short communications
Front-end XY-slits assembly for the SPring-8 undulator beamlines
aJAERI-RIKEN SPring-8 Project Team, The Institute of Physical and Chemical Research (RIKEN), Kamigori-cho, Ako-gun, Hyogo 678-12, Japan
*Correspondence e-mail: oura@sp8sun.spring8.or.jp
A front-end XY-slits assembly has been designed for the SPring-8 undulator beamlines. This assembly can handle the high from the undulator, its grazing-incidence L-shaped configuration employing an enhanced heat-transfer technology.
1. Introduction
The front-end XY-slits assembly (FE-SLIT) is one of the high-heat-load components in the SPring-8 undulator beamlines (Sakurai et al., 1995; Kitamura, 1996). The main objective of the FE-SLIT is to reduce the heat load on the Be windows and first optical element by cutting off the off-axis part of the undulator beam. It consists of a set of L-shaped slits equipped with precision horizontal and vertical linear actuators and associated bellows for lateral movement, as shown in Fig. 1.
The L-shaped slit will be impinged by a very intensive undulator beam with 451 kW mrad−2 peak in the worst case. To handle such a high the L-shaped slit has been designed to be of a grazing-incidence configuration so as to stretch out the heat load on the slit-blade. In order to reduce the heat load on the slit-blade further, the assembly is installed downstream of the pre-slit assembly (Takahashi et al., 1998), whose circular-aperture exit is 4 mm in diameter, and the graphite filters assembly (Sakurai et al., 1995). Even if the aperture size is limited by the pre-slit, the FE-SLIT should handle almost 5 kW of radiated power when the undulator gap is set to be minimum.
In this article, we describe the design features of the FE-SLIT and its thermomechanical properties. The installation of the assembly and its utilities are also described.
2. Design features of the FE-SLIT
The design of the shape of the grazing-incidence slit-blade is based on the APS-L5 slit (Shu et al., 1995), with suitable modifications for direct cooling. The grazing-incidence angle is fixed at 1.57° on the vertical blade surface. This angle will decrease the peak power from 540 to 14.8 W mm−2 on the slit-blade. The slit-blade is made of a single Glidcop block and is formed as part of the inner wall of the vacuum vessel. Both ends of the Glidcop block are explosively bonded to stainless-steel UHV joints. Fig. 2 shows a conceptional view of the FE-SLIT. The upstream XY-slit consists of left-vertical and lower-horizontal blades. The downstream XY-slit consists of right-vertical and upper-horizontal blades. Each slit block has two cooling channels and is water-cooled; the convective heat transfer coefficient is enhanced by the use of copper mesh brazed into the cooling channels (Kuzay et al., 1991). Each slit block is mounted on an X translation (transverse to the beam and in a horizontal axis) stage and a Z translation (transverse to the beam and in a vertical axis) stage to allow the vessels to be positioned with respect to each other such that a rectangular aperture size can be adjusted. The FE-SLIT can be manipulated into a variety of aperture shapes with sizes from 10 × 8 down to 0 mm (H) × 0 mm (V) by controlling the XZ actuators. As we can see in Fig. 2, the aperture size can be set to enable most photons of fundamental radiation to pass through the FE-SLIT but most waste photons to be shut off. The aperture size of the FE-SLIT, however, should be determined to secure safety in utilizing downstream components such as Be windows and/or a first optical component (Oura et al., 1998).
Each actuator, which is motor-controlled, has a datum switch and two limit switches. These limit switches are set to avoid over-run resulting in accidental irradiation onto parts of the downstream beamline components. The resolutions of linear motion of the X-stage and the Z-stage are 1 and 0.25 µm pulse−1, respectively.
3. Thermomechanical analysis
As was mentioned in §1, the FE-SLIT should handle almost 5 kW of radiated power in the most severe condition, even if the pre-slit restricts the aperture to a circular diameter of 4 mm. In order to understand the thermomechanical properties of the XY-slit regarding maximum surface temperature and thermal stress, finite-element analyses were performed using ANSYS for some cases of full irradiation onto the vertical blade and partial irradiation, in which the aperture of the FE-SLIT was set to 2 × 1, 1 × 0.5 and 0 mm (H) × 0 mm (V). The dependence of the surface temperature and the thermal stress on the aperture size were studied.
In the ANSYS finite-element analyses, a model of the XY-slit was made for an L-shaped blade, whose length was 77% of the final design in the beam direction, having two cooling channels. The modelled XY-slit was assumed to be located at 30 m from the light source and 2 m downstream of the pre-slit assembly. The spatial distribution of the absorbed power was approximated by
where h and v are, respectively, the horizontal and vertical distances measured from the centre of the beam axis, PT is the total power, σv is the standard deviation of the spatial distribution of the radiated power in the vertical direction, Wh is the half width of the spatial distribution of the radiated power in the horizontal direction, and δ(h,v) is the δ function, which becomes unity when the coordinate (h,v) resides in the projection of the pre-slit on the XY-slit. Table 1 shows the material properties and the quantities used in the analyses. We assumed a convective heat-transfer coefficient of 15000 W m−2 K−1 in conservative calculation. The temperature of the cooling water was assumed to be 303 K.
|
3.1. Maximum surface temperature and thermal stress
3.1.1. Full irradiation onto the vertical blade
In the ANSYS calculation, the location of the beam centre was assumed to be at the centre of the vertical blade. This situation is, however, the special case in which the X-stage overruns due to an accident or it is actuated intentionally. The calculated maximum temperatures of the vertical blade and the surface of the cooling channel are about 510 and 373 K, respectively. The corresponding maximum equivalent stress, i.e. Mise's equivalent stress, is calculated to be about 254 MPa.
In the case of irradiation onto the horizontal blade, the footprint of the beam extends much more than in the case of vertical-blade irradiation, so that the resulting maximum temperature for the same cooling condition is less severe.
3.1.2. Partial irradiation
The results from the ANSYS calculation are summarized in Table 2. In the case of zero aperture size, all the photon beam is cut by the upstream XY-slit and the downstream XY-slit with the ratio 3:1, respectively. Although this situation is the most severe case possible in normal operation of the beamline, the corresponding maximum equivalent stress is less than the yield strength of Glidcop (SCM Metal Products, 1994).
|
When the aperture of the FE-SLIT is set to allow the most photons of fundamental radiation to pass through, the resulting maximum equivalent stress may be much less than the yield strength. In such a situation, however, we should notice that there is some thermal distortion along the L-shaped edge of the slit-blade leading to a small change in the aperture size of the FE-SLIT. Estimation for such a problem strongly depends on the condition of restraint. A rigorous analysis is in progress.
4. Installation of the assembly and its utilities
The centre of the FE-SLIT is located at 28.9 m from the undulator and the total length of the assembly is 1.935 m. A photograph of the FE-SLIT is shown in Fig. 3. This is installed inside the shielding wall and is located downstream of the pre-slit assembly and the graphite filters assembly so as to reduce waste photons on the slit-blade. A separate water-cooling system is mounted on the common base just beneath the assembly. This consists of two cooling manifolds and flexible stainless-steel hosing for connection to the slit bodies via Swagelok connectors. Each manifold has two pressure gauges and two different types of flow monitor. Fig. 4 shows a schematic drawing of the water-cooling system. The pressure gauges and one of the flow monitors give outputs of 4–20 mA DC. By using those outputs, we can monitor both water pressures (Pin and Pout) and the flow rate outside the shield wall. We can also obtain information about the pressure drop (Pin − Pout) due to the mesh brazed inside the cooling channel. After the installation of the first XY-slits assembly into the front end of the pilot undulator beamline (BL47XU), we measured the pressure drop of the cooling water as a function of the flow rate. Fig. 5 shows the relationship between the pressure drop and the flow rate. The flow rate in abscissa is for the gross water flowing in two cooling channels. The pressure drop is proportional to the square of the flow rate and the which depends on the Reynolds number. Since there are so many influencing factors, it is not easy to discuss quantitatively the curves in Fig. 5. We can suppose, however, a heat-transfer coefficient by making a comparison with the experimental data obtained by the APS group (Kuzay, 1996). For the cooling channel we expect that as much as a four- to fivefold increase in the heat-transfer coefficient relative to a plain channel is possible.
References
Kitamura, H. (1996). HOSHAKOU, 9, 403–412. (In Japanese.)
Kuzay, T. M. (1996). Private communication.
Kuzay, T. M., Collins, J. T., Khounsary, A. M. & Morales, G. (1991). ASME/JSME Thermal Engineering Proceedings, Vol. 5, pp. 451–459. ASME (Book No. 10309E-1991).
Oura, M., Sakae, H., Sakurai, Y. & Kitamura, H. (1998). J. Synchrotron Rad. 5, 609–611. Web of Science CrossRef CAS IUCr Journals
Sakurai, Y., Oura, M., Sakae, H., Usui, T., Kimura, H., Oikawa, Y., Kitamura, H., Konishi, T., Shiwaku, H., Nakamura, A., Amamoto, H. & Harami, T. (1995). Rev. Sci. Instrum. 66, 1771–1773. CrossRef CAS Web of Science
SCM Metal Products (1994). Catalogue of Glidcop, SCM Metal Products, Inc., Research Triangle Park, North Carolina, USA.
Shu, D., Brite, C., Nian, T., Yun, W., Haeffner, D. R., Alp, E. E., Ryding, D., Collins, J., Li, Y. & Kuzay, T. M. (1995). Rev. Sci. Instrum. 66, 1789–1791. CrossRef CAS Web of Science
Takahashi, S., Sakurai, Y. & Kitamura, H. (1998). J. Synchrotron Rad. 5, 581–583. Web of Science CrossRef CAS IUCr Journals
© International Union of Crystallography. Prior permission is not required to reproduce short quotations, tables and figures from this article, provided the original authors and source are cited. For more information, click here.