The PRT has stated its desire to emphasize ease of use, with performance maximized for the most common crystallographic problems such as Se-Met MAD experiments, ligand binding experiments on previously-solved structures, and hot projects requiring rapid access. Individual PRT members have specialized programs at X8-C ranging from industrial crystallography to kinetic crystallography. The technical capabilities of the beamline are listed below in some detail.
Flux: X8-C is unusual among the NSLS bending-magnet beamlines in that it has a parabolic collimating mirror placed upstream of the monochromator, which gives it a higher flux than most. A comparison among NSLS beamlines used for protein crystallography puts the relative flux through a 200-micron pinhole at 12 keV at roughly 5 times that of X12-B, X12-C, and X-9, and a factor of 18 below that of the X-25 wiggler line.
Wavelength range: Among the NSLS protein crystallography beam lines, X8-C has an unusually large wavelength range available. The practical range of the beamline is 2.5 to 0.65 Å (5 to 19 keV). These wavelengths are easily reachable under user control in a matter of minutes. This large wavelength coverage is potentially useful for high-resolution structures, large-unit-cell structures, very small or large crystals, and unusual anomalous scatters. Note that the mode of operation of the NSLS ring (2.5 GeV versus 2.8 GeV) makes a big difference in the spectrum at low energies. For example, in 2.8 GeV operation the flux at 5 Å is reduced by a factor of 10 versus 2.5 GeV operation while the flux at 0.7 Å is enhanced by about a factor of two. If your project needs one extreme of wavelength or another, you may wish to take this into account when choosing beam times.
MAD: The collimating mirror optics also deliver flux at a lower bandpass than a simple entrance-slit monochromator. For experiments such as Se-Met MAD, this results in an additional factor of 3 improvement when compared against the X12 and X25 beamlines which need to narrow the slits to deliver the low bandpass required. X8-C is also very stable in wavelength and has demonstrated a drift over monthlong shutdown periods of less than 1 eV. The lower heat loads on the monochromator compared to a wiggler line generally make for less drift. DOE-funded methods development on MAD is being done at X8-C, and user programs directly benefit from this research program. In the last few months of operation a number of MAD structures have been solved here, a few of them in real time with SOLVE. MAD structures have been solved at X8-C from Se, Br, and Sm, among other metals.
Detector: In January, 1999, X8-C accepted delivery of the first ADSC Quantum 4R detector. This is a 2304 by 2304-pixel mosaic design with a square active area 187 mm on a side. Direct coupling of the fiber optics to the CCD's gives it about 40% more light throughput than previous versions of the detector; the sensitivity is thought to be about 0.8 sigmas per 12 keV X-ray photon. This design also has an improved point-spread function Dark currents are less than one 12-keV X-ray equivalent per 5 minutes. Data transfer times are under 9 seconds in full-size, low-noise readout mode. Fast-readout and 2x2-binned readout modes are available. The former mode seems to give data statistics of nearly equal quality as the slow-readout mode for many crystallographic problems, at a savings of 3-4 seconds per transfer.
Camera: A MAR rotation camera is presently being used with the Quantum 4R. The crystal-detector distance can be adjusted in the range 50-300 mm. At the minimum distance, the limiting 2-theta angle is some 62o. Two pairs of manual slits on the MAR allow the beam size to be match to the crystal size or reduced to minimize spatial overlap of diffraction spots. The rotation rate should be kept less than 6 seconds per degree. If you have strongly-diffracting crystals that give overloads even at this rate (as we have seen in 300-micron-thick myoglobin crystals), we recommend that you collect at a shorter wavelength or with smaller slit sizes or both.
Maximum resolution: Besides the wavelength and crystal-detector distance dependence from Bragg's law, this value depends via the overlap limit on the size of the unit cell, slit size, and beam optics. Given the detector and camera geometries listed above, at present the maximum resolution obtainable in the best case for strongly diffracting crystals with small unit cells is approximately equal to the wavelength of the radiation used. An excellent 1.0 Å data set was collected at X8-C using 0.8-Å radiation, and the flux is reasonable to wavelengths as short as 0.65 Å, so data sets with resolutions as high 0.65 Å are possible at X8-C.
Largest unit cell: The size of the largest unit cell that can be measured is a complicated matter that depends on the slit size, diffraction power, resolution limit, wavelength used, divergence of the beam, and time available. At present, no offsetting of the detector is possible (but that is expected to change by January 2001). One practical limit arises from the 20-micron pixel size of the Quantum 4 detector, which implies that 50 microns is a practical minimum spot size to avoid spatial undersampling. At this slit size, crystals with unit cells in the range of 400 Å are probably feasible. A MAD structure from a crystal of 380 Å cell axis has been solved.
Cryo hardware: At present, the system accomodates cryocrystallography using either the Hampton or Yale standards over a broad range (3-40 mm) of needle lengths. Goniometers and magnet bases are available at the beamline, although it never hurts to bring your own if you use unusual hardware or have strong preferences. The cryosystem is an Oxford Cryostream with automatic fill from a 120-liter dewar. The system can go for approximately a week with unattended operation. A hutch cooler/dehumidifier system will be installed during August, 1999.
Computing: There are several of the latest workstations at the
beamline to support data collection, processing, display, and archival.
A 633-MHz DEC Alpha with 500 MB of memory and one 500-MHz Alpha with 384
Mb of memory are the primary machines for data processing. They can
receive, process, and store one 10 Mb Quantum 4 image file in under
two seconds. Re-reading the file and integrating it under DENZO or
similar software take on the order of 5-7 seconds each. A SGI O2
workstation is used to display graphics and run some common crystallography
packages at the beamline.
A Linux system runs the beamline controls and is a server for backup
media. There are currently about 50 GB of disk storage available
for user data storage.
DENZO/HKL program suit is provided for data anylysis. The curent version installed on SGI O2 and DEC Alfa is 1.96. It is provided by the authors free of charge under certain conditions.
It's important that before you visit X8-C you realize that the beamline and CCD can produce data very fast for strongly-diffracting crystals. For the myoglobin crystals mentioned above, the aggregate data production rate is close to 1 MB/second and can be sustained for hours. This is twice the rate at which a 8mm 8500-mode tape drive can back up the data. It doesn't make sense for you to have 2 days of beam time at X8-C, only to spend the next 4 days backing up your data. To help with this problem, we have a variety of backup devices available. Some of these devices can store frames at a rate of one second per frame.
Other important computer resources are a 100Mb/sec switched network, and printers at the beamline and apartment. The beamline hardware is controlled by a Linux computer system running the ACE program with the GRACE graphical front end. This may be familiar to those who have used it at X12-C, X25, and elsewhere around the NSLS.
Lab space: For those of you who bring crystals in a form other than frozen, there is a wet lab space associated with the beam line (room 1-164). It is well-equipped for crystallization and crystal mounting and includes half of a cold room.