A new high heat load x-ray tube

Arthur H. Iversen

Coriolis Corporation, 15315 Sobey Road, Saratoga, California 95070

Stephen Whi taker

Department of Chemical Engineering University of California at Davis, Davis, California 95616

Abstract

A new class of internally liquid cooled rotating anode x-ray tubes capable of high average and high peak power is proposed for medical applications. The principles of this new tube design are discussed and estimated performance levels are compared to levels attainable wi th conventional solid rotating anode medical x-ray tubes and with existing liquid cooled rotating anode x-ray tubes. Performance improvements of the proposed tube include essentially unlimited anode Heat Uni t loading, ana anode cooling rates, peak and continuous, that are more than an order of magnitUde greatel;" than is currently available. The new tube would enable present high average power demanding procedures such as CT and digital and conventional vascular angiography to be performed more efficiently and with higher patient throughput. New imaging techniques requiring even higher average power levels such as energy subtraction, slit scanning, x-ray spectrum optimization and special scatter rejection methods would become more clinically practical.

Limitations of present medical x-ray tube performance

In medical radiography, system performance hinges largely on the operating characteristics of the installed x-ray tube. Rotating anode x-ray tubes have not altered in any substantial manner since their commericalization in 1932. Increases in anode size and rotational speeds have prov ided rela ti vel y modest increases in both peak and aver age power. However, demand s on x- ray tube performance have risen dramatically in recent years. Starting wi th vascular examinations and other demanding special procedures and more recently wi th CT and digital subtraction angiography, average power requirements have increased resulting in tube performance and life problems. Tube manufacturers have been impelled to expend large sums to increase tube anode ratings from the previous limit of 400,000 heat units to the current 1 plus million heat units. This is reflected in a virtual doubling of tube prices. Based on the foregoing, it would be expected that the technology and costs required to increase tube performance to a 10 million heat unit rating would be formidable.

New imaging techniques with promise for advances in diagnostic capability and quality will require further increases in the average power capabili ties of x-ray tubes. For example, energy subtraction has been shown to have clinical potential in conjunction with digital radiography. Filtering techniques for x-ray spectral optimization can improve image quality, i.e. contrast, while reducing patient dose~ However, this technique requires about a 3X increase in average power~ Slit scanning devices will also place heavy demands on average power because of the inefficient use of the x-ray output of the tube resulting in a 6X or greater increase in average powers requirments.

Conventional rotating anode x-ray tubes operate under two sets of constraints; one is the sing Ie exposure technique limi t dictated by focal track and focal spot temperature limi ts, and the second is the heat storage, (Heat Unit (HU) rating) and cooling characteristics of the anode. The permissible single exposure technique is further modified by the temperature profile of the anode at any point during a procedure.

Tube operating limitations require system performance compromises which balance tube life with focal spot size and thus image quality as well as patient throughput. A typical 0.6mm focal spot, 40 KW tube, i.e. a cold anode maximum single exposure of 100KV, 400mAand O.ls, could handle over 90% of the patient load. In heavy duty applications, a 1.2mm (100 KW) tube is used because of unacceptably short tube life when a 0.6mm tube is used. This short life results because for thick body parts the 0.6mm focal spot tube must be operated at or near its cold anode specified rating. Focal spot temperature calculations demonstrate the problem and show the dramatic drop in focal spot temperature, at a given load, by using a larger focal spot. For example, at a 40KW technique factor on a 4" diameter anode, 13· target angle rotating at 10,000 RPM, focal spot temperatures are obtained using the equation,

TABLE 1

ATo (c) = 2376 w 3/2

where w is the focal spot size in mm,

Focal Spot Size mm

0.6 1.2 2.0

Focal Spot Temperature

5112· C 1807" C

840' C

/, r I - f '+)'+ ,~~. n ''4;

It can be seen that increasing the focal spot size from 0.6mm to 1.2mm provides a dramatic reduction in focal spot temperature, resulting in lower operating temperatures and consequent longer tube Ii fee However, the price paid is greater image unsharpness resulting in a less diagnostic radiograph. The temperature for the 0.6mm focal spot is shown to be above the mel ting point of tungsten. However, NEMA standards allow the focal spot size to be larger than the specified nominal size. A 40% increase in size along both axes would reduce the temperature from S16SoC to 3118"C for the 0.6mm focal spot tube, whereas the 1.2mm focal spot temperature would reduce to 1100"'C. A typical upper temperature limit for the focal spot is 2600· Cr

Furthermore, as cumulative anode heat load increases, the permissable focal spot temperature decreases correspondingly. This in turn dictates tube derating. In CT, for example, tube ratings are reduced by about 70% from their peak instantaneous rating. If the peak instantaneous power of a given tube could be used for all exposures, CT scan times could be shortened substantially. This would reduce patient motion unsharpness in high resolution scans. Similar benefits would be obtained in certain special procedures where the choice of a smaller focal spot to provide higher resolution could be made.

A further desirable tube characteristic would be higher anode heat (HU) loading and more rapid anode cooling. Incremental improvements in this area of tube performance are being obtained by the use of anodes that are more massive or made of exotic materials. Improvements in anode heat storage and faster cooling rates permit heavier focal spot loading during a procedure and also reduce the cooling time between exposures. In CT, 20 slices are typically taken on a patient and the average procedure time is 50 minutes. If the average cooling time between scans is 30 seconds, then 10 of the 50 minutes, or 20%, is required for anode cool ing. A further benefi t of reduced cooling time might be the development of new diagnostic techniques requiring a series of rapid CT slices.

Performance of current liquid cooled rotating anode x-ray tubes

Much improved tube performance with regard to average power capability, anode coling rates and bearing performance can be obtained from internally liquid cooled rotating anode x-ray tubes. Tubes of this type have been used for over 30 years in x-ray crystalography. The mcDor manufacturers are Rigaku in Japan and Elliott in England. A review of the specification~for these tubes shows that heat transfer efficiency at the anode to liquid heat exchange surface is generally in the range of 100w/em~to 500 w/em~ The largest tube, the Rigaku RU-IOOO, produces 60KWcontinuous duty with a Imm X 10mm focal spot, but requires a 40 em (16") diameter anode to do so. The Soow/em-dissipation capability dictates the size of the anode. A more efficient heat transfer surface would permit a smaller anode. By comparison, medical rotating anodes are generally 10 cm (4 ") in diameter. In addi tion, because of the inefficient heat transfer at the anode surface, the most efficient heat transfer medium, namely water, is needed to obtain the stated cooling efficiencies. However, water poses certain design contraints. Because it is conductive, water cooled anodes are operated at ground potential. Therefore tubes of this design cannot be converted to medical use without replacing the split potential generators, cables, etc., an expensive proposition. Moreover, water generated corrosion creates maintenance problems.

It is apparent that if liquid cooled rotating anode x-ray tubes are to be useful in medical applications a more efficient heat transfer surface must be developed to reduce the size of the anode and also a coolant other than water, i.e. a dielectric, must be used so that the anode can be operated at elevated voltages.

A solution to the above requirements is to be found in the Rotodyne heat transfer concepts. The proposed Rotodyne X-Ray Tube enables high peak power to be achieved along wi th high average power in an efficient, long-lived design. This contrasts with moderate peak and average power capabilities of current liquid cooled rotating anode x-ray tubes, and the high peak and low average power ratings of medical rotating anode x-ray tubes. Whereas present liquid cooled rotating anode x-ray tubes are power limited at the maximum anode-to-liquid-coolant heat exchange rate, the Rotodyne tube has as its power limit the melting point of the anode at the focal spot. The Rotodyne tube derives its substantial increase in output power from a greater than ten-fold improvement in liquid cooling efficiency.

Construction of liquid cooled rotating anode x-ray tubes

Referring now to Fig. 1, the basic structure of a liquid cooled rotating anode x-ray tube will be described. A hollow anode 20 attaches to a hollow rotating shaft 21. A rotational fluid seal 22 is mounted at the end of hollow shaft 21. A stationary cupped cylindrical attachment 23 wi th entrance duct 24 is mounted to rotational fluid seal 22. A stationary tube 25 disposed concentrically with, and extends through, stationary hollow cupped cylindrical attachment 23. A hermetic seal is provided between attachment 23 and stationary tube 25. Stationary tube 25 extends longitudinally, and concentrically, within hollow rotatable shaft 21 into the hollow rotatable anode 20. A stationary septum 26 is mounted on hollow stationary tube 25, and disposed within hollow anode 20. Hollow anode 20 is rotatably coupled to stationary septum 26 by a rotational bearing 27 and a fin shaped radial support and centering structure 28 attached to the inner, stationary segment of bearing 27.

A rotatable bearing member 29 including an inner rotating segment 30 and outer stationary segment 31 is utilized to rotatably couple rotatable shaft 21 to a mounting member 32 and to a vacuum envelope 33. Inner rotating segment 30 of rotatable bearing member 29 is fastened to the outside diameter of hollow rotatable shaft 21. Outer stationary segment 31 of rotatable bearing member 29 is fastened to mounting member 32 and vacuum envelope 33. Suitable rotatable high vacuum sealing means 34 such as ferrofluidic seal, is incorporated in bearing 29 to vacuum seal stationary member 31 to rotatable shaft 21 to facilitate provision of a vacuum within vacuum envelope 33, surrounding anode 20.

An electron g\ln 36 is mounted wi thin vacuum envelope 33. Electron gun 36 provides an electron beam 37 focused upon electron beam track 38 on the exterior periphery of anode 20. Illumination of anode 20 by beam 37 causes generation of x-rays which exit through a vacuum tight x-ray transparent window 39 in vacuum envelope 33.

A pulley 40, or other means, is connected to a suitable motor by a belt (not shown) to provide rotational drive to shaft 21, and thus, anode 20. A port 35 is provided in envelope 33 for attachment to means, not shown, to obtain or maintain the necessary vacuum within the evacuated space 46. The vacuum may be generated by an Ion pump.

The basic structure of Fig. 1, having been described above, functions as follows. Cooled fluid from an external heat exchanger and pump assembly (not shown) is pumped into the x-ray tube through duct 24. The coolant then travels toward the anode 20 between the outer diameter of stationary inner tube 25, and the inner diameter of rotatable hollow shaft 21. The coolant then passes along inside input face 41 of anode 20, and outside of input face 42 of septum 26, until it reaches the anode heat exchange surface 43.

Septum 26 serves to direct the entire coolant flow into close proximity to the anode heat exchange surface by providing a narrow channel between the septum 26 and anode heat exchange surface 43. The width of the septum 45 is typically greater than the width of the electron beam track and is generally centered wi th the electron beam track. The spacing between the septum and the anode heat exchange surface is designed to maintain optimum flow and heat exchange conditions. The geometry is always such that the enti re heat exchange surface of the abode is simultaneously and continuously exposed to coolant flow. In this manner, the entire heat exchange surface is continuously cooled and hot spots cannot develop due to interrupted coolant availability, Thus, optimum heat transfer is obtained and maintained.

Having passed over the anode heat exchange surface 43 to point 44, the heated coolant now passes the outboard faces of the anode inside surface and septum, past support fins 28 and out through the inside of stationary tube 25. From there, the coolant proceeds to the external heat exchanger pump (not shown) and back to the x-ray tube.

Bearing design and tube processing considerations

A fundamental design change takes place in the rotating bearings and related vacuum seals when going from solid anode medical tubes to hollow liquid cooled anodes. By placing the bearing wi thin the vacuum envelope, conventional medical tubes avoid a rotating vacuum seal. However, the result is that,exotic lubricatjon is required, i.e. silver, and substantial clearances, approximately .003" radial and .012" axial, are required for thermal expansion. This results in the common anode vibration poroblem. Anode cooling is by radiation from th~ anode qurface and by conduction through the bear ing. Bearing temperatures can range in the 500 - 600· C range and the result is short bearing life. In CT, bearing life may well be an important tube failure mode. Here the centrifugal force of gantry rotation combined wi th the cantilevered support of the anode places additional deforming and life shorting force on the hot bearings. For example, a 1 second 360 fast scan gives rise to a 2g force on an 18" rotation radius. At the top of the gantry the force on the anode is Ig whereas it is 3g's at the bottom. This can also give rise to a variable deflection of the hot anode shaft with consequent focal spot movement.

When going from solid anode medical tubes to liquid cooled tubes, the bearing and anode

3

problems are simpli fied. With liquid cooled rotating anode x-ry tubes, the large shaft diameter combined wi th the substantially room temperature operation of the shaft, which resul ts from the internal flow of coolant, provides a very stiff structure and assures negligible anode displacement. Furthermore, high quality bearings such as are used in aircraft engines may be employed thereby virtually eliminating the vibration related problems of high noise levels and shortened life. Bearing life can now be realized in the thousand hour range in contrast to the current ten or more hour category. However, a more severe problem, namely the design of a rotating vacuum seal, must be resolved. Until recently, oil seals wi th their attendant shortcomings were used? The incorporation of ferrofluid rotating vacuum seals has substantially improved tube performance and Ii fee Ferrofluid seals used in liquid cooled rotating anode x-ray tubes for x-ray li thography in the manufacture of IC' shave 10" (25 em) diameter anodes rotating at 8 ,OOORPM~. These tubes have twice the surface speed of a medical 4" (lOem) diameter anode rotating at 10,000 RPM. Triplettqat Intel Corporation reports 3 years of failure free performance of ferrofluid seals in a liquid cooled rotating anode x-ray tube used for x-ray lithography in the manufacture of 1 and 4 megabit bubble memories.

The incorporation of "0" ring vacuum seals in current liquid cooled rotating anode designs severely restricts the tube processing technology that maybe used. The use of elastomers (i.e.,_ "0" rings) for vacuum seals, which generally have a high vacuum temperature limit of less than 200 C, severely inhibits tube bake-out methods and restricts electron beam processing techniques. The result is that large vacuum pumps are used to remove the copi us amounts of gases that evolve during operation. A further limitation of this design is that without special care, high voltage instabili ty''c:an be -a problem. High field gradients exi st between the closely spaced cathode and anode (focal spot) and as the gases evolve from the hot focal spot, there is a transient localized high pressure in the high voltage gradient region. When this pressure reaches a critical value, arc-over occurs. The vacuum pumping port is generally located some distance away from the cathode region and the gases must diffuse to the port before the pump, no matter what its size, can be effective.

The use of "hard vacuum" fabrication and processing techniques such as braze and heliarc seals, which can typically withstand temperatures of 800° C, instead of "0" rings, results in a much cleaner tube wi th minimal high vol tage instabili ty. The ferrofluid seal is hermetic and possess a very low vapor pressure'~ Because of its very small exposed surface area, the ferrofluid seal generates a negligible volume of gas'! Therefore, a small ion pump would be sufficient, such as 1 L/second, its size being principally dictated by the gas evolution from the anode during tube operation. With a properly processed tube, this gas level could be kept quite small.

The Rotodyne x-ray tube

The most efficient liquid cooling of a surface is achieved under the condition known as nuclea te boil ing. Nuclea te bubbles are formed at nuclea ti ng si tes, generally pi ts or cav i ties on the heat exchange surface. The higher the heat flux, the more vigorous is the boiling. However, when the heat flux reaches a value such that nucleate bubbles are formed faster than they can be carried away, the bubbles merge to form a vapor blanket known as film boiling. The vapor blanket is a poor conductor of heat and causes the heated surface temperature to ri se rapidly resulting in burn out. This phenomenon is known as the critical heat flux. Thus, it is the formation, departure and interaction of nuclea te bubbles that largely determines the heat transfer characteristics of a liquid cooled surface. The Rotodyne tube achieves its superior cooling by designing heat exchange surfaces that incorporate one or both of two separate and distinct physical phenomenon. One is a modified version of swirl flow and the other is a predetermined geometry and distribution of nucleate boiling cavitation sites. The Rotodyne heat transfer surfaces are designed for the optimum formation and rapid departure of nucleate bubbles thereby obtaining improved heat transfer.

Nucleating sites in the form of cavities prepared on the anode heat exchange surface are of such dimensions that nucleate bubbles of optimum size are formed. The cavities are spaced apart on the anode heat exchange surface such that at maximum heat flux, the bubbles do not interact to form destructive film boiling. In this manner, the maximum density of nucleating sites, i.e. best heat transfer, is obtained while also maximizing the critical heat flux (burn out).

Swirl flow cooling has been described in detail in the classic paper of Gambill and Greene~a In their experiment, a hollow tube of metal was heated by passing electric current through it from electrodes clamped at each end. To cool the tube, water was injected at high pressure tangentially to the inside surface of the tube and at an angle such that the water follows a helical path i.e. it "swirls" along the inside surface of the tube as it travels towards the discharge end. Efficient cooling is achieved by two mechanisms; one is the high shear velocity between the liquid and the tube wall, and the other is the pressure gradient generated by the centrifugal force of the liquid as it traces its helical path down the inside of the tube. The pressure gradients more rapidly break loose the nuclea te bubbles thereby increasing heat removal by a factor of 4X to 5X compared to linear flow. Gambill and Greene obtained a maximum heat dissipation of 17,400 w/ema • To convert an industry standard 100KW peak tube to a 100 KW continuous duty tube having a 1.2mm focal spot on a 10 cm (4") diameter anode rotating at 10,000RPM, a dissipation of 6,000 w/ema. is

required. It is seen that the needed heat dissipation is about 1/3 that obtained by Gambill and Greene, thus providing a reasonable margin of safety.

The Rotodyne tube achieves the advantages of swirl flow cooling in a unique manner. The high shear velocity (52m/sec.) between the liquid and the anode surface is obtained by the rotation of the anode. The desired pressure gradients are created in the liquid by curving the anode heat exchange surface thereby generating a centripital acceleration. Further means for optimum implementation of the swirl flow phenomenon are incorporated into the anode design and are best illustrated by reference to Fig. 2.

The principles of operation of this preferred embodiment include utilizing the rotation of anode 20 to pump the coolant by attaching centrifugal flow pump vanes 58 to the surface of the input face 41 of the anode 20. That is, the rotating anode is converted into a centrifugal flow pump. In this manner high pressure, about 200 PSI, is generated in the reg ion of the anode heat exchange surface 43 where it is desired. One of the major benefits is that the rotating liquid seal (not shown) can then be operated at low or zero pressure, thereby insuring long life and eliminating leakage. Upon leaving the tip 62 of the centrifugal flow pump vane 58, the liquid is engaged by the tip 64 of the curved axial flow pump vane 66. The axial flow pump vanes 66, which are mounted on stationary septum 26, serve to redirect the flow of the liquid such that it traverses the path of anode rotation at essentially 90·. The axial flow pump vanes 66 are curved in such a manner as to avoid or minimize undesirable liquid flow characteristics such as cavitation. The anode heat exchange surface 4:3 is curved in a manner such that the high veloci ty liquid flow, generated by the

~entrifugal flow pump vanes attached to the anode, generates the desired centripi tal acceleration p. ,and in turn the pressure gradient, where V is the liquid velocity and R is the radius of curvature 'of the anode heat exchange surface. The anode heat exchange surface 43 is also prepared with cavities of predetermined dimensions and distribution in order to further improve heat transfer by the optimum formation of nucleate bubbles. Having traversed the anode heat exchange surface 43, the liquid is redirected by the curved axial flow pump vane surface on the discharge side to be smoothly engaged by exhaust turbine vanes 76 which are attached to the anode discharge face 78. The exhaust turbine vanes insure the smooth flow of discharge liquid down the anode discharge face 78, with minimal circumferential velocity thereby minimizing back pressure due to centrifugal force arising from rotating liquid. The liquid then exits from discharge tube 25 to an external heat exchange system.

Expected performance of the Rotodyne tube

Having described one of several configurations of the Rotodyne Tube, the expected performance is plotted on a standard single exposure radiographic rating chart, Fig.3, for a 1.2mm focal spot, 100KW tube. At 125KV, a common CT tube vol tage technique, Curve A (sol id 1 ine) shows the normal drop in tube current wi th increasing exposure time for a conventional tube starting wi th a cold anode. For typical CT scan times of 2, 3 and 8 seconds, the permissible beam current for a cold anode start drops from its 840 rnA peak to about 350 rnA at 2 second, 270 rnA at 3 seconds and 180 rnA at 8 seconds. These are reductions of about 60%,70% and 80% respectively. These ratios remain much the same for all focal spot si zes. with the Rotodyne tube, a beam current of about 800 rnA could be maintained for all exposure times as shown by the broken line of Curve A. As the conventional anode heats up with sequential exposures, the single exposure rating curve lowers as illustrated by solid Curve B. The lower permissable beam current with a hot solid anode further increases the ratio of the Rotodyne tube operating current, always a constant 800 rnA, to that available from a conventional tube. This advantage translates directly into shorter exposure times, or alternatively, into smaller focal spot sizes when using the Rotodyne tube.

Curve A (broken line) shows the expected performance of the Rotodyne tube at 125 KV. wi th a thin wall in the anode heat exchange region, i.e. Imm MO, the thermal time constant is about 0.02 seconds. Thus, the anode operates like a solid anode for the first 0.02 seconds and then levels off at about 95% of the peak instantaneous power when equilibrium is reached on the liquid cooled side of the anode. This operating power level may be used for any arbi trary exposure time resulting in what is essentially an unlimited anode Heat Unit rating. At other focal spot sizes the continuous operating power is also at 95% of peak instantaneous power.

In a continuously cooled tube, the anode cooling rate is equal to the average power rating and is therefore focal spot size dependent for a given set of anode parameters. For example, a 1.2mm focal spot, 100KW Rotodyne tube would have a continuous cooling rate of -

100 (KV) X 1000 (rnA) X 60 (sec/min) X 1.35 (3 phase) X .95 (%of peak) = 7,695,000 HU/min.

This shows a significant improvement over typical peak cooling rates of less than 200,000 HU/min. for conventional solid anode tubes.

A useful es!:imate of. thelRotgdvne tube performance in a cl inical si tuatio.lL may.b

be had bv comparIson Witu conventiona tu es In cT. -comparIsons are made ot severa1. CT tOI es WIth a Rotodyne tube of the same focal spot geometry for both tube performance and scan times .'.3

Focal Spot Size(mm)

And Target Ang Ie

seimens 1. 6 ( 4) Rotodyne 1. 6 (4)

GE .7X.9(U') (maxiray 125B)

, Rotodyne .7X.9(10) (4 )

• Eimac .6 (12_ 1/2) (B-160-) (HA465) Rotodyne .6 ( 4)

Peak Power (KW)

30/50

146 (1)

24 (2)

70(1)

(3 )

40(1)

TABLE 2

Anode Capaci ty

Heat Units (HU)

1,350,OOO

Unlimited

Unlimited

Unlimited

Max Anode

Cooling Rate (HU/Min)

325,000

11,826,OOO(1)

150,000

5,700,000(1)

150,000

3,080,000(1)

(1) peak and continuous (2) 4 sec. rat i ng? (3)not listed (4)standard 10,000 RPM 4" dia., and 13' anode is assumed.

High Resolution Scan Time

(Max. )

8 (Sec.)

O.9

8

3

8

2.5

It is assumed that the highest resolution scan corresponds to the highest mAs. Rotodyne scan times are calculated from the highest mAs listed divided by the mA corresponding to l20KV for the Seimens and Eimac tube and 80KV for the GE tube. 80KV was used for the GE tube as this then approximates the 24KWmax. allowed for the GE tube. From the above, it can be seen that scan times can be significantly shortened thereby reducing patient motion. In other procedures such as angiography, smaller focal spot sizes may be used to improved image quality.

The preceeding are just a few examples of current radiographic procedures that could benefit from a continuous duty liquid cooled rotating anode x-ray tube. The essentially unlimited Heat Unit capability of the Rotodyne tube assures that any power demanding procedure; present, experimental or yet to be devised is likely to be practical clinically.

Acknowlegements

The authors gratefully acknowledge the critical reading and valuable suggestions provided by Robert Gould of the Universi ty of California, San Francisco and Robert Jennings of the National Center for Devices and Radiological Health (NCDRH). The authors are also indebted to Thomas Shope and Robert Wagner of NCDRH for their valuable comments. This work is being supported by the FDA under SBIR Grant 1 R43 FD 01237-01

References

1. Wagner et al. - The Bottom Line in Radiologic Close Reduction Recent and Future Developments in Medical Imaging SPIE Vol. 206 pgs. 60-66, 1979.

2. Previtte - Erbium Filtration in Iodine Contrast Studies Radiologic Technology Vol. 53, #5, pgs. 399-405, 1982.

3. Sorenson et al. - Rotating Disk Device for Slit Radiography of the Chest Radiology 134: 227-231, January 1980.

4. Asle et al. - Factors Affecting the Design and Operation of Rolling Element Bearings in Rotating Anode X-Ray Tubes 22nd ASLE Annual Meeting, Toronto, May 1-4, 1967.

5. H. J. Queisser - Topics in Applied Physics X-Ray Optics Vol. 22,-pgs. 26-28, Springer-Verlag, N.Y. 1977.

,

6. U.S. Patent 414,326,144

7. Vacuum Rotary Feedthroughs for High Vacuum Applications Microelectronics Manufacturing and Testing March 1982.

8. B. Fay - High Power 13.3 ~ X-Ray Source for Submicron Lithography J. Vac. Sci. Technol. Vol. 19, *4 Nov/Dec 1981.

9. B. B. Triplett - X-Ray Lithography for VLSI Intel Magnetics, Inc. 3000 Oakmead Village Drive Santa Clara, CA 95051

10. Okada - ~ New Transmission Target for the X-Ray Lithography Source J. Vac. Sci. Technol. 17(5) page 1236, Sept/Oct 1981.

11. Ray et ale - Mass Spectrometric Studies of Material Evolution from Magnetic Liquid Seals Re~ Sci. Instrum. 51 (10) October 1980.

12. Gambill et ale - Chern. Eng. Prog., Vol. 54 #10 pgs 68-76, nctober 1958.

13. Journal of CAT, pgs 423-428, April 1982.

36

40

FIG. I.

to e e e

66

FIG. 2.

CI)

150 140 1:300 120 110

0

0

0 0

100 0 ~ ~ 90

~ 80 <X 70

0

j 60

:E 50

0 0 0 0

0 0

40 30 20 100

0

RADIOGRAPHIC RATING CHART

-.

i'--

/

'--...!

"""-

FOCAL SIZE: r. 2 m m

10-" A

,- --.. --- - - -. -. ~,d I

.... ~ ..... ~ !oo. f'..

B- I"-r--..." t'-... ........ ~ ......... ~, ~ r-....

0 .001.002 .005 .01 .02 .05 .1 .2 .5 I 2 5 10

1/120 1/301 1;10 I 3/10 I 1/60 1/20 1/5 1/2

MAXIMUM EXPOSURE TIME IN SECONDS FIG- 3

2 o