ISSN: 0195-5373 e-ISSN: 2708-521X Vol . 42 No. 1 Jan/Feb. 2021

www.at-spectrosc.com

Cover Feature: Jinhua Li, Rui Pei, Fangfang Teng, Hao Qiu, Roald Tagle, Qiqi Yan, Qiang Wang, Xuelei Chu , and Xing Xu

Micro- XRF Study of the T roodontid Jianianhualong tengi Reveals New Biological and Taphonomical Signals

Journal Overview The Atomic Spectroscopy (Print ISSN 0195−5373; Online ISSN 2708-521X) is a peer-reviewed international journal started in 1962 by Dr. Walter Slavin. It is dedicated to advancing the analytical methodology & applications, instrumentation & fundamentals, and the science of reference materials in the fields of atomic spectroscopy. Publishing frequency: Six issues per year.

Editor in Chief Editorial Board Xian-Hua Li () Beibei Chen (Wuhan University, China) Miguel Ángel Aguirre Pastor (Universidad de Executive Editor Biyang Deng (Guangxi Normal University, Alicante, Spain) Wei Guo (China) China) Mingli Chen (Northeastern University, China) Associate Editors Carsten Engelhard (University of Siegen, Ming Xu (Research Center for Eco-Environmental Germany) Sciences, CAS, China) Michael Dürr (Germany) Chao Li (National Research Center for Muharrem INCE (Munzur University, Turkey) Wei Hang (China) Geoanalysis, China) Mustafa Soylak (Erciyes University, Turkey) Zhaochu Hu (China) Chao Wei (National Institute of Metrology, Qian Liu (Research Center for Editorial Assistants China) Eco-Environmental Sciences, CAS, China) Chaofeng Li (Institute of Geology and Anneliese Lust (USA) Qiu-Li Li (Institute of Geology and Geophysics, Geophysics, CAS, China) Fadi. Abou-Shakra (USA) CAS, China) Chengbin Zheng (Sichuan University, Kenneth R. Neubauer (USA) Rong Qian (Shanghai Institute of Ceramics, CAS, China) Ming Li (China) China) Érico M. M. Flores (Universidade Federal de Robert J. Thomas (USA) Rui Liu (Sichuan University, China) Santa Maria, Brazil) Stefano Pagnotta (University of Pisa, Italy) Editorial Office Fuyi Wang (Institute of Chemistry, CAS, China) Steven Ray (State University of New York at Address: No. 68 Jincheng Street, East Lake Haizhen Wei (Nanjing University, China) Buffalo, USA) High-tech Development Zone, Wuhan, Honglin Yuan (Northwest University, China) Svetlana Avtaeva (Institute of Laser Physics SB 430078, P. R. China. Ignazio Allegretta (Università degli Studi di RAS, Russia) Tell: +8627-67883452 Bari, Italy) Taicheng Duan (Changchun Institute of Applied Fax: +8627-67883456 Jacob T. Shelley (Rensselaer Polytechnic Chemistry, CAS, China) E-mail: [email protected] Institute, NY, USA) Wen Zhang (China University of Geosciences, Homepage: www.at-spectrosc.com Jan Kratzer (Academy of Sciences of the Czech Republic, Czech) China) Paper Submission Jianbo Shi (Research Center for Xiaodong Wen (Dali University, China) Submission by E-mail: Eco-Environmental Sciences, CAS, China) Xiaomin Jiang (Sichuan University, China) [email protected] Jianfeng Gao (Institute of Geochemistry, CAS, Xiaowen Yan (Xiamen University, China) 2020 Subscription/Claims China) Xuefei Mao (Chinese Academy of Agricultural Jing Liang ( Haiguang Instrument Co., Sciences, China) Printed Copy: U.S. $120.00 per year LTD., China) Yan Luo (University of Alberta, Canada) E-mail: [email protected] Kazuaki Wagatsuma (Tohoku University, Yanbei Zhu (National Metrology Institute of Article Indexing Services Japan) Japan, Japan) Ke Huang (Sichuan Norman University, China) Indexed in Web of Science & Scopus Ying Gao (Chengdu University of Technology, Kelly LeBlanc (National Research Council 2019 SCI Impact Factor: 1.292 China) Canada, Canada) Yongliang Yu (Northeastern University, China) Publisher & Copyright Lianbo Guo (Huazhong University of Science Yueheng Yang (Institute of Geology and Atomic Spectroscopy Press Limited and Technology, China) Geophysics, CAS, China) Flat H29, 1/F, Phase 2 Kwai Shing Ind. Bldg., Liang Fu (Yangtze Normal University, China) Yu-Feng Li (Institute of High Energy Physics, Lu Yang (National Research Council Canada, No. 42-46 Tai Lin Pai Road, CAS, China) Kwai Chung, N.T., 999077, Hong Kong, Canada) Zheng Wang (Shanghai Institute of Ceramics, Luyuan Zhang (Institute of Earth Environment, P. R. China CAS, China) ASPL holds copyright to all material CAS, China) Zhenli Zhu (China University of Geosciences, Man He (Wuhan University, China) published in Atomic Spectroscopy unless China) Meng Wang (Institute of High Energy Physics, otherwise noted on the first page of the paper. Zhi Xing (Tsinghua University, China) CAS, China)

In This Issue:

Micro-XRF Study of the Troodontid Dinosaur Jianianhualong tengi Reveals New Biological and Taphonomical Signals # Jinhua Li, Rui Pei, Fangfang Teng, Hao Qiu,e Roald Tagle, Qiqi Yan, Qiang Wang, Xuelei Chu, and Xing Xu...... 1 Excitation Behavior of Copper Ionic Emission Lines During the 3d94p - 3d94s Transition in the Glow Discharge Plasma with Xenon in Comparison to Using Argon and Krypton Kazuaki Wagatsuma...... 12 Spark Discharge-LIBS: Evaluation of One-Point and Multi-Voltage Calibration for P and Al Determination Alan Lima Vieira, Edilene Cristina Ferreira, Dário Santos Júnior, Giorgio Saverio Senesi, and José Anchieta Gomes Neto...... 18 Determination of Elemental Impurities in Iron-nickel-based Superalloys by Glow Discharge Mass Spectrometry F. F. Hu, C. H. Wang, J. D. Li, P. Y. Liu, H. Liu, and L. Zhang...... 25 Leaching of Gallium from Coal Fly Ash, Alumina and Sediment Samples with an Acid Mixture for its Determination by ICP-OES A. C. Sahayam, G. Venkateswarlu, and S. Thangavel...... 32 Influence of Spot Size on LA-ICP-MS Ablation Behavior for Synthetic Calcium Tungstate and Silicate Glass Reference Material NIST SRM 610 Yuting Xiao, Jian Yang, Jun Deng, Wei Wang, Yuqiu Ke, and Yijian Sun...... 36

# Front Cover Article ISSN 0195-5373/e-ISSN 2708-521X; CODEN ASPND7 JAN/FEB. 2021 42(1), 1-42.

www.at-spectrosc.com Micro-XRF Study of the Troodontid Dinosaur Jianianhualong tengi Reveals New Biological and Taphonomical Signals

Jinhua Li,a,b,† Rui Pei,c,d,e,† Fangfang Teng,f Hao Qiu,a Roald Tagle,g Qiqi Yan,g Qiang Wang,c,d Xuelei Chu,a and Xing Xuc,d,* a Key Laboratory of Earth and Planetary Physics, Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing 100029, P. R. China b Laboratory for Marine Geology, Qingdao National Laboratory for Marine Science and Technology, Qingdao 266061, P. R. China c Key Laboratory of Vertebrate Evolution and Human Origins, Institute of Vertebrate Paleontology and Paleoanthropology, Chinese Academy of Sciences, Beijing 100044, P. R. China d CAS Center for Excellence in Life and Paleoenvironment, Beijing 100044, P. R. China e State Key Laboratory of Palaeobiology and Stratigraphy, Nanjing Institute of Geology and Palaeontology, Chinese Academy of Sciences, Nanjing 210008, P. R. China f Xinghai Paleontological Museum of Dalian, Dalian 116023, Liaoning, P. R. China g Bruker Nano GmbH, Am Studio 2D, 12489 Berlin, Germany Received: November 15, 2020; Revised: November 28, 2020; Accepted: November 28, 2020; Available online: December 01, 2020. DOI: 10.46770/AS.2021.01.001

ABSTRACT: Jianianhualong tengi is a key taxon for understanding the evolution of pennaceous as well as of troodontid theropods. It is known by only the holotype, which was recovered from the Lower Cretaceous Yixian Formation of western Liaoning, China. In this study, we carried out a large-area micro-X-Ray fluorescence (micro-XRF) analysis of the holotype of Jianianhualong tengi via a Brucker M6 Jetstream mobile XRF scanner. The elemental distribution measurements of the specimen show an enrichment of typical bone-associated elements, such as S, P and Ca, which allows to visualize the fossil structure. Additionally, the bones are enriched with several heavier elements, such as Sr, Th, Y and Ce relative to the surrounding rocks. The enrichment is most likely associated to secondary mineralization and the phosphates from the bones. Interestingly, the plumage shape correlates with an enrichment in elements, such as Cu, Ni and Ti, consistent with the findings of a previous study1 on Archaeopteryx using synchrotron imaging. Elemental variations among the skeleton, the unguis and the sheath blade further indicate their possible compositional or ultrastructural differences, providing new biological and taphonomic information on the fossilized keratinous structures. An in-situ and nondestructive micro-XRF analysis is currently the most ideal way to map the chemistry of meter-sized fossils and has so far been mainly restricted to small samples. Micro-spatial chemical analysis of larger samples usually required a synchrotron facility. Our study demonstrated that a laboratory-based large-area micro-XRF scanner can provide a practical tool for the study of large specimens, thus allowing to collect full chemical data in order to obtain a better understanding of evolutionary and taphonomic processes.

INTRODUCTION feathers and bones, and bring our understanding of this major transition to a new level.1, 5-12 The chemistry of exquisitely Morphological studies on the skeleton and the plumage have preserved fossil animals, including several iconic flying/gliding forged a solid link between non-avialan coelurosaurians and birds, theropods, have been investigated to reveal information on their building a well-accepted framework for understanding the paleobiology and the fossilization process.1, 6, 9, 10, 12-14 Various dinosaur-bird transition.2-4 Recent works started to look beneath chemical imaging techniques, e.g. Fourier-transform infrared the surface and into the ultrastructure and chemistry of fossil (FTIR) spectroscopy, Raman spectroscopy, X-ray spectroscopy, www.at-spectrosc.com/as/article/pdf/202101001 1 At. Spectrosc. 2021, 42(1), 1-11 and secondary-ion mass spectroscopy (SIMS), have been preserved in tuffaceous sandstones, DLXH 1218 is compressed employed to track the molecular, elemental and isotopic and preserved on a tuffaceous shale slab, even though the bones information.1, 5, 6, 8-13, 15-18 Unfortunately, previous studies all are not entirely flattened (Fig. 1a). The shale slab was broken into focused on non-destructive chemical imaging of small-sized several pieces while being collected and was glued together when specimens (e.g. Archaeopteryx, Confuciusornis), limited by the the fossil was prepared. DLXH 1218 contains a nearly complete analytical instruments6, 8, 11, 12 or the chemical analyses of small skeleton associated with soft tissues, including feathers (Fig. 1a). samples destructively taken from fossils.9, 10, 13, 16, 17 Up to now, The skeleton is black in color. At least two layers of the matrix can systematic non-destructive chemical imaging on meter-sized be observed on the slab: an overlying white layer (WL) and an specimens of non-avialan has not been carried out. underlying brown layer (BL, possibly rich in organic matter) that preserve the fossil. The immediate areas surrounding the skeleton, The Middle-Upper and the Lower Cretaceous of possibly soft remains (SR) of the specimen, have a brighter color western Liaoning and the neighboring areas have yielded than the remaining parts of the matrix consisting of a brown layer abundant avialans and non-avialan dinosaur specimens with (Fig. 1a). The size of the slab is approximately 90 cm long and 70 exquisite preservation of the soft tissues that enlightened our cm wide. DLXH 1218 measures ~100 cm in preserved skeletal understanding of the origin and early evolution of feathers and bird body length and ~117 mm in femoral length. DLXH 1218 is flight.3, 19-22 The unique preservation of these specimens also estimated to be ~112 cm in total length with a fully preserved tail. provides a good opportunity to detect the hidden ultrastructural, chemical and taphonomic information on feathers and bones.8-10, Micro-XRF imaging and data analysis 16, 23 The troodontid theropod Jianianhualong tengi is represented by an exquisitely preserved specimen from the Lower Cretaceous The elemental maps of the specimen DLXH 1218, were obtained Yixian Formation.24 Using regular tools and laser-stimulated using the Bruker M6 Jetstream mobile X-ray fluorescence (XRF) fluorescence imaging, it revealed to have asymmetrical scanner which permits large surface scanning with a lateral pennaceous feathers, suggesting asymmetry, a resolution up to 50 µm. The analysis was performed at the Dalian synapomorphy of the Paraves, a clade including birds and their Xinghai Museum for in-situ studies (Fig. S1a). The Bruker M6 close relatives. Compared to other Mesozoic paravians with Jetstream consists of a mobile measuring head that can scan the asymmetrical pennaceous feathers, Jianianhualong tengi is over surface of a sample on a XY-motorized stage (Fig. S1b). The one meter long and apparently not adapted for flight as indicated maximum range of the XY-motorized stage is 800 mm × 600 mm by its short forelimbs.24 Besides the novelty of feathers, (H × W). Mounted on the measuring head is a Rh-target X-ray tube Jianianhualong tengi is also featured by a mosaic combination of (30 W maximum power, usually operated at 50 kV and a primitive and derived troodontid features.24 maximum current 0.6 mA). The X-ray beam is guided to the sample through polycapillary focusing optics. The beam diameter In this study, we performed several micro-X-Ray fluorescence can be regulated by adjusting the distance between the X-ray analyses (micro-XRF) using the portable Bruker M6 Jetstream X- source and the sample surface, the standard settings are between ray fluorescence (XRF) scanner of the holotype of Jianianhualong 100 µm and 500 µm. The X-ray signal from the sample is collected tengi to reveal the chemistry of the whole specimen as well as with a 30 mm2 XFlash silicon drift detector with an energy some of its relevant features. The mobility of the instrument resolution <145 eV at Mn-Kα. The instrument is equipped with provides ideal conditions (detailed in the Materials and Methods two magnifying optical cameras to document the area of analysis. section) for in-situ micro-XRF measurement and high- The portability of the instrument allows to bring the instrument to performance elemental distribution analysis of large samples (over the object and to measure the samples directly on-site, reducing 0.6 m2) for the first time with a high spatial resolution in the 100 the risk of transport damage of valuable samples. This provides µm range.25-27 The analyses revealed new information for ideal conditions for an in-situ micro-XRF measurement for high understanding the paleobiology of this troodontid dinosaur and the performance element distribution analysis of large samples with taphonomy of this important fossil. high spatial resolution in the 100 µm range.25-27

The specimen was placed horizontally on a table and properly MATERIALS AND METHODS oriented (Fig. S1b), then scanned fully (the whole specimen) and partially (several sub-areas of interest with higher resolution). Full- Troodontid dinosaur Jianianhualong tengi area XRF elemental maps of DLXH 1218 are shown in Fig. 1 and The holotype, the only known Jianianhualong tengi specimen, Fig. S2. The acquisition conditions were 1500 × 1090 pixel for a was discovered from the Lower Cretaceous Yixian Formation (the total of 1.635 million pixel/map, and the acquisition time per pixel middle section of the Jehol group) of Baicai Gou, Yixian County, was 15 ms with a pixel size of 500 µm. The total measurement western Liaoning, China. The holotype (DLXH 1218) is now time was 9 h 30 min (Fig. 1). The feather and bone chemistry were housed at the Dalian Xinghai Museum (DLXH). Unlike other further studied via in-situ micro-XRF imaging on several sub- troodontid specimens from the Yixian Formation that are 3D- areas of interest using higher spatial resolution. The acquisition www.at-spectrosc.com/as/article/pdf/202101001 2 At. Spectrosc. 2021, 42(1), 1-11

Fig. 1 Overall elemental distribution of the holotype of Jianianhualong tengi revealed via micro-XRF imaging. (a) Light photo of Jianianhualong tengi (DLXH 1218). The skeleton of the fossil is represented in black color. The matrix contains an overlaying white layer (WL), a brown layer (BL), and an immediate layer surrounding the skeleton, possibly soft remains (SR) of the specimen. (b)-(i) Micro-XRF detail maps (false-color images) of K-Kα (b), Fe- Kα (c), Mn-Kα (d), Ti-Kα (e), Ca-Kα (f), P-K (g), S-Kα (h), and Sr-Kα (i). More Micro-XRF detail maps (false-color images) of other elements are shown in the supplementary material (Fig. S1). (j) Combined map of four elements. Green, K; aqua, Mn; red, Fe; magenta, Sr. (k) The corresponding XRF spectrum shows the detectable elements of Al, Si, P, S, K, Ca, Ti, Ce, Mn, Fe, Ni, Cu, Zn, Ga, Zn, As, Th, Rb, Sr, Y and Zr. The Rh and Ar peaks originate from the Rh-target microfocus-X-ray tube and ambient air, respectively. conditions for the tail measurements were 393 × 193 pixel for a total of 191.580 k pixel/map, 20 ms/pixel with the pixel size of 400 total of 75.849 k pixel/map, 100 ms/pixel with the pixel size of 400 µm, and 1 h 25 min total measurement time (Fig. 4). µm, and 2 h 17 min total measurement time (Fig. 2). The The XRF spectra were collected, deconvoluted and examined acquisition conditions for the pelvis and the hind limb based on the 2D maps of the XRF data, with each pixel in the map measurements were 848 × 584 pixel for a total of 495.232 k holding an XRF spectrum that can be analyzed. XRF raw data is pixel/map, 10 ms/pixel with the pixel size of 250 µm, and 1 h 57 the collection of counts of element-specific fluorescent X-ray min total measurement time (Fig. 3). The acquisition conditions signals received by the XRF instrument detector. Therefore, for the cranium measurements were 484 × 555 pixel for a total of chemical elements were identified based on the peaks in the XRF 268.620 k pixel/map, 20 ms/pixel with the pixel size of 200 µm, spectra corresponding to the characteristic X-ray emission lines of and 1 h 51 min total measurement time (Fig. 4). The acquisition each element. The sum of the counts of a particular peak or an X- conditions for the pes measurements were 515 × 372 pixel for a www.at-spectrosc.com/as/article/pdf/202101001 3 At. Spectrosc. 2021, 42(1), 1-11

Fig. 2 (a) Light photo of a small region of the Jianianhualong tengi tail. (b)-(f) Micro-XRF detail maps (false-color images) of Fe (b), Ti (c), Cu (d), Mn (e), Ni (f) distribution within bone materials and feathers of the Jianianhualong tengi tail. ray emission line or family of X-ray emission lines is often in imaging proportion to the concentration of the relevant element.28 In order The XRF measurements performed provide an overall distribution to compare the chemical composition of the specimen further, of at least 20 recognizable elements in the specimen (Fig. 1k). various regions of interest (ROI) were selected for Sorting of the elements by their main phase shows that the light semiquantitative analysis (Tables S1-S4). In addition, principal rock where the fossil is embedded is rich in Si, K, Ca, Ti, Mn, and component analysis (PCA) was applied to improve the XRF Fe (note that all elements are most likely present as oxides, but the spectrum for each component.29 oxygen cannot be detected here) (Fig. 1, Fig. S2). The elements It should be noted that the given analytical conditions could not present and their respective intensities suggest a clay rich provide true values of the concentration of each element. Usually, sedimentary rock with a carbonate fraction. a quantitative XRF analysis can be done based on empirical The areas with high Mn are scattered on the white layer models (standard-based) or Fundamental Parameters (FP). (represented by the black spots under normal light). This is most Empirical models require matrix-matched standards, but no likely related to pyrolusite mineralization (MnO2). The darker appropriate standards for this application exist at this time. FP parts in the visual image of DLXH 1218, some of them associated models require all elements including carbon to be analyzed, but to feathers, correlate with an increase in Fe and As. The Fe and As the carbon concentration could not be assessed since carbon levels are very low in the skeleton. They are most concentrated in cannot be measured in an open beam XRF system because of air the brown layer and less concentrated in the overlying white layer absorption. In addition, DLXH 1218 is not entirely flat. As a result, (Fig. 1c, Fig. S2f). The soft tissue remnants including the feathers the distance between the object and the instrument as well as the (represented by the light brown areas close to the skeleton under orientation between the object and the excitation-detector plane visible light) exhibit lower Fe levels than the immediate can be different at any given location. Therefore, semi-quantitative surrounding matrix (the brown layer) (Fig. 1c). Rubidium and Ga data processing was used here to assess the relative elemental are possibly controlled by the matrix. They are high in the white concentrations between ROIs. layer of the matrix, slightly lower in the brown layer of the matrix, and extremely low in the skeleton (Fig. S2d, S2e).

RESULTS AND COMPARISONS Zinc is distributed primarily in the white layer of the matrix and the cement-like material (Fig. S2i). The various rock fragments of Overall elemental distribution revealed via full-area XRF www.at-spectrosc.com/as/article/pdf/202101001 4 At. Spectrosc. 2021, 42(1), 1-11

Fig. 3 (a) Light photo of a small region of the Jianianhualong tengi hip. (b)-(f) Micro-XRF detail maps (false-color images) of Fe (b), Ti (c), Cu (d), Mn (e),

Ni (f) distribution within bone materials and feathers of the Jianianhualong tengi hip. the fossil is glued together with a cement-like material, containing reported in synchrotron studies of fossil feathers.1, 6, 12 The Fe level some traces of Zn. The Zn level is low in the skeleton and the soft seems relatively lower in the feathers than of the immediate tissue including the feather remnants, which contradicts the results surrounding body tissue, while Ti, Cu, Mn and Ni seem relatively observed in the Thermopolis Archaeopteryx where the bone higher in the feathers rather than in the immediate surrounding materials have a higher Zn level than the feathers and the matrix.12 body tissue (Fig. 2). Notably, unlike the striped pattern seen on the As a contrast, the coliiformes bird has a higher Zn level in the tail feathers (Fig. 2d), the pattern of Cu is not clearly associated feathers, while the matrix and the bones contain less Zn.30 with the feathers on the pelvis (Fig. 3d). The Cu and Mn levels apparently vary for feathers on different parts of the body. The tail The bone is, as expected rich, in Ca and P, corresponding feathers have richer Cu than the pelvis feathers, while the Mn level mineralogically to apatite (Fig. 1f, g). The apatite mineral appears is higher in the pelvic feathers than in the tail feathers (Tables S1- strongly enriched several incompatible elements such as: Th, Sr, S3). However, these variations in the feather samples are not found and Y as well as some REE mainly Ce (Fig. S1, Tables S1-S4). for Fe, Ti and Ni. Even though Zn as well as Ca, Fe, Cu and Mn Comparatively, Sr, Y, Th and Ce levels seem heavily associated are known to be typically associated with melanin pigmentation,1, with the skeleton and are relatively low in other regions including 6, 11, 32, 33 we did not find a connection between Zn and the feathers the slab matrix and the soft tissue remains, as mentioned above. A on DLXH 1218. high Ca level associated with Sr and Y was noticed in both vertebrate and invertebrate fossil materials.31 Cu and Ni appear also distributed in the bone materials, while Fe, Ti and Mn are not significantly enriched in the bones (Figs. 2, 3). Elements associated with the feathers Copper and Ni are distributed in the white layer of the matrix and slightly more concentrated in the feathers. Both elements are less The regions where feather remains can be observed show an accumulated in the bone materials and in the brown layer of the enrichment and correlation pattern of several elements including matrix, and high concentrations appear in the glue which was used Mn, Ti, Ni and Cu (Fig. 1, Fig. S1). This is most visible for the for fossil restoration (Fig. S2g, S2h). A high level of Cu and Ni is feather region in the detail scan of the tail see in Fig. 2. Whereas also known to be in the feathers of many fossil and living taxa, an integration of transitional elements such as Mn, Cu and Ni into such as , Archaeopteryx, Gansus and other fossils and the organic component appears more obvious, is the enrichment of living birds.1, 6, 8, 12, 30, 32 In all of these specimens and DLXH 1218, Ti in Fig. 2c is more difficult to explain, especially since the the Cu level is generally higher in the feathers than in the bones, correlation of Ti with the feather structure is much clearer than for and the Cu concentration usually varies for the different feather the other elements as can be seen in the scan from the tail (Fig. 2). samples of the individual specimen. 6, 12, 30, 32 As shown in Fig. 2a, 2c-2f and Fig. 3a, 3c-3f, the distributions of Ti, Cu, Mn, and Ni match well with the stripe Sr- and Ca-XRF imaging the morphology and preservation of pattern of the feathers. The enrichment of many of these elements the skeleton associated to the feather-structure has also been previously www.at-spectrosc.com/as/article/pdf/202101001 5 At. Spectrosc. 2021, 42(1), 1-11

Fig. 4 Micro-XRF detail maps (false-color images) of Ca and Sr distribution within the cranium (a)-(c), the pes (d)-(f), and the pelvis and the tibiotarsus (g)- (i) of the Jianianhualong tengi. The first column indicates the corresponding light photos, the second for Ca distribution maps, the third for Sr distribution maps. Arrows indicate regions with calcium removal.

The elements Sr and Ca were found in the bone materials, but Ca level is present ventral to the distal dorsal process of the right is also observed in the soft tissue remains and on the slab. In ischium possibly due to a breakage in that bone. On the other hand, contrast, Sr seems to be heavily and uniformly present in the the Sr map shows the complete shape of the distal dorsal process skeleton, but at very low levels in other regions (Figs. 1, 4). The of the ischium (Fig. 4), resulting from a relatively stable Sr level Sr/Ca ratio is generally consistent for most parts of the skeleton for that area. (0.26~0.41, Tables S1-S4). Fig. 4 shows a comparison of the XRF Ambiguously shaped dark-colored matter is preserved near the maps of Ca and Sr in the cranium, the pes and the hind limb. hind limbs of DLXH 1218, close to the proximal portion of the Though no novel morphologies were revealed via chemical tibiotarsus and the distal region of the pes (Fig. 4d-4f, Tables S3, imaging, it is clear that, compared to the Ca-maps, the Sr-maps S4). The high Sr levels and the high Sr/Ca ratio (~0.59) match the significantly improve the skeletal image by having a better shape of the dark-colored regions, reaffirming that these are contrast and making the sutures and outlines of the bones more remains of smashed bone materials rather than imprints of feathers visible. This is possibly due to the fact that Sr is only rich in the or other non-skeletal structures, especially considering the skeleton, or has a better lateral resolution than Ca, or both. proximal portion of the tibiotarsus is severely fragmented and the Furthermore, comparisons between the Ca and Sr elemental maps pedal digits are damaged, which makes the bone materials easily may also provide insight into the taphonomy and the process of scattered around the damaged bone. fossilization. It should be pointed out that a careful examination of the For instance, the pedal digits, the right metatarsus and the right specimen does not reveal new morphology that has not already calcaneum (Sr/Ca ratio ~0.83) show distinct Ca loss compared to been observed under normal light. No trace of cartilage remains other parts of the skeleton, while the Sr level remains stable. Since were associated with the sternum with regard to the Ca, Sr or other there is damage to the bone, it is possible that this makes Ca easier elemental maps (Fig. S4), which supports the hypothesis that to be carried away during fossilization (Fig. 4). Similarly, a low Ca sternal plates are absent in troodontids.34 www.at-spectrosc.com/as/article/pdf/202101001 6 At. Spectrosc. 2021, 42(1), 1-11

Fig. 5 (a) Light photo of a small region of the Jianianhualong tengi claw. The bony claw (C), the unguis (SU) and the sheath blade (SB) can be observed based on light photo imaging (a). (b)-(j) Micro-XRF detail maps (false-color images) of K (b), Fe (c), Mn (d), Ca (e), P (f), Sr (g), Th (h), and Y (i) distribution within the bone and sheath materials of the Jianianhualong tengi claw. (j) Combined map of three elements. Aqua, Mn; red, Fe; magenta, Sr. (k) XRF spectra originating from the four small regions of interest (ROI) circled in different colors in the panel (j). ROI 1 and 2 are white and brown regions of the matrix, respectively, ROI 4 indicates the region of the bony claw, and ROI 3 is the region containing the sheath blade.

Detailed examination of the claw sheaths of DLXH 1218 (Figs. Structure and chemistry of the claw sheath 5, 6) revealed a chemical pattern that was not observed in previous Interestingly, the claw sheaths showed elemental signatures that studies.12, 35, 36 No differences were noticed between the bony and were previously not noted by using chemical imaging (Figs. 5, 6). keratinous structures of the claw in the Thermopolis In DLXH 1218, the sheaths surrounding the distal portion of the Archaeopteryx,12 but the thickened dorsal rim of the sheath manual and pedal distal phalanges are preserved. The sheaths have (unguis) of DLXH 1218 exhibits slightly lower levels of Sr, Th elevated Ca, P, Sr, Th, and Y levels relative to the matrix, but very and Y than the bony digit (Figs. 5g-i, Table S3), while the Ca level low K, Fe and Mn levels, which matches the pattern of other bones is similar to that in the bony parts (Figs. 5e, Table S3). However, (Fig. 5, Table S3). A high P profile of the sheaths was also noticed the flattened sheath blade exhibits even lower Sr, Th, Y and Ca in the Thermopolis Archaeopteryx, Shuvuuia and Citipati.12, 35, 36 than the unguis. The Sr/Ca ratio is also much lower in the sheath The Ca level is relatively low in the sheath as a result of richer Ca blade than in the unguis, but this ratio resembles that of the feathers, in the matrix of the Thermopolis Archaeopteryx,12 yet a high Ca even though the Ca concentration is still higher in the sheath blade level is present in the sheaths of Citipati and Shuvuuia12, 35, 36 as in than in the feathers (Fig. 6, Table S3). Unlike the bony ungual Jianianhualong tengi, indicating preservation as calcium digits, claw sheaths are modified scales in reptiles and birds, phosphate. The S level is higher in the sheaths than in the matrix composed of beta keratins over alpha keratins,37, 38 yet the of DLXH 1218 as in the Thermopolis Archaeopteryx and elemental signature of the unguis resembles that of the bony claw Shuvuuia.12, 35 www.at-spectrosc.com/as/article/pdf/202101001 7 At. Spectrosc. 2021, 42(1), 1-11

Fig. 6 (a) Regions of interest associated with the claw of Jianianhualong tengi, and (b) Sr/Ca ratios of each region. SB, sheath blade; SU, unguis; SL, skeleton and bony claw. rather than the sheath blade and the keratinous feathers. Previous showed interesting elemental abnormalities (Fig. 1). The Ca, P and studies showed that extant archosaurian unguis developed from S levels in this crack are similar to that of the skeleton, while the denser corneous materials than the sub-unguis.37, 38 Calcification Sr, Th, Y and Ce levels are very low in this crack in comparison to of keratin has also been observed occasionally in dense and the bones. Unlike other cracks that were filled with glue during hardened keratinous structures in extant animals39, 45 and therefore, fossil preparation, the As, Cu, Ni, Ga and Zn levels are also low in a similar calcification process of the keratins may also be present this crack (Fig. 1, Fig. S2). Therefore, we infer that this crack was in this Mesozoic theropod in vivo, which caused the enrichment of likely formed during fossilization, and the water flux carried the Ca and a higher Sr/Ca ratio in the unguis than in the sheath blade. lighter elements of Ca, P and S from the animal remains into this crack. On the other hand, the heavier elements, such as Sr, Th, Y and Ce, more easily accumulated in the bone materials and were DISCUSSION not carried away by water flux. This may further indicate that at least part of these heavier elements in the skeleton are original, as Information revealed by micro-XRF chemical imaging can be supported by the fact that Sr participates significantly in Applying large-area micro-XRF on DLXH 1218 helped to reveal bioapatite mineralization throughout the life of an animal.40 the surface elemental distribution in the biological remains of the specimen and its surrounding geological matrix. Unlike chemical Comparisons with chemistry of other fossil materials mappings in previous studies that focus on the lighter elements,12, A comparison of the chemical mapping of DLXH 1218 and other 30 micro-XRF imaging is also able to detect the distribution of vertebrate fossils not only shows consistent patterns of chemical heavier elements. Generally, the distribution of Ti, Cu, Mn and Ni accumulation in the bones and feathers, but also indicates that is associated with the soft remains, which match well with the elements behave differently depending on the chemical pattern of the feathers. The heavier elements of Sr, Th, Y and Ce composition of the surrounding rocks, as previously reported.8, 31, are strongly associated with the skeleton. Though no novel 41 Rossi and co-workers14 concluded that the elemental signature morphology was uncovered by chemical mapping, compositional of vertebrate fossils is tissue-specific and controlled by many variations were found in some structures to reflect either biological biological and environmental factors, which is in agreement with or taphonomic features. For instance, the Sr variation among the our findings for DLXH 1218, namely, that the similarities and unguis, the sheath blade and the bony claw likely indicate the differences in the elemental distribution in the feathers and bones compositional differences among these structures. The Ca can be compared with finding for other vertebrate fossils. Since variation in some parts of the skeleton (e.g. between pedal digits the composition of bones is representative of a particular and the metatarsus) possibly showed Ca loss during fossilization. depositional environment and most likely also the local condition Combining chemical information of the lighter and heavier where fossilization occurred, this variation of local compositional elements can provide insight into the material exchange of the fingerprint could be used to assign poorly or not correct fossil remains and the surrounding rocks during fossilization. documented fossil to certain location. For this purpose, additional Notably, a crack-like structure posterior to the proximal caudal systematic work is needed. www.at-spectrosc.com/as/article/pdf/202101001 8 At. Spectrosc. 2021, 42(1), 1-11 Specifically, S, P and Ca are the most commonly examined forming bone bioapatite, and relatively high Sr levels have also elements in chemical studies based on fossil materials.31,42 been found in many fossil bone materials.30, 40, 41, 43 Different Consistently, micro-XRF with DLXH 1218 showed that S, P and chemical behaviors of Sr and Y in calcium phosphates and Ca are associated with the organic matter (both the skeleton and carbonates have been suggested by Gueriau and co-workers31,43 soft tissues), while the Ca levels are slightly higher in the skeleton, based on fish and non-vertebrate fossils, yet the distribution of Sr a pattern similar to previously examined fossil specimens.1, 12, 32 and Y are congruent in DLXH 1218, as the bones are calcium Comparatively, the Ca level in the matrix of DLXH 1218 is phosphates only. relatively low, different from the Thermopolis Archaeopteryx, where the matrix rock is calcium-rich.12 The claw sheaths Sr/Ca variation in different body structures and the preserved in DLXH 1218 have elevated Ca and P levels, congruent fossilization of the claw sheath with the patterns found in the Thermopolis Archaeopteryx, In DLXH 1218, the bony skeleton, claw sheaths and feathers show Shuvuuia and Citipati specimens.12, 35, 36 various Sr/Ca ratios. The Sr/Ca ratios in most parts of the skeleton Some elements, such as Ca, Zn, Mn, Cu and Ni, have been are between 0.25 and 0.40. In contrast, damaged bones usually associated with melanin pigmentation in modern and fossil experience significant calcium removal; while the Sr level remains feathers.1,6, 11, 18, 30, 32, 33 In the fossils of Jianianhualong, Anchiornis, relatively consistent, causing the Sr/Ca ratio generally to be above Archaeopteryx, Gansus, and other extinct and living birds, the Cu 0.40 in these damaged parts. This may suggest that the bone level usually varies with different feather samples of an individual surface helped to preserve the Ca in the bone’s postmortem or and is generally higher in the feathers than in the bones.6, 12, 30, 32 during fossilization, yet the detailed process remains to be Though in DLXH 1218, the distributions of Ca, Cu, Mn and Ni discovered. The Sr/Ca ratio in the unguis of DLXH 1218 is around match the plumage shapes, no obvious enrichment of Zn was 0.20, slightly lower than in the bony parts; while the sheath blade observed with either the feathers or the bones, which is also has a Sr/Ca ratio of around 0.10, similar to that of the feathers and different from the Thermopolis Archaeopteryx, Confuciusornis generally lower than in other parts of the specimen, including the and the fossil coliiformes bird (AMNH FARB 30806) where the matrix. feathers and the bone materials have elevated Zn levels.6, 12, 30 The unguis of DLXH 1218 further resembles the bony skeleton Instead, the highest level of Zn is present in the glue added in by having relatively high levels of Th, Ce and Y, while the sheath DLXH 1218. In DLXH 1218, aside from the matching pattern of blade contains lower levels of Th and Ce, and undetectable Y. Mn and the pelvic feathers, high Mn level areas are also scattered Nevertheless, the presence of P, Ca, Sr, Th and Ce in the claw on the sediments, similar to the Thermopolis Archaeopteryx and sheath and the skeleton suggests that the keratinous unguis and the the Green River feathers (BHI-6358, BHI-6319), where the Mn bony skeleton might share a similar process of fossilization. Saitta level is generally higher in the matrix than in the feathers,12,32 but and Vinther 44 proposed two major modes for the keratin is unlike that in Confuciusornis, Gansus and the fossil coliiformes preservation in fossils: (a) predominantly organic in anoxic, fine- specimens, where Mn is rich in the feathers but absent in parts of grained sediments via melanosomes or amorphous melanin the matrix.6, 30, 32 A high Ti level is also found in the feathers related components and (b) predominantly as calcium phosphate preserved in DLXH 1218, a pattern that has not been reported in in coarse-grained, relatively oxic sediments. Chemical mapping of other fossil materials, although a variation in Ti level has been DLXH 1218 indicates that a combination of both modes could detected between two Gansus feathers and in contrast, Ti is absent happen. The feathers of DLXH 1218 fossilized in mode (a) as in some modern bird feathers.32 typical for most Jehol feather material. The claw sheaths of DLXH Though a high Fe level is also known in some melanized 1218 fossilized as calcium phosphate as in mode (b), but in anoxic, structures, the fossil feathers preserved in DLXH 1218 show a fine-grained sediments like other Jehol fossils. relatively lower level of Fe than in the surrounding areas, similar Even though previous studies36, 44 cannot determine whether the to that in the Thermopolis Archaeopteryx.12 DLXH 1218 calcium phosphate in the fossilized keratinous structure was resembles the coliiformes bird in having a very low Fe level in endogenous or secondarily precipitated, our results support that bones, while the bone of the Thermopolis Archaeopteryx is more the fossilized claw sheaths are at least partially associated to the iron-rich than its feathers. Barden and co-workers32 noted that the secondary mineralization of phosphate, with a similar process to feathers of the modern marabou stork and white-naped crane did the fossilization of the bones. The variation of Ca and P among not contain noticeable Fe with EDS (energy dispersive X-ray each part of the claw also supports that endogenous phosphates are spectrometry) analyses, and the Fe level was low and varied in two likely present in the fossilized sheath claws. Chemical mapping fossilized Gansus feathers. We suspect that the plumage iron level revealed that more calcium phosphate was preserved in the is somewhat influenced by the matrix. fossilized unguis than in the sheath blade in DLXH 1218 (Table Strontium is strongly present in the bones preserved in DLXH S3). Developmental studies showed that in extant archosaurians, 1218, which is not surprising as Sr commonly participates in the unguis is developed from denser keratinous materials than www.at-spectrosc.com/as/article/pdf/202101001 9 At. Spectrosc. 2021, 42(1), 1-11 those in the sub-unguis37, 38 and therefore, the enrichment of QQ analyzed the data. All authors contributed ideas and discussed calcium phosphate in the fossilized unguis compared with the rest the results. Li JH, Pei R and Xu X wrote the manuscript. of the sheath may be associated with the in vivo calcification of the Notes hardened unguis of DLXH 1218, much like the calcification of † dense keratinous structures documented in other extant animals.39, J.-H. Li and R. Pei contributed equally to this work. 45 The authors declare no competing financial interest.

CONCLIUSIONS ACKNOWLEDGMENTS This study performed large-area Micro-XRF scanning of the This study was supported financially by the National Natural holotype Jianianhualong tengi and, for the first time, obtained the Science Foundation of China (grants 41890843, 41688103, chemical mapping of the whole specimen with a resolution at the 41920104009, 41621004 and 41972025), the Laboratory for micron scale. A detailed analysis of the elemental distribution and Marine Geology, Qingdao National Laboratory for Marine compositional variations of the skeleton, feathers and the Science and Technology (grant MGQNLM201704), and the State surrounding matrix demonstrated that the main elements of the Key Laboratory of Palaeobiology and Stratigraphy (Nanjing organisms such as S, P and Ca, as expected, were found to be Institute of Geology and Palaeontology, CAS) (No.193122). We associated with both the skeleton and the soft tissue remains, while thank Mrs. Wang Lixia for inspiring us with this project. We thank Ca was relatively higher in the former than in the latter. Some Mr. Meng Liang (DLTV-4), Mr. Wu Baojun (UCAS) and Mrs. elements, such as Cu, Mn, Ni and Ti, were detected to match the Zhang Jiang (CCTV-10) for their assistance in the organization of plumage shapes, while other elements, such as Sr, Th, Y, and Ce, the research project. We also thank Mrs. Gao Xia and Zhou Jian were strongly associated with the skeleton. Elemental variations at the DLXH museum for their assistance in the experiment. We within the keratinous claw sheath indicate the possible are grateful for Mrs. Su Hui and Sun Liqing at the Boyue compositional or ultrastructural difference between the unguis and Instruments (Shanghai) Co., Ltd. for their efficient job in the the sheath blade, and the calcium phosphate in the fossilized claw instrument coordination. J.H.L. benefited from discussions with sheath may be both authigenic and allogenic. A careful colleagues in Office 442 at IGGCAS. examination of the chemical signature of the bones, the claw sheaths and feathers of the Jianianhualong tengi (DLXH 1218) provides new information on the taphonomy of this important fossil and also of the paleobiology of this key species for REFERENCES understanding the evolution of paravians as well as troodontids. It 1. P. L. Manning, N. P. Edwards, R. A. Wogelius, U. Bergmann, H. E. also gave important clues for further study on a few small regions Barden, P. L. Larson, D. S. Wings, V. M. Egerton, D. Sokaras, R. A. Mori and W. I. Sellers, J. Anal. At. Spectrom., 2013, 28, of interest that can be sampled without damaging the specimen. 1024-1030. https://doi.org/10.1039/C3JA50077B 2. K. Padina and L. M. Chiappe, Biol. Rev., 1998, 73, 1-42. ASSOCIATED CONTENT https://doi.org/10.1111/j.1469-185X.1997.tb00024.x 3. X. Xu, Z. Zhou, R. Dudley, S. Mackem, C.-M. Chuong, Please contact the corresponding author for the Supporting G. M. Erickson and D. J. Varricchio, Science, 2014, 346, 1253293. Information (Fig. S1-S6 and Table S1-S4). Supplementary figures https://doi.org/10.1126/science.1253293 1-6, additional information on the XRF analysis; Supplementary 4. S. L. Brusatte, R. J. Butler, P. M. Barrett, M. T. Carrano, D. C. Evans, G. T. Lloyd, P. D. Mannion, M. A. Norell, D. J. Peppe, tables 1-4 and associated figures, net intensity (cps) of each P. Upchurch and T. E. Williamson, Biol. Rev., 2015, 90, 628-642. element of regions of interest. https://doi.org/10.1111/brv.12128 5. U. Bergmann, P. L. Manning and R. A. Wogelius, Annu. Rev. Anal. Chem., 2012, 5, 361-389. AUTHOR INFORMATION https://doi.org/10.1146/annurev-anchem-062011-143019 6. R. Wogelius, P. Manning, H. Barden, N. Edwards, S. Webb, Corresponding Author W. Sellers, K. Taylor, P. Larson, P. Dodson and H. You, Science, 2011, 333, 1622-1626. https://doi.org/10.1126/science.1205748 . *X. Xu 7. Q. Li, K.-Q. Gao, J. Vinther, M. D. Shawkey, J. A. Clarke, L. D’Alba, Q. Meng, D. E. G. Briggs and R. O. Prum, Science, Email address: [email protected] 2010, 327, 1369-1372. https://doi.org/10.1126/science.1186290 Author Contributions 8. Y. Pan, L. Hu and T. Zhao, Nat. Sci. Rev., 2018, 107, 1040-1053. https://doi.org/10.1093/nsr/nwy107 Li JH, Xu X and Teng FF designed the research, Li JH, Tagle R 9. Y. Pan, W. Zheng, A. E. Moyer, J. K. O’Connor, M. Wang, and Yan QQ performed the experiments, Li JH, Pei R and Yan X. Zheng, X. Wang, E. R. Schroeter, Z. Zhou and M. H. Schweitzer, Proc. Natl. Acad. Sci. USA, 2016, 113, E7900-E7907. www.at-spectrosc.com/as/article/pdf/202101001 1 0 At. Spectrosc. 2021, 42(1), 1-11 https://doi.org/ 10.1073/pnas.1617168113 27. S. Saverwyns, C. Currie and E. Lamas-Delgado, Microchem. J., 10. Y. Pan, W. Zheng, R. H. Sawyer, M. W. Pennington, X. Zheng, 2018, 137, 139-147. https://doi.org/10.1016/j.microc.2017.10.008 X. Wang, M. Wang, L. Hu, J. O’Connor and T. Zhao, 28. D. S. Bright, and D. E. Newbury, J. Microsc., 2004, 216, 186-193. Proc. Natl. Acad. Sci. USA, 2019, 116, 3018-3023. https://doi.org/10.1111/j.0022-2720.2004.01412.x http://doi.org/10.1073/pnas.1815703116. 29. S. Aida, T. Matsuno, T. Hasegawa and K. Tsuji, 11. P. L. Manning, N. P. Edwards, U. Bergmann, J. Anné, W. I. Sellers, Nucl. Instrum. Meth. B, 2017, 402, 267-273. A. van Veelen, D. Sokaras, V. M. Egerton, R. Alonso-Mori and https://doi.org/10.1016/j.nimb.2017.03.123 K. Ignatyev, Nat. Commun., 2019, 10, 1-13. 30. V. M. Egerton, R. A. Wogelius, M. A. Norell, N. P. Edwards, https://doi.org/ 10.1038/s41467-019-10087-2 W. I. Sellers, U. Bergmann, D. Sokaras, R. Alonso-Mori, K. Ignatyev, 12. U. Bergmann, R. W. Morton, P. L. Manning, W. I. Sellers, S. Farrar, A. van Veelen, J. Anné, B. van Dongen, F. Knoll and P. L. Manning, K. G. Huntley, R. A. Wogelius and P. Larson, J. Anal. At. Spectrom., 2015, 30, 627-634. Proc. Natl. Acad. Sci. USA, 2010, 107, 9060-9065. https://doi.org/10.1039/C4JA00395K https://doi.org/10.1073/pnas.1001569107 31. P. Gueriau, C. Jauvion and C. Mocuta, Palaeontology, 2018, 61, 13. J. Anné, N. P. Edwards, R. A. Wogelius, R. T.-D. Allison, I. S. William, 981-990. https://doi.org/10.1111/pala.12377 A. van Veelen, U. Bergmann, D. Sokaras, R. Alonso-Mori, K. Ignatyev, 32. H. E. Barden, R. A. Wogelius, D. Li, P. L. Manning, N. P. Edwards M. E. Victoria and L. M. Phillip, J. Roy. Soc. Interface, 2014, 11, and B. E. van Dongen, PLoS One, 2011, 6, 25494. 20140277. https://doi.org/10.1098/rsif.2014.0277 https://doi.org/ 10.1371/journal.pone.0025494 14. V. Rossi, M. E. McNamara, S. M. Webb, S. Ito and K. Wakamatsu, 33. M. Niecke, M. Heid and A. Krüger, J. Ornithol., 1999, 140, 355- Proc. Natl. Acad. Sci. USA, 2019, 116, 17880-17889. 362. https://doi.org/10.1007/bf01651032 https://doi.org/10.1073/pnas.1820285116 34. X. Zheng, J. O’Connor, X. Wang, M. Wang, X. Zhang and Z. Zhou, 15. P. L. Manning, P. M. Morris, A. McMahon, E. Jones, A. Gize, Proc. Natl. Acad. Sci. USA, 2014, 111, 13900-13905. J. H. S. Macquaker, G. Wolff, A. Thompson, J. Marshall, K. G. Taylor, https://doi.org/10.1073/pnas.1411070111 T. Lyson, S. Gaskell, O. Reamtong, W. I. Sellers, B. E. v. Dongen, M. Buckley and R. A. Wogelius, P. Roy. Soc. Lond B. Bio., 2009, 276, 35. E. T. Saitta, I. Fletcher, P. Martin, M. Pittman, T. G. Kaye, L. D. True, M. A. Norell, G. D. Abbott, R. E. Summons, K. Penkman and 3429-3437. https://doi.org/10.1098/rspb.2009.0812 J. Vinther, Org. Geochem., 2018, 125, 142-151. 16. B. Jiang, T. Zhao, S. Regnault, N. P. Edwards, S. C. Kohn, Z. Li, https://doi.org/10.1016/j.orggeochem.2018.09.008 R. A. Wogelius, M. J. Benton and J. R. Hutchinson, Nat. Commun., 36. E. Moyer Alison, W. Zheng and H. Schweitzer Mary, P. Roy. Soc. 2017, 8, 14779. https://doi.org/10.1038/ncomms14779 Lond B. Bio., 2016, 283, 20161997. 17. Y. C. Lee, C. C. Chiang, P. Y. Huang, C. Y. Chung, T. D. Huang, https://doi.org/10.1098/rspb.2016.1997 C. Wang, C. Chen, R. Chang, C. Liao and R. R. Reisz, Nat. Commun., 37. L. Alibardi, Ann. Anat., 2020, 231, 151513. 2017, 8, 1-8. https://doi.org/ 10.1038/ncomms14220 https://doi.org/10.1016/j.aanat.2020.151513 18. N. P. Edwards, A. Van Veelen, J. Anné, P. L. Manning, U. Bergmann,

W. I. Sellers, V. M. Egerton, D. Sokaras, R. Alonso-Mori, 38. L. Alibardi, Anat. Sci. Int., 2020, 84, 189-199. https://doi.org/10.1007/s12565-009-0015-4 K. Wakamatsu, S. Ito and R. A. Wogelius, Sci. Rep-UK, 2016, 6, 34002. https://doi.org/10.1038/srep34002 39. A. Book, and R. H. Champion, Progress in the biological sciences 19. Q. Ji, P. J. Currie, M. A. Norell and S. A. Ji, Nature, 1998, 393, in relation to dermatology-2. Cambridge, Cambridge University Press, 1964. 753-761. https://doi.org/10.1038/31635 20. X. Xu, X. L. Wang and X. C. Wu, Nature, 1999, 401, 262-266. 40. W. E. Cabrera, I. Schrooten, M. E. De Broe and P. C. D'Haese, https://doi.org/10.1038/45769 J. Bone Miner. Rev., 1999, 14, 661-668. https://doi.org/10.1359/jbmr.1999.14.5.661 21. X. Xu, X. T. Zheng, C. Sullivan, X. L. Wang, L. D. Xing, Y. Wang, X. M. Zhang, J. K. O'Connor, F. C. Zhang and Y. H. Pan, Nature, 41. M. B. Goodwin, P. G. Grant, G. Bench and P. A. Holroyd, 2015, 521, 70-73. https://doi.org/10.1038/nature14423 Palaeogeogr. Paleoclimat. Paleoecol., 2007, 253, 458-476. https://doi.org/10.1016/j.palaeo.2007.06.017 22. M. Wang, J. K. O’Connor, X. Xu and Z. Zhou, Nature, 2019, 569, 256-259. https://doi.org/10.1038/s41586-019-1137-z 42. M. H. Schweitzer, Annu. Rev. Earth Pl. Sc., 2011, 39, 187-216. https://doi.org/10.1146/annurev-earth-040610-133502 23. Q. Zhao, M. J. Benton, S. Hayashi and X. Xu, Acta Palaeontol. Pol., 43. P. Gueriau, C. Mocuta, D. B. Dutheil, S. X. Cohen, D. Thiaudiere, 2019, 64, 323-334. https://doi.org/10.4202/app.00559.2018 S. Charbonnier, G. Clement, L. Bertrand and O. Consortium, 24. X. Xu, P. Currie, M. Pittman, L. Xing, Q. Meng, J. Lü, D. Hu and PLoS One, 2014, 9, 86946. C. Yu, Nat. Commun., 2017, 8, 14972. https://doi.org/ 10.1371/journal.pone.0086946 https://doi.org/10.1038/ncomms14972 44. E. T. Saitta, and V. Jakob, Palaeontol. Electron., 2019, 22.3.2E, 25. M. Alfeld, J. V. Pedroso, M. van Eikema Hommes, G. Van der Snickt, 1-30. https://doi.org/10.26879/1017E G. Tauber, J. Blaas, M. Haschke, K. Erler, J. Dik and K. Janssens, 45. P. R. Blakey, C. Earland, and J. P. G. Stell, Nature, 1963, 198, 81. J. Anal. At. Spectrom., 2013, 28, 760-767. https://doi.org/ 10.1039/C3JA30341A https://doi.org/10.1038/198481a0 26. J. R. Duivenvoorden, A. Käyhkö, E. Kwakkel and J. Dik, Heritage Sci., 2017, 5, 6. https://doi.org/10.1186/s40494-017-0117-6

www.at-spectrosc.com/as/article/pdf/202101001 1 1 At. Spectrosc. 2021, 42(1), 1-11 Excitation Behavior of Copper Ionic Emission Lines During the 3d94p - 3d94s Transition in the Glow Discharge Plasma with Xenon in Comparison to Using Argon and Krypton

Kazuaki Wagatsuma*

Institute for Materials Research, Tohoku University, Katahira 2-1-1, Sendai 980-8577, Japan

Received: October 25, 2020; Revised: November 16, 2020; Accepted: November 16, 2020; Available online: November 28, 2020.

DOI: 10.46770/AS.2020.204

ABSTRACT: The emission spectrum of copper in the wavelength range of 200-230 nm was investigated when xenon, instead of argon or krypton, was employed as the plasma gas in glow discharge plasma (GDP). The Cu II emission lines, which were assigned to the 3d94p - 3d94s transition of the copper ion, were observed, but their emission intensities were different depending on the plasma gas employed. The Cu II 224.700-nm line was studied intensively in an argon GDP. As was previously reported, the excitation mechanism of this line is an asymmetric charge-transfer collision with the argon ion. Here, a coincidence in the total excitation energy enables the corresponding energy level of the 3d94p 3 P2 to be selectively populated through the charge-transfer collision. On the other hand, in a xenon GDP, the emission intensities of the Cu II 224.700-nm line as well as the other Cu II lines were not so strong, but a statistical relationship was found in the plots of the reduced emission intensity of these Cu II lines versus the excitation energy. This result implies that these Cu II lines are not emitted by a non-thermal process, such as charge-transfer collision, but by a thermal excitation process, such as electron collision in the xenon GDP. The first ionization potential of xenon is more than 3 eV lower than that of argon; therefore, it is almost impossible to excite the 3d94p levels of the copper ion by the charge-transfer collision with xenon ion. A similar result to the xenon GDP was obtained with the krypton GDP because the ionization potential of krypton was still insufficient to obtain these high-lying excited levels.

INTRODUCTION analyte atom. The nature of the plasma gas plays a very important role in determining the excitation/ionization processes4,5 and thus Glow discharge plasma have been utilized for one of the excitation the analytical performance in GD-OES.6,7 Unlike other plasma sources in optical emission spectrometry, known as glow sources, the GDP is easily maintained without any instrumental discharge optical emission spectrometry (GD-OES). GD-OES is modification when several gases, such as argon, neon, krypton, employed for the direct analysis of solid samples because this helium, and their mixtures, are used as the plasma gas, while argon excitation source has several benefits for the analytical application, is mostly used in conventional GD-OES. Many studies have been such as rapid sampling through cathode sputtering, minimal published regarding the features of the GDP using different plasma sample pretreatment, and a wide concentration range in the gases and their mixtures, indicating that the plasma gas quantification.1-3 dominantly determines the spectrum pattern of the analyte The emission characteristics of the glow discharge plasma elements excited by the GDP.8-13 Only few studies on the xenon (GDP) vary greatly and depend on the type of plasma gas plasma have been published not only in scientific papers but employed which is little affected by the substrate material of the analytical notes in GD-OES. The probable reason is that xenon gas www.at-spectrosc.com/as/article/pdf/2020204 12 At. Spectrosc. 2021, 42(1), 12-17 is very expensive and its benefits have not yet been established, difference before and after the collision.16 The charge transfer especially for analytical applications. However, in comparison to collision most likely takes place when the surplus energy, ΔE, is other plasma gases, use of xenon GDP may give new interesting very small; in other words, an energy resonance is needed.16 In information on the excitation mechanism of emission lines. Thus, such a case, a particular excited level can be ionized/excited this paper focuses on the excitation of the 3d94p excited levels of selectively because the gas species have to provide the internal the copper ion in a xenon GDP. energy to be corresponding to the sum of the excitation energy and the ionization potential of the colliding partner. Therefore, an Previous studies have shown that selective excitation to a appropriate combination between the emission line and the plasma particular excited level occurs in the excitation of an emission gas can improve the detection sensitivity in GD-OES when the spectrum in GDP.14 This effect is caused by an energy transfer resonance condition is fulfilled for the corresponding excited from the gas species, called a collision of the second kind, whereas energy level. The charge-transfer process was also found in a collision of the first kind means a transfer of kinetic energy like excitations of other metallic elements in GDPs.14 My previous a collision with a fast electron.15 An asymmetric charge-transfer studies indicated that different emission lines in iron, cobalt and process16 is a typical second-kind collision and a major excitation nickel were observed between argon and krypton plasmas, which channel for causing the emission spectra excited by GDPs. The can be caused by charge transfer collisions to particular 3d84p 3d94p - 3d94s transition of singly ionized copper contributes to an excited levels of these singly ionized atoms.19,20 abnormal spectral pattern through the charge-transfer process in Ar GDP. Steers et al. first reported that the Cu II 224.70-nm line In this paper, the Cu II emission lines assigned to the 3d94p - has a very high emission intensity in the Ar GDP.17 This effect can 3d94s transition were measured in detail when xenon was be caused by selective excitation to a particular excited level of the employed as the plasma gas in GD-OES. The spectrum pattern is 9 3 copper ion, 3d 4p P2 (8.23 eV), through an asymmetric charge- analyzed using the transition probability for each Cu II line in transfer collision between an Ar ion and a Cu atom.17 In this comparison to the argon and krypton plasmas. It will also be 5 2 collision, the ground-state levels of the argon ion, the 3p P3/2 investigated to what extent the charge-transfer process works for 5 2 18 (15.76 eV) and the 3p P1/2 (15.90 eV), work as an energy donor excitation of the Cu II lines because the ionization potential of as well as an electron acceptor. This type of collision is generally xenon is more than 2 eV lower than that of argon. The main represented by the following equation: Mg + Ar+ → M+* + Arg + objective of this study was to suggest the probable excitation ΔE, where the superscripts g, +, and * mean a ground state, an ionic mechanism for the Cu II lines in the xenon GDP. state, and an excited state, respectively, and ΔE is the energy

Table 1. Assignment and Transition Probability of Cu II Lines Employed in This Study Assignment23 Transition probability24 Wavelength (nm) Upper (eV) Lower (eV) (× 108 sec-1) 9 1 9 3 Cu II 201.558 3d 4p P1 (9.1245) 3d 4s D1 (2.9754) 0.23 9 1 9 3 Cu II 202.548 3d 4p D2 (9.0944) 3d 4s D1 (2.9753) 0.84 9 3 9 3 Cu II 203.585 3d 4p D1 (9.0633) 3d 4s D1 (2.9753) 3.96 9 1 9 3 Cu II 203.713 3d 4p F3 (8.9167) 3d 4s D2 (2.8326) 1.39 9 3 9 3 Cu II 204.379 3d 4p D3 (8.7830) 3d 4s D3 (2.7187) 1.92 9 3 9 3 Cu II 205.498 3d 4p D2 (8.8634) 3d 4s D2 (2.8326) 1.72 9 3 9 3 Cu II 208.530 3d 4p F2 (8.6623) 3d 4s D3 (2.7187) 0.057 9 3 9 3 Cu II 210.479 3d 4p D2 (8.8639) 3d 4s D1 (2.9753) 0.98 9 1 9 1 Cu II 211.212 3d 4p P1 (9.1245) 3d 4s D2 (3.2563) 4.09 9 1 9 1 Cu II 212.297 3d 4p D2 (9.0944) 3d 4s D2 (3.2563) 2.52 9 3 9 3 Cu II 212.604 3d 4p F2 (8.6623) 3d 4s D2 (2.8326) 1.64 9 3 9 1 Cu II 213.437 3d 4p D1 (9.0633) 3d 4s D2 (3.2563) 0.017 9 3 9 3 Cu II 213.598 3d 4p F4 (8.5213) 3d 4s D3 (2.7187) 4.19 9 3 9 3 Cu II 214.897 3d 4p F3 (8.4862) 3d 4s D3 (2.7187) 0.64 9 3 9 3 Cu II 217.941 3d 4p F2 (8.6623) 3d 4s D1 (2.9753) 2.37 9 1 9 1 Cu II 218.963 3d 4p F3 (8.9167) 3d 4s D2 (3.2563) 1.10 9 3 9 3 Cu II 219.226 3d 4p F3 (8.4862) 3d 4s D2 (2.8326) 2.83 9 3 9 1 Cu II 221.026 3d 4p D2 (8.8639) 3d 4s D2 (5.2563) 1.48 9 3 9 3 Cu II 221.810 3d 4p P1 (8.4204) 3d 4s D2 (2.8326) 3.08 9 3 9 3 Cu II 222.887 3d 4p P0 (8.5361) 3d 4s D1 (2.9753) 3.69 9 3 9 1 Cu II 224.262 3d 4p D3 (8.7830) 3d 4s D2 (3.2563) 2.17 9 3 9 3 Cu II 224.700 3d 4p P2 (8.2347) 3d 4s D3 (2.7187) 3.34 9 3 9 3 Cu II 227.626 3d 4p P1 (8.4204) 3d 4s D1 (2.9754) 0.54 9 3 9 3 Cu II 229.437 3d 4p P2 (8.2347) 3d 4s D2 (2.8326) 0.21 www.at-spectrosc.com/as/article/pdf/2020204 13 At. Spectrosc. 2021, 42(1), 12-17 EXPERIMENTAL The Grimm-style glow discharge lamp employed in this study has been described in my previous paper.21 Based on the original

Grimm model,22 the lamp comprised a hollow anode of 8.0 mm in diameter and a planner cathode (sample), and the distance between the electrodes was adjusted to 0.3 - 0.5 mm. High-purity xenon

(99.999 %), argon (99.99995 %), and krypton (99.9995%) were employed as the plasma gas. The lamp was evacuated to less than 2.0 Pa with two oil rotary pumps, and then any one of these plasma gases was introduced to flow continuously during the measurements. The flow control of the plasma gas was carried out with a set of gas valves consisting of a ball (on/off) valve and a needle valve, which were inserted into each gas line. The pressure of the plasma gas was measured with a Pirani gauge, whose readings had been corrected for each gas by using a Baratoron- type capacitance manometer, at the vacuum port of the lamp.

The emission spectra were measured with a Czerny-Turner mounting spectrometer (P-5200, Hitachi Corp., Japan), equipped Fig. 1 Correlation in the excitation energy between energy levels of singly ionized copper and the ionization levels of xenon, argon, and krypton. with a photomultiplier (R-955, Hamamatsu Photonics, Japan). The discharge power was supplied with a dc power supply device (HEOPT-1B60-L1, Matsusada Precision Ltd., Japan). All of the excited energy levels of copper and the first ionization level as well measurements were conducted in constant voltage mode. A pure as the metastable levels of the plasma gases employed: xenon, copper plate (purity, 99.99 %) was prepared as the sample, which krypton, and argon. The energy levels are drawn based on the was polished with waterproof emery paper (No. 600) and then database compiled by Moore.18,23,25 The energy scale is rinsed with ethanol. Before measurement, pre-discharge was represented by ‘total excitation energy’ that sums the excitation carried out for a few minutes to remove the surface contaminants. energy in the copper ion and the first ionization potential (7.73 eV).23 As shown in Fig. 1, the excited energy levels of the 3d94s and the 3d94p electron configuration23 are located in the total RESULTS AND DISCUSSION excitation energies of 10.4 - 11.0 eV and 15.9 - 16.8 eV, respectively. Copper ionic lines, which are assigned to optical Wavelength table of analyzed Cu II lines transitions from the 3d94p to 3d94s energy levels, are observed in The optical transitions between the 3d94p and 3d94s electron the GDP-excited spectra; however, their relative intensities are configurations of singly ionized copper23 give rise to many drastically dependent on the plasma gas employed, as described emission lines of Cu II in the wavelength range of 195 - 250 nm. later. 10 1 The ground state is the 3d comprising a singlet level of S0, and A charge-transfer collision for an excited level of copper ion is the 3d94s and the 3d94p are the second and the third excited shown using the following general equation: configurations, respectively.23 In this study, 24 Cu II lines ranging g 10 2 + +* g from 201.6 - 229.5 nm were measured, as listed in Table 1. Their Cu (3d 4s S1/2, 0.00 eV) + G → Cu + G + E (Eq. 1) excitation energies of the upper energy level range from 8.23 to where G is a plasma gas and the superscripts g, +, *, and m indicate 9.12 eV. The electron configuration and the energy levels of the ground, ionic, excited, and metastable states, respectively, and ΔE copper ion are cited in a data book compiled by Moore.23 Kono corresponds to the energy difference before and after the collision. and Hattori measured relative lifetimes of the 3d94p excited levels As pointed out by Steers,17 the Cu II 224.700-nm line has very of the copper ion using a delayed-coincidence technique, then high intensity in an Ar GDP, which is well explained by an estimated the transition probabilities of the Cu II lines from these asymmetric charge-transfer collision with the argon ion. Fig. 1 excited levels with an accuracy of 10 %,24 as cited in the fourth 5 2 includes the ground state levels of the argon ion, e.g., 3p P1/2 column of Table 1. In this study, this set of the transition 2 (15.93 eV) and P3/2 (15.76 eV) on the scale of total excitation probability was employed as a standard for the sensitivity energy.18 The reason for the excitation for the Cu II 224.700-nm correction between the different Cu II lines. line is that there is a similarity in the total excitation energy 9 3 Correlation of excited energy levels between singly ionized between the corresponding upper level of the 3d 4p P2(8.2347 5 2 copper and plasma gases eV) and the 3p P1/2 of argon ion. Such a correspondence of their excitation energies can be graphically understood in Fig. 1. In my Figure 1 illustrates a correlation diagram between the ground and www.at-spectrosc.com/as/article/pdf/2020204 14 At. Spectrosc. 2021, 42(1), 12-17 previous study it was reported that a different group of Cu II lines, which was assigned to the 3d95s - 3d94p transition of the copper ion, was intensively observed in a Ne GDP by the charge transfer collision with the neon ion.26 Furthermore, previous researchers found very intense emission lines of other metallic elements using various plasma gases, such as argon, neon, krypton, and helium, in which a charge-transfer process with the plasma gas was involved.27,28 However, the first ionization potential of xenon, e.g., 5 2 2 25 the 5p P1/2 (13.43 eV) and P3/2 (12.13 eV), is more than 3 eV lower than that of argon, as shown in Fig. 1. This study shows the keen interest taken to establish the effect of plasma gases used for the excitation of the Cu II lines. Fig. 2 Dependence of the reduced emission intensity (the normalized Normalized intensity of copper emission lines for different intensity divided by the transition probability) of Cu II lines on the plasma gases excitation energy in an argon GDP. Discharge condition: 28 mA /500 V at 670-Pa Ar. The net emission intensity of 24 Cu II lines (as listed in Table 1) was measured for three different plasma gases. Each discharge condition was fixed in a voltage constant mode: 27 mA/650 V at 600-Pa Xe, 34 mA/600 V at 600-Pa Kr, and 28 mA /500 V at 670- Pa Ar. These conditions were selected so that the sputtering rate could be similar to each other. Their actual sputtering rates were estimated from the weight loss before and after the 10-min discharges, resulting in a relative value of 0.87 : 1.0 : 0.85 for Xe : Kr : Ar. The net intensities of the Cu II lines were averaged from triplicate measurements and their relative standard deviations were within 5%, except for several weak lines. The intensity data of the Cu II 213.437-nm line was excluded for the subsequent analysis because it overlapped with a neighboring atomic line of copper at Cu I 213.428 nm. Moreover, the net intensities were normalized Fig. 3 Dependence of the reduced emission intensity (the normalized with the intensity of the Cu II 201.558-nm line (9.1245 eV) for intensity divided by the transition probability) of Cu II lines on the excitation energy in a xenon GDP. Discharge condition: 27 mA/650 V at each plasma gas, and the normalized intensities were finally 600-Pa Xe. corrected by the difference of the sputtering rate.

Reduction of the normalized intensity by transition charge-transfer process would work insufficiently for excitations probability of other 3d94p excited levels requiring larger energies due to a lack The normalized intensity of the Cu II lines was divided by the of the total excitation energy. For instance, there appears an energy 1 corresponding transition probability (gA) to compensate for the shortage of 0.93 eV to excite the 4p P1 level for the Cu II 201.558- different frequencies for each optical transition. A value of nm and the Cu II 211.212-nm lines. Therefore, assuming that their (Intensity / gA), defined as ‘reduced emission intensity’, will be excitations are caused by a thermal process, such as electron discussed in relation to the excitation energy. collision, the charge-transfer collision additionally contributes to 3 the excitation of the 4p P2 level by a factor of 200-300, as seen in First, the result of the Ar GDP, which has already been Fig. 2. recognized to include the selective excitation of the Cu II lines, is represented in Fig. 2. This excitation-energy dependence of the A similar analysis to the Ar GDP was carried out in a Xe GDP. reduced emission intensity clearly indicates an abnormal enhanced Fig. 3 shows a variation in the reduced emission intensity of the 9 3 population of the 3d 4p P2 (8.2347 eV) and a normal population Cu II lines as a function of the excitation energy in the Xe GDP. of the energy levels having excitation energies of more than 8.8 Differing from the result of the Ar GDP (see Fig. 2), the plots seem eV. A charge-transfer process, in which a collision with the 3p5 to be slightly varied with an increase in the excitation energy along 2 P1/2 (15.93 eV) of the argon ion is involved, well explains this with a negative linear relationship (a dotted line in Fig. 2), although selective excitation because of a good match in the total excitation their deviations are somewhat large. This result implies that the energy between the colliding partners. In this reaction, the Cu II lines would be mainly excited by any thermal collision difference in the total excitation energy is estimated to be -0.035 (transfer of the kinetic energy) in the Xe GDP and that there are eV, thus realizing an energy resonance condition. Further, the no particular channels for their excitations as in the Ar GDP. As www.at-spectrosc.com/as/article/pdf/2020204 15 At. Spectrosc. 2021, 42(1), 12-17 combinations with the excited levels in the range of 8.3-8.6 eV, ΔE becomes almost zero and energy resonance is expected. As

compared to the result with Xe GDP (see Fig. 3), the reduced intensity of the Cu II lines in the lower excitation energy look to be slightly elevated in the Kr GDP graph. However, there is no

clear evidence in Fig. 4 that any selective excitation among these Cu II lines may occur. A reason for this is that these metastable levels might be less populated than the ground-state level in the plasma.

CONCLUSIONS

Fig. 4 Dependence of the reduced emission intensity (the normalized This paper represents the excitation feature of the Cu II lines intensity divided by the transition probability) of Cu II lines on the derived from the 3d94p - 3d94s transition in a Xe GDP, in excitation energy in a krypton GDP. Discharge condition: 34 mA/600 V at comparison to Ar and Kr GDPs. As pointed out in previous 600-Pa Kr. studies, a very intense emission line, the Cu II 224.700-nm line, was observed in the Ar GDP. The major excitation mechanism is illustrated in Fig. 1, it is not possible to cause the charge-transfer an asymmetric charge transfer collision with the argon ion, in process from the Xe ion to reach the 3d94p excited levels of the which there is a coincidence in the excitation energy between the 9 3 copper ion. In this case, there is much less excitation energy (2.5 - corresponding energy level of the 3d 4p P2 and the ground-state 3.4 eV) in the charge-transfer collision for their excitations. In level of the argon ion. However, when the plasma gas was changed addition, other Cu II lines, which were assigned to the 3d95s (13.39 from argon to xenon, no intense Cu II lines were found in the GDP - 13.68 eV) - 3d94p transition and assigned to further high-lying spectra, but a linear relationship among the reduced intensities of excited states, were very faint or were not able to be observed in the Cu II line was found. In the Xe GDP, they were not emitted by the Xe GDP. Accordingly, few Cu II lines with strong intensity a non-thermal process, such as charge-transfer collision, but by a were not found in the Xe-GDP spectra, principally because there thermal excitation process, such as electron collision. The major were no effective excitation channels for the charge-transfer reason for this is that the first ionization potential of xenon is more 9 collision with the xenon ion. than 3 eV lower than that of argon, thus not enabling the 3d 4p excited levels to be populated through the charge-transfer collision Finally, the excitation behavior of the Cu II lines was with the xenon ion. The results obtained with the Kr GDP was investigated when krypton was employed as the plasma gas. Fig. similar to that with the Xe GDP, because the ionization potential 4 shows a dependence of their reduced intensities on the excitation of krypton was still insufficient to obtain the excited levels. Such energy in a Kr GDP. This variation is analogous to that of the Xe large differences of the intensity plots imply that the charge GDP (see Fig. 3), implying that a similar excitation mechanism transfer process mainly determines the emission spectrum from would work for excitations of the Cu II lines. The first ionization high-lying excited energy levels, like the 3d94p energy levels of 5 2 2 potential of krypton, e.g., 4p P1/2 (14.00 eV) and P3/2 (14.66 the copper ion, so also in other metallic elements. eV),23 is more than 1 eV higher than that of xenon; however, they cannot work as energy donors in the charge-transfer collision to produce the 3d94p excited levels of the copper ion. Fig. 1 shows AUTHOR INFORMATION that the internal energy of the krypton ion is short for this collisional reaction by 1 - 2 eV. On the other hand, the first excited Corresponding Author 9 2 2 state of the copper atom, comprising 3d 4s D5/2 (1.389 eV) and 2 *K. Wagatsuma D3/2 (1.642 eV), is located more than 1 eV above the ground-state level.23 They might act as metastable energy levels for any Email address: [email protected] collision process in GDPs, because they belong to the same even 10 2 Notes parity as the ground state of the copper atom, the 3d 4s S1/2 (0.00 eV). In such a case, the metastable levels may take part in a charge- The authors declare no competing financial interest. transfer collision as given in the following general equation:

m 9 2 2 2 + +* g Cu (3d 4s D3/2 and D5/2) + G → Cu + G + E (Eq. 2)

If krypton is employed as G, it is possible to cause a charge- transfer collision to produce the 3d94p excited levels of the copper ion due to a good matching in the total excitation energy. In some www.at-spectrosc.com/as/article/pdf/2020204 16 At. Spectrosc. 2021, 42(1), 12-17 REFERENCES 13. Z. Weiss, E. B. M. Steers, and J. C. Pickering, Spectrochim. Acta Part B, 2015, 110, 79-90. 1. R. Payling, D. Jones, and A. Bengtson, Glow Discharge Optical http://dx.doi.org/10.1016/j.sab.2015.05.013 Emission Spectrometry, John Wiley and Sons, Chichester, U.K., 14. S. Mushtaq, J. Anal. At. Spectrom., 2020, 35, 1814-1826. 1997. https://doi.org/10.1039/D0JA00001A 2. T. Neils and R. Payling, Glow Discharge Optical Emission 15. A. von Engel, Ionized gases, Clarendon Press, London, U.K., 1965. Spectroscopy - A Practical Guide, The Royal Society of Chemistry, 16. O. S. Duffendach and J. G. Black, Phys. Rev., 1929, 34, 35-43. Cambridge, U.K., 2003. https://doi.org/10.1103/PhysRev.34.35 3. R. K. Marcus and J. A. C. Broekaert, Glow Discharge Plasmas in 17. E. B. M. Steers and R. J. Fielding, J. Anal. At. Spectrom., 1987, 2, Analytical Spectroscopy, John Wiley and Sons, Chichester, U.K., 239-244. https://doi.org/10.1039/JA9870200239 2003. 18. C. E. Moore, Atomic Energy Levels Vol. I, NBS Circular 467, 4. K. Wagatsuma and K. Hirokawa, Anal. Chem., 1985, 57, 2901-2907. Washington, D.C., USA, 1948. https://doi.org/10.1021/ac00291a036 19. K. Wagatsuma and H. Honda, Spectrochim. Acta Part B, 2005, 60, 5. K. Wagatsuma and K. Hirokawa, Spectrochim. Acta Part B, 1987, 1538-1544. https://doi.org/10.1016/J.SAB.2005.10.004 42, 523-531. https://doi.org/10.1016/0584-8547(87)80031-9 20. K. Wagatsuma, Anal. Bioanal. Chem., 2009, 393, 2067-2074. 6. K. Wagatsuma, K. Hirokawa, and N. Yamashita, Anal. Chim. Acta, https://doi.org/10.1007/S00216-009-2700-5 1996, 324, 147-154. https://doi.org/10.1016/S0003-2670(97)00048-2 21. K. Wagatsuma and K. Hirokawa, Surf. Interface Anal., 1984, 6, 167-170. https://doi.org/10.1002/SIA.740060404 7. K. Wagatsuma, Anal. Sci., 2003, 19, 325-327. https://doi.org/10.2116/analsci.19.325 22. W. Grimm, Spectrochim. Acta Part B, 1968, 23, 443-454. https://doi.org/10.1016/0584-8547(68)80023-0 8. K. Wagatsuma, Anal. Sci., 2010, 26, 303-309. https://doi.org/10.2116/analsci.26.303 23. C. E. Moore, Atomic Energy Levels Vol. II, NBS Circular 467, Washington, D.C., USA, 1952. 9. S. Mushtaq, V. Hoffmann, E. B. M. Steers, and J. C. Pickering, J. Anal. At. Spectrom., 2012, 27, 1423-1431. 24. A. Kono and S Hattori, J. Opt. Soc. Am., 1982, 72, 601-605. https://doi.org/10.1039/C2JA10359A https://doi.org/10.1364/JOSA.72.000601 10. S. Mushtaq, E. B. M. Steers, V. Hoffmann, Z. Weiss, and 25. C. E. Moore, Atomic Energy Levels Vol. III, NBS Circular 467, J. C. Pickering, J. Anal. At. Spectrom., 2016, 31, 2175-2181. Washington, D.C., USA, 1958. https://doi.org/10.1039/C6JA00231E 26. K. Wagatsuma and K. Hirokawa, Spectrochim. Acta Part B, 1991, 11. S. Mushtaq, E. B. M. Steers, J. C. Pickering, and K. Putyera, 46, 269-281. https://doi.org/10.1016/0584-8547(91)80028-2 J. Anal. At. Spectrom., 2014, 29, 681-695. 27. K. Wagatsuma, Z. Physik, 1996, D37, 231-239. https://doi.org/10.1039/C3JA50332A https://doi.org/10.1007/S004600050032 12. Z. Weiss, E. B. M. Steers, J. C. Pickering, and S. Mushtaq, 28. K. Wagatsuma, J. Anal. At. Spectrom., 1996, 11, 957-966. Spectrochim. Acta Part B, 2014, 92, 70-83. https://doi.org/10.1039/JA9961100957 http://dx.doi.org/10.1016/j.sab.2013.12.006

www.at-spectrosc.com/as/article/pdf/2020204 17 At. Spectrosc. 2021, 42(1), 12-17 Spark Discharge-LIBS: Evaluation of One-Point and Multi- Voltage Calibration for P and Al Determination

Alan Lima Vieira,a Edilene Cristina Ferreira,a Dário Santos Júnior,b Giorgio Saverio Senesi,c and José Anchieta Gomes Netoa,* a São Paulo State University - UNESP, Analytical Chemistry Department, P.O. Box 355, 14801-970, Araraquara - SP, Brazil b Federal University of São Paulo - UNIFESP, Chemistry Department. 09913-030, Diadema - SP, Brazil c CNR-Istituto per la Scienza e Tecnologia dei Plasmi (ISTP) sede di Bari, Via Amendola, 122/D - 70126 Bari, Italy

Received: October 31, 2020; Revised: November 12, 2020; Accepted: November 12, 2020; Available online: November 30, 2020.

DOI: 10.46770/AS.2020.202

ABSTRACT: Spark discharge (SD) laser-induced breakdown spectroscopy (LIBS) is a technique suitable to overcome the low energies of lasers by reheating the plasma, increasing the emission intensities and to perform single-standard calibration. A calibration method called one-point and multi-voltage calibration (OP-MVC), which requires two different voltages applied to both the standard and the sample, is proposed for use with SD-LIBS. The performance of this method was compared to that of the one- point and multi-lines calibration (OP-MLC) and the slope ratio calibration (SRC) methods for LIBS determination of Al in certified reference plant leaves and P in commercial fertilizers. No statistical differences at the 95% confidence level were observed between the Al and P concentrations determined by OP-MVC LIBS for the Al certified values and the P values measured by high-resolution continuum-source flame atomic absorption spectrometry (HR-CS

FAAS). The limit of detection (LOD) for P was 0.60 wt% P2O5 and 35.1 mg kg-1 for Al. The relative standard deviation (n=3) was typically 7% for Al and in the 4 - 10% range for P.

INTRODUCTION decreased 10 times for Pb, Mg and Sn in soil15 and improved to 0.028 ppm for Cu in onion leaves.23 The interest in laser-induced breakdown spectroscopy (LIBS) for elemental analysis has increased greatly in recent years due to its Besides sensitivity, calibration is also considered a main issue in capability to analyze a wide range of samples.1 However, despite LIBS analysis due to the matrix effects.9,24 Several univariate and its advantages, the low sensitivity and calibration of this technique multivariate calibration methods have been applied to overcome are still challenging.2-4 these effects, including partial least squares (PLS),9,25 artificial neural networks (ANN),9,25 principal component regression Among several approaches attempted to achieve LIBS signal (PCR),9,25 external calibration (EC),18,24 internal standardization enhancement,2,3,5,6 spark discharge (SD) was found to be simple (IS),24 calibration-free (CF),26,27 one-point calibration (OPC),27,28 and relatively low-cost.7 Since its first proposal in 1962,7 SD-LIBS CF inverse,27,29 C-sigma,27,30 and single-sample calibration has been used to analyze a variety of samples8 in various sectors (SSC).31 In particular, the SSC method does not require the of agriculture9, such as soil,10-17 phosphate rock,10,18 rice,19 honey,20 calculation of plasma temperature, electron number density or coal21,22 and onion leaves.23 For instance, when using the SD other experimental parameters.26-30 method, the signal-to-background (S/B) ratio increased up to 7 times in the analysis of Pb and As in soil,11 up to 3 times for As, The use of a single standard to produce several signal intensities Al, Ba, Ca, Co, Fe, Mg, Pb, Si, Sr, Ti and V in soil,13 and up to 12 (multi-points), which avoids the preparation of a set of solid times for P in fertilizer.18 Furthermore, the detection limit standards with a similar matrix to test the samples at different www.at-spectrosc.com/as/article/pdf/2020202 18 At. Spectrosc. 2021, 42(1), 18-24 analyte concentrations, is very attractive. The one-point and multi- prepare all solutions, obtained with a Millipore Rios 5® reverse line calibration (OP-MLC) method proposed by Hao et al.32 used osmosis and a Milli-Q Academic® deionizer system (Millipore several wavelengths of Mn, Cr, Ni and Ti to analyze alloy steel, Corporation, Bedford, MA, USA). Hydrochloric acid, nitric acid, achieving the relative errors of 9, 22, 21 and 36%, respectively. and 30 wt% hydrogen peroxide were purchased from Merck Nunes et al.33 used the slope ratio calibration (SRC) method33 to (Darmstadt, Germany). determine B, Ca, Cu Mg, Fe, Mn, P and Zn in plant leaves with an Digestion was performed in a Multiwave microwave oven equation equivalent to single-point calibration. An increase of the (Anton Paar, Graz, Austria). The optimized program involving emission intensities as a function of the applied voltage was power/ramp time/hold time consisted in the following steps: step observed in the SD-LIBS analysis of Al and Cu in air,34 As, Pb, Si 1, 1000 W/15 min/0 min; step 2, 1000 W/0 min/10 min; step 3, 0 and Mn in soil,11,35 Si in silicon crystals,36 P in fertilizers18 and Pb W/0 min/20 min (ventilation). After cooling, the digests were and Fe in copper alloys.37 Quantitative analysis by SD-LIBS transferred into polypropylene flasks and diluted with deionized usually employs calibration with a fixed single voltage and water to a final volume of 30 mL. standards at various concentrations,9-23 whereas the use of a single standard and various applied voltages has rarely been reported in Instrumentation the literature. The LIBS system used for Al and P determinations consisted of a This study aimed to evaluate the performance of one-point and Q-switched ND:YAG laser (Big Sky Ultra 50, Quantel USA, multi-voltage calibration (OP-MVC) for the SD-LIBS Bozeman, MT, USA), operating at 1064 nm, 9 ns of pulse determination of Al and P in plant leaves and fertilizers, duration, 10 Hz repetition rate, energy power of 48.7 ± 0.4 mJ, respectively, in comparison with high-resolution continuum measured by laser power and energy meter (FieldMaxII-P, source atomic absorption spectrometry (HR-CS AAS), OP-MLC Coherent, Inc., Santa Clara, CA, USA). The laser beam was and SRC. This method required the measurement of the intensities focused onto the sample surface by a plano-convex lens (12-cm (I) of the standard (Cstd) and the unknown sample (Csample) at two focal length), thus producing a spot diameter of about 300 µm, voltages, so that Csample can be calculated (see Equation 3 below). resulting in a laser fluence of ca. 69 J cm-2 and an irradiance of 0.59 GW cm-2 delivered to the sample. The emission spectra were collected by an optical fiber bundle at an angle of 45° with respect EXPERIMENTAL to the laser beam and transferred to four spectrometers of the HR2000+ (Ocean Optics, Dunedin, FL, USA), operating at a Samples, standards, reagents and procedures spectral range from 200 at 620 nm, with optical resolution of 0.5 Six fertilizer samples containing different P concentrations (P2O5 nm full width at half maximum (FWHM). The analysis was wt%) were obtained at a local market in Araraquara city (São performed using 1-ms integration time and 1-µs Q-switched delay. Paulo State, Brazil). An amount of 1000 g of each sample was The LIBS system included an automatic x-y direction sample quartered, and aliquots of 250 g were ground manually with a pistil holder and a video camera inside the sampling chamber in order in an agate mortar to obtain a fine powder, which was then dried to help the analyst to control the sample position. The laser was at 105° C for 12 h. Calcium carbonate and NIST SRM 1568a Rice rasterized over a sample surface of about 3 mm2 and broadband Flour were employed as blanks for P and Al, respectively, and to spectra from 10 single-pulse ablations were collected and mapped calculate the limit of detection (LOD) and the limit of on discrete XY spatial coordinates. All measurements were quantification (LOQ). The 1515 Apple Leaves, 1570a Spinach performed at atmospheric air pressure. Leaves, 1573a Tomato Leaves and 120c Phosphate Rock standard The SD-LIBS system employed two cylindrical, pure tungsten reference materials (SRM) from the National Institute of electrodes of 100-mm length and 2.6-mm diameter, with tips Standards and Technology (NIST, Gaithersburg, MD, USA) were arranged at a distance of 4 mm between them and 2 mm above the employed for the evaluation of calibration and accuracy. Disc sample surface. The electrodes were connected to a single-spark pellets of 250 mg powdered fertilizer samples, SRM plant leaves generator operating in the voltage range from 2.0 to 4.5 kV. More and blanks were prepared daily using the SL-10/15 hydraulic press information about the SD system can be found elsewhere.18 The (Solab, Piracicaba, Brazil), operating at a pressure of 7.5 ton cm-2 peak heights of the atomic lines of Al I at 308.22 nm and P I at for 3 min. 214.91 nm were obtained using Excel spreadsheet For comparative purposes, P was determined by HR-CS AAS “Chromatograms and Spectra Handling”38 and Microsoft Excel after acid digestion of the sample. In particular, 200 mg of each for data processing and calculating the concentrations. A scheme fertilizer sample and phosphate rock SRM was accurately of the experimental setup is shown in Fig. 1. weighed, transferred to a microwave flask, and 3 mL of HCl and The concentrations of the sample digests, prepared as described HNO3 solution at a ratio 1:3 (reverse aqua regia), 2 mL of above, were determined by HR-CS AAS using an Analytik Jena deionized water and 1 mL of hydrogen peroxide were added. High ContrAA 300, equipped with a xenon short-arc lamp XBO 301 purity de-ionized water (resistivity 18.2 MΩ cm) was used to www.at-spectrosc.com/as/article/pdf/2020202 19 At. Spectrosc. 2021, 42(1), 18-24

Fig. 1 Scheme of the experiment.

(퐼 +퐼 ).(푉 −푉 ) (GLE, Berlin, Germany) and operating in hot-spot mode as the 퐴 = 1 2 2 1 (Eq. 4) 2 continuum radiation source was used. The instrument was If the emission intensity is proportional to the concentration, equipped with a high-resolution double-Echelle grating then also its area is. Thus, if the ratio between the intensities in Eq. monochromator (< 2 pm per pixel in the far ultraviolet range) and 3 is replaced by the ratio between the corresponding areas, the a charge-coupled device (CCD) array detector. All measurements concentration of the analyte in the sample can be calculated were performed in triplicate using optimized operating according to: conditions.39 푠푎푚 푠푎푚 푠푡푑 (퐼1+퐼2) 퐶 = 퐶 푠푡푑 (Eq. 5) Fundamental aspects (퐼1+퐼2) The LOD and LOQ can be calculated according to: In atomic emission spectrometry, the emission response at a LOD = CBlk + 3SBlk (Eq. 6) given wavelength (퐼푗푖 ) is directly related to the analyte (휆푖푗) 32 LOQ = CBlk + 10SBlk (Eq. 7) concentration (C) and the excited-state transition energy (Ej), according to: where CBlk and SBlk are, respectively, the average concentration

푗푖 퐹퐶퐴푗푖푔푗 and the standard deviation of 10 measurements of the 퐼 = 푒−퐸푗/푘퐵푇 (Eq. 1) (휆푖푗) 푄(푇) concentration in the blank samples.40 where F, Aji, gj, Q(T), kB, and T are, respectively, the instrumental In OP-MVC, the uncertainties of the analyte concentration proportionality constant, transition probability, excited-state degeneracy, partition function, Boltzmann constant and (푆퐶푠푎푚 ) depend on the errors associated to the intensities of both temperature. the sample (푆퐼푠푎푚 ) and the standard (푆퐼푠푡푑), and on the confidence Based on this equation, for a certain analyte with the interval of the CRM standard (푆 ), which can be estimated by concentration C in a given medium, the emission intensity, I, at a 퐶푠푡푑 given wavelength, can be described as: the error propagation according to:

2 2 2 I = mC (Eq. 2) 푆퐶 푆퐶 푆퐼 푆퐼 푠푎푚 = √( 푠푡푑) + ( 푠푎푚) + ( 푠푡푑) (Eq. 8) 퐶푠푎푚 퐶푠푡푑 퐼푠푎푚 퐼푠푡푑 where m is the proportionality constant of the LIBS instrument. Comparative strategies based on single-standard calibration If a calibration standard with a similar matrix is measured by the The performance of OP-MVC was compared to that of the SRC33 same instrument in identical plasma conditions, the ratio between and OP-MLC32 methods in determining P in fertilizers and Al in the intensities is proportional to the ratio between the respective plant materials. In particular, the following equation was used for concentrations, then the concentration of the analyte in the sample SRC: can be calculated according to: 푆푎푚푝푙푒 퐼푠푎푚 푏 퐶푠푎푚 = 퐶푠푡푑 (Eq. 3) 퐶푆푎푚푝푙푒 = 퐶푆푡푑 (Eq. 9) 퐼푠푡푑 푏푆푡푑 The proposed OP-MVC method requires that both the standard where bsample and bstd are the slopes of the standard and the sample, and the sample be submitted to SD using two different voltages which were obtained using the 5, 10, 15, 20 and 30 cumulative

(V1, V2) consecutively, so that the two measured intensities, I1 and laser pulses with the SD-LIBS operating at 4.5 kV. The analytical

I2, define a rectangular trapeze with an area A, given by: lines P I at 214.91 nm and Al I at 308.22 nm were used for P and Al, respectively. www.at-spectrosc.com/as/article/pdf/2020202 20 At. Spectrosc. 2021, 42(1), 18-24 The following equation was used for OP-MLC: earlier information,18 the P I line at 214.91 nm was selected for P determination in fertilizers, and the Al I line at 308.22 nm was 퐶푆푎푚푝푙푒 = 퐶푆푡푑 훼푆푎푚푝푙푒 .푆푡푑 (Eq. 10) chosen for Al determination in plants. Despite the higher intensity where αSample.Std is the slope obtained by plotting the intensities of observed for the line of Al I at 309.27 nm, this line was not used the sample (y-axis) versus those of the standard (x-axis) measured due to possible spectral interferences caused by the Mg I lines at for different emission lines of the analyte. In particular, the 309.10, 309.29 and 309.68 nm.41 These lines might become emission lines of P I at 213.62 nm, 214.91 nm and 215.41 nm, and critical for samples rich in Mg, such as NIST SRMs 1573a and the line of Al I at 308.22 nm, which was selected after considering 1515 which contain 12,000 mg kg-1 (not certified) and 2710 ± 120 its lower interference, were employed. However, as OP-MLC mg kg-1 of Mg, respectively. requires multiple lines, different wavelengths located at the wing The effect of the applied voltage at the values of 2.0, 2.5, 3.0, of the Al I core line at 308.22 nm, i.e., 308.01, 308.07, 308.12, 3.5, 4.0 and 4.5 kV on the emission intensities of the P I line at 308.17, 308.22 nm, were also used. The SD-LIBS employed 4.5 214.91 nm and the Al I line at 308.22 nm was evaluated for the kV because the Al signals were very low when only LIBS is used. NIST SRM 120c and NIST SRM 1515 samples, respectively (Fig. 3). In both cases, the plots of the line intensity against the applied RESULTS AND DISCUSSION voltage were linear with correlation coefficients (r) of 0.989 and 0.999 for Al and P, respectively. The NIST SRM 1570a and 1573a The P lines profile in the spectra of NIST SRM 120c Phosphate in the case of Al (Fig. 3a) and all fertilizer samples in the case of P Rock (33.34 ± 0.06 wt% P2O5) and the Al lines profile in the (Fig. 3b) were analyzed only at higher voltages, i.e., 3.5, 4.0 and spectra of NIST SRM 1515 Apple Leaves (284.5 ± 5.8 mg kg-1 Al) 4.5 kV, at which the highest line intensity of the analyte, i.e., a acquired by LIBS and SD-LIBS are showed in Fig. 2. Based on greater sensitivity, was achieved. In all cases, for both Al and P, the intensity versus voltage curves were linear with r > 0.97.

The extended temporal evolution of the plasma, and the increase

of plasma temperature and electronic density achieved by reheating the plasma, improved the atomization and excitation processes.10 In particular, the highest increase in the signal

intensity for both elements, i.e., 6.7-fold for P (NIST SRM 120c) and 4.4-fold for Al (NIST SRM 1515), was achieved at the maximum applied voltage, i.e., 4.5 kV, which yielded a satisfactory relative standard deviation (RSD) for both Al and P. Thus, further experiments aiming to optimize data collection were carried out at the voltage of 4.5 kV.

Notwithstanding the high content of P in fertilizers and the re- excitation with 4.5-kV high voltage discharge, no self-reversal Fig. 2 LIBS (dashed line) and SD-LIBS (continuous line) spectra for (a) P effect was observed for the peak profiles of P in the concentration in NIST SRM 120c Phosphate Rock and (b) Al in NIST SRM 1515 Apple range from 4.8 to 33.34 wt% P2O5 (Fig. 4a). Furthermore, the plots Leaves. The voltage applied in SD-LIBS was 4.5 kV. of the intensity of the emission lines of P I at 213.62, 214.91 and

Fig. 3 Plots of line intensity vs. voltage (n = 3) for (a) line Al I at 308.22 nm in three NIST SRM plant leaves and (b) line of P I at 214.91 nm in NIST SRM 120c and fertilizer samples. The error bars represent the standard deviation of the measurements. www.at-spectrosc.com/as/article/pdf/2020202 21 At. Spectrosc. 2021, 42(1), 18-24 showed that the average intensity for both the Al and P lines remained constant for replicated spectra  5 and that a satisfactory precision (RSD: 6% for P and 17% for Al) was achieved for replicate spectra  10. Thus, the total of 10 replicate spectra was chosen for data acquisition as a compromise between precision, sensibility and sample throughput.

After optimization, the proposed calibration method was applied to determine P in fertilizers and Al in plant leaves using one or two voltages. In particular, the NIST SRM 120c Phosphate

Rock was used as the unique standard to determine the P content in the five commercial fertilizer samples, which were also analyzed by HR-CS FAAS. The data in Table 1 show good

agreement with the paired t-test at the 95% confidence level Fig. 4 SD-LIBS spectra of solid standards containing 0 (black), 4.8 (red), between the data achieved by OP-MVC LIBS and those obtained 11.9 (light blue), 22.9 (magenta), 27.8 (green) and 33.34 wt% P2O5 (dark by HR-CS FAAS. The RSD values for P were in the 4.3 – 9.6% blue) at an applied voltage of 4.5 kV (a), and (b) linear working range of P range. I lines at 213.62, 214.91 and 215.41 nm. For Al determinations, the NIST SRM 1515 Apple Leaves was used as the unique standard to analyze NIST SRM 1570a Spinach

Leaves and NIST SRM 1573a Tomato Leaves, whereas the NIST SRM 1570a Spinach Leaves was used as the unique standard to analyze NIST SRM 1515 Apple Leaves. The results obtained by OP-MVC LIBS were in good agreement with the certified values with a paired t-test at the 95% confidence level (Table 1). The recoveries were in the range of 94 – 107% and the RSD values in the 6.0 – 6.9% interval.

The use of two voltages, V1: 3.5 kV and V2: 4.5 kV applied sequentially, achieved good performance in precision and accuracy, which may be attributed to the linear relationship existing between the line intensity and the applied voltage (Fig. 3).

Furthermore, the use of three voltages appeared not to alter the Fig. 5 Effect of the number of replicate spectra on the emission line results significantly, but it limited the analytical throughput due to intensities of Al I at 308.22 nm in NIST 1515 Apple Leaves SRM (○) and the increased time needed to charge the capacitors. Calibration P I at 214.91 nm in NIST 120c Phosphate Rock SRM (●) recorded at an using only one applied voltage at 4.5 kV, which provided the best applied voltage of 4.5 kV. Circles correspond to the average values and error bars to standard deviations. signal-to-noise ratio, was also tested. For the P and Al data, the relative standard deviations were in the range of 4.3 – 9.5% and 6 – 11%, respectively (Table 1). However, the precision achieved for 215.41 nm versus P concentration yielded straight lines (Fig. 4b) both P and Al was improved when calibration was performed with correlation coefficients  0.997. As these results were using two voltages (3.5 and 4.5 kV), which provided also better obtained using the maximum voltage discharge, the OP-MVC discrimination between the nominal and the actual applied voltage. method yielded emission intensities in the linear range. Generally, SD-LIBS studies do not investigate the influence of re-excitation For comparative purposes, the performance of the SRC and OP- by high-voltage discharge on self-absorption. The plasma MLC methods was also tested for P and Al determination by LIBS temperature homogeneity increase achieved by SD-LIBS without (Table 1). The results obtained for Al by OP-MVC were in changing the ablation rate would explain the reduction of P self- agreement with those achieved by SRC and OP-MLC at the 95% absorption in the plasma periphery. A similar effect was observed confidence level, but the RSD values for OP-MVC (in the range by Tang et al.42 who used microwave to increase LIBS sensitivity. of 6.0 – 7.0%) were better than those for SRC and OP-MLC (in the range of 8.3 – 15.8% and 23.0 – 26.7%, respectively). However, The effect of the number of replicate spectra recorded at the the results obtained for P were in agreement with those achieved emission intensities of P I at 214.91 nm and Al I at 308.22 nm was by SRC, but not by OP-MLC, at the 95% confidence level. The evaluated by acquiring 3, 5, 10, 15, 20, 25 and 30 spectra for the RSD values of P obtained by OP-MVC, SRC and OP-MLC NIST SRM 120c Phosphate Rock and NIST SRM 1515 Apple ranged from 4.3 – 9.5%, 4.0 – 14.1% and 0.6 – 15.8%, respectively. Leaves at the voltage of 4.5 kV. The results obtained (Fig. 5) Using OP-MVC, SRC and OP-MLC, the LOQs for P (expressed www.at-spectrosc.com/as/article/pdf/2020202 22 At. Spectrosc. 2021, 42(1), 18-24 -1 Table 1. Comparative Results (Mean ± Standard Deviation, n=3) for the Determination of P (wt% P 2O5) in Fertilizers and Al (mg kg ) in Plant Leaves Obtained by OP-MVC (with One and Two Voltages) and SRC and OP-MLC Methods OP-MVC Comparative Standard used Sample/SRM V1 = 3.5 kV SRC OP-MLC values V1 = 4.5 kV for calibration V2 = 4.5 kV P Sample 1 19.0 ± 0.3 a 17.3 ± 1.4 17.7 ± 0.9 20.5 ± 2.9 15.0 ± 0.1c SRM-120c Sample 2 6.5 ± 0.2 a 7.0 ± 0.8 6.9 ± 0.3 6.4 ± 0.4 7.9 ± 0.3 c SRM-120c Sample 3 17.2 ± 0.4 a 16.5 ± 1.6 17.1 ± 1.1 16.2 ± 1.7 14.0 ± 1.3 c SRM-120c Sample 4 9.2 ± 0.1 a 9.1 ± 0.7 9.4 ± 0.9 9.5 ± 1.1 11.4 ± 1.8 c SRM-120c Sample 5 19.0 ± 0.1 a 18.8 ± 1.2 18.5 ± 0.8 19.7 ± 0.8 16.8 ± 1.3 c SRM-120c Al NIST 1570a 310 ± 15 b 322 ± 14 316 ± 2 290 ± 46 343 ± 90 NIST 1515 NIST 1515 284.5 ± 5.8 b 288 ± 27 280 ± 18 295 ± 25 265 ± 61 NIST 1570a NIST 1573a 598.4 ± 7.1 b 586 ± 56 591 ± 41 579 ± 48 685 ± 183 NIST 1515 a Comparative value determined by HR-CS FAAS; b Certified values; c Not in agreement with comparative values at the 95% confidence level. as wt% P2O5) were, respectively, 2.0%, 2.3% and 2.7%, and for method was also evaluated with respect to the SRC and OP-MLC Al (mg kg-1) 117, 126 and 269. methods. Although SRC and OP-MLC are simple and performance calibration methods, they need high sensitivity LIBS Different LIBS configurations may affect the quality of the data systems depending on the specific analyte and sample. Further, obtained for P using OP-MLC, as the emission line of P I at 213.62 SRC can be performed with one emission line, while OP-MLC is nm may be interfered by those of Cu and Zn, which are commonly limited to elements that feature several lines. In general, all of these present in fertilizers, thus deteriorating the calibration curve calibration methods employ a single solid calibration standard, so obtained for OP-MLC. In particular, the correlation coefficients that difficulties related to the preparation of several solid standards obtained by plotting the P signal for the sample in the y-axis and (e.g., external calibration) with physicochemical properties similar that of the NIST SRM 120c Phosphate Rock in the x-axis varied to the samples are minimized. from 0.946 to 0.998, with the lower values possibly due to these spectral interferences. The significantly higher RSD values Finally, the proposed OP-MVC method cannot be considered as obtained by OP-MLC for Al in plant materials with respect to limited to LIBS analysis of Al and P in plants and fertilizers, but those for P in fertilizers (Table 1) might be due to the much lower its performance appears promising to be tentatively extended to content of Al (mg kg-1) compared to P (wt%). Furthermore, the other analytes and samples in various matrices. wavelengths employed for Al were located at the wing of the Al I line at 308.22 nm, i.e., the lines at 308.01, 308.07, 308.12, 308.17, and 308.22 nm. Although this procedure is not commonly used for AUTHOR INFORMATION MLC, it was a strategy adopted to overcome the limited number of non-interfered Al I lines. In conclusion, the differences among Corresponding Author the results obtained with these various methods might be due to *J. A. Gomes Neto the different instrumental strategies used to increase the signal intensities, such as cumulative laser pulses (SRC), several lines Email address: [email protected] (OP-MLC) and voltages (OP-MVC). Notes

The authors declare no competing financial interest. CONCLUSIONS In this work, SD-LIBS was successfully used in combination with ACKNOWLEDGMENTS the SSC calibration method of OP-MVC to determine Al and P in plant leaves and fertilizers, respectively, achieving suitable The authors thank the São Paulo Research Foundation (FAPESP, accuracy, precision and LOQs. The required SD extra device used Grant # 2019/07537-6) for financially supporting this work and should not be seen as a serious shortcoming as it is a simple, low the Conselho Nacional de Desenvolvimento Científico e cost, and rugged tool capable of enhancing the signal intensities Tecnológico (CNPq) for the fellowship to A.L.V. (grant# for any LIBS system, thus improving the performance especially 141977/2016-7), and researchship to J.A.G.N. (grant # of those featuring limited sensitivity. 302414/2017-7) and E.C.F. (grant # 308200/2018-7).

For comparative purposes, the performance of the OP-MVC www.at-spectrosc.com/as/article/pdf/2020202 23 At. Spectrosc. 2021, 42(1), 18-24 REFERENCES Z. Wang, J. Anal. At. Spectrom., 2019, 34, 1047-1082. https://doi.org/10.1039/C9JA00016J 1. S. Musazzi, and U. Perini, Laser-Induced Breakdown Spectroscopy, 23. A. Jabbar, M. Akhtar, S. Mehmmod, M. Iqbal, R. Ahmed and Theory and Applications. Springer-Verlag, Berlin, Germany, 2014. M. A. Baig, Spectrochim. Acta B., 2019, 162, 105719. 2. X. Fu, G. Li and D. Dong, Front. Phys., 2020, 68, 10.3389. https://doi.org/10.1016/j.sab.2019.105719 https://doi.org/10.3389/fphy.2020.00068 24. D. V. Babos, A. I. Barros, J. A. Nóbrega, and E. R. Pereira-Filho, 3. Y. Li, D. Tian, Y. Ding, G. Yang, K. Liu, C. Wang and X. Han, Spectrochim. Act. B., 2019, 155, 90-98. Appl. Spectrosc. Rev., 2018, 53, 1-35. https://doi.org/10.1016/j.sab.2019.03.010 https://doi.org/10.1080/05704928.2017.1352509 25. M. Martinez and M. Baudelet, Anal. Bioanal. Chem., 2020, 412, 4. M. Martinez and M. Baudelet, Anal. Bioanal. Chem., 2020, 412, 27-36. https://doi.org/10.1007/s00216-019-02195-1 27-36. https://doi.org/10.1007/s00216-019-02195-1 26. A. Ciucci, M. Corsi, V. Palleschi, S. Rastelli, A. Salvetti and 5. D. W. Hahn and N. Omenetto, Appl. Spectrosc., 2012, 66, 347-419. E. Tognoni, Appl. Spectrosc., 1999, 53, 960-964. https://doi.org/10.1366/11-06574 https://doi.org/10.1366/0003702991947612 6. Y. Liu, M. Baudelet and M. Richardson, J. Anal. At. Spectrom., 27. E. Grifoni, S. Legnaioli, G. Lorenzetti, S. Pagnotta, F. Poggialini, 2010, 25, 1316-1323. https://doi.org/10.1039/C003304A and V. Palleschi, Spectrochim. Acta B., 2016, 124, 40-46. https://doi.org/10.1016/j.sab.2016.08.022 7. F. Brech and L. Cross, Appl. Spectrosc., 1962, 16, 59-64. 28. G. Cavalcanti, D. Teixeira, S. Legnaioli, G. Lorenzetti, L. Pardini 8. W. D. Zhou, Y. H. Guo and R. R. Zhang, Front. Phys., 2020, 15, and V. Palleschi, Spectrochim. Act. B., 2013, 87, 51-56. 52201. https://doi.org/10.1007/s11467-020-0969-1 https://doi.org/10.1016/j.sab.2013.05.016 9. J. Peng, F. Liu, F. Zhou, K. Song, C. Zhang, L. Ye and Y. He, 29. R. Gaudiuso, M. Dell'Aglio, O. De Pascale, A. Santagata and Trends Anal. Chem., 2016, 85, 260-272. A. De Giacomo, Spectrochim. Acta B., 2012, 74-75, 38-45. https://doi.org/10.1016/j.trac.2016.08.015 https://doi.org/10.1016/j.sab.2012.06.034 10. A. A. Bol’shakov, X. Mao and R. E. Russo, J. Anal. At. Spectrom., 30. C. Aragón and J. A. Aguilera, J. Quant. Spectrosc. Radiat. Transf., 2017, 32, 657-670. https://doi.org/10.1039/C6JA00436A 2014, 149, 90-102. https://doi.org/10.1016/j.jqsrt.2014.07.026 11. L. I. Kexue, W. Zhou, Q. Shen, J. Shao and H. Qian, 31. R. Yuan, Y. Tang, Z. Zhu, Z. Hao, J. Li, H. Yu, Y. Yu, L. Guo, Spectrochim. Acta. B., 2010, 65, 420-424. X. Zeng and Y. Lu, Anal. Chim. Acta., 2019, 1064, 11-16. https://doi.org/10.1016/j.sab.2010.04.006 https://doi.org/10.1016/j.aca.2019.02.056 12. M. V. Belkov, V. S. Burakov; A. De Giacomo, V. V. Kiris, S. N. 32. Z. Q. Hao, L. Liu, R. Zhou, Y. W. Ma, X. Y. Li, L. B. Guo, Y. F. Lu Raikov, and N. V. Tarasenko, Spectrochim. Acta. B., 2009, 64, and X. Y. Zeng, Opt. Express, 2018, 26, 22926-22933. 899-904. https://doi.org/10.1016/j.sab.2009.07.019 https://doi.org/10.1364/OE.26.022926 13. K. Li, W. Zhou, Q. Shen, Z. Ren, and B. Peng, 33. L. C. Nunes, F. R. P. Rocha and F. J. Krug, J. Anal. At. Spectrom., J. Anal. At. Spectrom., 2010, 25, 1475-1481. 2019, 34, 2314-2324. https://doi.org/10.1039/C9JA00270G https://doi.org/10.1039/B922187E 34. O. A. Nassef and H. E. Elsayed-Ali, Spectrochim Acta. B., 2005, 60, 14. V. S. Burakov, S. N. Raikov, N. V. Tarasenko, M. V. Belkov and 1564-1572. https://doi.org/10.1016/j.sab.2005.10.010 V. V. Kiris, J. Appl. Spectrosc., 2010, 77, 595-608. https://doi.org/10.1007/s10812-010-9374-9 35. K. Li, W. Zhou, Q. Shen, Z. Ren, B. Peng, L. I. Kexue, W. Zhou, Q. Shen, Z. Ren and B. Peng, J. Anal. At. Spectrom., 2010, 25, 15. X. F. Li, W. Zhou, K. Li, H. Qian, R. Huiguo and R. Zhijun, 1475-1481. https://doi.org/10.1039/b922187e Opt. Comm., 2012, 285, 54-58. https://doi.org/10.1016/j.optcom.2011.08.074 36. W. Zhou, X. Su, H. Qian, K. Li, X. Li, Y. Yu and Z. Ren, J. Anal. At. Spectrom., 2013, 28, 702-710. 16. W. Zhou, K. Li, X. Li, H. Qian, J. Shao, X. Fang, P. Xie and W. Liu, https://doi.org/10.1039/c3ja30355a Opt. Lett., 2011, 36, 2961-2963. https://doi.org/10.1364/OL.36.002961 37. Y. Jiang, R. Li and Y. Chen, J. Anal. At. Spectrom., 2019, 34, 1838-1845. https://doi.org/10.1039/C9JA00169G 17. X. F. Li, W. D. Zhou and Z. F. Cui, Front. Phys., 2012, 7, 721-727. https://doi.org/10.1007/s11467-012-0254-z 38. A. L. Vieira, M. G. Nespeca, W. D. Pavini, E. C. Ferreira and J. A. Gomes Neto, Chemom. Intell. Lab. Syst., 2019, 194, 103816. 18. A. L. Vieira, T. V. Silva, F. S. I. de Sousa, G. S. Senesi, D. S. Júnior, https://doi.org/10.1016/j.chemolab.2019.103816 E. C. Ferreira and J. A. Gomes Neto, Microchem. J., 2018, 139, 322-326. https://doi.org/10.1016/j.microc.2018.03.011 39. M. A. Bechlin, F. M. Fortunato, R. M. Da Silva, E. C. Ferreira, and J. A. Gomes Neto, Spectrochim. Acta Part B., 2014, 101, 240-244. 19. M. P. Rodríguez, P. M. Dirchwolf, T. V. Silva, R. N. Villafañe, https://doi.org/10.1016/j.sab.2014.09.012 J. A. Gomes Neto, R. G. Pellerano and E. C. Ferreira, Food Chem., 2020, 331, 127051. https://doi.org/10.1016/j.foodchem.2020.127051 40. J. Miller, and J. C. Miller, Statistics and Chemometrics for Analytical Chemistry. Pearson, Harlow, U.K., 2010. 20. M. G. Nespeca, A. L. Vieira, D. Santos Júnior, J. A. Gomes Neto and E. C. Ferreira, Food Chem., 2020, 311, 125886. 41. NIST Atomic Spectra Database https://doi.org/10.1016/j.foodchem.2019.125886 (https://physics.nist.gov/PhysRefData/ASD/). 21. Z. Hou, Z. Wang, J. Liu, W. Ni and Z. Li, Opt. Express, 2014, 22, 42. Y. Tang, J. Li, Z. Hao, S. Tang, Z. Zhu, L. Guo, X. Li, X. Zeng, 12909-12914. https://doi.org/10.1364/OE.22.012909 J. Duan, and Y. Lu, Opt. Express., 2018, 26, 12121. https://doi.org/10.1364/OE.26.012121 22. S. Sheta, M. S. Afgan, Z. Hou, S. C. Yao, L. Zhang, Z. Li and

www.at-spectrosc.com/as/article/pdf/2020202 24 At. Spectrosc. 2021, 42(1), 18-24 Determination of Elemental Impurities in Iron-nickel-based Superalloys by Glow Discharge Mass Spectrometry

F. F. Hu,a C. H. Wang,a,* J. D. Li,b P. Y. Liu,b H. Liu,a and L. Zhanga a Guobiao (Beijing) Testing & Certification Co., Ltd., Beijing 101407, P.R. China b China United Test & Certification Co., Ltd., Beijing 101407, P.R. China

Received: November 27, 2020; Revised: December 20, 2020; Accepted: December 20, 2020; Available online: December 25, 2020.

DOI: 10.46770/AS.2020.215

ABSTRACT: A method for the determination of impurities in iron-nickel-based superalloys by high-resolution glow discharge mass spectrometry is described. The optimum discharge conditions were investigated to obtain stable discharge and good sensitivity. The interference was separated in high resolution mode of the instrument. The calibration relative sensitivity factors (RSF) for 12 elements were obtained with the matrix-matched certified reference material, IARM Ni909-18. The reference material was analyzed by using the calibration RSFs. The relative error of the measured value was within 16.4% compared with the certified value, and the relative standard deviation (RSD) was less than 7.0%. The 12 elements in the iron-nickel-based superalloy were determined using the calibration RSFs (quantitative analysis) and the standard RSFs (semi-quantitative analysis) in the instrument, including C, N, O and S. The results of the verification methods, such as inductively coupled plasma optical emission spectrometry (ICP-OES), inductively coupled plasma mass spectrometry (ICP-MS), high frequency combustion infrared absorption and inert gas pulse infrared thermal conductivity, were closer to quantitative analysis results. The elements without certified values in the reference material were determined by using the standard RSFs. There was no significant difference between the results of GDMS and the verification by ICP-MS. The validation revealed that the proposed analytical approach achieves reliable results for the rapid determination of several impurities, such as metals as well as non-metals, even the gas elements C, N, O and S.

INTRODUCTION The American Society for Testing and Materials (ASTM), one of the international authoritative standard systems, has developed A superalloy is mainly an iron-, nickel- or cobalt-based metallic testing methods for different elements in superalloys.5-8 These material, which works under the conditions of certain stress factors methods cover the major, minor and trace elements in nickel-based and at high temperatures of above 600 oC.1 Because of its superalloys. However, their deficiency is that it only covers few characterization of having high-temperature resistance, oxidation elements and slightly higher detection limits. The International resistance and corrosion resistance, superalloys are widely used in Organization for Standardization (ISO) methods for the chemical the aerospace and energy industries.2-3 With the increasing composition of superalloys are mostly for single elements.9-11 demand for high-performance aeroengine and turbine designs, the Obviously, the detection efficiency is low. Most of the Chinese performance of superalloys is also expected to be higher. Elements standard analysis methods for superalloy compositions are and their concentrations in superalloys influence their application chemical analysis methods, but only a few instrumental methods 4 and performance. Therefore, the accurate determination of trace are developed.12-15 The pretreatment of classical chemical methods elements in superalloys is very important. However, the is complex, and the detection efficiency is low. Moreover, some of components of superalloys are complicated and, therefore, the elemental detection methods are based on the analysis of requires their accurate determination. steel.14-15 www.at-spectrosc.com/as/article/pdf/2020215 25 At. Spectrosc. 2021, 42(1), 25-31 The reported methods for the determination of trace elements in Table 1. Elements and Concentrations in Certified Reference Material superalloys include flame atomic absorption spectrometry (IARM Ni909-18) 16 Content Uncertainty Content Uncertainty (FAAS), inductively coupled plasma optical emission Element Element w/% w/% w/% w/% spectrometry (ICP-OES),17 inductively coupled plasma mass Al 0.009 0.002 N 0.0026 0.0003 18-20 spectrometry (ICP-MS), glow discharge mass spectrometry B 0.0013 0.0005 Nb 4.6 0.1 (GDMS), 21-23 etc. The detection efficiency of FAAS is low C 0.006 0.001 Ni 37.7 0.2 because of single element determination and the detection limit of Co 13.1 0.1 O 0.004 0.001 ICP-OES is slightly higher for trace elements. The mass Cr 0.01 0.004 P 0.002 0.001 Cu 0.007 0.004 S 0.0018 0.0004 spectrometric interference in superalloys with a relatively simple Fe 42.4 0.3 Si 0.42 0.01 matrix is slight and therefore, the detection limit of ICP-MS can Mg 0.00012 0.00007 Ta 0.006 0.003 basically meet the detection requirements. However, for Mn 0.03 0.003 Ti 1.62 0.03 superalloys with more complex matrices, ICP-MS cannot meet the X. Yu21-22 and K. Su23 have done some work on the demand. determination of trace elements in superalloys by glow discharge Glow discharge mass spectrometry (GDMS) is a method for the mass spectrometry. They focused on mass spectrometric direct analysis of solid samples. The sample serves as the cathode interference and depth analysis in superalloy, but did not discuss for a glow discharge in inert gas, usually Ar. Atoms are released RSF and quantitative analysis. In this study, the experimental from the sample surface by cathode sputtering and then ionized in conditions were optimized. The standard RSFs of some elements the negative glow region by the Penning ionization and electron in the instrument were corrected by using a matrix-matched iron- impact ionization process. Since the atomization and ionization of nickel-based superalloy certified reference material (IARM a sample in the GD source are carried out in two different regions, Ni909-18), and the trace elements in the sample were analyzed by namely the cathode dark area near the surface of the sample and high resolution glow discharge mass spectrometry, including C, O, the negative glow region near the anode, GDMS has the advantage N, and S. The elements and concentrations in the certified of low matrix effects. Moreover, it is fast, offers high sensitivity, reference material (IARM Ni909-18) are listed in Table 1. For high resolution, and can be used for qualitative or quantitative other elements without definite value in the certified reference analysis of almost all elements in the periodic table. GDMS is material, semi-quantitative results were obtained by using the considered to be the most effective method for the direct standard RSF in the instrument. The methods of ICP-OES, ICP- determination of trace and ultra-trace elements in solid conductive MS, high frequency combustion infrared absorption and inert gas materials24 and is widely used in the analysis of metals, alloys, pulse infrared thermal conductivity were used to verify the results semiconductors, etc.25-32 obtained by GDMS.

In an analysis, the interference affects the accuracy of the results, so mass spectrometric interferences should be fully considered in EXPERIMENTAL GDMS analysis,21-22, 27, 33-35 especially for a complex matrix. The main interferences in GDMS analysis are heterotopic interference, Analysis by GDMS. A high-resolution glow discharge mass polyatomic ion interference and multi-charge ion interference. The spectrometer (Thermo Fisher, USA, Model: Element GD) was interferences can be reduced or eliminated by selecting a non- used for this study, which was equipped with a Grimm-type fast interference isotope, establishing a mathematical correction flow glow discharge ion source. The purity (volume fraction) of equation, adding a collision reaction cell system, using a high- the argon discharge gas is more than 99.9999%. The glow resolution instrument, changing the discharge gas, improving the discharge ion source adopts semiconductor temperature control, purity of the discharge gas, cooling the ion source, and so on. which could be cooled or heated. The samples should have a smooth surface and were loaded onto a sample holder with an The relative sensitivity factor (RSF) is actually a correction inner diameter of 18.8 mm. The sample was sputtered, atomized, factor, which relates the measured relative ion currents to the ionized, and then focused before entering the magnet. After the actual relative concentrations in the analyzed sample. Because of mass separation by the ions’ momentum in the magnetic field, the the low matrix effects as mentioned above, GDMS analysis can be electrostatic analyzer served for an energy separation to achieve considered being almost independent of matrix. When no suitable high resolution capabilities required for the separation of certified reference material can be found, the RSF obtained from polyatomic interference species from the analyte signals. Finally, one matrix reference material can be used to analyze another the ions on their respective mass were counted by the detection 25-27 matrix sample. In recent years, scholars have done more system. The mass analyzer separated the ions according to their 36-37 research on RSF. However, the experimental results show that mass and charge ratio. The Element GD has three detector modes, some elements in some matrices cannot provide satisfactory namely the counting and the analog detectors obtained by a results after being corrected by RSF obtained from different matrix secondary electron multiplier (SEM), and the Faraday detector. 28-29 reference materials. The resolution slit allowed three stage-fixed resolutions, i.e., low www.at-spectrosc.com/as/article/pdf/2020215 25 At. Spectrosc. 2021, 42(1), 25-31 (R=300), medium (R=4000) and high (R=10000) resolution. Table 2. The Optimized Parameters for GDMS Instrument control and data acquisition were handled by Element Parameter Value GD software. The peak jump mode was chosen in data acquisition. Discharge voltage (V) 1200 A brass block was used for mass calibration of the instrument over Discharge current (mA) 30 Discharge gas flow rate (mL/min) 430 the entire mass range. The detectors' calibration was done on a Peltier (℃) 15 day-to-day basis by a tantalum block. The measurements were Focus (V) -1026 carried out with a mass step of 0.01 amu and 120 number of data X-Deflection (V) 1.93 points per peak. The standard RSF (also called typical RSF), Y-Deflection (V) 0.23 Mass resolution 4000, 10,000 derived from a variety of metals and alloys, is built into the instrument software. (Agilent 7700, USA) to determine Li, Be, B, Na, Mg, Al, P, K, Ca, Sc, V, Cr, Cu, Zn, Ga, Ge, As, Se, Rb, Sr, Y, Zr, Mo, Ru, Rh, Pd, Sample preparation and measurement details. First, the Ag, Cd, In, Sn, Sb, Te, Ba, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, grease on the surface of the iron-nickel-based superalloy was Ho, Er, Tm, Yb, Lu, Hf, Ta, W, Re, Os, Ir, Pt, Au, Hg, Tl, Pb and cleaned with ethanol, then cleaned with deionized water. Dilute Bi. These elements were determined by external calibration using aqua regia was used to corrode the surface of the superalloy standard solutions (1 mg/mL, National Center for Analysis and sample to remove possible contamination, then cleaned with Testing of Non-ferrous Metals and Electronic Materials, P.R. deionized water. The sample was then dried with nitrogen and China). subsequently loaded into the GDMS system using a flat sample holder. The discharge parameters were optimized in constant Determination of C and S by high frequency combustion current mode to obtain enough and stable counts per second for infrared absorption method. The concentrations of C and S were 56Fe+ and 58Ni+. The optimized discharge conditions of the GDMS measured using a high frequency infrared carbon and sulfur for the iron and nickel matrix signal are listed in Table 2. The analyzer (LECO CS-844, USA). The flow velocity of oxygen superalloy was pre-sputtered for 10 minutes to remove the trace [purity≥99.99% (volume fraction)] was 3 L/min. The power of residues on the surface using a discharge current of 30 mA and a the high frequency furnace was 2.2 kW, and the frequency was 18 discharge gas flow rate of 430 mL/min. The main elements, such MHz. The purge time of the high frequency infrared carbon sulfur as Fe and Ni, were detected by Faraday detector, the impurity analyzer was 10 ~ 15 s, the delay time 20 s, the analysis time elements were determined by SEM detector. Four repetitive 40~50 s, and the comparator level 1~3. An amount of (0.500 ± measurements were recorded for all isotopes. 0.005) g iron-nickel-based superalloy sample was weighed into the crucible, and 0.70 g pure iron and 1.50 g tungsten tin flux were Determination of impurities by ICP-OES. A 0.10 g amount added to determine the C and S concentrations. (accurate to 0.0001 g) of iron-nickel-based superalloy sample was weighed into a 250 mL polytetrafluoroethylene beaker, then 3 mL Determination of O and N by inert gas pulse infrared hydrochloric acid, 1 mL nitric acid and 1 mL hydrofluoric acid, thermal conductivity method. The content of O and N was heated to 220 ℃ on an electric hot plate to dissolve the sample measured by using an oxygen and nitrogen analyzer (LECO completely, then the beaker was removed and left standing to cool. ONH836, USA). The purity of helium, the carrier gas, was more The contents were transferred into a 100 mL volumetric flask, then than 99.99% (mass fraction). The power of the oxygen and diluted to the mark with deionized water, and mixed. The sample nitrogen analyzer was set at 5.0 kW. An amount of 0.500 g iron- solution was nebulized into the inductively coupled plasma optical nickel-based superalloy sample was accurately weighed and put emission spectrometer (Agilent 725 ES, USA) to determine Mn into the nickel basket for analysis. by the matrix matching method. The pure metals [Fe, Ni Co, Ti, and Nb, purity≥99.99% (mass fraction)] were added with the same mass as the main component of the superalloy, dissolved, RESULTS AND DISCUSSION and the Mn standard stock solution (1 mg/mL) added after Isotope selection and mass spectrometric interference. The appropriate dilution (National Center for Analysis and Testing of isotope selection of elements to be measured was based on the Non-ferrous Metals and Electronic Materials, P.R. China). principle of high abundance and less interference. In GDMS Determination of impurities by ICP-MS. The iron-nickel-based analysis, the mass spectrum interference mainly came from the superalloy sample (0.1 g, accurate to 0.0001 g) was dissolved into discharge gas and the matrix elements. For the elements that may a 250 mL polytetrafluoroethylene beaker by adding 3 mL interfered, the isotopes without interference were selected or hydrochloric acid, 1 mL nitric acid and 1 mL hydrofluoric acid, determined in the appropriate resolution mode. The medium then heating. The solution was transferred to a 100 mL volumetric resolution of the Element GD was 4000 and the high resolution flask, Cs internal standard solution was added, then diluted to the 10,000, which could separate most of the interferences. The mark with deionized water, and mixed. The sample solution was elements to be measured, the mass, interference ions, the nebulized into the inductively coupled plasma mass spectrometer resolutions required and resolution mode are listed in Table 3. For www.at-spectrosc.com/as/article/pdf/2020215 25 At. Spectrosc. 2021, 42(1), 25-31 Table 3. Isotope, Mass Interference and Resolution flow rate were adjusted to obtain stable discharge and enough Resolution Medium or sensitivity. Element Mass Interference required high resolution K 39 38ArH+ 5689 High resolution The argon flow rate was set at 410 mL/min. The signal Ge 72 56Fe16O+ 9257 High resolution intensities of Fe and Ni were investigated with a discharge current 57Fe18O+, As 75 5781, 5817 High resolution of 25 mA, 27 mA, 29 mA, 30 mA and 31 mA. With an increase in 58Ni17O+ Br 79 40Ar38ArH+ 5406 High resolution discharge current, the atom sputtering rate from the sample 64Ni18O+, increased, along with the signal intensity. The results are shown in Se 82 7862, 3455 High resolution 40Ar40ArHH+ Fig. 1. When the current was increased to 30 mA, the increasing 60 40 + Mo 100 Ni O 6985 High resolution trend of the signal intensity decreased. Moreover, the discharge Ru 101 61Ni40O+ 8310 High resolution Medium time was shortened because the sample deposition caused a short Cs 133 93Nb40Ar+ 3623 resolution circuit. Therefore, the optimum discharge current for superalloy was set at 30 mA.

The discharge current was set at 30 mA. The signal intensities of Fe and Ni were investigated when the argon flow rate was 410 mL/min, 420 mL/min, 430 mL/min and 440 mL/min. The argon flow rate determined the number of argon molecules involved in the collision and the ionization per unit time. When the discharge current was fixed, the argon flow rate increased and the discharge voltage decreased, which means that the sputtering rate decreased. The more argon molecules in unit time, the higher the ionization efficiency. When the ionization efficiency was greater than the sputtering rate, the signal strength increased. When the ionization Fig. 1 Effect of discharge current on Fe and Ni signal intensity. efficiency was lower than the sputtering rate, the signal strength decreased. The results are shown in Fig. 2. When the argon flow rate reached 430 mL/min, the signal intensity of Fe decreased.

Therefore, the optimum argon flow rate for the iron-nickel-based superalloy was at 430 mL/min. The optimized discharge conditions of the GDMS are listed in Table 2.

Quantitative analysis of GDMS. In order to obtain quantitative

results, a concentration dependence of element-specific and matrix-specific sensitivity factors (named relative sensitivity factors, RSFs) has to be known. The element concentration in the

matrix can be expressed as: C I /A X =RSF × X X (Eq. 1) C (X/M) I /A M M M Fig. 2 Effect of argon flow rate on Fe and Ni signal intensity. where CX and CM represent the concentration of the analyte and

the matrix; RSF(X/M) is the relative sensitivity of element X in the Table 4. Comparison of the contnt of C, N, O, and S in Background, matrix; IX and IM are the ion beam signals of the element X and the Reference Material and Superalloy Sample matrix, respectively; and AX and AM are the isotope abundances of Backgroud Reference Sample Element element X and the matrix, respectively. The concentrations of C, w/(μg/g) w/(μg/g) w/(μg/g) N, O and S in 6N pure copper were detected as the background by C 25 1254 3096 N 30 1003 433 GDMS sputtering for a long time until stable results were obtained. O 100 1963 2116 Under the optimized discharge conditions above, the S 0.1 20.3 7.1 concentrations of C, N, O and S in the certified reference material the elements without interference, the most abundant isotopes (IARM Ni909-18) and the iron-nickel-based superalloy sample were selected for determination. were detected calibrating by standard RSF. The results, listed in Table 4, show that the background of C, N, O and S was 10 to 100 Optimization of discharge conditions. Discharge current, times lower than that of the certified reference material and the discharge voltage and discharge gas flow rate are the three main sample. The detection of these four gas elements was effective. discharge parameters in GDMS analysis. One is fixed and the other two are linked.38 In this study, the maximum discharge Under the optimized discharge conditions above, the RSF of the voltage was set at 1200 V, and the discharge current and the argon 12 elements in the iron-nickel-based superalloy was obtained for www.at-spectrosc.com/as/article/pdf/2020215 25 At. Spectrosc. 2021, 42(1), 25-31 Table 5. Comparison of Standard RSF and Calibration RSF Fig. 3. It should be noted that the calibration RSF values of C, N, Calib. Calib. Element Std. RSF Elements Std. RSF O and Mg were obviously different from those of the standard RSF. RSF RSF Al 1.27 2.01 Mn 1.01 1.42 Under the same discharge conditions, the calibration RSF was B 6.49 5.02 N 1 0.03 used for four independent determinations of the reference material. C 9.27 0.5 O 1 0.02 The comparison between the measured values and the certified Cr 1.28 2.42 P 3.66 8.71 Cu 2.44 6.01 S 3.43 4.01 values are shown in Table 6. The relative standard deviation (RSD) Mg 1.51 18.12 Ta 1.24 2.77 of four independent determinations was less than 7.0%, and the relative error was within ± 16.4%. These results proved that the method is stable and accurate.

Under the same discharge conditions, the standard RSF (semi- quantitative) in the instrument and the calibration RSF (quantitative) were used for four independent determinations of an

iron-nickel-based superalloy. The methods of ICP-OES, ICP-MS, high frequency combustion infrared absorption (for C, S) and inert gas pulse infrared thermal conductivity (for O, N) were used to

verify the results of GDMS. The results of the GDMS quantitative analysis were closer to the four-verification method results than Fig. 3 Comparison of the standard RSF and the calibration RSF. that of the semi-quantitative analysis, as shown in Table 7. However, there was no order of magnitude difference between the analysis results of semi-quantitative and quantitative analysis, Table 6. Comparison of Measured Values and the Certified Values of except for C, N, O, and Mg. The big difference between the the Reference Material analysis results of semi-quantitative and quantitative for N, O, Mg Average RSD Certified Relative Error Element was that the calibration RSF was obviously different from the w/(μg/g) (n=4) /% w/(μg/g) w/(μg/g) Al 92 0.5 90 2.25 standard RSF in the instrument, as discussed above. The RSD of B 13.2 1.1 13 1.84 four independent determinations was between 0.5%~12.7%, C 59.8 0.8 60 -0.3 which showed good precision. Cr 101.6 0.4 100 1.59 Cu 69.4 0.8 70 -0.88 Semi-quantitative analysis for other elements. Semi- Mg 1 6.9 1.2 -16.4 quantitative determination for other elements without definite Mn 304.3 0.3 300 1.42 value in the iron-nickel-based superalloy reference material was N 25.5 1 26 -1.89 O 40.4 0.8 40 1.03 carried out by using standard RSF in the instrument. The results of P 20.3 1.3 20 1.7 the GDMS semi-quantitative analysis were verified by ICP-MS, S 17.6 2.3 18 -2.28 as shown in Table 8. Because the detection limit of ICP-MS was Ta 60.6 0.7 60 1.08 not as low as of GDMS, the results of many elements in the ICP- MS analysis were only given as below detection limit. Generally Table 7. Comparison of GDMS Quantitative Analysis, Semi- speaking, there were no significant differences between the results quantitative Analysis and Verification Test Results for 12 Elements of GDMS and ICP-MS. Some element results of GDMS and ICP- RSD Semi- ICP- Quant. C, S O, N MS are even close, such as for Sb. Element (n=4) quant. MS/OES w/(μg/g) w/(μg/g) w/(μg/g) /% w/(μg/g) w/(μg/g) Al 358 0.8 235 327 / / B 110 0.5 154 106 / / CONCLUSIONS C 160 0.5 3096 / 167 / Cr 42 0.8 22 45 / / A high-resolution glow discharge mass spectrometric method for Cu 30 1.3 12 32 / / the quantitative and semi-quantitative analysis of trace elements in Mg 0.52 12.7 0.046 0.46 / / iron-nickel-based superalloy was successfully developed. The Mn 5587 0.2 3977 5182 / / method involves use of a matrix-matched certified reference N 13 3.7 433 / / 10 O 37 2 2116 / / 41 material, IARM Ni909-18. Quantitative analysis was carried out P 13 0.6 6.1 11 / / by using the calibration RSFs, which were obtained from 12 S 8.5 0.7 7.1 / 8 / elements in the certified reference material. A comparison of the Ta 14 1.8 6.1 13 / / results obtained with GDMS, ICP-OES, ICP-MS, high frequency the matrix-matched certified reference material (IARM Ni909- combustion infrared absorption and inert gas pulse infrared 18). The comparison between the calibration RSF value and the thermal conductivity shows good agreement between the methods. standard RSF value in the instrument is shown in Table 5 and Semi-quantitative analysis was carried out by using standard RSF www.at-spectrosc.com/as/article/pdf/2020215 25 At. Spectrosc. 2021, 42(1), 25-31 Table 8. Comparison of GDMS Semi-quantitative Analysis and ICP- superalloys. Beijing, Metallurgical Industry Press, 1987. MS Verification Test Results for Additional 52 Elements 5. ASTM E1473-16, Standard test methods for chemical analysis of Semi- ICP- Semi- ICP- nickel, cobalt, and high-temperature alloys. Element quant. MS Element quant. MS 6. ASTM E2594-14, Standard test method for analysis of nickel alloys w/(μg/g) w/(μg/g) w/(μg/g) w/(μg/g) by inductively coupled plasma atomic emission spectrometry. Li <0.05 <1 Cs <0.05 <0.5 7. ASTM E 1834-11, Standard test method for analysis of nickel alloys Be <0.05 <0.5 Ba <0.05 <0.5 by graphite furnace atomic absorption spectrometry. Na 0.12 <3 La <0.05 <0.5 K 0.2 <5 Ce <0.05 <0.5 8. ASTM E2823-11, Standard test method for analysis of nickel alloys Ca <0.05 <1 Pr <0.05 <0.5 by inductively coupled plasma mass spectrometry. Sc <0.05 <1 Nd <0.05 <0.5 9. ISO 9388:2009, Nickel alloys - determination of phosphorus content V 0.83 <1 Sm <0.05 <0.5 - molybdenum blue molecular absorption spectrometric method. Zn 0.17 <1 Eu <0.05 <0.5 10. ISO 11437-2:2010, Nickel alloys - determination of trace element Ga 7.35 6.3 Gd <0.05 <0.5 content by electrothermal atomic absorption spectrometric method - Ge 2.13 7.8 Tb <0.05 <0.5 part 2: determination of lead content. As 8.64 11 Dy <0.05 <0.5 Se 1.78 1.6 Ho <0.05 <0.5 11. ISO 22725:2011, Nickel alloys - determination of tantalum - induc- Rb <0.05 <0.5 Er <0.05 <0.5 tively coupled plasma atomic emission spectrometric method. Sr <0.5 <0.5 Tm <0.05 <0.5 12. HB 5220. National Defense Science, Technology and Industry Y <0.5 <0.5 Yb <0.05 <0.5 Commission of P. R. China. 2008. Zr 0.5 0.77 Lu <0.05 <0.5 13. GJB 8781. Commission of Science, Technology and Industry for Mo 2.29 3.4 Hf <0.05 <0.5 National Defence of the P. R. China. 2015 Ru <0.05 <0.5 W 1.31 5 14. GB/T 20127. National Standardization Administration of P. R. Rh <0.05 <0.5 Re <0.05 <0.5 China. 2006. Pd <0.05 <0.5 Ir <0.05 <0.5 Ag 4.72 1.8 Pt <0.05 <0.5 15. GB/T 223. National Standardization Administration of P. R. China. Cd 0.36 <0.5 Au <0.05 <0.5 2019. In <0.05 <0.5 Hg <0.05 <1 16. T. Lv and Y. Guo, Chin. J. Inorg. Anal. Chem., 2020, 10(2), 11-14. Sn 5.11 9.1 Tl <0.05 <0.5 https://doi.org/10.3969/j.issn.2095-1035.2020.02.003 Sb 1.99 2 Pb <0.05 <0.5 17. F. Luo, Chin. J. Inorg. Anal. Chem., 2020, 10(4), 59-62. Te <0.05 <0.5 Bi <0.05 <1 https://doi.org/10.3969/j.issn.2095-1035.2020.04.013 in the instrument. No significant differences between the results of 18. S. M. Mo, A. C. Li, X. Zhang, C. D. Qiu, C. B. Dong, and L. H. Zhu, Chem. Anal. Meter., 2020, 29(3), 129-132. GDMS and ICP-MS were found. The validation reveals that the https://doi.org/10.3969/j.issn.1008-6145.2020.03.028 proposed approach achieves reliable results for the rapid 19. G. W. Yang, Y. X. Hou, Q. B. Liu, and X. J. Li, Metall. Anal., 2019, determination of several impurities, such as metals as well as non- 39(4), 1-6. metals, and even for the gas elements of C, N, O and S. https://doi.org/10.13228/j.boyuan.issn1000-7571.010600 20. D. Na, Y. Sun, M. Wang, and H. Li, Chin. J. Inorg. Anal. Chem., 2019, 9(6), 54-58. AUTHOR INFORMATION https://doi.org/10.3969/j.issn.2095-1035.2019.06.012 21. X. Yu, X. J. Li, and H. Z. Wang, Metall. Anal., 2011, 31(11), 1-6. Corresponding Author https://doi.org/10.13228/j.issn.1000-757.2011.11.004 22. X. Yu, X. J. Li, H. Z. Wang, Metall. Anal., 2007, 27(4), 11-17. . *C. H. Wang https://doi.org/10.13228/j.issn.1000-757.2007.04.002 Email address: [email protected] 23. K. Su, X. W. Wang, and K. Putyera, Metall. Anal., 2011, 31(11), 18-23. Notes https://doi.org/10.13228/j.issn.1000-757.2011.11.006 24. C. D.Quarles Jr, J. Castro, and R. K. Marcus, Encyclopedia of The authors declare no competing financial interest. Spectroscopy and Spectrometry (Third Edition), 2017, p30-36 https://doi.org/10.1016/B978-0-12-374413-5.00056-7 25. W. Vieth and J. C. Huneke, Anal. Chem., 1992, 64, 2958-2964. REFERENCES https://doi.org/10.1021/ac00047a014 26. S. Raparthi, J. Arunachalam, N. Das, and A. M. S. Murthy, Talanta, 1. G. L. Chen, Superalloys. Beijing, Metallurgical Industry Press, 2005, 65, 1270-1278. 1988. https://doi.org/10.1016/j.talanta.2004.09.001 2. R. C. Reed, The superalloys: Fundamentals and applications. 27. R. Shekhar, M. A. Reddy, S. Thangavel, Y. Sunitha, A. C. Sahayam, Cambridge Univ. Press, 2008. and S. Kumar, At. Spectrosc., 2020, 41(3), 103-109. https://doi.org/10.46770/AS.2020.03.002 3. Y. T. Chen, Y. J. Chang, H. Murakami, S. Gorsse, and A. H. Yeh, Scr. Mater., 2020, 187, 177-182. 28. M. D. Sabatino, A. L. Dons, J. Hinrichs, and L. Arnberg, https://doi.org/10.1016/j.scriptamat.2020.06.002 Spectrochim. Acta B, 2011, 66, 144-148. https://doi.org/10.1016/j.sab.2011.01.004 4. Z. C. Xu and P. L. Ma, Effect and control of trace elements in 29. G. Chen, A. J. Ge, S. J. Zhuo, and P. L. Wang, www.at-spectrosc.com/as/article/pdf/2020215 25 At. Spectrosc. 2021, 42(1), 25-31 J. Chin. Mass Spectro. Soc., 2007, 28(1), 36-39. J. Chin. Mass Spectrom. Soc., 2006, 27(Supplement), 25-26. https://doi.org/10.3969/j.issn.1004-2997.2007.01.008 35. X. Yu, X. J. Li, and H. Z. Wang, Int J Mass Spectrom, 2007, 262, 30. Y. X. Hou, X. B. Liu, G. W. Yang, X. J. Li, and J. Y. Hu, 25-32. PTCA (Part B: Chem. Anal.), 2020, 56(3), 315-319. https://doi.org/10.1016/j.ijms.2006.10.001 https://doi.org/10.11973/lhjy-hx202003011 36. X. J. Wei, Z. Qin, P. H. Xiong, L. P. Wang, H. L. Zhang, D. C. Deng, 31. M. D. Sabatino, Measurement, 2014, 50, 135-140. and J. S. Liao, Spectrochim. Acta B, 2019, 154, 43-49. https://doi.org/10.1016/j.measurement.2013.12.024 https://doi.org/10.1016/j.sab.2019.01.004 32. C. Modanese, L. Arnberg, and M. D. Sabatino, Mater. Sci. Eng. B, 37. A. Bogaerts, K. A. Temelkov, N. K. Vuchkov, and R. Gijbels, 2014, 180, 27-32. Spectrochim. Acta B, 2007, 62, 325-336. https://doi.org/10.1016/j.mseb.2013.10.010 https://doi.org/10.1016/j.sab.2007.03.010 33. X. Yu, X. J. Li, and H. Z. Wang, PTCA (Part B: Chem. Anal.), 2010, 38. K. Wagatsuma , T. Saka, M. Yamaguchi, and K. Ito, J. Anal. Atom. 46(2), 206-210. Spectrom., 2002, 17(10), 1359-1362. https://doi.org/CNKI:SUN:LHJH.0.2010-02-041 https://doi.org/10.1039/b204747k 34. B. L. Rong, Y. Zhen, L. B. Tang, and R. B. Ji,

www.at-spectrosc.com/as/article/pdf/2020215 25 At. Spectrosc. 2021, 42(1), 25-31 Leaching of Gallium from Coal Fly Ash, Alumina and Sediment Samples with an Acid Mixture for its Determination by ICP-OES

A.C. Sahayam,a,b,* G. Venkateswarlu,a and S. Thangavela a National Centre for Compositional Characterisation of Materials (NCCCM), Bhabha Atomic Research Centre (BARC), ECIL (P.O.), Hyderabad 500062, India b Homi Bhabha National Institute (HBNI), Mumbai, India

Received: September 25, 2020; Revised: October 24, 2020; Accepted: October 24, 2020; Available online: October 29, 2020

DOI: 10.46770/AS.2020.183

ABSTRACT: Leaching of gallium (Ga) from coal fly ash, alumina and sediment is reported for its determination by inductively coupled plasma optical emission spectrometry (ICP-OES). A mixture of acids (H2SO4, HNO3 and HF) was used for the leaching process which was carried out overnight at room temperature or heated for 1 h on a hot plate. HF is essential for the

quantitative recovery of Ga from coal fly ash. Leaching with either HCl or HNO3 in combination with HF also yielded quantitative

recoveries of Ga from coal fly ash, but not from other matrices. After leaching, HF and HNO3 were evaporated and HCl was added to the acid leach before dilution for analysis. When mixed acid leaching was employed, leaching of Ga was 92-102%. For accuracy, the method was applied to a sediment sample, a candidate material prepared as the in-house CRM, and the values obtained were well within the uncertainty reported. The present procedure was validated using the solvent extraction method. The values by the two methods were in close agreement according to the Student’s t-test with a confidence level of 95%. The limit of detection was 0.45 mg Kg-1. The analytical methodology was applied for the determination of Ga in three coal fly ash samples, two from thermal power plants and a NIST CRM 1633b (not certified for Ga) as well as an alumina sample from an aluminum industry facility. A red mud sample, by-product of Bayer’s process of aluminum extraction, was also analyzed using

HCl/HF and HNO3/HF leach.

INTRODUCTION inorganic constituents of varying compositions depending on the grade of coal. Production of coal fly ash is very high, considering Gallium (Ga) is a chemical element that is relatively rare and has that 40% of the electrical power is generated through combustion been used largely in the electronics industry. Ga does not have its of coal globally.5 India is among the top three countries in the own ore and in nature occurs in trace amounts as Ga(III) world that utilizes coal-fired power plants.6 The majority of coal compounds, principally in bauxite (aluminum matrix) and zinc fly ash is unutilized and stored in ash ponds and surface impounds, ores.1,2 Due to the limited natural resources and the growing increasing the threat of contamination to both underground water demand, there may be global competition for the production of Ga and also air. Constituents of coal fly ash can be extracted based on and hence, alternate resources need to be evaluated for economic their utility and also commercial value. Ga is one such constituent reasons.3 Ga can also be found in coal, with approximately 10 and due to its efficient optical transitions and also electron mobility, million tons of Ga contained in the coal resources on earth.4 The it is recognized as a very important component in the preparation content of Ga in coal varies regionally. Coal fly ash is a solid of compound semiconductors used in photovoltaic cells.7 Prior to residue obtained after combustion of the coal in thermal power engaging in large-scale recovery, the coal fly ash samples need to plants. This by-product is toxic due to the presence of organic and be screened for their Ga content to establish the commercial value www.at-spectrosc.com/as/article/pdf/2020183 32 At. Spectrosc. 2021, 42(1), 32-35 for extraction. A simple analytical methodology is needed to carry a red mud sample. out such screening. Another important source of Ga is bauxite, an aluminum ore. Gallium is extracted from the by-product of aluminum extraction process and is present at the mg kg-1 level EXPERIMENTAL both in coal fly ash and in alumina. Hence, a preliminary screening Instruments procedure applicable to both types of samples would help in selecting a suitable lot for extraction that yields economic benefits. An inductively coupled plasma optical emission spectrometer (ICP-OES), Model ULTIMA 2, Horiba Jobin Yvon, France, Most of the recent papers on coal fly ash are dedicated to protect equipped with a concentric pneumatic nebulizer and cyclonic the environment from toxic constituents or the bulk recovery of spray chamber, was used for all measurements. The spectrometer the precious metal from coal fly ash.8,9 Zhao et. al.7 reported is equipped with a Czerny-Turner monochromator using a 2400 sulfuric acid leaching of Ga from coal fly ash for the subsequent grooves/mm holographic grating. The wavelength (294.364 nm) sequential recovery of Ga using P507 and Cyanex 272. Gutiérrez selected for Ga was based on sensitivity and lack of spectral et. al.10 reported leaching of Ga with HCl for sequential extraction interference. using Amberlite LA-2 and LIX 54N. Different acid extractions are carried out with the ultimate aim of bulk extraction using foam, Reagents and samples Cyanex, phosphate-based extractions, etc. Analytical reagent grade chemicals were used throughout. HCl, Gallium is determined by instrumental techniques such as HNO3, H2SO4 and HF were procured from Merck, India. Milli-Q atomic absorption spectrometry (AAS), inductively coupled water (Millipore Corporation, USA) with 18MΩ cm resistivity plasma optical emission spectrometry (ICP-OES) and inductively was used for the preparation of all solutions. The stock standard coupled plasma mass spectrometry (ICP-MS). Direct solution (1000 mg L-1) of Ga was procured from Merck, Germany. determination of Ga by these techniques is hampered by the matrix The coal fly ash samples were obtained from thermal power elements that affect recovery and precision, and also the restriction plants. The alumina and red mud samples were obtained from a on total dissolved salts in the sample solution for analysis by ICP. local aluminum producer. A sediment was used as a candidate for Slurry sampling electrothermal-AAS is the simplest method for reference material, and the value reported is recommended by the the determination of Ga in coal fly ash and does not require any producer by adopting a different procedure. All samples were sample treatment. However, the relative standard deviations (RSD) processed as received. reported are higher despite the fact that quantifications are carried out against aqueous standard calibration since there is no matrix Analytical procedure 11 effect. Solvent extractions are extensively used for the extraction An accurately weighed sample (0.25 g) was taken into a PFA - of Ga from 6M to 9M HCl solution since Ga forms anionic GaCl4 container with screw cap. Added were 1 mL each of H2SO4, HNO3 12-14 at higher acidities. It requires total dissolution of the sample and 0.5 mL HF, then left standing overnight or heated on a hot and also uses toxic organic solvents, such as isopropyl ether, plate at about 100 C for 1 h in closed conditions. After opening diethyl ether, acetylacetone, 5, 4-methylpentan-2-ol, tributyl the screw cap, the acids (HF and HNO3) were evaporated on a hot phosphate, etc. plate. The same can also be carried out on a platinum crucible. In Recently, the ultrasonic assisted dispersive liquid-liquid such case, it can be heated until the appearance of thick H2SO4 microextraction method was developed for the solvent extraction fumes. After cooling, 1 mL HCl was added to the residue, heated of Ga for its determination in vanadium titanium magnetite.15 for about 5 min and diluted with Milli-Q water to 25 mL along Similarly, the determination of Ga in alumina is also important to with the undissolved solids. The same procedure was applied for recover the precious metal. Gallium in Al is determined along with all samples. Blank solutions were also prepared as above, other elements after separation of aluminum from alumina excluding the sample. For only the coal fly ash and red mud solution by precipitation for its determination by ICP-OES.16 samples, a combination of either HCl/HF or HNO3/HF was used

Multi-stage leaching steps are reported using HNO3/HF/HClO4 for the same quantity of sample, then following the same analytical and Zheng Fang reported an HCl leach method with fewer procedure as above. 7,17 recoveries. A simpler procedure with easily adaptable steps and Complete dissolution of coal fly ash and sediment applicable to both CFA and alumina is useful for preliminary screening of the different sample lots for the recovery of Ga. To 0.25 g of sample taken into a platinum crucible, 1 mL H2SO4 and 2 mL HF were added and left standing overnight. HF was The aim of the present work was to develop a simple leaching evaporated on a hot plate until H2SO4 fumes appeared. To the method for the preliminary screening of coal fly ash, alumina and residue, 25 mL HCl was added and diluted to 50 mL with Milli-Q other samples for Ga prior to large-scale recovery. Conditions for water which resulted in a clear sample solution. If required, the quantitative recovery and analysis of different coal fly ash additional HCl was added and diluted to a higher volume. In case samples, alumina and a sediment sample are reported, along with of the NIST coal fly ash sample and to ensure the complete www.at-spectrosc.com/as/article/pdf/2020183 33 At. Spectrosc. 2021, 42(1), 32-35 dissolution of the very minute amount of residue left, it was fused Table 1. Amount of HF Required to Leach Ga from 250 mg of Coal Fly with lithium metaborate and dissolved in a minimum amount of Ash (CFA) in Presence of 1 mL each H2SO4 and HNO3 HCl, then was added to the sample solution. Sample Volume of Concentration of Ga in Recovery -1 a Solvent extraction of gallium No. 48% (w/v) HF ( μL) NTPC CFA (mg kg ) (%) 1 100 9.0 60.4 Ga in sample solutions taken in 6M HCl was extracted into diethyl 2 200 12.6 84.5 ether with equal volumes of aqueous and organic layers. Ga 3 300 13.5 90.6 extracted into diethyl ether was stripped back into 0.01M HCl, 4 400 15.1 101.0 5 500 15.2 102.0 evaporated on a water bath to volatilize residual diethyl ether, and 6 600 13.5 90.6 analyzed by ICP-OES. The blank solution was also prepared as aRecovery with respect to value obtained by ICP-OES, 14.9 mg kg-1. above. Table 2. Comparison of Concentrations Obtained by Present and Solvent Extraction Methods RESULTS AND DISCUSSION Sample Leaching method Solvent extraction Gallium does not have its own ore and is present at trace levels in NTPC-CFA 15.3 ± 0.3 14.1 ± 1.1 a bauxite, an aluminum ore. Hence, Ga is extracted from the by- IMMT-CFA 44.4 ± 2.1 45.0 ± 1.2 NIST-CFA 1633b 57.5 ± 1.8 55.2 ± 4.0 a product (obtained at an aluminum industry facility) or from other GSI Sediment (SSS1) 22.4 ± 2.3 24.4 ± 3.9b sources such as coal fly ash where it is present at the tens of mg Alumina 38.0 ± 1.0 39.0 ± 1.0 -1 kg level. Lots containing higher concentrations of Ga are Uncertainties are expressed as standard deviation of three measurements. preferred for better yield. Hence, secondary sources of Ga need to aanalyzed directly; bindependent method. be analyzed for Ga prior to bulk extraction. Direct determination Along with 1 mL each of H2SO4 and HNO3, 500 μL HF was of Ga from these samples after total dissolution is difficult by added to 0.25 g sample for quantitative leaching of Ga. Nitrate instrumental techniques for want of sensitivity and also due to vapors appeared after the addition of HF. All of the acids, except matrix interferences. For the determination of Ga, solvent H2SO4, were evaporated and 1 mL HCl was added since Ga has an extraction is normally adopted to pre-concentrate the analyte, affinity to HCl. The leach and the solid were transferred into a while simultaneously reducing matrix interferences. However, the polypropylene (PP) vial and diluted to 25 mL with Milli-Q water procedure is tedious as it involves total dissolution of the sample, for analysis by ICP-OES. External calibration was used for followed by solvent extraction using toxic organic solvents. quantification. As there is no coal fly ash CRM available and Moreover, residual organic solvents will extinguish the plasma if certified for Ga, a CRM candidate sediment material was analyzed, analysis is carried out by ICP-based techniques. Evaporation of the whose Ga concentration is indicated by the CRM producer, for residual solvent is an additional step. Acid leaching is a better accuracy. The concentration obtained by the present method and alternative to solvent extraction. Though leaching methods with the same reported by the CRM producer was within the different combinations of acids are reported, they either result in uncertainty mentioned (Table 2). fewer recoveries or are used for bulk extraction. Since the method was validated for sediment, a similar silicate Acid Leaching of Gallium matrix, the method was applied to three coal fly ash samples and Initially, leaching of Ga from coal fly ash was reported by using one alumina sample. The alumina sample was selected in order to 7M H2SO4. However, the recovery of Ga was found to be very explore the possibility of applying the method to alumina also. The poor, <5%. Hence, a combination of acids for leaching was concentrations obtained are shown in Table 2. The values obtained attempted with coal fly ash, obtained from the National Thermal on real world samples are validated by using the solvent extraction Power Corporation (NTPC), by oxidizing mineral acids, H2SO4 method after total dissolution. For total dissolution of coal fly ash, and HNO3 for initial attack of the samples, and resulting in very normally NaOH fusion or microwave digestion methods are poor recovery of Ga (30%). Based on previous experience and to adopted to ensure complete dissolution. In this study, a simple increase recovery, a small amount of HF was added.18,19 It was method was adopted using a mixture of acids and a hot plate for found that addition of 100 μL HF improved the recovery of Ga to heating as mentioned in the experimental section. 60% (for 0.25 g of CFA). Hence, HF was optimized ranging from For solvent extraction, sample solutions of coal fly ash and 100 to 600 μL for 0.25 g of coal fly ash sample. The recoveries are alumina taken in 6M HCl were used. A suitable aliquot was taken calculated based on true value of Ga obtained by ICP-OES (14.9 and Ga extracted with an equal volume of diethyl ether. Gallium mg kg-1). As shown in Table 1, the recovery of Ga increased with from the ether layer is back-extracted into 0.01M HCl solution. the HF concentration and maximum recovery was obtained The solution was analyzed for Ga after evaporating the ether, since between 400 to 500 μL. For further studies, the concentration of the presence of an organic solvent will extinguish the plasma 500 μL HF was used. during ICP-OES analysis. The results obtained by solvent extraction www.at-spectrosc.com/as/article/pdf/2020183 34 At. Spectrosc. 2021, 42(1), 32-35 Table 3. Comparison of Concentrations of Ga Obtained by Leaching ACKNOWLEDGMENTS with Different Acids The authors would like to thank Dr. Sanjiv Kumar, Head, NCCCM, Leaching method Total Sample for his support and encouragement. HCl + HF HNO3 + HF dissolution NTPC-CFA 15.6 ± 0.4 15.3 ± 0.3 14.9 ± 0.1 NIST-CFA 1633b 45.7 ± 2.0 49.5 ± 0.7 55.2 ± 4.0 GSI Sediment (SSS1) 15.4 ± 0.8 16.8 ± 0.3 24.4 ± 3.9a REFERENCES Alumina 40.0 ± 2.0 31.5± 0.5 39.0 ± 1.0 1. R. R. Moskalyk, Miner. Eng., 2003, 16, 921-929. Redmud 95.0 ± 2.3 89.6 ± 3.0 98.0 ± 3.0 https://doi: 10.1016/j.mineng.2003.08.003 Uncertainties are expressed as standard deviation of three measurements. 2. F. A. de Santana, J. T.P. Barbosa, G. D. Matos, M. G.A. Korn, and are close to the developed method with a 95% confidence level S. L.C. Ferreira, Microchem. J., 2013, 110, 198-201. according to the Student’s t-test. The results are shown in Table 2. https://dx.doi.org/10.1016/j.microc.2013.03.011 The limit of detection, calculated as the signal equal to three times 3. M. Frenzel, M.P. Ketris, T. Seifert, and J. Gutzmer, Resour. Policy, 2016, 47, 38-50. the standard deviation (six measurements) + blank, was 0.45 mg https:// doi: 10.1016/j.resourpol.2015.11.005 Kg-1. The concentration of Ga in the samples analyzed was in the 4. R.S. Blissett and N.A. Rowson, Fuel, 2012, 97, 1-23. -1 range of 15-98 mg kg . https://doi.org/10.1016/j.fuel.2012.03.024 In order to reduce the number of acids for leaching Ga, a 5. https://en.wikipedia.org/wiki/Electricity_generation 6. https://www.statista.com/statistics/530569/installed-capacity-of- combination of HCl/HF and HNO3/HF was tested. It was found coal-power-plants-in-selected-countries/ that the recoveries are quantitative in case of coal fly ash and red 7. Z. Zhao, L. Cui, Y. Guo, H. Li, and F. Cheng, Chem. Eng. J., 2020, mud. Red mud was considered since the above procedure could 381, 122699. https://doi.org/10.1016/j.cej.2019.122699 not be applied due to the complete dissolution in mixed acids. The 8. R. N. Lieberman, Y. Knop, X. Querol, N. Moreno, C. Mũnoz-Quirós, procedure was further simplified by using either of the acid Y. Mastai, Y. Anker, and H. Cohen, J. Hazard. Mater., 2018, 344, combinations; however, it is suitable only for coal fly ash and red 1043-1056. https://doi: 10.1016/j.jhazmat.2017.11.047 mud (Table 3). 9. A. R. K. Gollakota, V. Volli, and C.-M. Shu, Sci. Total Environ., 2019, 672, 951-989. https://doi.org/10.1016/j.scitotenv.2019.03.337 10. B. Gutiérrez, C. Pazos, and J. Coca, Waste Manag. Res., 1997, 15, CONCLUSIONS 371-382. https://doi.org/10.1177/0734242X9701500405 11. X.-Q. Shan, W. Wang, and B. Wen, J. Anal. At. Spectrom., 1992, 7, A simple method is reported using mixed acid leaching of Ga from 761-764. https://doi.org/10.1039/JA9920700761 coal fly ash, sediment and alumina. The leach solution was 12. N. H. Nachtrieb and R. E. Fryxell, J. Am. Chem. Soc., 1949, 71, analyzed for Ga by ICP-OES. The wide scope of the method, 4035- 4039. https://doi.org/10.1021/ja01180a047 applicable to silicate matrices and also alumina, makes it a 13. S. V. Mahamuni, P. P. Wadgaonkar, and M. A. Anuse, J. Serb. Chem. promising application for preliminary screening of Ga. The Soc., 2010, 75, 1099-1113. https://doi:10.2298/ JSC090630072M accuracy of the method was verified by analyzing sediment 14. I. Mihaylov and P. A. Distin, Hydrometallurgy, 1992, 28, 13-27. samples by an independent laboratory using an independent https://doi.org/10.1016/0304-386X(92)90062-5 method. Validation of the method was carried out by solvent 15. T. Cui, X. Zhu, L. Wu, and X. Tan, Microchem. J., 2020, 157, 104993. https://doi.org/10.1016/j.microc.2020.104993 extraction. The method is further simplified by reducing one acid, 16. A. L. Souza, S. G. Lemos, and P. V. Oliveira, Spectrochim. Acta B, however it is suitable only for coal fly ash and red mud. The 2011, 66, 383-388. https://doi.org/10.1016/j.sab.2011.03.001 proposed method is simple and can be adopted for screening 17. Z. Fang and H. D. Gesser, Hydrometallurgy, 1996, 41, 187-200. samples for Ga content. https://doi.org/10.1016/0304-386X(95)00055-L 18. C.-Y. Kuo, S.-J. Jiang, and A. C. Sahayam, J. Anal. At. Spectrom., 2007, 22, 636-641. https://doi.org/10.1039/B701112A AUTHOR INFORMATION 19. P. Mamatha, G. Venkateswarlu, A. V. N. Swamy, and A. C. Sahayam, Anal. Methods, 2014, 6, 9653-9657. Corresponding Author https://doi.org/10.1039/C4AY01914H * A. C. Sahayam

Email address: [email protected] Notes The authors declare no competing financial interest.

www.at-spectrosc.com/as/article/pdf/2020183 35 At. Spectrosc. 2021, 42(1), 32-35 Influence of Spot Size on LA-ICP-MS Ablation Behavior for Synthetic Calcium Tungstate and Silicate Glass Reference Material NIST SRM 610

Yuting Xiao,a Jian Yang,a Jun Deng,a Wei Wang,a Yuqiu Ke,a,b,* and Yijian Suna,* a Faculty of Materials Metallurgy and Chemistry, Jiangxi University of Science and Technology, Ganzhou 341000, P.R. China b Center of Analysis and Testing, Jiangxi University of Science and Technology, Ganzhou 341000, P.R. China

Received: November 23, 2020; Revised: December 08, 2020; Accepted: December 08, 2020; Available online: December 13, 2020.

DOI: 10.46770/AS.2021.01.006

ABSTRACT: Laser ablation behaviors are typically influenced by the laser operating parameters like the spot size, which has been well studied for silicate glass reference materials but not for samples, such as scheelite, which has the chemical composition

of CaWO4. In this work, the ablation behaviors of synthetic CaWO4 single crystal CaW-1 were studied and compared with those of the well-studied NIST SRM 610 silicate glass reference material. The results showed that LA-ICP-MS analysis of both the CaW-1 and NIST SRM 610 can obtain desired ablation craters and expected signal intensity ratios (R) with a spot size of 60, 44, or 32 μm, while it is not possible with the spot size of 10, 16, or even 90 μm due to the tapered craters or below-expected ablation efficiency/aerosol transport efficiency. Elemental fractionation was found for both CaW-1 and NIST SRM 610 at the small spot size. A spot size of ≥32 μm is preferred for CaW-1, and even for scheelite regardless of pulse number, while a spot size of 16 μm is desirable for NIST SRM 610 with the pulse number of 200, but a spot size of 44 μm is recommended as the pulse number increases to 300.

INTRODUCTION MS (e.g. aerosol ionization and subsequent ion extraction). To minimize laser-induced elemental fractionation, various technical Laser ablation-inductively coupled plasma-mass spectrometry approaches have been proposed after conducting experimental (LA-ICP-MS) has become a widespread technology used for research on LA-ICP-MS ablation behaviors, such as investigating direct solid sample analysis. It not only has the advantages of ICP- the morphology of ablation craters and aerosol particles, MS, such as high sensitivity, low limit of detection and full redesigning ablation cell geometry, optimizing transport tube elemental coverage, but also offers the capability of high spatial materials, and performing comparative studies of infrared resolution in situ microanalysis.1-7 Besides, there is no need for the femtosecond lasers and ultraviolet nanosecond lasers.13-17 Besides, preparation of sample solutions, thus avoiding contamination and the laser operating conditions, such as spot size, repetition rate, reducing polyatomic ion interferences (e.g. 140Ce16O+ on 156Gd+, fluence, pulse duration and energy, have been well studied for 60Ni40Ar+ on 100Ru+).8-12 silicate glass samples or reference materials (e.g. NIST SRM However, elemental fractionation is a significant issue for LA- 610).18-22 ICP-MS because it can induce analytical errors. This fractionation Scheelite is a common accessory mineral that has the chemical may result from matrix effects related to laser-sample interactions, composition of CaWO4. It forms in various types of hydrothermal aerosol transport processes, and processes occurring in the ICP- deposits and always incorporates abundant trace elements, such as www.at-spectrosc.com/as/article/pdf/202101006 36 At. Spectrosc. 2021, 42(1), 36-42 the rare earth elements (REEs) of Nb, Pb and Mo, via substitution Table 1. Typical Operating Conditions of LA-ICP-MS for Ca2+ or W6+ in the crystal lattice, which renders scheelite to Laser ablation system ArF excimer laser exhibit distinct trace element geochemistry and thus serves as a Wavelength, nm 193 fingerprint for deposit types, metallogenic settings and ore- Pulse duration, ns 15 -2 forming fluid compositions.23, 24 As a result, it is essential to Energy density, J cm 6 Repetition rate, Hz 6 determine the elemental concentrations in scheelite accurately. To Spot size, μm 10, 16, 32, 44, 60, 90 minimize analytical errors resulting from elemental fractionation, Pulse number 200 25-27 the matrix-matched calibration strategy is proposed. The Carrier gas (He) flow rate, L min-1 0.9 matrix-matched calibration standard CaW-1, which also has a ICP-MS instrument Agilent 7700x matrix composition of CaWO4, has been synthesized in our RF power, W 1550 previous work.28 However, due to the different physical properties Auxiliary gas (Ar) flow rate, L min-1 1.05 -1 and chemical composition between calcium tungstate and silicate Plasma gas (Ar) flow rate, L min 15 Dwell time per isotope, ms 50 glass, the laser ablation settings for silicate glass may not be Sampling depth, mm 7.5 suitable for LA-ICP-MS analysis of scheelite. For this reason and Detector mode Dual to further understand the ablation mechanisms of materials with Measured isotopes 44Ca, 89Y, 139La, 140Ce, different matrices, a systematic study of the ablation behaviors of 141Pr, 146Nd, 147Sm, scheelite is necessary. 153Eu, 157Gd, 159Tb, 163Dy, 165Ho, 166Er, The aim of this work is to investigate the influence of the LA- 169Tm, 172Yb, 175Lu ICP-MS operating parameters and the spot size on the ablation behavior of scheelite. Because it has the same chemical morphology of the ablation craters was investigated using a composition as natural scheelite and a relatively homogeneous Hitachi SU8010 field emission scanning electron microscope (FESEM) at the State Key Laboratory of Biogeology and distribution of REEs, the CaWO4 calibration standard CaW-1 was selected as a representative sample of scheelite. The ablation Environmental Geology, China University of Geosciences 30 behaviors, including signal intensity ratios (with and without (Wuhan). internal standardization) and the elemental fractionation index The NIST SRM 610 silicate glass reference material was (EFI) among different spot sizes, were systematically studied. For purchased from the National Institute of Standards and comparison, the laser ablation behaviors of CaW-1 were compared 31 Technology (NIST), USA. Synthetic CaWO4 single crystal with those of the NIST SRM 610 silicate glass reference material CaW-1 with homogeneous distribution of REEs was prepared under different spot sizes. using the Czochralski technique and is discussed in our previous 28 work. Accordingly, a multi-REE-doped CaWO4 single crystal was grown at the Jiangxi University of Science and Technology, EXPERIMENTAL P.R. China. The crystal was prepared following a two-step An Excimer 193 nm laser ablation system (MicroLas Laser procedure: First, polycrystalline materials by a solid-state reaction System GmbH, Germany) coupled to an Agilent 7700x ICP-MS in a muffle furnace with stoichiometric CaCO3, WO3, Na2CO3 and (USA) was used for the LA-ICP-MS analyses at the State Key rare earth element oxide (at a nominal concentration of ca. 250 μg Laboratory of Biogeology and Environmental Geology, China g-1) were prepared; second, the polycrystalline materials were University of Geosciences (Wuhan). High purity He at a constant heated up to ca. 1630 °C (power: 4750 W) and kept at this flow rate of 0.9 L min-1 was used in the ablation cell as the carrier temperature for 2 h. The crystal was grown using a-cut pure -1 gas and merged with argon (make-up gas) behind the ablation cell. CaWO4 seeds at a pulling rate of 1 mm h and a rotation rate of 6 The NIST SRM 610 silicate glass reference material was used for rpm. The as-grown CaWO4 single crystal was named CaW-1. routine tuning to obtain the maximum signal intensity of 89Y+, Both samples, the CaW-1 and NIST SRM 610, were analyzed 139La+, 157Gd+, and 175Lu+ and to maintain the 238U+/232Th+ ratio with a spot size of 10, 16, 32, 44, 60, and 90 μm, respectively, for close to 1 to ensure low oxide formation. Low oxide production the comparative investigation of the influence of spot size on laser- was assured by the m/z 248/232 ratio (representing sample interactions. 232Th16O+/232Th+), which was consistently <0.5%. The optimized operating conditions and measurement parameters are summarized in Table 1. A repetition rate of 6 Hz and a fluence of RESULTS AND DISCUSSION -2 6 J cm were used. Each analysis consisted of 10 seconds of Signal intensity ratios background signal acquisition, followed by ~32 seconds of ablation. The LA-ICP-MS data processing, off-line selection and Signal sensitivity as a crucial parameter for the accurate integration of the background and the analyte signals were quantification in LA-ICP-MS is matrix-dependent and, therefore, performed using an in-house program, ICPMSDataCal.29 The causes significant elemental fractionation in different samples. www.at-spectrosc.com/as/article/pdf/202101006 37 At. Spectrosc. 2021, 42(1), 36-42 Moreover, laser-induced fractionation may arise under small spot or proportionally changed), and the volume of the ablated size conditions.32 For these reasons, the influence of spot size on materials can be calculated according to Eq. 1. By contrast, the laser ablation was investigated for CaWO4 single crystal CaW-1 experimental values of R32-16 (4.10~4.59) were a little higher than and silicate glass reference material NIST SRM 610 with a the theoretical value (4.00 for R32-16). Since the crater shape for 32 repetition rate of 6 Hz and a fluence of 6 J cm-2. μm was found to be regular, the reason for higher experimental

values of R32-16 can be ascribed to the lower signal intensity of 16 In theory, the ablation crater is circular and its volume can be μm (I16) due to the tapered crater and, therefore, a smaller crater calculated according to Equation 1: volume and less ablated materials. This phenomenon was further 2 푉 = 휋푟 ℎ (Eq. 1) verified for R16-10 as disproportional and less materials were ablated from the sample, and a more seriously tapered crater was where V is the volume of the ablation crater, r is the radius of the found for 10 μm (I10), which renders the experimental values ablation crater (i.e., half of the spot size), and h is the height/depth (3.01~3.47) much larger than the theoretical value (2.56). of the crater. For the same samples, the matrix composition Surprisingly, the experimental values of R90-60 (1.76~1.83) were remains unchanged and the ablated mass (m) can also be much lower than the theoretical value (2.25). The reason for this is calculated based on the density (ρ) of the sample according to unclear, but we speculate that the ablation efficiency or aerosol Equation 2: transport efficiency for 90 μm may not reach the expected value 푚 = 푉휌 (Eq. 2) because a large number of aerosol particles may accumulate in front of the sample or within the trench, which led to shielding as As the sensitivity (S) keeps constant for one homogeneous 33 reported. It thus resulted in lower signal intensity (I90) and lower sample, the signal intensity can be calculated in Equation 3 as R90-60 when the spot size of 90 μm was employed, which may be follows: corrected by increasing the gas flow rate, but it needs further 퐼 = 푚푆 (Eq. 3) studies to verify. Based on these results, it can be concluded that a spot size of 60, 44, or 32 μm is desired for LA-ICP-MS analysis Under the same operating conditions, except for spot size, h, ρ, of the CaWO4 single crystal, while 16, 10, and even 90 μm are and S were theoretically treated as constant for one sample, and then under different conditions of spot size, the ratio (R) of unfavorable spot sizes due to the tapered craters or the below- intensity I can be obtained with Equation 4 as follows: expected ablation efficiency/aerosol transport efficiency. 2 2 퐼푎 휋푟푎 ℎ휌푆 푟푎 For silicate glass reference material NIST SRM 610, similar 푅푎−푏 = = 2 = 2 (Eq. 4) 퐼푏 휋푟푏 ℎ휌푆 푟푏 results as described above were found. The experimental values of where ra and rb are the radii of the ablation craters (i.e., half of the R60-44 (1.85~1.94) and R44-32 (1.95~2.04) for all REEs mainly spot size) and the subscripts a and b represent the values of the agreed with the theoretical values (1.86 for R60-44 and 1.89 for R44- spot size. 32), which further indicates that the crater shapes for 60, 44, and 32 As listed in Table 2, the theoretical values of the signal intensity μm are regular and the volume of ablated materials can be ratio (R) were calculated according to Eq. 4. For CaW-1, it was calculated according to Eq. 1. However, the experimental values of R32-16 (5.65~6.10) and R16-10 (3.39~4.03) were much higher than found that the experimental values of R60-44 (1.77~1.85) and R44-32 (1.87~1.97) for all REEs mainly agreed with the theoretical values the theoretical value (4.00 for R32-16, and 2.56 for R16-10). This can be attributed to a lower signal intensity resulting from the tapered (1.86 for R60-44 and 1.89 for R44-32). This demonstrates that the crater shapes for 60, 44, and 32 μm are mainly regular (cylindrical crater for 16 μm (I16) and 10 μm (I10), which is even

Table 2. Theoretical and Experimental Values of Signal Intensity Ratios (R) Under Different Spot Sizes

Signal Theoretical Sample intensity Y La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu value ratio (R)

Synthetic R90-60 2.25 1.76 1.77 1.75 1.79 1.78 1.76 1.78 1.79 1.78 1.78 1.77 1.80 1.80 1.83 1.77

CaWO4 single R60-44 1.86 1.82 1.80 1.78 1.80 1.79 1.82 1.82 1.80 1.80 1.80 1.82 1.82 1.79 1.77 1.85

crystal CaW-1 R44-32 1.89 1.89 1.92 1.95 1.92 1.94 1.89 1.92 1.87 1.93 1.92 1.92 1.97 1.92 1.89 1.88

R32-16 4.00 4.19 4.10 4.08 4.33 4.35 4.30 4.31 4.39 4.35 4.48 4.31 4.15 4.36 4.59 4.21

R16-10 2.56 3.27 3.01 3.15 3.06 3.12 3.16 3.01 3.05 3.14 3.22 3.16 3.16 3.44 3.23 3.47

NIST SRM R90-60 2.25 1.94 1.98 2.01 2.01 2.01 1.98 2.03 1.98 2.00 1.99 1.99 1.97 1.96 1.99 1.96

610 R60-44 1.86 1.94 1.91 1.85 1.88 1.90 1.91 1.87 1.91 1.92 1.90 1.92 1.94 1.93 1.91 1.92

R44-32 1.89 1.97 1.98 2.04 2.00 1.96 2.00 1.99 2.02 1.99 1.97 1.97 1.95 1.98 1.99 1.99

R32-16 4.00 5.77 5.67 5.76 5.70 5.92 5.65 5.98 5.74 5.92 5.91 5.95 6.10 6.07 6.03 6.05

R16-10 2.56 3.49 3.66 3.49 3.47 3.39 4.00 3.48 3.47 3.63 3.81 3.77 3.63 3.79 4.03 3.82

www.at-spectrosc.com/as/article/pdf/202101006 38 At. Spectrosc. 2021, 42(1), 36-42

Fig. 1 SEM micrographs of laser ablation craters for CaW-1 (with a spot size of (a) 32 μm, (b) 16 μm, and (c) 10 μm) and NIST SRM 610 (with a spot size of (d) 32 μm, (e) 16 μm, and (f) 10 μm). more severe than for the CaWO4 single crystal CaW-1. Internal standardization

Furthermore, the experimental values of R90-60 (1.94~2.03) were For LA-ICP-MS analysis, the internal standard can be used to lower than the theoretical value (2.25), but a little higher than that calibrate the difference of the ablated material mass. The low of the CaWO4 single crystal CaW-1 (1.76~1.83). This abundance isotope of matrix elements (e.g., 44Ca in calcite) is demonstrates that different laser-sample interactions occurred for usually employed for this purpose. In this study, 44Ca was the CaWO4 single crystal vs. the NIST SRM 610 silicate glass, employed as the internal standard, and the ratio (R) of intensity I where for the latter more efficient laser ablation occurred, thus it was re-calculated to 1 according to Equation 5: reduced the exceptional phenomenon for R90-60. In general, a spot 푖 퐼푆 2 푖 2 퐼푆 퐼푎/퐼푎 휋푟푎 ℎ휌푆 /휋푟푎 ℎ휌푆 size of 60, 44, or 32 μm was desired for LA-ICP-MS analysis of 푅푎−푏 = 푖 퐼푆 = 2 푖 2 퐼푆 = 1 (Eq. 5) 퐼푏/퐼푏 휋푟푏 ℎ휌푆 /푟푏 ℎ휌푆 the silicate glass sample, and the laser-sample interactions where superscript i represents element and IS the internal standard occurring for CaWO4 single crystal and NIST SRM 610 silicate (i.e. 44Ca in this work). glass were distinct. As shown in Fig. 2, after internal standardization, the The above-mentioned results for CaW-1 and NIST SRM 610 experimental values of R were calculated to be approximately 1.00 were verified by SEM micrographs of the laser ablation craters. independent of spot size. It indicates that 44Ca can be used to Since the craters for both samples with a spot size of 90, 60, 44, compensate and calibrate the difference of the ablated material and 32 μm are similar and regular, only the SEM micrograph of mass for both the CaWO4 single crystal and the NIST SRM 610 32 μm was displayed in Fig. 1 (a and d) to compare with those of silicate glass. More importantly, it can be concluded that 16 and 10 μm. As shown in Fig. 1, the most seriously and satisfactory results may also be obtained after internal moderately tapered craters were found for 10 μm (Fig. 1c and f) standardization when different spot sizes must be selected for LA- and 16 μm (Fig. 1b and e), respectively. It led to an unsatisfactory ICP-MS analysis. For example, (a) a large spot size was ablation volume for 16 and 10 μm and, thus, an undesired R32-16 preferential for real samples that contained very low and R16-10. In addition, rim growth around the craters due to the concentrations or had a heterogenous distribution of the analytes heat accumulation effect33 was found to be more obvious for CaW- to guarantee enough sensitivity and sampling representativeness, 1 (Fig. 1b and c) than for NIST SRM 610 (Fig. 1e and f), which while a medium or small spot size is encouraged for external indicates a slower heat transmission rate for CaWO4 single crystal standards to protect the ICP-MS detector and to save valuable than for NIST SRM 610 silicate glass. Moreover, more seriously reference material. (b) A small spot size should be employed for tapered craters can be seen for NIST SRM 610 (Fig. 1e and f) the analysis of microfossils to obtain high spatial resolution, while compared to those of CaW-1 (Fig. 1b and c). All of these a medium-to-small spot size is preferential for laser ablation on phenomena are caused by matrix effects, which may induce external standards for accurate quantification. (c) The laser beam analytical errors if NIST SRM 610 were used as the external focusing on the sample surface may vary each time for various standard to calibrate the elemental concentrations in CaWO4 reasons, which causes fluctuation of actual spot size on the single crystal and even scheelite. www.at-spectrosc.com/as/article/pdf/202101006 39 At. Spectrosc. 2021, 42(1), 36-42 size of 16 μm is desirable, but an extremely small spot size (10 μm) is not recommended for quantitative LA-ICP-MS analysis of

CaWO4 single crystal (and even scheelite) and NIST SRM 610 silicate glass, even though internal standardization is performed.

Elemental fractionation

To further verify the influence of spot size on the ablation behavior of CaW-1 and NIST SRM 610, the elemental fractionation index (EFI) was calculated according to Equation 6 as set out by Fryer et al.34: 퐼푖 퐼푖 퐸퐹퐼 = ( ) /( ) (Eq. 6) Fig. 2 Signal intensity ratios (R) under different spot sizes after internal 퐼퐼푆 푡2 퐼퐼푆 푡1 44 standardization with Ca. where Ii and IIS are the intensity of element i and the internal

standard IS, respectively; t1 and t2 are the first 16 sec. and the second 16 sec. during each LA-ICP-MS analysis.

As listed in Table 3, it was found that the EFI remained constant (ca. 1.00) for CaW-1 when the spot size was larger than 32 μm. However, obvious elemental fractionation (EFI ˃1.10 or ˂0.90) was found for Ce (1.26), Nd (1.18), Sm (1.27), Eu (1.11), Ho (1.14), and Tm (1.13) when a spot size of 10 μm was used, which

was also found for Nd (1.17), Sm (1.19), and Yb (1.11) when a spot size of 16 μm was used. For NIST SRM 610, more severe elemental fractionation was found for almost all REEs (except Eu)

as the EFI ranged from 0.20 for Dy to 0.70 for Nd when a spot size of 10 μm was used, while elemental fractionation was hardly found when the spot size increased to larger than 16 μm. It Fig. 3 Elemental fractionation index of REEs in CaW-1 (a) and NIST SRM indicates that a small spot size (i.e., 10 μm) is not suitable for 610 (b) under different spot size and pulse number. quantitative LA-ICP-MS microanalysis of both the CaWO4 single crystal and the NIST SRM 610. sample surface. For these cases, the elemental fractionation possibly caused by different spot sizes and ablation material mass This was also confirmed by performing additional experiments can be eliminated by internal standardization, which is significant with the same operating parameters, except for increasing the for improving the results of LA-ICP-MS analysis. It should number of laser ablation pulses from 200 to 300. As shown in Fig. definitely be noted that in both samples the R16-10 for some REEs 3a, when 300 pulses were performed on CaW-1, a significant (mainly for heavy REEs) was larger than 1.10, such as 1.13 for Tm elemental fractionation was found with the spot sizes of 10 μm and 1.14 for Lu in CaW-1, and 1.16 for Sm, 1.17 for Yb, 1.10 for and 16 μm since the EFI were lower than 0.90 for all REEs. As for Dy and Lu in NIST SRM 610. From this perspective, a small spot NIST SRM610 (Fig. 3b), when 300 pulses were used, more

Table 3. EFI for REEs in CaW-1 and NIST SRM 610 under Different Spot Sizes Sample Spot size (μm) Y La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu CaW-1 10 1.02 1.08 1.26 1.07 1.18 1.27 1.11 0.99 1.06 0.95 1.14 0.98 1.13 1.05 1.04 16 1.02 1.02 0.99 1.08 1.17 1.19 1.06 1.04 1.06 1.06 1.04 1.09 1.06 1.11 1.05 32 1.00 1.02 1.01 1.02 1.05 1.02 0.99 1.00 1.01 0.99 1.01 1.01 1.00 0.97 1.01 44 1.04 1.05 1.05 1.05 1.07 1.03 1.03 1.04 1.05 1.04 1.06 1.04 1.03 1.01 1.07 60 1.03 1.03 1.03 1.04 1.02 1.05 1.03 1.04 1.07 1.06 1.07 1.07 1.06 1.04 1.04 90 1.01 1.00 1.00 0.99 1.03 1.00 1.02 1.01 1.02 1.02 1.01 1.02 1.01 1.03 1.01 NIST 10 0.54 0.39 0.49 0.64 0.70 0.52 0.95 0.51 0.40 0.20 0.60 0.35 0.45 0.65 0.44 SRM 16 1.01 1.02 1.01 1.01 0.99 1.07 1.03 0.98 1.03 1.02 1.04 0.99 1.02 1.07 1.01 610 32 1.00 1.00 1.00 0.99 1.02 0.98 0.99 1.03 1.01 0.98 1.02 1.01 1.00 1.00 1.01 44 1.07 1.04 1.00 1.00 1.01 1.02 1.02 1.06 1.06 1.05 1.06 1.05 1.06 1.04 1.06 60 1.01 1.02 1.00 1.00 1.01 1.03 1.01 1.01 1.01 1.02 1.02 1.00 1.02 1.01 1.02 90 1.02 1.04 1.02 1.03 1.03 1.03 1.04 1.03 1.02 1.04 1.03 1.02 1.04 1.01 1.02

www.at-spectrosc.com/as/article/pdf/202101006 40 At. Spectrosc. 2021, 42(1), 36-42 significant elemental fractionation was found for all REEs not only ACKNOWLEDGMENTS with the spot size of 10 μm, but also with the spot sizes of 16 and 32 μm. This can be attributed to the fact that more pulses increase This work was supported by the National Natural Science the depth/diameter ratio of the ablation crater and, therefore, Foundation of China (No. 41603025), the Natural Science increase the down-hole fractionation.35,36 In conclusion, elemental Foundation of Jiangxi Province (No. 20192BAB203024, fractionation is found for both the CaW-1 and the NIST SRM 610 20181BAB211009), the Research Foundation of the Education under a small spot size, which becomes more severe when the Bureau of Jiangxi Province of China (No. GJJ180428, GJJ180461, number of laser ablation pulses increases due to down-hole GJJ170546), and the PhD Research Startup Foundation of Jiangxi fractionation. Combining the results as mentioned above, a spot University of Science and Technology (No. jxxjbs17066, size of ≥32 μm is preferential for CaW-1 and even scheelite jxxjbs17041, jxxjbs17051). regardless of pulse number, while a spot size of 16 μm is desirable for NIST SRM 610 with the pulse number of 200, but a spot size of 44 μm is recommended as the pulse number increases to 300. REFERENCES 1. L. Guo, Q. Li, Y. Chen, G. Zhang, Y. Xu, and Z. Wang, J. Anal. At. Spectrom., 2020, 35, 1441-1449. https://doi.org/10.1039/d0ja00054j CONCLUSIONS 2. P. Bohleber, M. Roman, M. Šala, and C. Barbante, J. Anal. At. Spectrom., 2020, 35, 2204-2212. In this work, comparative studies on the LA-ICP-MS ablation https://doi.org/10.1039/D0JA00170H behaviors of silicate glass reference material NIST SRM 610 and 3. Y. Zhu, C. Y. Li, M. Chen, G. X. Zhang, C. W. Mao, H. M. Kou, and synthetic calcium tungstate calibration standard CaW-1 were Z. Wang, At. Spectrosc., 2019, 40, 49-54. performed under different laser ablation spot sizes. LA-ICP-MS https://doi.org/10.46770/AS.2019.02.003 analysis on both the CaW-1 and NIST SRM 610 can obtain desired 4. Z. P. Yang, S. E. Jackson, L. J. Cabri, P. Wee, H. P. Longerich, and M. Pawlak, J. Anal. At. Spectrom., 2020, 35, 534-547. ablation craters and expected signal intensity ratios (R) with a spot https://doi.org/10.1039/c9ja00285e size of 60, 44, or 32 μm, while it is not possible with the spot size 5. W. Guo, R. Wang, W. X. Wang, and Y. E. Peng, of 10, 16, and even 90 μm due to the tapered craters or below- J. Food Compos. Anal., 2020, 91, 103517. expected ablation efficiency/aerosol transport efficiency. Besides, https://doi.org/10.1016/j.jfca.2020.103517 the difference of ablation material mass caused by different spot 6. H. Cui, W. Guo, L. L. Jin, Y. E. Peng, and S. H. Hu, sizes can be eliminated by internal standardization as the values of J. Anal. At. Spectrom., 2020, 35, 592-599. R were normalized to ca. 1.00, except with an extremely small spot https://doi.org/10.1039/c9ja00340a size. Elemental fractionation was found for CaW-1 and NIST 7. W. Guo, L. L. Jin, S. H. Hu, and Q. H. Guo, J. Agric. Food Chem., 2017, 65, 3407-3413. https://doi.org/10.1021/acs.jafc.7b00535 SRM 610 under the small spot size. A spot size of ≥32 μm is 8. W. Guo, S. H. Hu, J. A. Zhao, S. S. Jin, W. J. Liu, and H. F. Zhang, preferential for CaW-1 and even scheelite regardless of pulse Microchem. J., 2011, 97, 154-159. number, while a spot size of 16 μm is desirable for NIST SRM 610 https://doi.org/10.1016/j.microc.2010.08.003 with the pulse number of 200, but a spot size of 44 μm is 9. Y. Q. Ke, J. Z. Zhou, L. Qiao, M. H. Zhang, W. Guo, L. L. Jin, and recommended as the pulse number increases to 300. S. H. Hu, Anal. Methods, 2019, 11, 2129-2137. https://doi.org/10.1039/c8ay02664e 10. Y. T. Li, W. Guo, Z. C. Hu, L. L. Jin, S. H. Hu, and Q. H. Guo, J. Agric Food Chem., 2019, 67, 935-942. AUTHOR INFORMATION https://doi.org/10.46770/as.2018.01.001 11. W. Guo, S. H. Hu, X. J. Wang, J. Y. Zhang, L. L. Jin, Z. L. Zhu, and H. F. Zhang, J. Anal. At. Spectrom., 2011, 26, 1198-1203. Corresponding Author https://doi.org/10.1039/c1ja00005e * Y.-Q. Ke 12. X. Lin, W. Guo, L. L. Jin, and S. H. Hu, At. Spectrosc., 2020, 41, 1-10. https://doi.org/10.46770/as.2020.01.001 Email address: [email protected] 13. M. Hola, V. Konecna, P. Mikuska, J. Kaiser, and V. Kanicky, Spectrochim. Acta Part B, 2010, 65, 51-60. * Y.-J. Sun https://doi.org/10.1016/j.sab.2009.11.003 Email address: [email protected] 14. M. Hola, J. Ondracek, H. Novakova, M. Vojtisek-Lom, R. Hadravova, and V. Kanicky, Spectrochim. Acta Part B, 2018, Notes 148, 193-204. https://doi.org/10.1016/j.sab.2018.07.001 15. R. Kovacs and D. Günther, J. Anal. At. Spectrom., 2008, 23, The authors declare no competing financial interest. 1247-1252. https://doi.org/10.1039/b803789b 16. T. Luo, Y. Wang, M. Li, W. Zhang, H. H. Chen, and Z. C. Hu, At. Spectrosc., 2020, 41, 11-19. https://doi.org/10.46770/as.2020.01.002 www.at-spectrosc.com/as/article/pdf/202101006 41 At. Spectrosc. 2021, 42(1), 36-42 17. N. J. Saetveit, S. J. Bajic, D. P. Baldwin, and R. S. Houk, 27. M. V. Galiova, K. Stepankova, R. Copjakova, J. Kuta, L. Prokes, J. Anal. At. Spectrom., 2008, 23, 54-61. J. Kynicky, and V. Kanicky, J. Anal. At. Spectrom., 2015, 30, https://doi.org/10.1039/b709995a 1356-1368. https://doi.org/10.1039/c4ja00347k 18. F.-X. d'Abzac, F. Poitrasson, R. Freydier, and 28. Y. Ke, J. Zhou, X. , Y. Sun, J. Shao, S. You, W. Wang, Y. Z. Tang, A.-M. Seydoux-Guillaume, J. Anal. At. Spectrom., 2010, 25, and C. Tu, J. Anal. At. Spectrom., 2020, 35, 886-895. 681-689. https://doi.org/10.1039/b913584g https://doi.org/10.1039/D0JA00015A 19. A. El Korh, Spectrochim. Acta Part B, 2013, 86, 75-87. 29. Y. Liu, Z. Hu, S. Gao, D. Günther, J. Xu, C. Gao, and H. Chen, https://doi.org/10.1016/j.sab.2013.04.009 Chem. Geol., 2008, 257, 34-43. 20. T. Vaculovic, T. Warchilova, Z. Cadkova, J. Szakova, P. Tlustos, https://doi.org/10.1016/j.chemgeo.2008.08.004 V. Otruba, and V. Kanicky, Appl. Surf. Sci., 2015, 351, 296-302. 30. Y. H. Fang, Z. Q. Chen, S. Kershaw, H. Yang, and M. Luo, https://doi.org/10.1016/j.apsusc.2015.05.136 Global Planet. Change, 2017, 152, 115-128. 21. J. Fietzke and M. Frische, J. Anal. At. Spectrom., 2016, 31, 234-244. https://doi.org/10.1016/j.gloplacha.2017.02.011 https://doi.org/10.1039/c5ja00253b 31. https://www.nist.gov 22. X. H. Liao, T. Luo, S. H. Zhang, W. Zhang, K. Q. Zong, Y. S. Liu, 32. Z. Hu, Y. Liu, L. Chen, L. Zhou, M. Li, K. Zong, L. Zhu, and S. and Z. C. Hu, J. Anal. At. Spectrom., 2020, 35, 1071-1079. Gao, J. Anal. At. Spectrom., 2011, 26, 425-430. https://doi.org/10.1039/c9ja00404a https://doi.org/10.1039/c0ja00145g 23. Y. Fu, X. M. Sun, H. Y. Zhou, H. Lin, L. Y. Jiang, and T. J. Yang, 33. J. J. Gonzalez, A. Fernandez, D. Oropeza, X. Mao, and R. E. Russo, Ore Geol. Rev., 2017, 80, 828-837. Spectrochim. Acta Part B, 2008, 63, 277-286. https://doi.org/10.1016/j.oregeorev.2016.08.030 https://doi.org/10.1016/j.sab.2007.11.035 24. W. W. Zhao, M. F. Zhou, A. E. Williams-Jones, and Z. Zhao, 34. B. J. Fryer, S. E. Jackson, and H. P. Longerich, Can. Mineral., 1995, Chem. Geol., 2018, 477, 123-136. 33, 303-312. https://doi.org/10.1016/j.chemgeo.2017.12.020 35. T. Luo, Z. C. Hu, W. Zhang, Y. S. Liu, K. Q. Zong, L. Zhou, 25. Y. Q. Ke, Y. J. Sun, P. J. Lin, J. Z. Zhou, Z. F. Xu, C. F. Cao, Y. J. F. Zhang, and S. H. Hu, Anal. Chem., 2018, 90, 9016-9024. Yang, and S. H. Hu, Microchem. J., 2019, 145, 642-647. https://doi.org/10.1021/acs.analchem.8b01231 https://doi.org/10.1016/j.microc.2018.11.016 36. T. Luo, X. D. Deng, J. W. Li, Z. C. Hu, W. Zhang, Y. S. Liu, and 26. A. Ugarte, N. Unceta, C. Pecheyran, M. A. Goicolea, and J. F. Zhang, J. Anal. At. Spectrom., 2019, 34, 1439-1446. R. J. Barrio, J. Anal. At. Spectrom., 2011, 26, 1421-1427. https://doi.org/10.1039/c9ja00139e https://doi.org/10.1039/c1ja10037h

www.at-spectrosc.com/as/article/pdf/202101006 42 At. Spectrosc. 2021, 42(1), 36-42 . owners FOUR Qs respective their of ARE BETTER property the are trademarks THAN QQQ ther o

All

. Inc

, PerkinElmer of trademark registered a is

® PerkinElmer

. reserved rights

All

96703

. Inc

, PerkinElmer

2020

©

Copyright

Perfect for challenging applications in semiconductors, geosciences, biomonitoring, and more, the cutting-edge NexION® 5000 ICP-MS is the industry’s first four- quadrupole system, combining the simplicity of a reaction/collision cell with multi- quadrupole technology that transcends traditional triple quad. That means superior interference removal, stability, and matrix tolerance – all in a system that virtually eliminates routine maintenance, for unsurpassed instrument uptime. The NexION 5000 multi-quadrupole ICP-MS : This is performance to the power of four.

For more information on our NexION 5000 ICP-MS, visit www.perkinelmer.com/nexion5000 Copyright © 2020 PerkinElmer, Inc. 59397_400372_02 All rights reserved. PerkinElmer® is a registered trademark of PerkinElmer, Inc. All other trademarks are the property of their respective owners. EXPAND YOUR RANGE EXTEND YOUR RESOURCES EXTEND YOUR RESOURCES Avio ICP-OES 500 Formore information, visit perkinelmer.com/avio500 ICP-OES system. of ownership. Superior It’s performance. the range of what you throughp accomplish. High can enabled View by Dual technology, comes it togetall ICP, backgroundt simultaneous correction, and high the lowest with samples. argon And consumpt di$cult your accomplishlab can more, evenwit when dealing your and worksuperi Withsensitivity demands. high on and performance fasteri return the high-quality e Avio® ICP-OES 500 combines you the productivity with low cost of ownership The New Avio 500 ICP-OES - High throughput everything you an in want everything nvestment her to expand or resolution, hroughput h the most ut. Low cost ion of any need with need with