1. Introduction
Macromolecular crystallography (MX) is the primary technique to determine the atomic structure of biological molecules. The MX beamline plays an important role in this field. With continuing advances in beamline instruments and experimental methods, the data collection on MX beamline has become routine [1]. However, a signification portion of samples from challenging projects still diffract poorly. The samples require more specialized experimental conditions, such as a micron-sized beam to match the size of the crystal or of the good diffraction region, a high-precision or specially configured diffractometer, or a noise-free detector to improve the signal-to-noise ratio. To provide these conditions, more attention needs to be paid to the instrument design and integration of the endstation devices.
BL17U1 is the first dedicated MX beamline at Shanghai Synchrotron Radiation Facility (SSRF). In the eight years since it opened to users in 2009, this beamline has produced more than 2200 macromolecular structures [2]. This beamline employs an in-vacuum undulator as the source. The optical setup is designed to be simple and stable to achieve a high-flux focused beam with a reasonable beam size and small divergence [3]. The endstation of the beamline receives continuous partial upgrades; for example, an active beamstop was developed [4], and a microcollimator is now used to rapidly select micron beams, but the overall performance of the endstation devices is difficult to improve.
The upgrade reported here aims to improve the overall performance of the BL17U1 endstation. The key parameters of the devices will be improved. For example, the sphere of confusion (SOC) of the diffractometer is expected to reach ≤1 µm peak to peak. The EIGER X 16M detector (Dectris Ltd.) will allow data collection at up to 133 Hz. To exploit the full speed of the detector, the reproducibility and precision of the sample stages will be specified to meet the requirements of the centering and data collection schemes. The reliability of beamline operation, apart from the improvement of the instrument performance, is also considered carefully.
The installation was carried out during the summer shutdown, and final testing was conducted in November 2017. Before the goniometer was installed, it was assembled and tested in the development lab, where the vital components underwent a site acceptance test. On the basis of the offline measurement result, the assembly was revised several times to obtain the best performance. During the shutdown, the previous goniometer was removed, and the home-integrated one was transported to the site. The EIGER X 16M was delivered to the beamline in October and installed immediately. The upgraded endstation was tested using test samples. The design and realization of this upgraded endstation are presented in this paper.
2. Hardware
The goniometer and detector are supported on separate tables to eliminate their effects on each other. The goniometer includes a single rotatory axis, a video system for observing the sample, and the shutter and beam condition components. The configuration of the hardware is suitable for manual and robotic sample changes. To reduce the effect of thermal drift, the temperature and humidity are further controlled in the endstation.
2.1 Goniometer
As shown in Figure 1, the newly built ultrahigh-precision diffractometer consists mainly of an air-bearing goniometer with a tilt error motion smaller than < 2.4 μrad, a high-precision sample centering system with a repeatability of ±25 nm, an on-axis microscope with high magnification (12 × 10) and a long working distance (~33 mm), a backlight, an ultrafast shutter, a movable beamstop with an intensity monitor, changeable collimators, and other associated equipment such as a cryostream and fluorescence detector. Owing to the advantages of the ultrahigh-precision air-bearing goniometer, high-precision sample centering mechanism, and high-magnification on-axis microscope, the newly built goniometer can achieve high accuracy and high speed in finding and positioning ultrasmall (<5 μm) crystals at the beam center.
An ABRT 260 air-bearing direct-drive rotary stage (Aerotech, Inc.) was used for the high-precision goniometer. This stage has superior angular positioning, velocity stability, and error motion performance along with an impressive payload capacity and outstanding radial and axial stiffness. Further, this stage not only has excellent axial and radial error motions, but also has outstanding tilt error motion, which is very important for our goniometer. In the synchronous model, the tilt error motion is <2.4 µrad, which means that the SOC could be less than ±0.5 µm, as the axial distance is less than 0.2 m. Figure 2 shows the test results of the ABRT 260 air-bearing stage using a DT6220-CS02 capacitive sensor (Micro-Epsilon), which show that the horizontal and vertical SOC are both < ±0.5 μm.
The high-precision and high-speed sample centering system uses the sample_x, sample_y, and sample_z motions. The sample_x and sample_y motions move the sample in the vertical plane, and the sample_z motion moves the sample in the lateral direction. As shown in Figure 1, there is a two-dimensional movement stage on top of the air-bearing rotary stage (along the axial direction), and these two motions are labeled sample_x and sample_y. SLC series piezo drive positioners from SmarAct are used for the sample_x and sample_y motions because of the high resolution of up to 1 nm or less of the new picoscale interferometer. There is also a two-dimensional stepper stage below the air-bearing rotary stage, the horizontal and vertical motions of which are defined as sample_z and gonio_vert, respectively. The resolution of these two stages is 0.5 μm. The sample_x, sample_y, and sample_z motions are used to move the sample to the point of the beam center, and the gonio_vert motion is used to move the rotation axis of the rotary stage to the microscope crosshair or beam center.
For the customized on-axis microscope from Qioptiq Company, the magnification of the eyepiece is 10×, and the magnification of the objective lens is changeable, with a maximum of 12× for a digital camera. The resolving power of the on-axis microscope is measured using a high-resolution target, 1951 USAF. The microscope can clearly distinguish element 1 in group 9 under high magnification, as shown in Figure 3, indicating that its resolution could be as low as 1 µm. Thus, this on-axis microscope is highly suitable for crystals smaller than 5 μm.
In order to align the center of the on-axis microscope with the center of the X-ray beam, a five-axis motion system with x, y, and z movement and pitch and yaw rotation is chosen, as shown in Figure 4. In the endstation of the SSRF-BL17U1 beamline, the actual X-ray beam shape can be observed using an yttrium aluminate garnet plate. A crosshair is drawn to represent the center of the on-axis microscope. Thus, the beam center and on-axis microscope center can be easily aligned by changing x, z, the pitch, and the yaw. The y direction motion is used for adjusting the focus of the on-axis microscope. In order to make the image much clearer, a backlight is positioned after the sample holder by a rotary actuator. During data collection, the backlight is moved out of the X-ray beam.
In order to simultaneously block the transmitted X-rays and monitor the intensity of the X-rays after the sample, a beamstop with an intensity monitor was designed and manufactured, and is shown in Fig. 1. A changeable collimator with pinholes of different sizes (10, 20, 50, 100, and 200 μm) is installed before the sample to provide beams of different sizes for diffraction experiments. In addition, a fast shutter is mounted before the on-axis microscope along the X-ray beam to control the exposure time in diffraction experiments. Figure 5 shows the installation of the newly built high-precision goniometer with the new EIGER X 16M detector.
2.2 Temperature and Humidity Control
A dry and thermally stable environment is very important for microbeam MX experiments. There are many instruments inside the BL17U1 endstation hutch (EH), such as the high-precision diffractometer and EIGER X 16M detector, that require temperature and humidity control in the BL17U1 EH with a temperature stability of ±0.1°C and a low relative humidity of less than 30%.
The temperature in the experimental hall at SSRF is already controlled at 25 ± 3°C, but the humidity requirement has not been considered. In particular, the humidity is very high in the plum rainy season in Shanghai from April to June. In MX experiments, a moist environment will cause icing not only of the crystal samples, but also of some devices such as the robot gripper. To reduce the thermal fluctuation and to obtain a low-humidity environment, a dedicated air-conditioned facility has been designed and installed for the BL17U1 EH.
The low-temperature chilled water method adopted for the air conditioning system of the temperature and humidity control equipment is shown in Figure 6(a). The water temperature at the outlet is -3°C, and the temperature of the return water is 2°C. To allow for a safety margin in the refrigeration conditions, three coil refrigeration setups are used, with a refrigeration temperature of approximately 7°C for each coil refrigeration setup. Finally, it is easy to obtain an air flow temperature lower than 5°C, which is the dew point temperature, to ensure a relative humidity of less than 30%.
Then two-stage temperature control is adopted for the high-precision control system. Temperature control is achieved primarily by the equipment itself, and the temperature is controlled within 24.5 ± 0.2°C when the temperature is set to 25°C. The secondary precision control unit, which was installed at the terminal of the wind pipe, is also used, and finally a temperature of 25.0 ± 0.1°C is obtained in the EH. The temperature setting is changeable, and a program optimizes the two-stage control automatically.
Vibration during operation of the air-conditioning system was considered, and it can be isolated. First, we used a flexible connection between the outlet of the fan and the air pipe. Second, the internal equipment, which consists of the fan, three coil refrigeration units, and motor, was placed on rubber shock absorbers. Thus, the vibration to the diffractometer through the ground is negligible.
The temperature- and humidity-controlled EH has operated well with stable performance for four months. The temperature and humidity data obtained for two days (Feb. 6–7, 2018) are shown in Figure 6(b). We found that the humidity inside the hutch is clearly affected by the status of the door (whether it is open or closed). However, the temperature is affected only slightly within a short door opening time, that is, less than 15 min. The reason is that the temperature difference between the SSRF hall and the hutch is very small. Therefore, the newly built air conditioning system has realized temperature and humidity control for the hutch, maintaining a constant temperature (23 ± 0.1°C) and low humidity (less than 30%RH) successfully.
2.3 Area Detector
The EIGER X 16M hybrid pixel area detector has an active area 311.2 mm in width and 327.8 mm in height. The small pixel size of 75 µm and the fact that the point spread function is limited to one pixel also make the detector suitable for large-unit-cell samples. The maximum frame rate is 133 Hz, whereas the rate for the region of interest can be as high as 750 Hz. The frame readout is continuous. The format of diffraction images is HDF5 [5], and images can be compressed for data transmission. All of these features are very important for the new data collection schemes.
The detector is installed on the original support table, which can drive the detector in six dimensions. A minimum distance of 120 mm can be obtained under the current setup. At this distance, the recordable resolution is 1.1 Å at an incident X-ray wavelength of 0.97 Å. The resolution will be 0.78 Å at 0.69 Å (X-ray energy of 18 keV) and 2.81 Å at 2.48 Å. The minimum distance could be further reduced after the redesign of the transfer board and the detector’s protective cover.
3. Control System
An attempt to hide the hardware details in the graphical user interface was made by developing several hardware services. The SSRF’s revised version of Blu-Ice has been implemented and provided to general users since 2010 [6]. Users will not see any operational difference after the hardware is upgraded. The integration of the diffractometer and EIGER detector are described below.
The hardware of the diffractometer control system consists mainly of an Aerotech multi-axis motion controller (Ensemble), a SmarAct modular control system (MCS), a Kohzu SC series motion controller, a Galil DMC series motion controller, and some independent external functional modules. The hardware structure of the goniometer control system is shown in Figure 7. Each controller serves the motion control of its own equipment and uniformly outputs useful signals to the logical I/O module of the Galil controller, which is programmable to handle the input signals and outputs accurate control signals to other devices or external functional modules.
Integration of the diffractometer is performed by different motors and video systems. Four motor systems are applied: 1) the Kohzu motor system is used for Inline_camera_vert, Inline_camera_horz, Inline_camera_pitch, Inline_camera_yaw, and Inline_camera_focus; 2) the SmarAct motor system is used to adjust the position of the collimator and beamstop, and also sample_x, sample_y; 3) the Aerotech motor system is used for sample_z, gonio_phi, and gonio_vert; 4) the EPICS system is used for the air-bearing stages. To improve the stability of the motor systems, a Kohzu distributed hardware server (DHS), SmarAct DHS, and Aerotech DHS are developed using the C++ programming language. Complex motion can be obtained by a combination of different DHSs or a combination of different DHSs with EPICS. For example, beam intensity adjustments are made by a combination of the Aerotech DHS, Kohzu DHS, and SmarAct DHS, and crystal centering is achieved by a combination of the Kohzu DHS and Aerotech DHS; scanning is obtained by calling a function of the Aerotech DHS. The video systems are divided into an inline camera video system, robot video system, and hutch video system. The online camera video system is used for crystal centering. The robot and hutch video systems are used for monitoring the status of the robot and hutch, respectively.
To integrate the EIGER detector into Blu-Ice, an EIGER DHS is developed using the C++ programming language. All the functions of the EIGER detector are integrated into the EIGER DHS, and the EIGER DHS communicates directly with the EIGER detector control unit; therefore, the Blu-Ice system can control the EIGER detector through the EIGER DHS.
Data can be collected from the EIGER detector using a single-image testing model; a testing image can be collected for initial screening of the quality of crystals. A C++ program was developed to read HDF5 images, and an HDF5 diffraction image is automatically shown in Blu-Ice. Salt crystals and poorly diffracting crystals are not included in large-wedge dataset collection. The beam time and automatic data processing time can be significantly reduced after initial screening using the single-image testing model. Moreover, data collection can be performed using a shutterless model for large-wedge (up to 3600°) data collection. A full dataset containing a master file and multiple HDF5 image files is obtained from the EIGER DHS. The packed HDF5 image number is automatically calculated by dividing the total data collection time by 20 s; therefore, each packed HDF5 image is generated from the EIGER detector with a 20 s serial interval. The advantages of using a 20 s serial interval for each packed HDF5 images are as follows: 1) users do not need to set up the frame number in each packed HDF5 image file, which can significantly save beam time; 2) a 20 s serial interval can provide the necessary time for HDF5 image processing to show in Blu-Ice for monitoring by users.
4. Data Collection
Compared to the original experimental setup, the current instruments offer more precise positioning during sample rotation and continuous readout from the detector. Owing to the features of the EIGER X 16M, such as the fast readout and the absence of readout noise, the ultrafine-φ slicing data collection strategy [7], which has proved to be capable of obtaining better data quality than fine-φ slicing [8], is recommended. For the EIGER X 16M, an optimal delta phi setting of 1/10 of the mosaicity is recommended, although this setting requires huge storage and powerful computational capabilities.
After equipment installation and control integration were completed, standard lysozyme samples were used to test the data quality of the beamline. The crystallization condition of the lysozyme was NaAc, pH 4.6, 8% NaCl at 18°C. Crystals were cryoprotected with a cryoprotectant before being frozen in liquid nitrogen. The wavelength was 0.97918 Å. The exposure parameters are a delta angle of 0.2° per frame and a 0.2 s exposure, and the total number of frames is 1800. Data were processed using the xia2 software suite. The diffraction image in Figure 8, which was obtained from the merged raw images, was offered for inspection by users. The crystallographic statistics are shown in Table 1.
| Parameter | Value |
|---|---|
| Data collection | |
| Space group | P43212 |
| a (Å) | 78.35 |
| b (Å) | 78.35 |
| c (Å) | 36.97 |
| α, β, γ (°) | 90, 90, 90 |
| Resolution (Å) | 36.97–1.35 (1.39–1.35)a |
| Rmeas (%) | 6.0 (127.5)a |
| <I/σI> | 17.8 (1.7)a |
| Completeness (%) | 99.7 (99.7)a |
| Redundancy | 12.2 (12.2)a |
5. Discussion and Conclusion
The first self-integrated diffractometer, a temperature and humidity control facility, and the EIGER X 16M give the BL17U1 endstation a new look. The beamline is now in user operation after fast commissioning. Stable operation during user operation proves that the design of the high-precision goniometer is successful. Users’ feedback on the new system will be recorded for use in developing the next version of the goniometer.
The EIGER detector not only speeds up data collection, but also makes new experimental schemes possible. The typical experimental setup parameters, such as the oscillation angle, exposure time, and beam attenuation, will be gradually transferred from the mode under the charge-coupled device detector to that under this EIGER detector. The huge amounts of raw data require greater storage capability and improved data analysis methods and resources. The beamline staff will cooperate with the user community to exploit the new experimental schemes to facilitate users’ projects.
Current advances in synchrotron radiation instrumentation for MX experiments
. Arch. Biochem. Biophys. 602, 21-31 (2016) doi: 10.1016/j.abb.2016.03.021Biosync: A Structure Biologist's Guide to High Energy Data Collection Facilities (2017)
,http://biosync.rcsb.org, AccessedThe macromolecular crystallography beamline of SSRF
. Nucl. Sci. Tech. 26, 010102 (2015) doi: 10.13538/j.1001-8042/nst.26.010102An active beamstop for accurate measurement of high intensity X-ray beams
. Nucl. Instrum. Meth. A 735, 584-586 (2014) doi: 10.1016/j.nima.2013.10.011Control and data acquisition system for the macromolecular crystallography beamline of SSRF
. Nuclear Techniques 35, 5-11 (2012) (in Chinese)EIGER detector: application in macromolecular crystallography
. Acta Cryst. D 72, 1036-1048 (2016) doi: 10.1107/S2059798316012304Optimal fine φ-slicing for single-photon-counting pixel detectors
. Acta Cryst. D 68, 42-56 (2012) doi: 10.1107/S0907444911049833