Abstract

Inclusions and other inhomogeneities in explosion bonded copper/steel high vacuum components for a synchrotron storage ring are easily detectable by ultrasonic immersion testing. A focussing 25 MHz probe detects flaws down to the grain noise level of 0.2 mm equivalent reflector size by pulse echo technique. Some natural flaws are solely observed through their masking of the back-wall echo.

Table of Contents
• Test object: Absorber in an Electron Storage Ring
• Inspection objective: deficient bonding
• Test Conditions
• Test Results
• Conclusions
• References

Test object: Absorber in an Electron Storage Ring

At Science Centre Berlin-Adlershof in Germany, the second Berlin Electron Storage Ring BESSY-II is currently under construction. In it, electrons are accelerated to nearly the speed of light, and stored for many hours in a polygonal round path. In the deflection magnets they will emit so-called Synchrotron Light - white electromagnetic radiation of high intensity and frequencies ranging from Infrared to X-rays. Therefore it is highly suitable for manifold experiments in physics and materials science. The energy that is emitted during each deflection is radiated fan-like into tangentially oriented beam lines placed around the vacuum chamber. While most of the energy is guided onto monochromator devices and further to the end user's rigs, a small fraction will miss the beam line and hit the chamber wall in the plane of the round track. Hence the Vacuum chamber needs an absorber rail with water cooling mounted circumferentially around the beam path to carry off the excess energy, due to the high electron current and energy density that is desirable for a most efficient operation of the photon source. The stainless steel water cooling conduit of the absorber rail, operated under normal atmospheric pressure, needs crucially a tight and complete joint to the inner absorbing surface, which lies in the ultra high vacuum region of the chamber and is made of copper for good heat conductivity.

Figure 1. Cross section of the absorber rail. The Ultrasonic test is carried out on the complete rectangular part. Afterwards the corners and the cooling pipe are milled off.

Inspection objective: deficient bonding

Inadequate bonding between the stainless steel cooling pipe and the copper absorber because of inclusions, etc., may cause damage to the component due to the different thermal expansion coefficients involved, together with possibly severe local overheating.

Additionally, the copper is partially machined down to the bonding layer between steel and copper, so as to obtain mounting edges for a later welding of the rail to the rest of the component (see fig. 1). Possible voids in the bonding layer will thus be opened to the vacuum, resulting in poor final vacuum accompanied by extended evacuation time. Such voids are thus referred to as 'virtual leakage'.

In the described case it was decided to use explosion cladding as the bonding procedure between steel and copper in order to avoid the disadvantages of soldering such as virtual leakage, which are reported in the literature.

Test Conditions

The large copper and steel plates undergo an explosion cladding process and are afterwards cut into rectangular 1 meter long aligned rails with smooth surfaces. Rail widths are 14 mm for steel and 10 mm for copper. The rails are bonded to a complete length of far more than 100 m and the final absorber rail is built over the circumference of the vacuum chamber. Because of the long rail length it is a precondition that the bonding be free of inhomogeneities of less than 0.5 mm. For that reason 100% nondestructive testing is required using mechanized ultrasonic immersion technique with a focussing 15 MHz probe. A C-scan with a typical scan resolution of 0.2 mm generated the C-scan images. The impulse echo technique was used for the evaluation of the bonding area and the back-wall echo amplitude was recorded simultaneously (Fig 2).

Fig 2: Evaluation of the interface echo and back-wall echo in immersion testing

A reference block containing a series of flat bottom holes was manufactured to determine the test sensitivity and the signal/noise level. The artificial flat bottom holes (FBH) of 1.0 to 0.2 mm diameter were drilled into the steel side just reaching the boundary. The test used a 25 MHz focussing probe focused onto the boundary. The examination was carried out using 0.1 mm scan resolution (pitch).

Test Results

The C-scan image 3 shows the results of the reference block, starting left with 1 mm flat bottom holes to 0.2 mm right. The upper part displays the reflection amplitude from the FBH at the boundary while the lower shows the amplitude of the back-wall echo (shadow of the FBH). The calculated signal to noise ratio for the direct reflected pulse echo mode was better than 10dB and for the back-wall echo (shadow method) better than 16 dB (both at 0.2 mm).

Figure 3. Ultrasonic results from a reference block containing flat bottom holes of 1.0 to 0.2 mm. The Interface echo (top), Back-wall echo amplitude (below)
Explosion bonding techniques generally show a ripple. This ripple can be used for the quality control of the cladding process. The ripple generates a variation only of the back-wall echo amplitude (fig. 3 bottom), however no direct reflection (fig. 3 top).

Image 4 shows the result of the 0.2 mm FBH in direct impulse echo technique (top) and the back wall echo shadow technique (bottom). Both display similar good results for the test reflector while only the back-wall echo shows the boundary ripple.

Figure 4. 0,2 mm FBH: Echo of the interface (top) and backwall echo amplitude (below)

Image 5 shows the 0.4 mm FBH together with a natural inclusion, again in the upper part the boundary echo and in the lower the back-wall echo . In both images the FBH shows a round display at the bottom left side and the natural flaw is displayed at the upper right side of each image. The FBH images in both pictures are similar while the image of the natural flaw appears very different depending on the method. To get a better idea about the detection probability of both methods, more artificial flaws were provided from the copper side: spark-eroded 0.12 mm and 0.21 mm holes with approximate spherical bottoms. conical drill holes of 0.3 mm diameter. conical steel inclusions of 0.3 mm diameter.

Figure 5. Echo of the interface (upper image part) and back-wall echo (below image part): 0.4 mm FBH (round, left), at right the natural flaw

Image 6 shows the inhomogeneities displayed by the result of the direct impulse echo technique (upper) and the result of the back wall echo shadow technique (lower). Only the steel inclusions shows a directly reflected echo, while all other inhomogenities were detected only by the back-wall echo method.

Figure 6. Various artifical defects, direct reflection (top) and back-wall echo (below)

Image 7 displays the 0.4 mm FBH in a B-scan image. From its upper part it is evident that copper exhibits higher grain noise than steel. Image 7 shows in its bottom the ripples as well as the FBH shadow evaluated by the back-wall echo amplitude. The direct reflection is visible in the lower third of the image.

Figure 7. B-Scan image (perpendicular cross section) through the 0.4 mm- FBH; on top the entry echo, on bottom the back-wall echo. The echo of the drill hole is visible in the lower third portion.

Image 8 displays the result of a big natural flaw of a clad rail component, reflection (upper) and shadow (lower), tested with a 15 MHz probe. Parts containing such big defects were rejected. A reference level of a 0.5 mm FBH was used for determination of the acceptance level, which was was set to -6 dB or 50% echo amplitude of the reference level.

Figure 8. Natural defect as direct echo result (top) and back-wall echo (below), with use of a 15 MHz focused probe

Conclusions

Inclusions and other inhomogeneities of down to 0.2 mm equivalent FBH in an explosion clad component of 5 mm austenitic steel or copper can be clearly detected. Conventional ultrasonic equipment together with a focussing 15 MHz or 25 MHz probe can be used for that task. The back-wall echo should be evaluated in conjunction with the evaluation of the directly reflected flaw echo since many natural flaws in its orientation and geometry are only detectable by the shadow of the flaw.

References

1. Yuheng, Li; Deming, Shu; Tuncer M. Kuzay: Explosion Bonding of Dissimilar Materials for Fabricating APS Front End Components. Advanced Photon Source, Argonne, National Laboratory, June 1994
2. Samal, P. : The Metal Science Of Joining. Proceedings of TMS lnternational Symposium, 296 (1991)
3. Anderson, D. et al.: Explosive Welding. The Welding Institute, Cambridge, Eng- land, 1976
4. Kleven, St.: Ultrasonic Inspection of Explosion Welded Titanium Clad Plate. Materials Evaluation Vol. 54 Nr.5, May 1996
We express our gratitude to Carl Hanser Verlag, München, for permission to publish this article from Materialprüfung 1996 Vol 38 No.10.

Author