Magnesium Ion Source for High Intensity Mass Spectrograph Citation Pi, Te Hsien (1944) Magnesium Ion Source for High Intensity Mass Spectrograph. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/HA7K-HZ57. https://resolver.caltech.edu/CaltechTHESIS:12052012-154742447 Abstract Magnesium ions were produced by bombarding magnesium vapor by electrons emitted tram oxide coated cathodes. The ions were then focussed into a thin sheet by means of an electrostatic lens. This ion sheet was sent into the magnetic lens which focussed the sheet into a small spot where they were collected. The ion current obtained at the collector was 120 microamperes. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Physics) Degree Grantor: California Institute of Technology Division: Physics, Mathematics and Astronomy Major Option: Physics Thesis Availability: Public (worldwide access) Research Advisor(s): Smythe, William Ralph Thesis Committee: Unknown, Unknown Defense Date: 1 January 1944 Record Number: CaltechTHESIS:12052012-154742447 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:12052012-154742447 DOI: 10.7907/HA7K-HZ57 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 7304 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 06 Dec 2012 17:22 Last Modified: 10 Nov 2023 19:36 Thesis Files Preview PDF - Final Version See Usage Policy. 3MB Repository Staff Only: item control page

MA.GNmiW ION SO~E FOR HIGH :rnTENSITY .MASS SPIDTROGRAPB Thesis by Te Hsien Pi In Partial Fulfillment of the Requirements for the Degree ot Doctor ot Philosophy California Institute ot Technology Pasadena, California 1944 SUhNARY. Magnesium ions were produced by bombarding magnesium vapor by electrons emitted tram oxide coated cathodes. The ions were then focussed into a thin sheet by means ot an electrostatic lens. This ion sheet was sent into the magnetic lens which focussed the sheet into a small spot where they were collected. The ion current obtained at the collector was 120 microamperes. :rnTIDDUCTION. Mass s1Xtctrographs can, 'tor the present discussion, be divided into two classes: (a) those intended 'tor accurate mass and abundance ratio measurements and (b) those that are built 'tor specimen collection. For class (a), high resolution and strict collimation are essential. required need not be very large. used. The number o't ions "Point" sources are generally When the intensity ot the ion current is too high, the beam will spread because ot the space charge ettect, and it is much more serious than in the ce.se o't electrons which tor the some energy have much larger veloei ty. When it comes to collecting a specimen o't, say, one milligram, the high intensity is necessary. The resolution need not be too high. resolution and high intensity are incompatible. High I t again a "point" source is used, there is a de'tinite upper limit tor the ion beam current beyond which no sharp beam can be obtained. Tuve 1 and his collaborators had shown experimentally that there is considerable spreading in projecting ion beams o't much more than 10 microamperes to distances much in excess ·or 20 cent~ters. In 1929, w. R. Sm;vthe and L. H. Rumbaugh 2 designed and built a magnetic lens which is capable ot focussing sheet ot ions ot the same e/m to a point. a Since the ions are sp»ead out into a sheet except close to the 'tocal point, the total beam current can greatly be increased. The ion 2 source used was KUnsman catalyst which was available tor a few elements. ;r. Evvard 3 for magnesium. tried the electron bombardment method The electron emitter he used was a tantalum strip ot dimensions 0.005" X 0.085" X 10". The two strips in parallel took about 40 amperes at 40 volts to get an emission of 100 milliemperes. He was able to get e.n ion current at the collector of 24 microamperes with good resolution. It seems that there are still· possibilities of greatly increasing the collector current. work. This is the main object of the present 3 In order to get large ion current at the collector, five requirements must be satisfied. (1) There must be plenty ot magnesium vapor in the ionization chamber. ( 2) There must be enough electrons to bombard the vapor with. (3) The tield just above the alit of the chamber must be strong in order to draw out all the ions that are pushed outotthe slit. ( 4) The lens must be able to focuss them into a thin sheet. ( 5) This ion sheet must be parallel and directly beneath the pole pieces so that they all can go into the magnetic lens. The first requirement was met without aey- ditticulty while the titth one chiefly concerns the alignment which the more critical the sharper the ion beam is. is An oxide coated cathode supplies ample electron emission which meets the requirement (2), The third and fourth ones depend upon the structQre of the electrostatic lens. These will be discussed in detail later. Tantalum was not a very efficient Emitter. 1.5 kw power to produce 100 ma electron current. It took Even though there was a water cooling jacket, the soapstone insulators which were rather close to the tantalum filaments sometimes got hot enough to be conducting. The tirst thing was to tind another more etticieiit emitter. O:d.de coated cathodes were then tried w1 thout much success at first. When the anode voltage was tirst applied, the emission was about 40 ma, but very soon, it dropped practically to zero. It was 4 tound that the cathode turned into gray instead ot white color. This contamination was due to the gas given out when the magnesium heater was heated. If the magnesium was heated tirst and then the cathode activated, it stayed active. Then it was thought that high frequency voltage across two cathodes would accelerate the electrons so that they could ionize the magnesium vapor. Since the ions are so much heavier, they should n~be influenced by the high frequency tield it the frequency is high enough. The arrangement was as shown in Fig. 1. -----B G Fig . I G was a grid made of copper wire. D.c. potential was applied between cathodes ~ G making G positive with respect to K. G was also connected to the positive terminal of the 6000 V d.c. generator. B was grounded. The potential ot A could be changed continuously by means of a potentiometer arrangement 5 so as to obtain the best focussing condition. Without a high frequency field between the two cathodes, a collector current ot 7 microamperes was obtained. About the same current could also be obtained with 100 volts high trequency voltage alone. When both were used, the collector current was always smaller. The main disadvantage of this arrange- ment was that the focussing field greatly attected the field which was responsible tor ion production. Another source was built where the main feature was just to isolate this interaction between the focussing field and the field for the ion production. This was the final foDn adopted. The description and dimensions are given in the following sections. The first lens tried with the above ionization chamber consisted of three parallel plates each with a slit in the middle of 5 lllll width (Fig. 2). Lens With 130 ma. cathode emission and proper focussing, the ion current to A was 320 microamperes. A very sharp image was found on A which was 2 centimeters above the top plate of the lens. the lens was too strong. Evidently The spacing between the three plates was changed to that shown in Fig. 3. With this lens, the largest ion current at the collector was 60 microamperes which was far too small in comparison with the total beam current from the slit. This meant that the beam was still divergent so that only the small central part can get through while about tour fifth of them hit the pole pieces. The focal length of the lens could be calculated but the calculations of the potential distribution were a little involved and there is another kind of lens which will give a stronger field near the ionization chamber alit. So this was abandoned. In order to test whether this small collector current was due to the detects in the focussing or the defects in the aligmn.ent of the different parts, an arrangement like that in 7 Fig. 4 was set up. Two de:rlecting plates D and D were 1 2 installed just on top ot the lens and a deflecting voltage Another pair ot plates P and P were put in 1 2 just below the magnetic pole pieces. The distance between was applied. P1 and P2 was equal to the ge.p between the pole pieces which was 6 nm. So if the beam was wider than that, they both sould receive ions perhaps not of the same amount if the beam was off to one side. The results of some tests showed that P1 and P2 received about the same ion currents which tact proved that the alignment was good and that the beam was too broad tor the gap between the pole pieces. When a large deflecting voltage was applied between D1 and D2, the beam could be directed to either P1 or P 2 which amounted to about 150 microamperes. Therefore the whole problem was at this stage reduced to an electron optical one. N 5 l H.·T + THE ION SOUIDE. The ion source consisted of three parts (see Fig. 5): (1) the turnace F tor producing magnesium vapor, ( 2) the cham- ber C where the Tapor was bombarded by electrons, o.nd (3) the electrostatic lens L by means ot which the ions were focussed into a sheet. (1) The furnace was designed and constructed by :r. Evvard. The two thick-walled magnesium tubes were heated in nouum to supply the vapor. The description can be found in his Ph. D. thesis, 1943. (2) The ionization chamber C was 1.9 em wide o.e em deep and 26 em long. In the middle of the chamber there were tour parallel 0.005" tungsten Wires held tight by springs. The top two were connected together to for.m the anode A1 and the two at the bottom tor.med the anode ~· On each side ot the tour wires, there was one oxide coated cathode also 26 em long. They were supported by two baked soapstone pieces which closed the two ends of the ionization chamber. The only escape tor the ion as well as the magnesium vapor was through the alit S upwards. The two cathodes ond the ioni- zation chamber were connected together electrically so that t.11 the electron emission had to go to the two anodes. anode ~ was connected to a higher d.o. potential ot -'J.. The than that The electric field thus set up between ~ tmd A and 2 that between A1 and the top of the ionization chamber helped IO Magnes.um Ion 5ourc.e One inch Fig. 5 11 to push the ions out of the chamber. The voltage was supplied tram a small motor generator set which provided 200 volts d.c. voltage. The generator was carefUlly insulated since the whole ionization chamber was at about 6000 volts above ground potential. Construction and Activation of the Oxide Coated Cathode. The cathode consisted of a nickel tube I~&de from 0.002" nickel sheet wrapped around a stupekoff tube of 0.25 em outside diameter. It was spot welded together along the seam. Later it was found that welding was unnecessary. A helical heating coil made of 0.01" tungsten wire was put through the stupekoff tube to form the heating element which has a cold resistance of about one ohm. The two heaters at operating temperature took about one hundred watts power. The nickel tube was first cleaned w1 th ~1 acetate and let dry. 50% barium and 50% strontium carbonate powder was suspended in ~l acetate. A little collodion was added for better adhesion to the nickel surface. The emulsion was applied by means of a small camel hair brush. The tube was rotated while the coating was applied so as to prevent the emulsion from nowing to the under side. the coating was too thick. When first tried, When it was fired in vacuum, parts of the cathode were activated while parts became black. Because of the different emissive power of the different parts, spots were developed along the cathode. hot This was aggravated 12 by the tact that the tungsten heater had a larger resistance at the hot section, so more power was dissipated there. At one ttme, the tungsten heater was burned out at that spot while other parts had not even reached the operating temperature. When the coating was thin, no such uneven heating ever developed. The whole vacuum chamber was first evacuated to a pressure ot about 5 X lo-5 mm Hg. The magnesium heater was started at 2 amperes heating current. When the pressure dropped to the original Talue, the heating current was increased to 2.7 amperes by small steps and was kept there. It the current was higher than that, the magnesium began to vaporize. The vapor might deposit on the cathode or on the inaulator. Both were highly undesirable. When the pressure again dropped to 5 X lo-5 mm Hg., the cathode heater was turned on at 1.5 amperes. cur~ent As the vacuum improved, the heating was increased until it reached 2.7 amperes. case was the vacuum allowed to go beyond 10-4 mm Hg. w1 th 100 volts on ~ and ~' 40 milliamperes was obtained. In no Now the emission current of about This small current was mainly due to space charge limitation. When the spectrogra)h was actually operating, the cathode heating current was increased to 3 amperes. During the activation period, no voltage on the anodes was necessary. When the chamber was tul1 of iona, the emission went up to as much as 700 milliamperes with only 13 130 volts on the anodes. By then the four anode wires were red hot. The above order of turning on the magnesium and cathode heaters was necessary. If the cathode was heated first and the magnesium heater later, the cathode always got contaminated even the vacuum was always better than 7 X 10-5 nm Bg. (3) The electros~atic lens . Take the center of the lena as the origin of the coordinate system with z-axis along the central straight path of the ions, y-axis perpendicular to the ion beam and the line source, and the x-axis parallel to the line source. If we are not very close to the ends of the lens along the x-direction, the potential function ~ satisfies the two dimensional Laplace equation (1) :: 0 if we can neglect the space charge effect. This can be solved by means of Schwartz transformation for most of the lenses. The differential equation 4 for the path of the ion is ( I+ (* t] () <P '-t ay - 0 (2) In general, the potential distribution ia a very complicated tunction of y and z . It is practically impossible to get an exact solution of equation (2). However, if we confine ourselves to the discussion of the paraxial ions, we can develope 14 ~(z,y) into a power series in y. Then assume t(z,y) = .L <t(z) )')) (3) l'\":0 Substituting this into {1) and equating the coefficients of y, we get ( 4) Therefore ~oZ ~ '''(z. ) y:1 + .... l: ~ ,J,''r. ) r ..hiv( ~ - To lz. Y + 3: To z) Y - ( 5) ··· ( 6) It we only keep the first power of y, equation (2) becomes tl,y ~z.--- t [It(*)~] cP. '(z.) J.y 2.cp o dz + [ rf ( *)'"] .J..''( ) 2..ep 'fo z. '/ - 0 (7) The paraxial rays never make a large angle with the z-axis, so _&_ << I ch. then dropping the subscript tor the potential function along the axis, we have 0 (8) as the equation of the path of the ion. In general, even this is still too complicated to integrate. Therefore we have to resort to step-by-step method 15 of integration. .t Incident).Y, this equation for the path of the ion is independent of e/m, therefore it will be the same for all charged particles . The coefficients of the equation depends on the function ¢ and its derivatives only. R. Ge.ns 5 gave E. simple appro:rlm.ate method of calculating the path. The potential distribution along the lens axis • ght line se~ents as shown is approximated by a series of st~ in Fig. 10. Along each segment t''(~) 1>' {:z.) :: ecms~. =0 Equation (8) becomes 0 (9) Integrated once, we hav~ =c (10) and twice 1 = Yo + lC f[cf-Jifo] d(J) (11) di where y 0 and ¢0 are the respective values of y and¢ at the beginning of the segment. At the intersection of two segments, (12) Fquations (10), (11), and (12) are the general formulae for the determination of the ion path and therefore the focal 16 length ot the electrostatic lens. A.tter :finding out that the lens consisting ot three parallel plates with slit did not give a well defined beam, it was decided to try another kind as shown in Fig. 6, instee.d ot improving the old one. The reason was that by making the L, L, lower edges ot the parallel plates ~ and L only one halt 2 centimeter a~ tram the slit, the field there would be muoh stronger than that could be obtained with the other type. The potential distribution was solved by Schwartz transtormation method in w. R. ~he, Static and Dynamic Electricity, pp. 294-295. The potential on the axis of the lens which is (13) 17 where e = ~ jlv 'l.. (see Fig. 7). In the present ease, a a b - l/2 om. z v - - - L - - - - - , Y - - - V= 0 -v Therefore The calculations are listed in Table I. Fig. a is a plot ot equation (14). However, in actual use, the two lower plates (Fig. 9) z L----~ v. y 18 .n t: ~ l.() --1 u ("(") _.._ 0 -+.. "''0 0 ~ ("f) u QJ w '40 V"' l.() X 01 <::( ~ c: 0 4: 0 c: N 00 ()"'") 0 li: ---+" ::3 _...Q i:: \r) -+> - 1/'1 A 0 ~ ~ c 0 :-... - <IJ -+- ~ 0 P- ~ N l{) 0 0 0 - ~ d -&-\> \[) d --t- 01 0 0 0 19 were connected to potential v1 and the upper pair to v • 2 Fig. 10 shows the distribution when V1 = 1000 and v2 • 5000 volts. The straight line segments that approximate to the actual curve are also shown. The potential at the points or intersect! one A, B, C, D, E, F, G, and H are read off from the graph which are listed in Table II. are calculated in Table ni. ~! The slopes Since only vol te.ge ratios are i nvolved 1n all the calculations, we can use any kind of unit for the voltages. In order not to have very large numbers , we shall use looo volts as one unit . Suppose we have an ion coming into the lens with a direction parallel to aDdis 0 . 6a from the z- axis . At the point A where the ion first enters into the lens field (~) ~ o. The I slope of the ion path immediately to the right of point A is given by equation (12) . From Tables I and II , at A ( ~)~ = ~::j.OOJ I o fs' (~! ), I 0 Therefore equation (12) gives (4i)~ = = ·~ 0 2. )( 1·00 [ · 108' - 0 J ·O!>'l-4 From equation (10) , we can calculate the constant C. - ·0324- . - ')../ I I ~r ~ 0 0 0 a 0 ~ 0 l.t) -e- 80 ('( 0 0 d 0 0 0~ N ('\1 I ("() I (JI LL 22 Once C is known, equation (11) will give the 'VSlue of y at B. y .:. · ~ .f -'-X • 03'2.4 [ ~ - fT.OO] or ·1 = ' 5 5"4 Then the above process will be repeated to get the values of y at the succeeding points C, D, •• • etc. Table IV gives the calculations up to the place where the ion just crosses the z- axis . The distance from this point along the z-a:rls to the center of the lens is called the focal length of the lena . When v1 • 1000 volts, v2 .,. 5000 volts , f • 0 . 67 em. Similarly Tables V, VI, and VII give the calculations for the case v1 • 2000 volts, and V2 = 5000 volta , where the focal length was found to be 4.14 em. So by keeping the potential on the two upper plates fixed and by changing the potential on the lower two plates, the focal length can be changed continuously. 23 Table I l,-1 Stl"' _, -a.z .110 .111 .110 .141 .230 .208 .228 .210 .278 27 .360 .321 .368 .327 .442 .40 36 .514 .416 .494 .429 .588 .45 40.5 .604 .460 .572 .478 .668 .50 45 .707 .500 .658 .524 .752 .55 49.5 .828 .538 .754 .568 .840 .60 54 .973 .573 .862 .610 .938 .65 58.5 1.15 .603 .983 .647 1.04 .70 63 1.39 .630 1.13 .681 1.15 .75 67.5 1.71 .654 1.31 .713 1.28 .eo 72 2.18 .673 1.52 .738 1.44 .85 76.5 2.94 .688 1.80 .759 1.63 .90 81 4.46 .699 2.20 .774 1.91 .925 83.25 5.98 .702 2.49 .778 2.08 .95 85.5 9.06 .705 2.90 .782 2.34 .975 87.75 18.00 .707 3.58 .785 2.78 -~y e ·7o714ne ·107sine .10 go .111 .20 18 .30 Sit~ ' 24 Table II z <P Jf A -3.38 1.00 1.00 B -2.08 1.14 1.07 0 -1.33 1.44 1.20 D - .58 2.18 1.48 E .58 3.82 1.95 F 11.33 4.56 2.14 G 2 .08 4.86 2.21 H 3.38 5.00 2.24 .068 .132 .276 .478 .181 .070 .031 Table III d+ dz. OOA .ooo AB .108 BO .400 CD .987 DE 1.413 EF . 987 FG .400 GH .108 ROO .ooo .108 .292 . 587 . 426 -.426 -.587 -.292 -.108 25 Table IV Y''1 Y'' 2 0 A .600 .ooo -.0324 -.0324 B .559 -.0304 -.101 -.108 0 .488 -.090 -.189 -.272 D .336 -.184 -.217 -.320 E .120 -.164 -.157 -.307 F .007 -.144 -.143 -.305 Table v z ~ JCf A -3.38 2.00 1.414 B -2.08 2.11 1.453 0 -1.33 2.33 1.526 D - .58 2.89 1.700 E .58 4.].)) 2.02'7 F 1.33 4.6'7 2.161 G 2.08 4.90 2.214 H 3.38 5.00 2.236 .039 .073 .174 .327 .134 .053 ;012 26 Table VI ~ Q)A .ooo AB .085 BC .293 CD .747 DE 1.05 EF .747 FG .293 GH .085 HCD .ooo .085 .208 .454 .303 -.303 -.454 -.208 -.085 Table VII '/ Y.' y; c A .ooo .ooo -.0129 -.0180 B . 583 -.0124 -.0411 -.0593 c .556 -.0398 -.0939 -.143 D .489 -.0841 -.1097 -.186 E .383 -.0916 -.0779 -.158 F .326 -.0731 -.0853 -.126 G .2180 - .0593 -.0534 - .120 H .252 -.0536 -.0515 27 With the above mentioned ionization chamber and electrostatic lens, 120 Rdcroamperes werer obtain d at the collector. The resolution is mown in Fig. 11 with the slit in tront of the collector narrowed down to 2 mm. 5 slit wid+h ~ 4 '1 mrn. \.... 0 ~ ~2 0 u -...... ./ 'Q.... __ ,_,.' __ ,..,.,. / / 0~--~------~------L------------ 2f 25 Mass Number· Fa 'J . I I 29 ACKNOWLEDGMENT. The writer wishes to express his gratefulness to Prof. w. R. Smythe for his direction and encouragement. He wishes to thank Dr. J . Evvard for showing him the operation of the original set up, and Dr. c. Y. Meng for valuable discussions. 30 1. M. A. Tuve, o. Dahl, and L. R. Ha:fstad, Pbys. Rev. 48, pp. 315-337, Aug. 1935. 2. L . H. Rumbaugh, The Development of High Intensity Magnetic Lena Type Mass Spectrometer, Ph. D. Thesis, 193?, and w. R. Sn:w"the, L . H. Rumbaugh, and s . s . West, Phys. Rev. 45, pp. 724-727' May' 1934. 3. :r. Evvard, The Production and Separation ot Magnesium Ions, Ph. D. Thais, 1943. 4. I . G. Malott and D. w. Epstein, Electraa Optics in Television, p . 80, 1938, McGraw-Hill . 5. R. Gans, Electron Paths in Electron Optics, z. tech . Physik, 18, pp. 41-48, 1937, and v. K. ZWorykin and G. A. Morton, Television, 1940, :John Wiley.