this technology for a variety of applications that was used as part of an FTIR spectrometer and the infrared benefit from a reliable cryogenic refrigerator without absorption spectrum of a sheet of low-density poly-ethy- moving parts and associated vibrations. The all-solid-state lene (LDPE) was measured (Fig. 8). This is the first optical cryocooler was enabled by four key elements: First, demonstration of a sensor cooled by solid-state optical a high-quality YLF:Yb crystal with a low α and high η b ext refrigeration used as part of a practical device. Figure 8 provided efficient cooling at the heart of the cryocooler. shows the absorbance spectrum of the same sample Second, the use of a coldfinger provided the design flex- measured with an equivalent HgCdTe sensor cooled with ibility required to incorporate the sensor payload, support liquid nitrogen to 77 K. The bandgap energy E = hc/λ of structure, and diagnostic components. Third, the results this HgCdTe sensor corresponds to a cutoff wavelength of show that hydrophobic silica aerogels can provide excel- λ = 16.7 µm. The primary noise source is due to statistical lent thermal insulation as well as sufficient mechanical fluctuations of the thermally activated dark current within strength to support the laser-cooled assembly. Fourth, we the detector element. The dark current I is proportional demonstrated that a thermo-mechanically reliable van- 3+ to the number of carriers that are thermally excited across der-Waals bond can be created between Yb -doped and the bandgap and thus varies approximately as I ∝exp undoped YLF, which allowed the incorporation of a (−E /k T ),with the corresponding noise amplitude vary- transparent thermal link without the use of optical g B f pffiffiffiffi ing as I . The detector noise amplitude is therefore adhesives. These approaches are not without new chal- expected to be ~11× greater at 135 K than at 77 K, which lenges. The relatively poor thermal conductivity of the is consistent with the observed lower signal-to-noise ratio YLF thermal link creates a larger than desired tempera- in the measurement using the laser-cooled sensor. ture gradient between the payload and cooling crystal, which limits the base temperature of the payload. Ther- Conclusions mal link materials that have significantly greater thermal We used solid-state optical refrigeration to cool a pay- conductivity and a CTE that is comparable to the cooling load to cryogenic temperatures for the first time, which crystal are required to realize the full potential of optical Absorbance Hehlen et al. Light: Science & Applications (2018) 7:15 Page 10 of 10 refrigeration. The silica aerogel supports performed well 9. Hehlen,M.P.etal. Preparation of high-purity LiF, YF3, and YbF3 for laser refrigeration. Proceedings of SPIE 9000, Laser Refrigeration of Solids VII (SPIE, San in the present static application. More studies are needed Francisco, CA, 2014). to assess their thermo-mechanical performance under 10. Melgaard, S. D., Albrecht, A. R., Hehlen, M. P. & Sheik-Bahae, M. conditions of mechanical shock and vibration. Solid-state optical refrigeration to sub-100 Kelvin regime. Sci. Rep. 6, 20380 (2016). 11. Seletskiy, D. V., Melgaard, S. D., Di Lieto, A., Tonelli, M. & Sheik-Bahae, M. Acknowledgements Laser cooling of a semiconductor load to 165 K. Opt. Exp. 18,18061–18066 The authors thank T. Williamson and K. Baldwin (Los Alamos National (2010). Laboratory) for depositing the Ag/Au mirror and T. Cardenas (Los Alamos 12. Coluccelli, N. et al. Diode-pumped passively mode-locked Yb:YLF laser. Opt. National Laboratory) for performing sample metrology. Work at UNM was Exp. 16,2922–2927 (2008). partially supported by the Air Force Office of Scientific Research (AFOSR) under 13. Meissner H. E. Composites made from single crystal substances: U.S. Patent award numbers FA9550-15-1-024 and FA9550-16-1-0362 (MURI). 5441803. 1995-05-15. 14. Li, D.,Hong,P., Vedula,M.& Meissner,H.E. Thermal conductivity investigation of Author contributions adhesive-free bond laser components. Proceedings of SPIE 10100, Optical Com- M.P.H., M.S-B, and R.I.E., conceived the approach and supervised the work. M.P. ponents and Materials XIV. (SPIE, San Francisco, CA, 2017). H. designed the thermal link, and E.R.L. designed the clamshell. J.M., A.G., and 15. Hamilton, C. E. et al. Adsorption of ambient moisture by silica aerogel. Fusion A.R.A. performed the laser-cooling experiments at UNM. A.G. first implemented Sci. Tech. 63,301–304 (2013). the Herriott cell. C.E.H. synthesized the silica aerogels at LANL. S.P.L. assisted 16. Aggarwal, R. L.,Ripin,D.J., Ochoa, J. R. &Fan,T.Y.Measurement of thermo- with the FTIR measurements. All authors participated in writing the article. We optic properties of Y Al O ,Lu Al O ,YAlO ,LiYF ,LiLuF ,BaY F ,KGd(WO ) , 3 5 12 3 5 12 3 4 4 2 8 4 2 would like to dedicate this paper to the memory of co-author Eric R. Lee who and KY(WO ) laser crystals in the 80-300 K temperature range. J. Appl. Phys. 4 2 passed away unexpectedly since the submission of the manuscript. He was 98, 103514 (2005). instrumental to the success of this work. 17. Gragossian, A., Meng, J. W., Ghasemkhani, M., Albrecht, A. R. & Sheik- Bahae, M. Astigmatic Herriott cell for optical refrigeration. Opt. Eng. 56, Conflict of interest 011110 (2016). The authors declare that they have no conflict of interest. 18. Parker, J. et al. Thermal links for the implementation of an optical refrigerator. J. Appl. Phys. 105, 03116 (2009). 19. Melgaard, S. D. Cryogenic optical refrigeration: laser cooling of solids below Received: 28 February 2018 Revised: 24 April 2018 Accepted: 24 April 2018 123 K. Ph.D. dissertation, University of New Mexico, Albuquerque, NM (2013). . 20. Dobrovinskaya,E.R., Lytvynov, L.A., & Pishchik, V. Sapphire: Material, Manu- facturing, Applications (Springer, New York, NY, 2009). . 21. Onyx Optics, Inc. Mechanism of adhesive-free bonding (AFB®) (Onyx Optics, Inc.: References Dublin, CA, 2008). 1. Epstein, R. I., Buchwald, M. I., Edwards, B. C., Gosnell, T. R. & Mungan, C. E. 22. Payne, S. A. et al. 752 nm wing-pumped Cr:LiSAF laser. IEEE J. Quantum Observation of laser-induced fluorescent cooling of a solid. Nature 377, Electron 28, 1188–1196 (1992). 500–503 (1995). 23. Barnes, N. P. &Gettemy,D.J.Temperature variationofthe refractive indicesof 2. Thiede, J., Distel, J., Greenfield, S. R. & Epstein, R. I. Cooling to 208 K by optical yttrium lithium fluoride. J. Opt. Soc. Am. 70,1244–1247 (1980). refrigeration. Appl. Phys. Lett. 86, 154107 (2005). 24. Muley,S.V.&Ravindra, N.M.Emissivityofelectronicmaterials,coatings, and 3. Seletskiy, D. V. et al. Laser cooling of solids to cryogenic temperatures. Nat. structures. JOM 66,616–636 (2014). Photonics 4,161–164 (2010). 25. Rettelbach, T., Säuberlich, J., Korder, S. & Fricke, J. Thermal conductivity of silica 4. Seletskiy, D. V.,Epstein,R.& Sheik-Bahae, M. Laser cooling in solids: advances aerogel powders at temperatures from 10 to 275 K. J. Non-Cryst. Sol. 186, and prospects. Rep. Prog. Phys. 79, 096401 (2016). 278–284 (1995). 5. Nemova, G. Laser Cooling: Fundamental Properties and Applications (Pan 26. Parmenter, K. E. & Milstein, F. Mechanical properties of silica aerogels. J. Non- Stanford, Singapore, 2016). Cryst. Sol. 223,179–189 (1998). 6. Hehlen,M. P., Sheik-Bahae, M.,Epstein,R.I., Melgaard, S.D.& Seletskiy, D. V. 27. Omranpour, H., Dourbash, A. & Motahari, S. Mechanical properties improve- Materials for optical cryocoolers. J. Mater. Chem. C. 1, 7471–7478 (2013). ment of silica aerogel through aging: role of solvent type, time and tem- 7. Epstein, R. I. & Sheik-Bahae, M. Optical Refrigeration: Science and Applications of perature. AIP Conf. Proc. 1593, 298 (2014). Laser Cooling of Solids (Wiley, Weinheim, 2009). 28. Hehlen, M. P.,Boncher, W.L.& Love, S.P. Design study of a laser-cooled infrared 8. Sheik-Bahae, ?tlsb?>M. & Epstein, R. I. Optical refrigeration. Nat. Photonics 1, sensor. Proceedings of SPIE 9380, Laser Refrigeration of Solids VIII.(SPIE,San 693–699 (2007). Francisco, CA, 2015). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Light: Science & Applications Springer Journals
Loading next page...
 
/lp/springer_journal/first-demonstration-of-an-all-solid-state-optical-cryocooler-MiQRKzlSTr
Publisher
Nature Publishing Group UK
Copyright
Copyright © 2018 by The Author(s)
Subject
Physics; Physics, general; Applied and Technical Physics; Atomic, Molecular, Optical and Plasma Physics; Classical and Continuum Physics; Optics, Lasers, Photonics, Optical Devices
eISSN
2047-7538
D.O.I.
10.1038/s41377-018-0028-7
Publisher site
See Article on Publisher Site

Abstract

Solid-state optical refrigeration uses anti-Stokes fluorescence to cool macroscopic objects to cryogenic temperatures 3+ without vibrations. Crystals such as Yb -doped YLiF (YLF:Yb) have previously been laser-cooled to 91 K. In this study, we show for the first time laser cooling of a payload connected to a cooling crystal. A YLF:Yb crystal was placed inside a Herriott cell and pumped with a 1020-nm laser (47 W) to cool a HgCdTe sensor that is part of a working Fourier Transform Infrared (FTIR) spectrometer to 135 K. This first demonstration of an all-solid-state optical cryocooler was enabled by careful control of the various desired and undesired heat flows. Fluorescence heating of the payload was minimized by using a single-kink YLF thermal link between the YLF:Yb cooling crystal and the copper coldfinger that held the HgCdTe sensor. The adhesive-free bond between YLF and YLF:Yb showed excellent thermal reliability. This laser-cooled assembly was then supported by silica aerogel cylinders inside a vacuum clamshell to minimize undesired conductive and radiative heat loads from the warm surroundings. Our structure can serve as a baseline for future optical cryocooler devices. Introduction that can provide truly vibration-free cooling to cryogenic temperatures All-optical cooling of a solid was first observed in 1995 . by Epstein et al. , and extensive developments over the Solid-state laser cooling is achieved using anti-Stokes past two decades in materials, characterization techni- fluorescence, a process in which the average wavelength ques, and optical designs have laid the groundwork for of the fluorescence (λ ) emitted by a material is shorter practical applications. Early work focused on the cooling than the wavelength (λ) of the laser used for excitation . 3+ 3+ of Yb -doped fluoride glasses (e.g., ZBLANP:Yb ), the The cooling efficiency η is given by the ratio of the cooling efficiency of which was later found to be limited cooling power (P ) and absorbed power (P ): cool abs 3+ by the substantial inhomogeneous broadening of the Yb P λ 2 cool crystal-field transition in the amorphous glass host . The η ¼ ¼ η η  1 ð1Þ c ext abs 3+ P abs much smaller inhomogeneous broadening in Yb -doped 3+ fluoride crystals (e.g., YLiF :Yb ) allowed for higher where η = η W /(η W +W ) is the external quantum cooling efficiencies, which helped enable the break- ext e r e r nr efficiency and η = α (λ)/[α (λ)+α (λ)] is the absorption through into the cryogenic regime in 2010 . This break- abs r r b efficiency. Here, W and W are the radiative and non- through has fueled further research into solid-state optical r nr 4, 5 radiative decay rates of the emitting state, respectively; η refrigeration , which is currently the only technology is the fluorescence escape efficiency; α (λ) is the resonant absorption coefficient of the cooling transition; and α (λ) is the parasitic background absorption coefficient. The Correspondence: Mansoor Sheik-Bahae (msb@unm.edu) radiative process of cooling by anti-Stokes fluorescence Los Alamos National Laboratory, P.O. Box 1663, Los Alamos, NM 87545, USA competes with a variety of non-radiative processes that Department of Physics & Astronomy, University of New Mexico, Albuquerque, NM 87131, USA reduce η and η by multiphonon relaxation of the ext abs © The Author(s) 2018 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to theCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; Hehlen et al. Light: Science & Applications (2018) 7:15 Page 2 of 10 emitting state (W ) and absorption by impurities (α ), cryocooler capable of refrigerating a HgCdTe infrared nr b 3+ respectively. Yb -doped solids have received particular sensor to 135 K. This device is the only solid-state cooling 3+ attention because Yb can be excited with high-power device that works in the cryogenic regime, i.e., in the lasers of ~1 μm wavelength and can emit with high temperature range of liquid cryogens, such as Xe, with a quantum efficiency due to its simple energy-level struc- boiling point of 166 K. The most common solid-state 3+ ture. However, laser cooling of Yb -doped solids to refrigerators, thermoelectric coolers, cannot achieve temperatures nearly as low as 135 K. cryogenic temperatures is only practical if η η exceeds ext abs 0.98–0.99, a challenging requirement that necessitates a Figure 1 shows a block diagram of the solid-state optical cooling crystal with exquisite purity . High-purity YLiF cryocooler architecture used in this study. The payload is 3+ crystals doped with 10 mol% Yb (YLF:10%Yb) have attached via a coldfinger, mirror, and thermal link to the recently been cooled to 91 K by exciting them with a laser laser-cooling material. This assembly (refrigerator cold 3+ tuned to the E4→E5 crystal-field transition of Yb at side) is mounted with a support element and surrounded 1020 nm . by a closely fitting clamshell (refrigerator hot side), which Most solid-state laser cooling studies to date have is lined with a low-emissivity coating. The breakthrough focused on the optical refrigeration of the laser-cooling performance of the present solid-state optical cryocooler material itself or a load that is transparent to fluores- is enabled by the materials and geometries of both the cence . For practical applications, however, it is neces- thermal link and support element combined with a high- 3+ sary for the laser-cooling material to refrigerate an performance YLF:10%Yb crystal. The design of various attached arbitrary external payload, such as a sensor or components and the performance of the optical cryo- electronic component. The primary challenges for cooler are discussed in detail in the following sections. advancing from a basic laser-cooling setup to a practical optical cryocooler device involve: (1) managing the Materials and methods numerous heat and radiation flows in the system and (2) Optical components providing a sturdy, thermally insulating support structure We used the YLF:10%Yb laser-cooling crystal that was for the laser-cooled assembly. In this article, we address cooled to 91 K in a previous study . This crystal was cut these challenges and describe the design and first so that the linearly polarized pump laser has E k ~ c because experimental demonstration of a solid-state optical the absorption cross section of the E4→E5 crystal-field Herriott cell TiNOX low-ε coating input mirror YLF:10%Yb crystal 1020 nm Herriott cell laser YLF thermal Focusing back reflector link lens Ag/Au mirror HgCdTe Cu cold sensor finger Silica aerogel support Cu clamshell Vacuum Loquid-cooled base chamber Fig. 1 Block diagram (not to scale) showing the components of the solid-state optical cryocooler. The laser-cooling material (blue) is placed inside an astigmatic Herriott cell to enable multi-pass excitation by the pump laser. The sensor payload (red) is connected by a coldfinger (black), mirror (gray), and thermal link (green) to the laser-cooling material. A support element (purple) provides the mounting of this assembly within a closely fitting clamshell (orange), which is lined with a low-emissivity coating (olive) and mounted onto a liquid-cooled base (dark red). The cryocooler is contained within a vacuum chamber (gray) 4.10 3.30 4.10 Hehlen et al. Light: Science & Applications (2018) 7:15 Page 3 of 10 transition in a YLF:Yb crystal at 1020 nm is ~1.7× greater yielded hydrophobic aerogels that reversibly adsorbed <2 12 15 in this geometry than that with E?~ c polarization . wt% water during long-term storage in ambient air . This The thermal link was fabricated from a Czochralski- process also allowed custom aerogel shapes to be directly grown undoped YLF crystal boule (AC Materials, molded during synthesis. We fabricated hydrophobic Tarpon Springs, FL), which was found to have low silica aerogels with a 0.1 g/cm density in the shape of 5 parasitic background absorption coefficients of mm diameter × 5 mm tall cylinders as the basic support −4 −1 5 5 element. The samples were found to absorb <2 × 10 of ~ ~ α E k ~ c ¼ 9:6 ´ 10 cm and α E?~ c ¼ 4:4 ´ 10 b b the incident power of a 1015-nm laser and were therefore −1 cm , as determined by exposing the sample to a 1020- expected to produce a negligible heat load when exposed nm laser (43 W) and measuring the resulting temperature to YLF:Yb fluorescence. increase relative to a reference heat load using a thermal The solid-state optical cryocooler was placed in a camera (Nanocore 640, L3 Communications Corporation, −7 vacuum chamber (5 × 10 Torr), and the YLF:10%Yb Garland, TX, USA). The YLF thermal link was fabricated crystal was excited by a linearly-polarized continuous- such that its ~ c axis was aligned parallel to the ~ c axis of the wave fiber laser (IPG Photonics, Inc., custom-made) YLF:10%Yb crystal at the respective mating interface that provided 47 W at 1020 nm. The temperature of (Fig. 1). This crystallographic alignment minimized the the coldfinger was monitored with a calibrated silicon thermal stresses induced by the anisotropic coefficients of diode (LakeShore Cryotronics, DT-670-SD) located thermal expansion (CTE) of YLF. The thermal link sur- in the coldfinger base. In addition, non-contact differ- face that was attached to the coldfinger was subsequently 3 ential luminescence thermometry (DLT) was used to coated with a metallic mirror by first etching the YLF in monitor the temperature of the cooling crystal by col- an argon plasma followed by sputter deposition of 200 nm lecting fluorescence with a multimode optical fiber of silver, followed by 200 nm of gold. This produced a inserted flush with the clamshell wall near the YLF:10% silver mirror with high reflectance on the inside of the Yb crystal. thermal link that was protected from oxidation by the gold layer. Figure 2a, b show the mirrored YLF thermal Thermal model 13, 14 ® link before and after Adhesive-Free Bonding (AFB The dynamics of the observed temperature changes of by Onyx Optics, Inc., Dublin, CA) to the YLF:10%Yb the cooling crystal and coldfinger follow an intricate crystal, respectively. The elevated temperatures during the interplay between the absorbed laser power, cooling effi- AFB process caused interdiffusion of the Ag and Au lay- ciency, parasitic losses (due to fluorescence absorption) at ers, as shown by the discoloration of the external gold interfaces, radiative loads, conductive loads, and useful layer in Figure 2b. No discoloration of the optically loads (e.g., heat lift from the sensor). A rigorous model important inside Ag mirror was observed. Furthermore, that numerically solves the 3-dimensional heat equation no mechanical failure was observed when temperature including the exact structure of the optical cooler could, cycling an equivalent assembly 22 times from 300 to 75 K in principle, reproduce the observed thermal dynamics. (at −6.9 K/min) and back to 300 K (at +21.6 K/min), However, we were able to determine the essential features demonstrating that a thermo-mechanically reliable AFB and underlying parameters by considering a simple 3- between undoped YLF and YLF:10%Yb was achieved. element model, as depicted in Figure 3, that consisted of Silica aerogels were synthesized by a sol–gel process, the YLF:Yb cooling crystal (at temperature T ) connected followed by drying in supercritical methanol, which to the coldfinger (at temperature T ) by a thermal link (at ab c 105° 3.97 0 10 20 010 20 75° 10.20 mm mm Fig. 2 Geometry, fabrication, and bonding of the YLF thermal link. a Fabricated YLF thermal link with a sputtered Ag/Au mirror on the cold- finger interface. b YLF thermal link AFB-bonded to the Brewster-cut YLF:10%Yb cooling crystal. c Cross sectional view and dimensions (in mm) of the δ= 15° geometry with a 90° inside angle. The dimension in the orthogonal direction is 4.10 mm. The orientation of the ~ c axis is indicated, and it is parallel to the ~ c axis of the YLF:10%Yb crystal 8.76 Hehlen et al. Light: Science & Applications (2018) 7:15 Page 4 of 10 area of the cooling crystal, link, and cold-finger, Cold respectively. ϵ  ϵ are the effective emissivities, ε/(1 + χ), x l finger Link of YLF in a tight clamshell coated with a low emissivity (ε ) TiNOX solar absorber (Almeco GmbH, Bernburg, Fluorescence abs T T T Germany) held at constant temperature T , and ϵ is the c f x l f Payload Heat conduction effective emissivity of the copper coldfinger. Here, χ= (1−ε )ε A /ε A (Ref. [11]), where ε and ε are the c l l c c l c Yb:YLF emissivities and σ is the Stefan-Boltzmann constant. β represents the fraction of the total fluorescence power that effectively leaks though the link and is consequently Fig. 3 Three-body thermal model of the optical cryocooler. The absorbed by the coldfinger. P is the useful thermal load lift model consists of a cooling crystal (YLF:Yb) at temperature T and a (payload heat lift) and small parasitic load from the sensor thermal link (YLF) with a mid-length temperature of T connected to a wires and aerogel supports. copper coldfinger at temperature T . These elements are placed in a vacuum clamshell whose walls are coated with low-emissivity materials at a constant (ground) temperature of T Results and discussion Cooling crystal and Herriott cell As showninEq. (1), the cooling power of a solid-state temperature T ), which in this case, was also constructed optical refrigerator linearly scales with the laser power from a YLF (undoped) crystal. absorbed by the cooling crystal (YLF:Yb), P = P η , abs in cpl Assuming that the adhesive-free bond between the which in turn depends on N , α ,and L .The tem- rt r x cooling crystal and thermal link introduces negligible perature dependence of α scales as α (T )∝[1+exp(δE / r r x g −1 −1 2 thermal resistance and parasitic loss, the equations gov- k T )] ,where δE ≈ 460 cm is the width of the F B x g 7/2 erning the temperature evolution of the three elements ground-state multiplet in YLF:Yb, and k is the Boltz- are approximated as follows: mann constant . The decrease of α at low tempera- tures requires N to be ~13× greater at 135 K than at rt dT 4 4 C ¼ ηðÞ T P ðÞ T þ ϵ σ T  T A x x abs x x x c c x 300 K to realize the same P .Torealizealarge P ,the abs abs dt ð2Þ Brewster-cut YLF:10%Yb crystal was placed inside an þKðÞ TðÞ T  T l x l x astigmatic Herriott cell formed by a concave spherical dT 4 4 C ¼KðÞ TðÞ T  T þ ϵ σ T  T A l l x l x l l c l input mirror (R = 50 cm) and a cylindrical back reflector dt ð3Þ (R = ∞, R = 50 cm) with the laser focused into the cell x y þKðÞ T T  T f l f l with a spherical lens (f = 5 cm) through a 400 μmdia- dT 4 4 C ¼ KðÞ T T  T þ ϵ σ T  T A meter center hole in the input mirror. We estimate f f l f l f f c f dt N ≈ 40 in the final configuration, which yields η > rt cpl þβ η ðÞ T P ðÞ T þ P x abs x lift −1 l ext 0.9999 with L = 1.11 cm and α = 0.12 cm at 1020 nm x r ð4Þ ~ and 135 K for E k ~ c. v v where C ¼ V C ðÞ T , with V and C denoting x;l;f x;l;f x,l,f x;l;f x;l;f Thermal link the known volume and heat capacity (J/K/m ), respec- Attaching the payload directly to the YLF:Yb crystal is tively, of the cooling crystal (x), link (l), and coldfinger (f). impractical. Although this would provide an excellent The temperature dependence of the YLF heat capacity is thermal contact, the payload would be directly exposed to taken from Ref. 16. P (T) = P η is the absorbed abs in cpl YLF:Yb fluorescence and absorb radiation that could power drawn from the incident power P (at λ = 1020 in exceed P . Therefore, a transparent thermal link was cool nm). Here, η = 1−exp(−2N α (λ,T)L ) is the laser cpl rt r x inserted between the YLF:Yb crystal and payload (Fig. 1) coupling efficiency, where N is the number of roundtrips rt to reduce the amount of fluorescence reaching the pay- the laser makes through the crystal of length L placed load while maintaining an efficient thermal path between inside the astigmatic Herriot cell and α (λ,T ) is the r x 18, 19 the payload and YLF:Yb crystal . The thermal, 3+ resonant absorption coefficient of the Yb crystal-field mechanical, and optical properties of the resulting inter- transition . K (T ) = 2κA/L is the effective thermal l l l face between the YLF:Yb crystal and thermal link require conductance of the link with a cross-sectional area of A careful consideration. The use of adhesives to bond the and an effective length L . The factor of 2 arises from the two elements can introduce optical absorption and cor- fact that T is the temperature of the link halfway between responding heating. We avoided this by bonding the the crystal and the coldfinger. κ(T ) is the temperature- cooling crystal and thermal link by van-der-Waals forces dependent thermal conductivity of the thermal link 13, 14 in an Adhesive-Free Bond (AFB®) . Although sapphire (undoped YLF). A , A , and A denote the total surface x l f is an attractive thermal-link material due to its low optical Radiative load Laser 140 mW 120 mW 100 mW 80 mW 60 mW 40 mW P = 20 mW Hehlen et al. Light: Science & Applications (2018) 7:15 Page 5 of 10 ab 0.10 10.0 Mirror 9.8 δ Inside angle 9.6 0.08 9.4 YLF thermal link YLF:Yb 9.2 35° 9.0 30° 8.8 0.06 20° 8.6 8.4 8.2 0.04 15° 8.0 10° 7.8 15° 7.6 0.02 5° 0° 7.4 7.2 0.00 7.0 0 5 10 15 20 25 30 35 020 40 60 80 100 120 140 P [mW] Tilt angle δ [deg] fluor Fig. 4 Optimization of the YLF thermal link geometry. a The fraction of fluorescence power reaching the link/mirror interface η for different tilt angles δ as obtained from a raytracing simulation (red circles) and a fit of the data to a single exponential function (black line). The inset shows a cross-sectional view of the kinked thermal link (green) attached to the YLF:Yb laser-cooling crystal (blue) and coated with a mirror on the interface that attaches to the coldfinger (see Fig. 1). The tilt angle δ is indicated. b Calculated heat load on the thermal link from thermal radiation P vs. the rad optical heat load due to fluorescence absorption at the link mirror P for links with different δ values. The calculation assumed link and clamshell fluor temperatures of 135 and 300 K, respectively, P =47 W, η = 0.999, η = 0.996, and R = 0.97. Link geometries with an inside angle of (90 −δ) and in cpl ext m 90 are shown as filled and open circles, respectively. Lines of constant total heat load P = P + P are indicated L fluor rad absorption and high thermal conductivity (κ ≈ 400 W/m K Given these constraints, we optimized the YLF thermal at 100 K ), our preliminary experiments exploring the link in the geometry of a short kinked YLF waveguide AFB of YLF:10%Yb and sapphire found frequent bond (Fig. 4a, inset). The goal was to minimize the total heat failures upon thermal cycling from 300 to 77 K. The stress load introduced by the thermal link P = P +P , L fluor rad at the AFB interface induced by a temperature change ΔT where P is the fraction of fluorescence power absorbed fluor is proportional to ΔTΔα, where Δα is the difference in the at the thermal link mirror surface (see Fig. 1) and P is rad coefficients of thermal expansion (CTE) of the two bon- the heat load due to radiative heat transport from the 21 −6 ded materials . The CTE of sapphire (α = 5.22 × 10 coated clamshell surface (warm) to the thermal link −1 −6 −1 20 K and α = 5.92 × 10 K at 296 K ) and YLF (α = (cold). Our measurements have shown that negligible heat b a −6 −1 −6 13.0…14.3 × 10 K and α = 8.0…10.1 × 10 at 300 is produced by the absorption of the fluorescence in the 17, 22 −6 −1 K ) indicate a Δα ≈ (8−9) × 10 K and a resulting bulk of the YLF thermal link. We performed an optical stress of 180–200 MPa for ΔT = 223 K, which exceeds the raytracing simulation of the light transport in the thermal tolerable stress of mechanically reliable AFB composites link using the FRED Optical Engineering software (Pho- of dissimilar materials by ~10×. We therefore chose to ton Engineering, LLC, Tucson, AZ, USA). A 4.1 × 4.1 use undoped YLF instead of sapphire as the thermal link mm link cross section was chosen to match the width of material, nominally yielding Δα≈0 and thus a stress-free the existing YLF:10%Yb crystal. The model used λ = adhesive-free bond between the YLF:10%Yb cooling 1004.7 nm as measured for YLF:Yb at 135 K, and a cor- crystal and the undoped YLF thermal link. However, this responding averaged refractive index of n = (2n + n )/3 o e results in a ~14× lower thermal conductivity of YLF (κ = = 1.456, where n = 1.448 and n = 1.471 , was used for a o e 24 W/m K and κ = 34 W/m K at 100 K ) than that of both undoped YLF and YLF:10%Yb. The fluorescence was sapphire. The undoped YLF thermal link, therefore, must represented by 10 rays that were uniformly and iso- be as short as possible to minimize the temperature dif- tropically emitted within the attached YLF:Yb crystal. As ference between the payload and YLF:Yb crystal, while shown in Figure 4a, an exponential decrease of the frac- still providing good fluorescence light rejection. tion of fluorescence power reaching the link/mirror P [mW] rad Hehlen et al. Light: Science & Applications (2018) 7:15 Page 6 of 10 interface η occurs when increasing the tilt angle δ; η P = 31.3 mW, of which 75% is due to residual mirror L L L does not significantly change for δ > 15°. This simulation absorption. Compared with a typical P = 470 mW (for cool yields P = P η η η (1 − R ), where R is the P = 47 W, η ≈ 1, and η = 0.01), the heat load intro- fluor in cpl ext L m m in cpl c reflectance of the mirror, which is set as 0.97 for the duced by the YLF thermal link is relatively small. sputtered silver layer used in this study, and P = 47 W, in η = 0.999, and η = 0.996 are assumed. The radiative Silica aerogel supports cpl ext heat load on the thermal link is approximately The laser-cooled assembly consisting of the YLF:Yb 4 4 P ¼ ε A σ T  T =ðÞ 1 þ χ . A 1-mm gap is assumed crystal, YLF thermal link, mirror, coldfinger, and payload rad l l c l between the thermal link and the TiNOX-coated clam- (Fig. 1) must be mounted within the closely fitting shell surfaces as well as ε =0.9 (from CaF ), ε = 0.05 clamshell structure in a manner that introduces minimal l 2 c (TiNOX solar absorber datasheet), T = 135 K, and T = thermal conduction from the clamshell to the cooled l c 300 K. Figure 4b plots P vs. P for thermal links with assembly, is mechanically reliable, and can be easily rad fluor different δ. The δ = 15° design achieves the lowest P , assembled. Earlier experiments using silica optical fibers whereas links with a smaller δ have a significantly larger with sharpened tips as support elements were difficult to P due to the greater η , and links with greater δ have a assemble and mechanically unreliable. The new approach fluor L slightly larger P due to their larger surface areas A and implemented in this study uses several silica aerogel rad l A . The link performance is slightly improved by provid- cylinders to support the cooled assembly under the ing a 90° inside angle (see inset Fig. 4a), a geometry that is coldfinger. Silica aerogels are open-celled mesoporous also more favorable for fabrication than geometries with SiO networks with low mass densities of typically smaller inside angles. The final thermal link design 0.01–0.2 g/cm as well as low optical absorption in the (Fig. 2c) is estimated to produce a total heat load of visible and near-infrared spectral range. Pure SiO ab YLF:Yb YLF:Yb crystal YLF crystal link YLF link HgCdTe sensor Cu coldfinger Cu coldfinger (1) (2) (3) 1′′ Cooled assembly Aerogel cylinder Clamshell #1 (4) (5) (6) Herriott cell mirrors Payload Clamshell #2 Clamshell #3 viewport Fig. 5 Solid-state optical cryocooler design and assembly. a Cooled assembly consisting of a YLF:Yb crystal, YLF thermal link, and copper coldfinger. b Image of the cooled assembly showing the HgCdTe sensor payload attached to the coldfinger. c Cryocooler assembly sequence consisting of the insertion of seven aerogel cylinders (1), placement of the cooled assembly (2), installation of side and top clamshell layers (3,4,5), and positioning and alignment of Herriott cell mirrors (6) Hehlen et al. Light: Science & Applications (2018) 7:15 Page 7 of 10 ab Coldfinger 280 YLF:Yb Initial slope: 0.31 K/mW 134.9 K 01 2 34 0 50 100 150 200 Time [h] Applied heat load [mW] Fig. 6 Performance of the solid-state optical cryocooler. a Temperature of the coldfinger (solid trace) and YLF:10%Yb crystal (dotted trace) as a function of time after turning on the 1020-nm laser in steps from 0, 12, 25, to 47 W. A steady-state coldfinger temperature of 134.9 K was reached after 4 h. b Heat load curve of the cryocooler pumped at 47 W, showing an initial slope of 0.31 K/mW aerogels have low thermal conductivities of ~0.004 W/m optical baffle that prevented residual fluorescence from K (at 300 K) and ~0.001 W/m K (at 100 K) in a vacuum , reaching the payload. making them attractive for use as support structures in The clamshell serves as the heat sink for the fluores- optical cryocoolers. The laser-cooled assembly (19.1 g cence emitted by the YLF:Yb crystal and must be designed total weight) rested on three 5 mm diameter × 5 mm tall to minimize the radiative heat loads [Eqs. (2)–(4)] it aerogel cylinders (58.9 mm total area), thus producing a imparts on the cooled assembly, which requires a geo- pressure of 3.18 kPa. Four additional aerogel cylinders metry that closely envelops the cooled assembly to were used around the outside perimeter of the coldfinger minimize the surface area and maximize χ. Figure 5c base to laterally secure the assembly. No mechanical shows the assembly sequence of the layered structure, degradation of the aerogel cylinders was observed under which allows for a stepwise installation of the clamshell at these conditions, which is consistent with the compressive a separation of 1 mm around the cooled assembly. We strength of silica aerogels, for which values of 1 MPa (0.25 used a TiNOX Energy solar absorber (Almeco GmbH, 3 26 3 27 g/cm ) and 0.047–0.11 MPa (0.1 g/cm ) have been Bernburg, Germany) glued to the inside surfaces of the reported. clamshell with silver epoxy to provide high absorbance of the YLF:Yb fluorescence and low thermal emissivity. The Laser-cooled assembly and clamshell clamshell also included a viewport that provided optical Earlier solid-state optical cryocooler designs envisioned access to the HgCdTe sensor payload. attaching the payload directly to the coated side of the 19, 28 thermal link , an approach that creates significant size Laser cooling of a HgCdTe sensor constraints for the placement of the support elements and Figure 6a shows the coldfinger temperature (measured potentially larger payloads. We therefore introduced a with a silicon diode) and the cooling crystal temperature copper coldfinger in the present cryocooler, allowing for (measured by DLT) as a function of time after turning on more space to accommodate larger payloads, aerogel the 1020-nm laser. The 1020 nm laser power was supports, as well as other instrumentation, such as a increased in steps from 0, 12, 25, to 47 W as a precau- temperature sensor and heating element. Figure 5a, b tionary measure to reduce the potentially large tempera- show the cooled assembly consisting of the YLF:Yb ture gradient and associated thermally induced stresses crystal/YLF thermal link unit and the HgCdTe infrared between the YLF:Yb crystal and coldfinger that develop at photo-sensor chip (InfraRed Associates, Inc.) attached to early times. This is evident in the stepwise decrease of the the copper coldfinger. The necked shape of the coldfinger YLF:Yb crystal temperature as measured by DLT. The near the thermal link interface was created to provide an coldfinger temperature reached 134.9 K after 4 h, Temperature [K] Cold finger temperature [K] Hehlen et al. Light: Science & Applications (2018) 7:15 Page 8 of 10 a b 300 296.0 Measurement Measurement Model Model 295.9 295.8 295.7 295.6 295.5 250 295.4 0 100 200 300 400 500 600 0 20 40 60 80 100 Time [s] Time [s] Fig. 7 a YLF:Yb cooling crystal temperature (T ) and b copper coldfinger temperature (T ) at early times. The measured (solid lines) and calculated x f (dotted lines) temperatures are shown as a function of time after turning on the 1020-nm pump laser at t = 0 with N = 1 rt representing the first ever demonstration of cooling a small temperature increase due to the small fluorescence payload to the cryogenic regime using solid-state optical leakage β at a rate given by ≈β η P (T )/C (T ) before l l ext abs c f c refrigeration. Furthermore, these data experimentally cooling from the crystal reverses this process in a time validate the various geometry and material choices for the scale approximated by τ ≈ K (T )/C (T ), as shown in l c l c thermal link, coldfinger, aerogel supports, and clamshell Figure 7b. Beyond this point, both the crystal and discussed above. cold-finger continue to cool and reach the steady- The results shown in Figure 6a were obtained without state condition with a temperature difference link link an electrical current flowing through the HgCdTe sensor, ðT  T Þ  2P =K ðT Þ, where P denotes the f x l l final load load i.e., without an external heat load introduced by the total heat load power (parasitic and useful) carried payload. A separate resistor installed in the coldfinger through the link. The data used in Figure 7 and the cor- base allowed for the application of well-defined heat loads responding fits were obtained for N = 1 (one laser rt and measurement of the corresponding increase in the roundtrip). The parameters obtained from these fits are coldfinger temperature. Figure 6b presents the heat load β ≈0.2%, η (T )≈1.2% at λ=1020 nm (corresponding to l c c −4 −1 curve obtained with the YLF:10%Yb crystal pumped at 47 η ≈0.992), α ≈2.1×10 cm , and ϵ  ϵ  0:38. The ext b x l W. The initial slope is 0.31 K/mW, a value that is char- values for η and α are in close agreement with pre- ext b 3+ acteristic for this particular cryocooler. Powering up the viously measured values for this 10% Yb -doped YLF HgCdTe sensor through a Fourier-Transform Infrared crystal. β also agrees with the calculated value from the (FTIR) spectrometer (Midac M4401) at the 134.9 K base raytracing model. T was fixed at 283.15 K, which is the temperature resulted in a temperature increase of ~2.5 K, final measured temperature of the clamshell. With the which corresponds to an estimated 8 mW heat load above parameters fixed, the model predicts a final cold- introduced by the active HgCdTe sensor. This value is in finger temperature T = 135.0 K with a temperature drop good agreement with the 8.8 mW heat load calculated of ΔT = T −T = 6.8 K across the link when the number f x from the 22.5 mA bias current and 17.3 Ω resistance (at of roundtrips is increased to N ≥30, as expected in the rt 134.9 K) of the HgCdTe sensor. Herriott cell arrangement. Considering the simplicity of The thermal model presented above [Eqs. (2)–(4)] our model, these calculated values are in excellent allows us to gain insight into the roles of the various terms agreement with our observed values of T = 134.9 K and in the observed temperature dynamics. The cooling ΔT = 6.1 K (Fig. 6a). The above calculations assumed a crystal temperature (T ) initially drops rapidly with the negligible heat lift (P = 0 mW) in comparison to the x lift slope given by ≈ −η (T )P (T )/C (T ) before the radia- existing radiative and fluorescence loads. Furthermore, c c abs c x x tive load (from the chamber) and conductive load the measured initial slope of 0.31 K/mW (Fig. 6b) is in (through the link) slow down this process, as shown in good agreement with ∂T /∂P ≈ 0.31 K/mW obtained f lift Figure 7a. The coldfinger, however, initially experiences a from the model calculations. Temperature [K] Temperature [K] Hehlen et al. Light: Science & Applications (2018) 7:15 Page 9 of 10 Laser-cooled HgCdTe sensor (135 K) Liquid-nitrogen cooled HgCdTe sensor (77 K) 2.0 1.5 1.0 0.5 0.0 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 –1 Wavenumbers [cm ] Fig. 8 FTIR spectroscopy using a laser-cooled HgCdTe infrared sensor. FTIR spectrum of a sheet of low-density poly-ethylene (LDPE) measured with the laser-cooled HgCdTe sensor (135 K, red trace) and, for comparison, an equivalent liquid-nitrogen cooled HgCdTe sensor (77 K, black trace). The larger relative noise observed at the extremes of the spectra is simply a result of the smaller signal (HgCdTe sensitivity and/or source brightness) in these regions; the noise amplitude is wavelength-independent FTIR spectroscopy with the laser-cooled HgCdTe sensor represents a breakthrough in this field and opens the door In a final experiment, the laser-cooled HgCdTe sensor to using this technology for a variety of applications that was used as part of an FTIR spectrometer and the infrared benefit from a reliable cryogenic refrigerator without absorption spectrum of a sheet of low-density poly-ethy- moving parts and associated vibrations. The all-solid-state lene (LDPE) was measured (Fig. 8). This is the first optical cryocooler was enabled by four key elements: First, demonstration of a sensor cooled by solid-state optical a high-quality YLF:Yb crystal with a low α and high η b ext refrigeration used as part of a practical device. Figure 8 provided efficient cooling at the heart of the cryocooler. shows the absorbance spectrum of the same sample Second, the use of a coldfinger provided the design flex- measured with an equivalent HgCdTe sensor cooled with ibility required to incorporate the sensor payload, support liquid nitrogen to 77 K. The bandgap energy E = hc/λ of structure, and diagnostic components. Third, the results this HgCdTe sensor corresponds to a cutoff wavelength of show that hydrophobic silica aerogels can provide excel- λ = 16.7 µm. The primary noise source is due to statistical lent thermal insulation as well as sufficient mechanical fluctuations of the thermally activated dark current within strength to support the laser-cooled assembly. Fourth, we the detector element. The dark current I is proportional demonstrated that a thermo-mechanically reliable van- 3+ to the number of carriers that are thermally excited across der-Waals bond can be created between Yb -doped and the bandgap and thus varies approximately as I ∝exp undoped YLF, which allowed the incorporation of a (−E /k T ),with the corresponding noise amplitude vary- transparent thermal link without the use of optical g B f pffiffiffiffi ing as I . The detector noise amplitude is therefore adhesives. These approaches are not without new chal- expected to be ~11× greater at 135 K than at 77 K, which lenges. The relatively poor thermal conductivity of the is consistent with the observed lower signal-to-noise ratio YLF thermal link creates a larger than desired tempera- in the measurement using the laser-cooled sensor. ture gradient between the payload and cooling crystal, which limits the base temperature of the payload. Ther- Conclusions mal link materials that have significantly greater thermal We used solid-state optical refrigeration to cool a pay- conductivity and a CTE that is comparable to the cooling load to cryogenic temperatures for the first time, which crystal are required to realize the full potential of optical Absorbance Hehlen et al. Light: Science & Applications (2018) 7:15 Page 10 of 10 refrigeration. The silica aerogel supports performed well 9. Hehlen,M.P.etal. Preparation of high-purity LiF, YF3, and YbF3 for laser refrigeration. Proceedings of SPIE 9000, Laser Refrigeration of Solids VII (SPIE, San in the present static application. More studies are needed Francisco, CA, 2014). to assess their thermo-mechanical performance under 10. Melgaard, S. D., Albrecht, A. R., Hehlen, M. P. & Sheik-Bahae, M. conditions of mechanical shock and vibration. Solid-state optical refrigeration to sub-100 Kelvin regime. Sci. Rep. 6, 20380 (2016). 11. Seletskiy, D. V., Melgaard, S. D., Di Lieto, A., Tonelli, M. & Sheik-Bahae, M. Acknowledgements Laser cooling of a semiconductor load to 165 K. Opt. Exp. 18,18061–18066 The authors thank T. Williamson and K. Baldwin (Los Alamos National (2010). Laboratory) for depositing the Ag/Au mirror and T. Cardenas (Los Alamos 12. Coluccelli, N. et al. Diode-pumped passively mode-locked Yb:YLF laser. Opt. National Laboratory) for performing sample metrology. Work at UNM was Exp. 16,2922–2927 (2008). partially supported by the Air Force Office of Scientific Research (AFOSR) under 13. Meissner H. E. Composites made from single crystal substances: U.S. Patent award numbers FA9550-15-1-024 and FA9550-16-1-0362 (MURI). 5441803. 1995-05-15. 14. Li, D.,Hong,P., Vedula,M.& Meissner,H.E. Thermal conductivity investigation of Author contributions adhesive-free bond laser components. Proceedings of SPIE 10100, Optical Com- M.P.H., M.S-B, and R.I.E., conceived the approach and supervised the work. M.P. ponents and Materials XIV. (SPIE, San Francisco, CA, 2017). H. designed the thermal link, and E.R.L. designed the clamshell. J.M., A.G., and 15. Hamilton, C. E. et al. Adsorption of ambient moisture by silica aerogel. Fusion A.R.A. performed the laser-cooling experiments at UNM. A.G. first implemented Sci. Tech. 63,301–304 (2013). the Herriott cell. C.E.H. synthesized the silica aerogels at LANL. S.P.L. assisted 16. Aggarwal, R. L.,Ripin,D.J., Ochoa, J. R. &Fan,T.Y.Measurement of thermo- with the FTIR measurements. All authors participated in writing the article. We optic properties of Y Al O ,Lu Al O ,YAlO ,LiYF ,LiLuF ,BaY F ,KGd(WO ) , 3 5 12 3 5 12 3 4 4 2 8 4 2 would like to dedicate this paper to the memory of co-author Eric R. Lee who and KY(WO ) laser crystals in the 80-300 K temperature range. J. Appl. Phys. 4 2 passed away unexpectedly since the submission of the manuscript. He was 98, 103514 (2005). instrumental to the success of this work. 17. Gragossian, A., Meng, J. W., Ghasemkhani, M., Albrecht, A. R. & Sheik- Bahae, M. Astigmatic Herriott cell for optical refrigeration. Opt. Eng. 56, Conflict of interest 011110 (2016). The authors declare that they have no conflict of interest. 18. Parker, J. et al. Thermal links for the implementation of an optical refrigerator. J. Appl. Phys. 105, 03116 (2009). 19. Melgaard, S. D. Cryogenic optical refrigeration: laser cooling of solids below Received: 28 February 2018 Revised: 24 April 2018 Accepted: 24 April 2018 123 K. Ph.D. dissertation, University of New Mexico, Albuquerque, NM (2013). . 20. Dobrovinskaya,E.R., Lytvynov, L.A., & Pishchik, V. Sapphire: Material, Manu- facturing, Applications (Springer, New York, NY, 2009). . 21. Onyx Optics, Inc. Mechanism of adhesive-free bonding (AFB®) (Onyx Optics, Inc.: References Dublin, CA, 2008). 1. Epstein, R. I., Buchwald, M. I., Edwards, B. C., Gosnell, T. R. & Mungan, C. E. 22. Payne, S. A. et al. 752 nm wing-pumped Cr:LiSAF laser. IEEE J. Quantum Observation of laser-induced fluorescent cooling of a solid. Nature 377, Electron 28, 1188–1196 (1992). 500–503 (1995). 23. Barnes, N. P. &Gettemy,D.J.Temperature variationofthe refractive indicesof 2. Thiede, J., Distel, J., Greenfield, S. R. & Epstein, R. I. Cooling to 208 K by optical yttrium lithium fluoride. J. Opt. Soc. Am. 70,1244–1247 (1980). refrigeration. Appl. Phys. Lett. 86, 154107 (2005). 24. Muley,S.V.&Ravindra, N.M.Emissivityofelectronicmaterials,coatings, and 3. Seletskiy, D. V. et al. Laser cooling of solids to cryogenic temperatures. Nat. structures. JOM 66,616–636 (2014). Photonics 4,161–164 (2010). 25. Rettelbach, T., Säuberlich, J., Korder, S. & Fricke, J. Thermal conductivity of silica 4. Seletskiy, D. V.,Epstein,R.& Sheik-Bahae, M. Laser cooling in solids: advances aerogel powders at temperatures from 10 to 275 K. J. Non-Cryst. Sol. 186, and prospects. Rep. Prog. Phys. 79, 096401 (2016). 278–284 (1995). 5. Nemova, G. Laser Cooling: Fundamental Properties and Applications (Pan 26. Parmenter, K. E. & Milstein, F. Mechanical properties of silica aerogels. J. Non- Stanford, Singapore, 2016). Cryst. Sol. 223,179–189 (1998). 6. Hehlen,M. P., Sheik-Bahae, M.,Epstein,R.I., Melgaard, S.D.& Seletskiy, D. V. 27. Omranpour, H., Dourbash, A. & Motahari, S. Mechanical properties improve- Materials for optical cryocoolers. J. Mater. Chem. C. 1, 7471–7478 (2013). ment of silica aerogel through aging: role of solvent type, time and tem- 7. Epstein, R. I. & Sheik-Bahae, M. Optical Refrigeration: Science and Applications of perature. AIP Conf. Proc. 1593, 298 (2014). Laser Cooling of Solids (Wiley, Weinheim, 2009). 28. Hehlen, M. P.,Boncher, W.L.& Love, S.P. Design study of a laser-cooled infrared 8. Sheik-Bahae, ?tlsb?>M. & Epstein, R. I. Optical refrigeration. Nat. Photonics 1, sensor. Proceedings of SPIE 9380, Laser Refrigeration of Solids VIII.(SPIE,San 693–699 (2007). Francisco, CA, 2015).

Journal

Light: Science & ApplicationsSpringer Journals

Published: Jun 6, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off