Full Terms & Conditions of access and use can be found athttps://www.tandfonline.com/action/journalInformation?journalCode=ghpr20

High Pressure ResearchAn International Journal

ISSN: 0895-7959 (Print) 1477-2299 (Online) Journal homepage: https://www.tandfonline.com/loi/ghpr20

Advanced integrated optical spectroscopy systemfor diamond anvil cell studies at GSECARS

Nicholas Holtgrewe, Eran Greenberg, Clemens Prescher, Vitali B. Prakapenka& Alexander F. Goncharov

To cite this article: Nicholas Holtgrewe, Eran Greenberg, Clemens Prescher, Vitali B. Prakapenka& Alexander F. Goncharov (2019): Advanced integrated optical spectroscopy system for diamondanvil cell studies at GSECARS, High Pressure Research, DOI: 10.1080/08957959.2019.1647536

To link to this article: https://doi.org/10.1080/08957959.2019.1647536

Published online: 01 Aug 2019.

Submit your article to this journal

View Crossmark data

Advanced integrated optical spectroscopy system fordiamond anvil cell studies at GSECARSNicholas Holtgrewea, Eran Greenberga, Clemens Prescherb, Vitali B. Prakapenkaa* andAlexander F. Goncharovc

aCenter for Advanced Radiation Sources, University of Chicago, Chicago, IL, USA; bPhoton Sciences, DeutschesElektronen-Synchrotron (DESY), Hamburg, Germany; cGeophysical Laboratory, Carnegie Institution ofWashington, Washington, DC, USA

ABSTRACTRaman and optical spectroscopy are versatile tools fornondestructive characterization of a wide range of properties ofnovel materials and minerals in situ at extreme and ambientconditions. These techniques are genuinely complementary toX-ray tools (diffraction and spectroscopy) in the probe energy,momentum transfer, and time scale, making concomitant X-rayand optical probes available for advanced sample analysis. Wehave built a state-of-the-art, user-friendly integrated Raman andoptical spectroscopy system at Sector 13 (GeoSoilEnviroCARS,University of Chicago, IL) of the Advanced Photon Source (APS),Argonne National Laboratory (ANL), where optical probes areavailable now in combination with high resolution in-situsynchrotron X-ray diffraction and spectroscopy tools (XRD, IXS,XES, NFS, and others) for extensive sample investigation. Theintegrated optical system enables a variety of techniquesincluding multi-colored (five laser lines: 266, 473, 532, 660, and946 nm) confocal Raman, fluorescence, and optical spectroscopyfrom ultraviolet (UV) to near infrared (IR) spectral ranges(266–1600 nm), and Coherent Anti-Stokes Raman spectroscopy(CARS) in combination with near IR double sided laser heating.

ARTICLE HISTORYReceived 21 May 2019Accepted 21 July 2019

KEYWORDSRaman spectroscopy; laserheating; diamond anvil cell

Introduction

Optical vibrational and electronic spectroscopy is a useful tool to probe electronic andatomic structure, chemical bonding, and vibrational, elastic, and thermodynamic proper-ties. Such spectroscopic measurements of materials applied under extreme conditionsprovide unique information about changes in structure, bonding, thermochemical andother properties. These investigations have a broad applicability to several disciplinesincluding studies of the Earth’s and planetary interiors, synthesis of novel materials,nanoscience/nanotechnology, physics and chemistry of materials under extremeconditions, etc. The spectroscopic techniques are uniquely complementary and oftenindispensable (e.g. in case of light elements) to synchrotron X-ray diffraction (XRD) and

© 2019 Informa UK Limited, trading as Taylor & Francis Group

CONTACT Nicholas Holtgrewe nicholas.holtgrewe@gmail.com Center for Advanced Radiation Sources, Universityof Chicago, Chicago, IL 60637, USA*Contact for system usage: prakapenka@cars.uchicago.edu.

HIGH PRESSURE RESEARCHhttps://doi.org/10.1080/08957959.2019.1647536

spectroscopy, as well as neutron diffraction methods that are commonly used to deter-mine the structure and a range of material properties. Raman spectroscopy is one ofthe major analytical tools to probe these material states and has been widely used forthe last four decades to study materials under extreme conditions of high pressure andvariable temperature (P-T ) and to characterize the materials at ambient pressure [1–8].This technique has important advantages of being noninvasive, nondestructive, fast,easily interpretable, and highly informative about the material’s structural, chemical,and electronic states.

The advent of synchrotron X-ray sources has made profound advances in understand-ing properties of geological and planetary materials under variable P–T and compositionalenvironments and in characterization of novel materials of interest to industry and tech-nology [9–11]. Synchrotron XRD facilities across the world have been a staple for samplecharacterization, however often this data is inconclusive, while complementary spectro-scopic techniques are unavailable at the time of X-ray experiments. Thus, most timesthis spectroscopic analysis is performed separately at a home facility with differentsamples due to the destructive/hysteretic nature of material compression and decompres-sion at the XRD beamlines. The GeoSoilEnviroCARS (GSECARS, University of Chicago)beamline (Sector 13) at the Advanced Photon Source (APS) in Argonne National Labora-tory (ANL) is committed to solving this issue by providing users a variety of advancedanalytical techniques to enhance data collection quality and efficiency at the facility.Here we address this problem by building an off-line Raman/optical integrated instrument,where the same samples being investigated by XRD and/or XES can be also probed byoptical techniques at identical or very similar P-T conditions. The advanced opticalsystem described in this work was constructed at GSECARS and consists of spectroscopictechniques optimized for (but not limited to) diamond anvil cells (DACs) to fully character-ize samples in-situ and in combination with the X-ray data collection. This paper will focuson the detailed descriptions of the optical layout, multi-colored confocal Raman system,Raman mapping features, near IR laser heating, Coherent Anti-Stokes Raman spectroscopy(CARS) system, and additional software tools for data analysis.

Overview

Raman and optical spectroscopy tools in applications to experiments at extreme con-ditions must have high spatial resolution and be very versatile. This is dictated by agreat variability of material properties at a wide range of extreme P-T conditions thatinclude changes in electronic and crystal structure, bonding length and type, amorphiza-tion, melting, etc. that all affect substantially the spectral characteristics and the optimalconditions of their collection. Moreover, the optical and vibrational properties ofdiamond anvils change drastically with pressure and temperature, making it necessaryto create appropriate provisions for overcoming the adverse effects (e.g. stress inducedfluorescence [12,13]). Many of these problems can be addressed and mitigated by choos-ing the most appropriate excitation wavelength, objective lens type, spectrometer grating,laser power, etc. [7,14]. For example, to avoid diamond and sample fluorescence caused bydefects in electronic levels, which is often within the visible spectrum, one can use eitherUV or near IR excitations [15,16]. To increase the efficiency of the Raman scattering oneshould choose as short of an excitation wavelength as possible, as it is proportional to

2 N. HOLTGREWE ET AL.

the fourth power of the excitation frequency, ν4exc. On the other hand, high energyphotons may destroy sample, while use of longer excitations wavelengths (low energyphotons) are less destructive and allows achieving higher spectral resolution routinely.To excite ruby fluorescence at high pressure a variety of different excitation wavelengthsmay be needed, and additionally time resolved measurements could be advantageous[17,18]. The vibrational excitations often reveal resonance behavior when the laserquantum energy becomes close in energy to the electronic excitation energy or the elec-tronic bandgap [19–21]. This resonance behavior can be utilized to determine changes inthe electronic structure of the materials under pressure including metallization via theband gap closure.

In designing our system, we had the above considerations in mind to make a provisionfor as broad as possible user community and users with different levels of expertise. Thesystem presented here enables Raman studies for a range of excitation wavelengthsvarying from UV (266 nm) through visible (473, 532 and 660 nm) to near IR (946 nm).These capabilities are supported by the choice of special optics, spectrometer diffractiongratings, and the detectors; all of which provide the Raman coverage up to at least4500 cm−1 for all excitation wavelengths. With the exception of the UV 266 nm excitation,the Raman spectra can be extended down to approximately 10 cm−1 due the use of solidstate volume Bragg filters [22]. One other extremely important feature of such operation isthe ability to perform Raman (and also CARS and optical spectroscopy) measurementsfrom the same sample position. In the system, this is realized via an automatic changeof the laser wavelength, which includes change of the interface optics for the laser exci-tation and signal collection optical paths. All these features enable the most versatileRaman operations. In addition to conventional spontaneous Raman spectroscopy, thenew system provides an access to multiplex CARS spectroscopy, which involves a non-linear three-wave stimulated mixing process, thus effectively increasing the efficiency ofthe scattering process via its coherent nature. This technique (as well as conventionalRaman) can be combined with a near IR laser for double-sided sample heating. Moreover,a white pulsed brilliant CARS supercontinuum light source can be used for optical trans-mission and reflectance spectroscopy to probe the materials electronic properties in situ[23–26].

Optical layout

The full layout of the instrument on a 5′ × 7′ optical table (Newport Integrity 2) is shown inFigure 1, a top-view of a 3D model of the table [27] (a) and an image of the table with theenclosure doors open (b). Note that all lasers are located along the far wall and are wellblocked by individual enclosures with interlocked shutters in addition to the enclosuredoors and ceiling. Each enclosure door is linked to an interlock system that blocks theexposed beam by disabling the laser shutters if the user attempts to open the enclosurewhile a beam is exposed. A description of each laser line path will follow in the next sub-sections, but the general sample area is broken into three sample stations: General Raman,Ultra-Violet (UV) Raman, and Advanced (Laser Heating/CARS). The user has the option tochoose between any of these three sample positions by using the horizontal movement ofthe sample stage (S1). The sample mounts are identical to DAC holder mounts used at theSector 13 and other APS beamlines (BKL-4 kinematic base with DAC holders positioning

HIGH PRESSURE RESEARCH 3

sample at 3.5 in. from base plate surface), allowing for easy transfer of the sample betweenAPS X-ray beamlines and the Raman spectrometer. Additionally, there are custom mountsfor any type of free-standing samples. Visualization of the sample is accomplished by semi-transparent pellicle mirrors (Figure 1(a), P) that guide the lights toward the sample and thecollected image toward a camera. Thepelliclemirrors near the objective lenses aremountedon motorized flip mounts to avoid the signal path during collection. Alignment of the

Figure 1. (a) 3D model of the optical layout for the advanced optical system. Each laser excitation pathis color coordinated: λex = 946 nm (gray), 660 nm (red), 532 nm (green), 473 nm (blue), 266 nm (cyan),Laser Heating (white), and CARS (orange and green dashed). For CARS the orange is the pulsed super-continuum path and the green is the pulsed 532 nm path. FM1: Motorized Flip Mirror, P: Pellicle Mirror(55:45), BS: Polarized Beamsplitter, DM: Dichroic Mirror (1064 nm), L: Lens, PH: Pinholes (diameter 50–75 µm), NF: Notch Filter (specific to excitation wavelength). (b) Photo of the optical system with theenclosures doors open.

4 N. HOLTGREWE ET AL.

pellicle mirrors was done visually with a standard sample (e.g. silicon) to make the sampleimage and laser spot confocal with the spatial filter (i.e. projection of pinhole onto thesample). Only minor adjustments of the spatial filter and sample imaging were optimizedfor maximum signal. A comprehensive list of the spatial filter sizes is shown in Table 1 foreach excitation wavelength and objective lenses. Detection of signals is accomplishedusing the Princeton Instruments Acton Series 2500 (focal length = 500 mm) spectrographin combination with three gratings: 1800 (UV enhanced), 1200, and 300 grooves/mm andtwo detectors: UV-enhanced PIXIS100 (200–1000 nm) and NIRvana (850–1600 nm). A com-prehensive chart showing the grating efficiencies, detector ranges, and Raman collectionranges (0–4500 cm−1 for each excitation laser) is shown in Figure 2. Additionally, themaximum powers for each excitation laser and percent attenuation due to optics in thesystem is shown in Table 2.

General Raman

The general Raman sample station is meant for users to take routine Raman measure-ments that do not require any advanced techniques (e.g. laser heating) but still providethe same high-quality data collection as the Advanced station. This is the most versatileposition as the user has the option for four different excitation lasers (single frequencynarrow band 946, 660, 532, and 473 nm with the bandwidth <0.03 nm) as well as fourdifferent long working distance objective lenses (5x, 10x, 20x, and 50x). The objectivelenses are mounted onto a kinematic Mitutoyo manual turret for convenient usage. Theoptical path of each excitation laser is shown in Figure 1(a) and is indicated by its assignedcolor: gray, red, green, and blue for 946, 660, 532, 473 nm laser wavelengths, respectively.Each line follows the same design where the excitation beam is guided toward a 300 mmtranslation stage holding an optical breadboard containing specific optical componentsfor individual laser frequencies. In Figure 1(a), only the 660 nm line is shown to be com-plete, however moving the translation stage in 3′′ steps to the left selects other laser exci-tations. Changing between excitation lasers is limited by warming up the laser andrepositioning of this translation stage, typically a few minutes or faster. Following the660 nm line, the laser beam is expanded by two lenses to a 6 mm diameter and guidedtoward the translation stage with protected silver mirrors (Note: all mirrors are protectedsilver mirrors unless indicated otherwise). On the translation stage, the beam is reflectedoff a notch filter (NF) (OptiGrate) and guided off the translation stage and toward thesample. The beam is then focused down onto the sample with the user selected objectivelens and the backscattered light is collected, following the same path as before back to thenotch filter. The notch filter is angled to where it blocks the excitation frequency optimally

Table 1. Spatial filter diameter (in µm) projected onto sample (measured using silicon).

Objective Lens Magnification946 nm

75 µm pinhole660 nm

50 µm pinhole532 nm

50 µm pinhole473 nm

50 µm pinhole266 nm

50 µm pinhole

5x 27 18 18 1810x 16 9 9 9 1020x 8 4.5 4.5 4.550x 3 2 2 220x (LH) 8 4.5 4.5 4.5

Note: LH = laser heating objective lens.

HIGH PRESSURE RESEARCH 5

but allows the Raman signal to pass. This signal is then sent through a spatial filter, com-posed of two achromatic convex lenses (L) and a pinhole (PH) optimized for the spectralrange of the signal of interest, and two additional notch filters before being focused intothe spectrograph with a 75 mm achromatic convex lens for detection. An example of allfour excitations being used is shown in Figure 3, where elemental sulfur is analyzed.This shows the low Raman frequencies, for both Stokes and Anti-Stokes frequencies, thesystem is capable of detecting (estimated to be less than 10 cm−1 for all the excitationwavelengths) and spectral resolution of the instrument. An example of ultra-high pressureapplication of the system is shown in Figure 4, where the diamond edge is located around1686 cm−1 corresponding to a pressure of 197 GPa and a few sample peaks are clearlyvisible. In addition to Raman measurements, the system is also capable of detectingruby fluorescence for pressure measurements with a <1 mW supplementary class-IIbroadband 532 nm laser that can enter the green beam path using pneumatic switches.

Polarized Raman

As of the time of this publication, the 660 nm laser line is the only line that has beenadapted for polarized Raman spectroscopy. The addition of a rotatable half-wave plate

Figure 2. Grating efficiency curves for the three gratings available in the spectrograph (1800, 1200, and300 grooves/mm). The detector ranges are displayed in block format at the top (PIXIS gray, NIRvanadark gray) and the Raman ranges (0–4500 cm−1) for each laser excitation are shown at the bottom.

Table 2. Summary of maximum power for each laser excitation line and percent attenuation of thebeam prior to reaching the sample.Excitation line (nm) Maximum power (mW) Attenuation due to optics (not including objective lens)

660 500 ∼20%532 2000 ∼20%473 350 ∼40%946 200 ∼50%266 200 ∼30%

6 N. HOLTGREWE ET AL.

near the output of the laser and a rotatable linear polarizer before the spectrograph allowsfor all four parallel/cross laser polarization orientations, using Porto’s notation [28]: X(YY)X,X(YZ)X, X(ZZ)X, and X(ZY)X. An example of LiNbO3 Raman signal using all these orien-tations is shown in Figure 5, which can be compared to previous observations [29]. Thisfeature is a vital tool for characterization of single crystals and liquids as it provides anassignment of various vibration modes due to their selectivity to polarization selectionrules (e.g. Ref. [30,31]).

UV Raman

The UV Raman path is separate from the other laser excitation paths due to the use ofspecial UV optics (i.e. all mirrors are aluminum enhanced mirrors and lenses are made

Figure 3. Raman spectra of elemental sulfur (S) taken using different excitation wavelengths λex = 946,660, 532, and 473 nm.

Figure 4. Raman spectrum of a metal hydride at 197 GPa. Collection details: 20 s exposure time, 300grooves/mm grating, 20x objective lens, λex = 660 nm, 50 mW, PIXIS detector.

HIGH PRESSURE RESEARCH 7

of quartz). The 266 nm beam is produced by frequency doubling (Coherent MBD200) theoutput of a 532 nm laser (Coherent VERDI V2) and is expanded and guided toward abroadband polarized cube beamsplitter, where the power is selected by rotating a half-waveplate prior to the cube. The beam is sent toward the quartz objective lens (MitutoyoPlan UV Infinity Corrected 10x) where it is focused down onto the sample and the back-scattered Raman signal is collected following the same path back to the polarized cubebeamsplitter. The signal is then sent through a spatial filter, similar to the generalRaman lines, and razor edge filter to reject the excitation line prior to entering the spectro-graph in a different entrance port than the other Raman lines. The Raman spectra can bemeasured down to 230 cm−1 for the UV Raman part of the system that is limited by avail-able UV laser line filters. An example of UV Raman spectra are diamond and sapphireshown in Figure 6. In the case for diamond, the background fluorescence is substantiallyreduced resulting in clear detection of the first and second ordered signals for diamonds,which are highly fluorescent in the visible spectral range. In addition to fluorescence avoid-ance, other benefits to UV Raman include increased scattering efficiency and resonanceRaman spectroscopy as mentioned above. It is worth noting that in case of high pressurestudies with DACs the upstream diamond anvil must be of the Type II variety, as to allowlow absorbance of the 266 nm excitation beam/collected signal. There are no specificlimits for free standing samples except special attention should be taken for laserpower control with available range of 1–200 mW as high energy UV photons maydestroy samples.

Raman mapping

In combinationwith all the spectroscopic techniques, there is the ability to perform a Ramanmap for a user-defined region. Themotorized samplemount is computer controlled with allthree XYZ axes and can thusmove the sample incrementally in a grid-like pattern (rectangu-lar or square). An example of one XY Raman map taken at 532 nm excitation is given in

Figure 5. Polarized Raman spectra of LiNbO3 measured in a backscattering geometry for different inci-dent and collected polarized light (Porto notation shown) (λex = 660 nm).

8 N. HOLTGREWE ET AL.

Figure 6. UV (λex = 266 nm) Raman spectra for diamond (left) and sapphire (right). For diamond, the UV Raman (purple) is compared with 473 nm excitation (blue),which contains the Raman signal with an undesirable fluorescence background.

HIGHPRESSU

RERESEA

RCH

9

Figure 7. The map is a grid (Figure 7, inset) where each square represents an integratedintensity distribution that is a user defined Raman shift frequency range (Figure 7, redbox), typically the full-width half maximum of an identifying feature (vibron, roton, etc.)of the sample. In Figure 7, the sample was an iron oxide at high pressure that was analyzedpost laser heating with the map representing the intensity of the Raman feature at∼150 cm−1. The map clearly indicates regions that were heated (purple spectrum) andunheated (blue spectrum), allowing the user to visualize the laser heated area. The spatialresolution of the Ramanmapping is dependent on the choice of both laser excitationwave-length and objective lens magnification, with the lower excitation wavelengths and highermagnifications giving the highest resolution.

Advanced (laser heating/CARS)

The third sample position (Figure 1(a)) is reserved for more advanced techniques, such asIR laser heating [32,33] and CARS [23,34]. It is composed of two objective lenses (Mitutoyonear IR 20x) allowing for both double sided laser heating and CARS which relies ontransmission geometry. All excitation lasers listed in the general Raman path can beused in the Laser Heating/CARS path by use of one flip mirror (Figure 1(a), FM1) thatpicks off the input laser. The laser heating is accomplished by a linear polarized 100Wnear IR (1064 nm) laser (IPG Photonics 1YLR-100-1064-LP) that is shaped into a flat-topprofile by a Pi-shaper (AdlOptica) and divided into two beams by a polarized cube beams-plitter. The power ratio between both sides of the sample can be adjusted by a single half-waveplate prior to the beamsplitter. These two beams are guided into the sample by a pairof 1064 nm dichroic mirrors (Semrock). Temperature detection is accomplished by usingthe corresponding Raman optical path, moving grating positions and motorized mirror

Figure 7. Example of a Raman map generated in the updated T-Rax software of laser heated LiFe5O8

ordered spinel at 15 GPa (heated to 1600 K, quenched to 300 K). The 7 × 7 grid (inset) represents theuser-defined map (2 µm step size) for a total of 49 Raman spectra (two spectra shown on the plot). Thenormalized intensity of each square (scale-bar left inset) is directly related to the Raman frequency at∼150 cm-1 (red box), showing the location of unheated (blue square) and heated regions of sample(purple square). Collection details: 10 s exposure time, 1200 grooves/mm grating, λex = 660 nm,100 mW, PIXIS detector.

10 N. HOLTGREWE ET AL.

flipping. For example, in Figure 1(a), if the temperature from the down-stream side of thesample is desired, the pellicle mirrors would flip out and mirror FM2 would flip into theoptical path. The collected thermal radiation signal would then be sent to the spectro-graph through the user defined laser line path.

Coherent Anti-Stokes Raman spectroscopy (CARS) is a technique of choice when stron-ger signals selective for the materials of interest are needed. This method can also be com-bined with laser heating to provide information that can be difficult to obtain withstandard Raman plus laser heating in a diamond anvil cell. The CARS signal is generatedthrough a three-wave mixing process between the wpump, wprobe, and wstokes frequencies[34]. In this system we utilize a pulsed supercontinuum light source with a 532 nm output(LEUKOS, 1 ns, 33 kHz), where the supercontinuum (wstokes) and 532 nm (wpump andwprobe) beams are overlapped spatially and temporally to achieve maximum intensity ofthe CARS signal. As the name implies, CARS probes the enhanced Anti-Stokes vibrationalsignal, which is always higher in energy than the excitation laser frequency and henceavoids background contamination from fluorescence and/or certain high temperatureconditions. An example of CARS signal taken in the optical system is high pressure nitro-gen (N2, 22 GPa) at room temperature, shown in Figure 8, with the N2 vibrons in agree-ment with previous reported data [35].

Software

The entire system is computer controlled using the Experimental Physics and IndustrialControl System (EPICS), which integrates with the data collection software LightField (Prin-ceton Instruments).

Preliminary data analysis is performed using the T-Rax software (http://www.clemensprescher.com/programs/t-rax). T-Rax is a Python based software which providesthe capability to:

(1) Obtain temperatures during laser heating by fitting a Planck curve to the spectro-radio-metric measurement signal for both sides and as a temperature series whenmulti-frame

Figure 8. CARS spectra at room temperature for nitrogen at 22 GPa (green curve) compared with spon-taneous Raman (blue curve). The peak at ∼1350 cm−1 is from the diamond anvil.

HIGH PRESSURE RESEARCH 11

data has been recorded. In order to estimate temperatures, themeasured signal has to becorrected for the optical system response. In T-Rax, twoprocedures are implemented. Thecalibrationof the systemcanbe achievedbymeasuring either a spectrumof a sourcewitha known intensity at each wavelength (etalon) or a source with black body radiation at aknown temperature. Different calibrations for e.g.multiple setups or different calibrationsources can be saved and restored on the fly for fast comparison.

(2) Obtain pressure in the sample chamber by fitting the position of ruby fluorescencepeaks or the position of the diamond edge, whereby the several different availablepressure calibrations can be selected.

(3) Compare Raman spectra using overlays or combined with the above-mentionedmapping feature.

In all the different modes of T-Rax the region of interest of the detected image in thespectrometer can be selected graphically. T-Rax makes heavy use of the existent scientificopen source Python infrastructure. The graphical user interface (GUI) is written using thePyQt5 library (www.riverbankcomputing.com/software/pyqt), which provides Pythonbindings for the Qt5 library (www.qt.io). We chose to use PyQt-Graph (http://www.pyqtgraph.org) for plotting images and spectra.

Conclusion

We have designed and built the integrated optical system combining in one instrumentRaman spectroscopy with five excitation wavelengths (266, 473, 532, 660, 946 nm),double sided IR laser heating, coherent anti-Stokes Raman scattering, and broadband(200–1600 nm) high-efficiency optical laser spectroscopy (with a supercontinuum lasersource above 400 nm). The system is optimized for sample characterization in situ atextreme conditions in the diamond anvil cell (DAC) but could be also used for free-stand-ing samples as well. Utilizing a compact custom-designed cryogenic high pressure system,that is currently under development at GSECARS, we will be able to study optical proper-ties of materials at temperatures down to 10 K.

The advanced optical system for materials and minerals research under extremepressure and temperature is fully operational and located at the GSECARS, Sector 13,APS, ANL. It is available for users through the APS GUP proposal system if it is a part ofbeamline proposal or upon submitting a request at the GSECARS website (https://gsecars.uchicago.edu). It is expected that the additional spectroscopic data acquiredwith this system will complement X-ray synchrotron data collected at various beamlinesat APS as well as enhance the user’s individual projects.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This work was supported by NSF MRI proposal (EAR-1531583). The system is located at GeoSoilEnvir-oCARS (The University of Chicago, Sector 13), Advanced Photon Source (APS), Argonne National Lab-oratory. GeoSoilEnviroCARS is supported by the National Science Foundation – Earth Sciences (EAR –

12 N. HOLTGREWE ET AL.

1634415) and Department of Energy-GeoSciences (DE-FG02-94ER14466). The Advanced PhotonSource, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOEOffice of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357. Thiswork was also supported by DOE – Geosciences and DOE Office of Science by ANL.

References

[1] McMillan PF. Raman Spectroscopy in Mineralogy and Geochemistry. Annu Rev Earth Planet Sci.1989;17:255–279.

[2] Hemley RJ, Porter RF. Raman spectroscopy at ultrahigh pressures. Scr Metall. 1988;22(2):139–144.

[3] Gillet P, Hemley RJ, McMillan PF. Vibrational properties at high pressures and temperatures. RevMineral. 1998;37:525–590.

[4] McMillan PF, Dubessy J, Hemley RJ. Chapter 7: Applications in earth, planetary and environ-mental sciences. In: Turrell G, Corset J, editors. Raman Microscopy: development and appli-cations. New York: Academic Press; 1996. p. 289–365.

[5] Goncharov AF, Struzhkin VV, Gregoryanz E, et al. Raman scattering of metals to very high press-ures. In: Hemley RJ, Chiarotti GL, Bernasconi M, editors. Proceedings of the international schoolof Physics, “Enrico Fermi” Course CXLVII. Amsterdam: IOS Press; 2002. p. 297–313.

[6] Goncharov AF, Struzhkin VV. Raman spectroscopy of metals, high-temperature superconduc-tors and related materials under high pressure. J Raman Spectrosc. 2003;34(7-8):532–548.

[7] Goncharov AF. Raman Spectroscopy at high Pressures. Int J Spectrosc. 2012;2012:617528.[8] Lin JF, Alp EE, Goncharov AF. Raman and nuclear resonant spectroscopy in geosciences. In:

Holland HD, Turekian KK, editors. Treatise on geochemistry. 2nd ed. Oxford: Elsevier; 2014. p.195–211.

[9] Duffy TS. Synchrotron facilities and the study of the Earth’s deep interior. Rep Prog Phys.2005;68(8):1811–1859.

[10] Shen G, Mao HK. High-pressure studies with x-rays using diamond anvil cells. Rep Prog Phys.2017;80(1):016101.

[11] Hemley RJ, Mao HK, Struzhkin VV. Synchrotron radiation and high pressure: new light onmaterials under extreme conditions. J Synchrotron Radiat. 2005;12:135–154.

[12] Liu J, Vohra YK. Fluorescence emission from high purity synthetic diamond anvil to 370 GPa.Appl Phys Lett. 1996;68(15):2049–2051.

[13] Goncharov AF, Gregoryanz E, Hemley RJ, et al. Spectroscopic studies of the vibrational and elec-tronic properties of solid hydrogen to 285 GPa. Proc Natl Acad Sci U S A. 2001;98(25):14234–14237.

[14] Goncharov AF, Struzhkin VV, Hemley RJ, et al. New techniques for optical spectroscopy at ultra-high pressures. In: Nellis WJ, Manghnani MH, Nicol MF, editors. Science and technology of highpressure: proceedings of AIRAPT-17. Hyderabad, India: Universities Press; 2000. p. 90–95.

[15] Asher SA, Johnson CR. Raman spectroscopy of a coal liquid shows that fluorescence interfer-ence is minimized with ultraviolet excitation. Science. 1984;225(4659):311–313.

[16] Le Pevelen DD. NIR FT-Raman. In: Lindon JC, Tranter GE, Koppenaal DW, editors. Encyclopedia ofspectroscopy and spectrometry. 3rd ed. Oxford: Academic Press; 2017. p. 98–109.

[17] Chen NH, Silvera IF. Excitation of ruby fluorescence at multimegabar pressures. Rev Sci Instrum.1996;67(12):4275–4278.

[18] Eggert JH, Moshary F, Evans WJ, et al. Ruby at high pressure. III. A pumping scheme for the Rlines up to 230 GPa. Physical Review B. 1991;44(14):7202–7208.

[19] Loubeyre P, Occelli F, LeToullec R. Optical studies of solid hydrogen to 320 GPa and evidencefor black hydrogen. Nature. 2002;416:613–617.

[20] Zhou W, Chen X-J, Zhang J-B, et al. Vibrational, electronic and structural properties of wurtziteGaAs nanowires under hydrostatic pressure. Sci Rep. 2015;4:6472.

[21] Wang Y, Jiang S-Q, Goncharov AF, et al. Synthesis and Raman spectroscopy of a layeredSiS2phase at high pressures. J Chem Phys. 2018;148(1):014503.

HIGH PRESSURE RESEARCH 13

[22] Erdogan T, Mizrahi V. Thin-film filters come of age. Photonics Spectra. 2003;37(7):94–96.[23] Goncharov AF, Dalton DA, McWilliams RS, et al. Development of ultrafast spectroscopic tech-

niques to study rapid chemical and physical changes in materials under extreme pressureand temperature conditions. MRS Online Proceedings Library Archive. 2012;1405.

[24] Goncharov AF, Lobanov SS, Tan X, et al. Experimental study of thermal conductivity at highpressures: Implications for the deep Earth’s interior. Phys Earth Planet Inter. 2015;247:11–16.

[25] Lobanov SS, Holtgrewe N, Lin J-F, et al. Radiative conductivity and abundance of post-perovs-kite in the lowermost mantle. Earth Planet Sci Lett. 2017;479:43–49.

[26] Jiang S, Holtgrewe N, Lobanov SS, et al. Metallization and molecular dissociation of dense fluidnitrogen. Nat Commun. 2018;9(1):2624.

[27] Onshape. www.onshape.com.[28] Arguello CA, Rousseau DL, Porto SPS. First-Order Raman effect in Wurtzite-type Crystals. Phys

Rev. 1969;181(3):1351–1363.[29] Schaufele RF, Weber MJ. Raman Scattering by Lithium Niobate. Phys Rev. 1966;152(2):705–708.[30] Steele JA, Puech P, Lewis RA. Polarized Raman backscattering selection rules for (hhl)-oriented

diamond- and zincblende-type crystals. J Appl Phys. 2016;120(5):055701.[31] Fateley WG, McDevitt NT, Bentley FF. Infrared and Raman Selection rules for Lattice Vibrations:

The correlation method. Appl Spectrosc. 1971;25(2):155–173.[32] Goncharov AF, Montoya JA, Subramanian N, et al. Laser heating in diamond anvil cells: devel-

opments in pulsed and continuous techniques. J Synchrotron Radiat. 2009;16(6):769–772.[33] Prakapenka VB, Kubo A, Kuznetsov A, et al. Advanced flat top laser heating system for high

pressure research at GSECARS: application to the melting behavior of germanium. High PressRes. 2008;28(3):225–235.

[34] Evans CL, Xie XS. Coherent Anti-Stokes Raman Scattering Microscopy: Chemical Imaging forBiology and Medicine. Annu Rev Anal Chem. 2008;1:883–909.

[35] Hellwig H, Daniels WB, Hemley RJ, et al. Coherent anti-Stokes Raman scattering spectroscopy ofsolid nitrogen to 22 GPa. J Chem Phys. 2001;115(23):10876–10882.

14 N. HOLTGREWE ET AL.