1 Introduction

Although the Shanghai Synchrotron Radiation Facility (SSRF) has operated successfully since May 2009, the BL14W1 beamline remains the only available beamline for X-ray absorption fine structure (XAFS) experiments at this facility. However, an increasing number of users from diverse institutes over the world have focused on XAFS techniques for advanced characterization. Thus, the demand for beamtime has significantly increased, and currently, only about one quarter of the applicants can be approved for conducting experiments on the BL14W1 beamline. Clearly, either additional XAFS beamlines are required, or the presently available beamtime must be used more efficiently.

High-throughput (HT) techniques allow multiple samples (up to hundreds) to be tested sequentially without interrupting the experiment process. During the past two decades, HT techniques have received significant attention in various fields of research such as the optimization of the conversion and selectivity of new catalysts [1-3], measurements of reaction kinetics, and spectroscopic methods including infrared spectroscopy [4], Raman spectroscopy [5], X-ray fluorescence [6], fluorescence microscopy [7], imaging polarimetry [8], nuclear magnetic resonance spectroscopy [9], and X-ray diffraction [10]. HT techniques have been particularly well established in many advanced synchrotron radiation facilities internationally, e.g., beamline X9B of the National Synchrotron Light Source (NSLS) for X-ray absorption spectroscopy [11], beamline X06DA of the Swiss Light Source for in-situ X-ray diffraction [12], and beamline 9.3 of the Synchrotron Radiation Source (SRS) for in-situ XAFS [13]. The implementation of HT techniques has significantly enhanced the beamtime efficiency of these experimental beamlines, particularly for in situ measurements.

Previously, conventional XAFS measurements conducted at the BL14W1beamline employed a control system based on the Experimental Physics and Industrial Control System (EPICS) software environment for the Bragg motor and LabVIEW software to control the data-acquisition system. LabVIEW communicates with EPICS through NI (National Instrument) Lab-VIEW’s Datalogging and Supervisory Control (DSC) module [14]. However, because LabVIEW and EPICS employ independent operational environments, data transmission between LabVIEW and EPICS may fail during scanning. Furthermore, the repetitious scanning employed by HT techniques may increase the possibility of this transmission failure.

In the present study, we sought to solve the above communication problem by applying a single system based on EPICS to integrate the control of all instruments (e.g., motors, counters, detectors) for HT XAFS measurements. Control System Studio (CSS) was selected to develop a graphical user interface (GUI) that allows experimenters to implement HT XAFS measurements. The successful operation of the proposed HT XAFS data-collection system was demonstrated by scanning different copper-ceria catalyst samples sequentially. The HT XAFS data-collection system is now operational at beamline BL14W1 of the SSRF.

2 Experimental

2.1 HT XAFS implementation at beamline BL14W1

In addition to general XAFS instrumentation (e.g., ion chambers and laser pointer), HT XAFS measurements also require a high-precision two-dimensional (2-D) translation stage comprised of SGSP33-100XY and SGSP80-20ZF stepping motor driven stages (OptoSigma Corp.) for the x and z directions, respectively, and an alumina multi-sample holder, both of which are schematically illustrated in Fig. 1. In experiments, an XAFS spectrum is collected by monochromatic X-rays sequentially passing through the 1^st ion chamber (I₀), pellet sample, and the 2^nd ion chamber (I₁). The X-ray counts before and after transmission through the sample are recorded by I₀ and I₁, respectively. Under the proposed HT data-collection mode, up to 8 samples can be loaded in the alumina holder, where the diameter of each cell is 10 mm, and the sample interval in the x-direction is 23 mm while the interval in the z-direction is 20 mm. The base of the sample holder is attached to the 2-D translation stage, which is controlled to sequentially introduce the different samples into the beam path. In experiments, we must first calibrate the position of the #1 cell using the laser pointer, which is pre-aligned with the X-ray beam. After calibration, the sample holder design and stage control assure the accurate placement of sample centers in the beam path. During the data scanning process, a local variable Num records the sequence of measured samples. The experiment is then terminated by the control system after completion of the final scan. All signals are transmitted to an operation interface (OPI) via a local area network (LAN). This data-collection process provides for the continuous scanning of 8 samples.

Fig. 1Full-screen View

HT XAFS measurement implementation, where monochromatic X-rays sequentially passing through the 1^st ion chamber (I₀), pellet sample, and the 2^nd ion chamber (I₁). The laser pointer serves to calibrate the location of the #1 cell of the sample holder.

2.2 Hardware design

2.2.1 Architecture

Fig. 2 presents a schematic representation of the Bragg motor control system (left panel) and the HT XAFS data-collection system (right panel). The entire system is based on EPICS, and is comprised of three segments: input/output controller (IOC), OPI, and channel access (CA) [15-17]. The HT XAFS data-collection system adopts two IOCs, where one is for the high-precision 2-D stage and double crystal monochromator (DCM) control [18], while the other is employed for the counter.

Fig. 2Full-screen View

Schematic of the Bragg motor control system (left panel) and the HT XAFS data-collection system (right panel)

The translation stage control system employed is the preexisting system at beamline BL14W1. The IOC of the system is comprised of a Versa Module Eurocard (VME) crate, MVME5500 single-board computer, and a MAXv-8000 8-axis motor controller. The IOC runs the VxWorks real-time operating system (RTOS), and the EPICS distributed control system. In addition, a stepper motor driver (SMD5002) is employed to convert the low-current control signal obtained from the MAXv-8000 into a higher-current signal to drive the stage motor. The MAXv-8000 control signals are transmitted to the stage motor by a step motor driver interface (SLS 2017). The OPI and IOC communicate via a LAN.

2.2.2 Data transmission devices

A voltage-to-frequency converter (V2F100) and a counter (NCT08-01) are respectively employed to convert and collect signals. The V2F100 is a two-channel unit with standard Bayonet Neill-Concelman (BNC) connector inputs and LEMO connector outputs. The counter provides two channels operating at frequencies of up to 100 MHz for transistor-transistor logic (TTL) input signals. Table 1 lists the main characteristics of the counter.

Table 1Full-screen View

Main characteristics of the NCT08-01counter

Index	Characteristic
1	8 independent channels
2	Up to 100 MHz (Min) count frequency
3	32 bit data
4	Timer accuracy: 0.005%
5	Timer resolution: 0.000001 s (1 μs)
6	Communication: LAN, USB

Show more

Fig. 2 also illustrates the signal transmission process. A small current generated in the ion chamber is converted into a usable voltage by a current amplifier (DLPCA-200), and then the voltage is transformed into a 100 MHz TTL signal by a voltage-to-frequency converter. Meanwhile, the counter counts pulse number of TTL, which is subsequently communicated to the OPI via the LAN.

2.3 Software design

2.3.1 Data collection system

Fig. 3 presents a schematic representation of the HT XAFS data-collection system. The GUI developed by CSS sends and receives messages to the IOCs via Ethernet. The sscan module controls the process of XAFS scan. Data storage is used for XAFS data storage. The motor of 2-D translation stage and counter are controlled by different IOCs. Specifically, the counter is controlled by a soft IOC running on a PC.

Fig 3Full-screen View

Schematic of the HT XAFS data-collection system

JavaScript and Python were chosen as the scripting environment. Python was used for defining how to calculate the number of positions in a single scan, and JavaScript was applied to define (1) how to select a scan model, (2) when to collect XAFS data, (3) when to switch a new sample, and (4) the total number of samples to be scanned. StreamDevice and asyn (an interface between EPICS drivers and device support) were used to generate EPICS device support for the counter.

2.3.2 Soft IOC of the counter

The entire soft IOC program for the counter is expressed by EPICS database records, StreamDevice protocol files, and the IOC startup command file (st.cmd) [19]. StreamDevice provides generic EPICS device support, and is not limited to a specific type of device or manufacturer, which makes recompilation unnecessary to support a new device. Instead, the StreamDevice protocol files for the counter are simply configured in plain ASCII text, which defines the commands understood by the counter and the replies it sends [20].

2.3.3 Operation mechanism of the sscan module

The sscan module is a key data-acquisition component for the HT XAFS data-collection system. Fig. 4 illustrates the sscan module interface. It can be seen that the Positioner (i.e., the Bragg motor) option SCAN MODE is selected as TABLE, indicating that the positioner locations are contained in an array, and the total number of values in the array is determined by NPTS (i.e., the total number of positions to visit). After initializing the parameters of the Positioners, DetTriggers, and Detectors, the sscan record commands the positioner to move to its starting position. Here, the Bragg motor moves to a preset energy point. Then, the sscan record continues to update the positioner location according to the locations given in the array, and, after each update, waits for the resulting callback. After positioning is complete, the sscan moduletriggers the detector. In other words, DetTriggers initiates the counter function and the sscan record collects Detectors data. The sscan record first evaluates the validity of the collected data before reading the detector signals. Then, the sscan module switches to perform another move, trigger, and read sequence. The algorithm does not stop until the sscan record completes all NPTS steps [21-22].

Fig. 4Full-screen View

The sscan module interface

2.3.4 GUI

CSS is a development platform based on Eclipse rich client platform (RCP) software, and is a collection of tools running on top of EPICS to monitor and operate large-scale control systems. Just as EPICS works with the individual segments of named data associated with the various instruments, denoted as PVs, and connections to control systems, the CSS core provides the necessary application program interface (APIs) for convenient operation. At the client level, CSS allows PVs to be accessed and modified from the network by means of the CA communication protocol used by EPICS to transfer information via a network. For complex dynamic behaviors that cannot be achieved by rules (i.e., widget properties), several scripts are attached to the widget in CSS to obtain the desired behaviors [23]. Herein, JavaScript and Python are selected as scripting environments. The widget accepts PVs as inputs, and its execution is triggered when the value is changed. In the HT XAFS data-collection system, motion control and data collection are all programmed in JavaScript and Python. Fig. 5 shows the control flow of the HT XAFS data-collection system.

Fig. 5Full-screen View

Control flow of the HT XAFS data-collection system

The GUI implements two scanning modes, denoted herein as successive scanning and selective scanning, which are selected in the Scan pane shown on the left-hand side of Fig. 6. Successive scanning collects XAFS data for a sequence of samples, whereas selective scanning collects XAFS data for a selected subset of samples. To facilitate the selection function, a check box is included within each green circular LED icon representative of one of 8 samples in the Scan pane shown in Fig. 6, and its selection includes that sample for XAFS measurements. The LED icon of a given sample lights up during data collection for that sample. As for Scan parameter pane, it defines energy section, energy step and integral time of XAFS measurement. Motor pane is for 2-D translation stage control. Choose and Calibrate pane is for choosing the element for XAFS measurements and calibrating the energy of monochromator. Status and XAFS spectra panes show the status of monochromator and XAFS spectrum, respectively.

Fig. 6Full-screen View

GUI of the HT XAFS data-collection system

3 Results and discussion

To demonstrate the operation of the HT XAFS data-collection system, we conducted experiments involving copper-ceria catalysts. The catalyst samples were prepared by a solution-based wet-chemical approach, and the synthetic parameters, such as molar ratio of Cu/Ce, precipitation pH value, and reaction temperature, were tuned delicately to generate nano-sized materials with distinct catalytic reactivities. XAFS is a powerful and indispensable tool for detecting the structural evolution of these slightly different powder samples. We selected 16 typical pellets of copper-ceria catalyst as our testing samples. Conventional XAFS measurements required 10 min for the collection of each spectrum (Ce L-III edge, −200 to +400 eV) plus 5 min between each successive scan for loading the new sample. In total, 250 min were required to complete the testing of all 16 catalysts. Moreover, the experimenter was required to remain at the XAFS station to operate the data-collection process.

In experiments employing the HT XAFS data-collection system, the copper-ceria catalysts were separated into two groups of 8 samples each. Each of the 8 powder pellets were continuously pre-scanned in a small range of energy before formally scanned to roughly determine the sample quality and the XAFS data-collection parameters. The XAFS spectrum of each sample was automatically recorded to a specific txt-format file after completion of the scan in successive scanning mode. Fig. 7 displays the Ce L-III edge XAFS spectra of these investigated catalysts. Clearly, the data-quality is very high and the profile is fairly smooth without any observable defects. Further analysis of the white line of the X-ray absorption near edge spectra (XANES) determined the different oxidation states of the Ce atom for each copper-ceria sample. The HT XAFS results were identical to those tested by the conventional XAFS data-collection mode, as indicated in Fig.7 (e), (f). However, the total collection time for the HT XAFS data-collection system was only 160 min, which is significantly less than that of the conventional XAFS data-collection mode (250 min). Additionally, the experimenter need not remain at the XAFS station after scanning is initiated, and its automatic functionality eliminates the possibility of subsequent human error.

Fig. 7Full-screen View

Ce L-III edge XAFS spectra of copper-ceria samples obtained by the HT XAFS data-collection system

More importantly, the above HT XAFS experiment of copper-ceria catalysts represents a milestone in the development of the BL14W1 beamline at the SSRF, because this was the first time that the entire XAFS system was integrated into a single EPICS system, where LabVIEW was excluded. This is of significant importance for future work, which can now be conducted under this stable data-collection system.

Further improved efficiency will be achieved by the implementation of the proposed HT XAFS data-collection system for in situ experiments. For instance, a typical in situ XAFS test includes the following steps: (1) load the catalyst and stabilize the reaction gas (1 h); (2) heat the sample to the target temperature and hold to initialize and complete the reaction (1.5 h × 3; assuming 3 different target temperatures); (3) collect XAFS data at each target temperature (1 h × 3); (4) cool down to room temperature and complete the test (1.5 h). Thus, each sample would require a total of 10 h, or 80 h (3.3 days) for 8 samples. However, the HT XAFS data-collection mode accomplishes this much more efficiently because all 8 catalysts are collectively loaded and the reaction gas stabilized (1 h), all 8 samples are collectively heated to each of the 3 target temperatures (1.5 h × 3), XAFS data is collected at each temperature for all 8 samples in succession (1 h × 3 × 8), and all 8 samples are collectively cooled to room temperature (1.5 h). The entire process would therefore require a total of 31 h (1.3 days), and 2 full days (48 h) of beamtime would be saved. Meanwhile, the reaction conditions, such as the temperature, pressure, gas composition, and space velocity, are strictly identical between the measured samples during HT XAFS data collection, which excludes the possibility of human error during data collection for a single batch of samples, and also improves the reliability of in situ experiments.

4 Conclusion

An HF XAFS data-collection system has been successfully developed at the BL14W1 beamline of the SSRF by employing specialized hardware (multi-sample holder, 2-D translation stage, voltage-to-frequency converter, and counter) and software (EPICS and CSS). Actual XAFS measurements confirmed the good performance of the implemented system. Beamtime is saved by eliminating the need to interrupt data collection to switch samples, which is particularly beneficial for in situ experiments. More importantly, the integration of all equipment control under a single EPICS system provides a very stable data-collection system, and serves as a good opportunity to upgrade our operational system for other XAFS experiments. Furthermore, the proposed HT XAFS data-collection system can also be employed for fluorescence studies because the hardware/software functionalities are similar.