You are on page 1of 19

IEEE TRANSACTIONS ON MAGNETICS, VOL. 21, NO.

4, JULY 1991

3601

Demagnetizing Factors for Cylinders
Du-Xing Chen, James A. Brug, Member, IEEE, and Ronald B. Goldfarb, Senior Member, IEEE

Abstract-Fluxmetric (ballistic) and magnetometric demagnetizing factors Nf and N , for cylinders as functions of susceptibility x and the ratio y of length to diameter have been evaluated. Using a one-dimensional model when y 2 10, Nf was calculated for - 1 5 x < and N,,, was calculated for x 00. Using a two-dimensional model when 0.01 5 y 5 50, an important range for magnetometer measurements,N , and Nf were calculated for - 1 5 x < 00. Demagnetizing factors for x < 0 are applicable to superconductors. For x = 0, suitable for weakly magnetic or saturated ferromagnetic materials, Nfand N,,, were computed exactly using inductance formulas.
Q)
+

I. INTRODUCTION HE study of demagnetizing factors for ellipsoids and their degenerate forms (spheres, infinite plates, infinite cylinders) dates from the work of Poisson [ 11 and was elaborated upon by Thomson [2], Evans and Smith [3], and Maxwell [4]. Experimental investigations on finite right circular cylinders began in the 1870’s, when ballistic galvanometers were first used in magnetic measurements of iron [5], [6]. The literature distinguishes between “magnetometric” and “fluxmetric” (or “ballistic”) demagnetizing factors N , and Nf[7]. N , refers to an average of magnetization over the entire specimen and is appropriate for magnetometer measurements of small samples. Nf refers to an average of magnetization at the midplane of the sample and is appropriate for measurements made with short search coils. Values of the demagnetizing factor were deduced from the shearing of the magnetic hysteresis loop [8], a procedure originally developed by Lord Rayleigh for ellipsoids [9], or by measuring the magnetization and the field at the cylinder’s side [ 101. Factors for cylinders of several aspect ratios were published [7], [8], [ l l ] , [12]. By the 1900’s, it was apparent that the values of the factors depended on the susceptibility x of the material [13]-[15]. Although it has been criticized from a pedagogical point of view [16], [17], the use of fictitious magnetic poles to calculate demagnetizing fields has been universal. The first theoretical treatment of magnetic pole distributions in finite cylinders was by Green [18]. An early model to attempt to explain experimental data considered point

T

Manuscript received March I I , 199 1, The authors are with the Electromagnetic Technology Division, National Institute of Standards and Technology, Boulder, CO 80303. D.-X. Chen is on leave from the Electromagnetism Group, Physics Department, Universitat Autbnoma de Barcelona, 08 I93 Bellaterra, Spain. J. A. Brug is on leave from the Thin Film Department, Hewlett-Packard Laboratories, Palo Alto, CA 94303. IEEE Log Number 9102083.

magnetic poles at each end of a cylinder [ 191. This simplistic model could be used only for long uniformly magnetized cylinders, and the results deviated significantly from experimental data on ferromagnetic samples. During the 1920’s and 1930’s, there were several theoretical papers on Nf for material with constant susceptibility x. The results were given as functions of x and the length-todiameter ratio y. These used one-dimensional models with approximations as needed to suit the computational techniques of the time. The first systematic theoretical cal’ culation of N for the high susceptibility case was done by Wurschmidt [20], [21]. He calculated Nf of cylinders using a one-dimensional model in which the cylinder had side surface poles and point end poles. He used Taylor expansions for the magnetization and the demagnetizing field at the midplane. The calculation was complicated, W. and he completed it only for the case y = 50 and x For the y and x dependence of Nf,he gave qualitative results using the first few terms of the expansion. A similar approach with simpler expressions was used by Neu00 mann and Warmuth [22], who calculated Nf for x a n d y 2 10. Stablein To obtain the susceptibility dependence of Nf, and Schlechtweg [23] used a quadratic approximation and two linear differential equations. The model was improved by substituting uniform end-surface poles for point end poles. Their results included 30 values of Nf for 10 I y 5 500 and 12.56 Ix c W . An extension of the y region to 0 was achieved by Warmuth [24]-[26], who fitted existing data and extrapolated graphically using the demagnetizing factor N of ellipsoids as a reference. The values of Nf calculated from the one-dimensional models were consistent with the data of ballistic measurements on soft magnetic materials. Bozorth and Chapin [27] compiled the results, which were later plotted in Bozorth’s book [28]. To obtain axial demagnetizing factors more accurately, especially for short cylinders, two-dimensional calculations are needed. The simplest case is x = 0, where N , and Nfas functions of y can be derived analytically. The approximation x = 0 applies to diamagnets, paramagnets, and saturated ferromagnets. N,,, for 25 values of y from 0.2 to 1000 were obtained accurately to four significant figures by Brown [29] from a calculation of self-inductance [30] and listed as a table in Brown’s book [31]. Crabtree [32] obtained the same values for the average demagnetizing factor by integration of the local field over the cylindrical volume. Moskowitz et al. extended
- + -

+

U . S . Government work not protected by U.S. Copyright

1.-

i-r

in which the magnetization is uniform.3602 IEEE TRANSACTIONS ON MAGNETICS.00.. are plotted. where both the magnetization M and the demagnetizing field Hd are uniform under a uniform applied field H. (1) N is the demagnetizing tensor. and expanded the electrostatic potential at the cylinder center [42]-[44]. in magnetic measurements of ferromagnetic. The details of the calculation were published by Aharoni. the number of y values for N .(m) can be deduced from his electric polarizability results because of the analogy between electrostatics and magnetostatics. and they are widely used. For x = 0. For susceptibilities other than 0. almost all books give results obtained before 1950. for large x. Taylor calculated electric and magnetic polarizabilities for conducting cylinders for 0.01 Iy I50. Using a similar approach with simpler base functions. Approximate values for N .1 Ix < 00) based on the one-dimensional model of Stablein and Schlechtweg [23].1 and x perfectly diamagnetic and ideally soft ferromagnetic materials. y . . sometimes inappropriately. and KaczCr and Klem extended it to hollow cylinders [34]. and more generally.05 Templeton et al. For the case x = 0. [48]. Sato and Ishii [37] obtained a simple expression to approximate N . and Nffor lop5 Iy 5 lo3.25 %. 11. JULY 1991 Brown’s method to cylinders of polygonal cross section [33]. there are no data on N. and z axes.1) can also be deduced from his results. 4. who also calculated the self-energy of cylinders [49]. [46]. respectively. the most accurate case.( . For 0. NO.. Unlike them. 27. For long cylinders (y > lo). The fact that the side and endpole densities have basically a 6-’13 dependence. and obtain more accurate results.00. was used to construct the set of polynomials. for a complete range of y and x.. the internal field is H where = H. Susceptibility x is traditionally assumed to be constant in the material and is therefore defined as M / H . model is also applicable to N . the variation of magnetization across the radius of the cylinder is negligible at the midplane. The . and there are no data for x < 0. Brug and Wolf [57] calculated the magnetization distribution in disks and obtained the local demagnetizing factor for materials that undergo phase transitions. Because his calculation for magnetic polarizability was for a uniform quasi-static but nonpenetrating applied field. For x # 0.1. and Nf for x = 0. a two-dimensional finite element method is used that takes into account the variation of magnetic pole density along the side and ends of the cylinder. calculate the demagnetizing field H d ( z ) directly at 25 points along the axis. where M is the magnetic moment per unit volume and H is the internal magnetic field. we use Taylor expansions for M ( z ) . [59]. more elaborate methods are used.. his convergence error was less than 0. According to Taylor. For y > 10. [41]. and Yamamoto and Yamada [54] obtained Nfand N. Chen and Li [38]. Kraus [56] determined the complete local demagnetizing tensor for uniformly magnetized cylin+ + ders. N. where 6 is the distance from the corner. d - NM. + H = H. were calculated by Vallabh Sharma using uniformly magnetized volume elements [36]. [52] calculated the distribution of magnetization and magnetic field in long cylinders with large susceptibilities using volume and boundary integral equations. Joseph and Schlomann [55] solved for local demagnetizing factors in uniformly magnetized cylinders and used a series expansion to account for nonuniform magnetization. and Nfare given for -1 Ix < 00. and those that exist have large errors because they were obtained by extrapolation. and 00. To estimate their error. [39] obtained Nf for x = 0 using magnetostatic potential calculations. different techniques have been used. For short and y cylinders ( < lo). for x . . Using experimental resistance network analogs. 03 for 0. and we calculate Nfas a function of y and x (.25 Iy 5 4 in both the longitudinal and transverse directions. N. If the three principal ellipsoid axes coincide with the x . VOL. and Nffor these susceptibilities were first treated by Taylor for perfectly conducting cylinders [40]. we calculate Nf and N . and superconducting materials. These types of graphs and tables appear in other books on magnetism and magnetic materials.1 % for the longitudinal direction. there are even less data. In this paper.3 1% [45]. there is no complete picture for the y and x dependence of Nfand N. Nf and N . Archer and Guancial [511 and Fawzi et al. [60]. weakly magnetic. Nffor x = 0 was calculated exactly by Joseph [35]. The work was based on their earlier magnetostatic analysis of the magnetization process in soft ferromagnetic cylinders with constant end-pole densities [47]. the deviations of their 12 computed Nf(oo) values were less than 0. there remain some problems. there is a lack of data on the x dependence of Nf. accurate for large y. FLUXMETRIC MAGNETOMETRIC AND DEMAGNETIZING FACTORS The demagnetizing correction is nontrivial for samples in open magnetic circuits. He developed a method introduced by Smythe that expressed charge densities on the side and ends in terms of a set of orthogonal polynomials. calculated axial Nffor x I y I250 [45].. 00 correspond to The susceptibilities x = . Compared to an approximate formula with 8 adjustable parameters. Okoshi [53] obtained Nffor x . N . Templeton and Arrott calculated the root-mean-square deviation of the normalized potential from 0 and found it to be less than 0. However. In summary. and Nfis insufficient for accurate interpolation. An exact correction can be obtained only for ellipsoids [4]. Values of N . . for x = 0. In Zijlstra’s book [58]. cylinders with nonuniform magnetization [50]. For x # 0. ferrimagnetic. Several papers have treated demagnetizing factors at points. we give 61 exact inductance calculations of N .

-(Hd)/M = 1 . Relationship between N . (1 1) For the entire volume.: DEMAGNETIZING FACTORS FOR CYLINDERS 3603 with N. The magnetometric demagnetizing factor N . Thus B = po(Ha Hd M ) . similar to the differential susceptibility dM/dH at the corresponding magnetic state. B. is defined by ki = (8) (9) 4a2/(4a2 + 12).2&/(pona2). (1 2 4 Thus the average demagnetizing field can be obtained from (10) and (12a) as (Hd) = +/(p~na*) M.I II I CHEN et al. N . to a longitudinally magnetized cylinder coincident with it [29]. we obtain the average B' at the midplane from the flux a0in the one-turn secondary loop of radius a: ( B ' ) = +0/(na2). the demagnetizing field is not uniform. An exact formula for the self-inductance L. Cohen [63] derived an exact general formula for the mutual inductance of two concentric coaxial thin solenoids (denoted by subscripts 1 and 2). Thus the cylinder can be modeled as a solenoid with the same M . In SI units. which is defined by k: = a 2 / ( a 2 Z2). so M = I/(2l). The flux density B in a magnetic material is related to the internal field H and the magnetization M : B = p O ( H M ) . 0 5 N I 1. the fluxmetric (or ballistic) demagnetizing factor Nf is defined as the ratio of the average demagnetizing field to the average magnetization at the midplane perpendicular to the axis. and its average Hd can be obtained from M and average B' using (10).E(km)I. the definition of N . + Nz = 1. and Nffor x = 0 and a wide range of y. In fact. = +/I. If the sample is located in a uniform applied field Ha along its axis. and the problem is that of the mutual inductance of a solenoid and a single-turn loop located at its midplane.. + N. and two kinds of susceptibility-dependent demagnetizing factors are defined. We take the solenoid as having one turn (n = l ) . in the limits l2 = I . with a2 = a l . of a thin solenoid of length 21. with respect to the B' field.) are the complete elliptic integrals of the first and second kind of modulus k.. (10) Nf FOR x = 0 DETERMINED BY INDUCTANCE CALCULATIONS Brown [29] showed how N . we can obtain the average B' from the average flux 9 in the solenoid as (B' ) = +/(m2). An exact formula for the mutual inductance L. and L. is limited to x = 0.) and E(k. An ideal thin solenoid carrying current I through n turns over a length 21 is equivalent. and number of turns n is [611 From ( l l ) . a cylinder in an axial field has a uniform magnetization M . radius a. as an alternative to (6) and (8). and (14a). (3) where N is called the demagnetizing factor. N . Nf is obtained when the length of one of the model solenoids approaches 0. m where the modulus k. 111. related to the applied and demagnetizing fields as defined for ellipsoids in (1). we define B' as the Amperian flux density: + Nf and N . We have used it successfully for these calculations.. and Nf may be calculated using the mutual inductance of two model solenoids of the same diameter. (for N.a 3 } . both N . Following Brown [29]. Nf and L. where H is. we have H = Ha where F(k. could be determined using a self-inductance calculation in which a uniformly magnetized cylinder was modeled as a solenoid. we obtain the final expression forthe magnetometric demagnetizing factor: N. (13 4 (144 The definition of self-inductance is L. For ferromagnetic or ferrimagnetic materials. In cylindrical samples. In [58]. are functions of the ratio y of cylinder length to diameter and the susceptibility x of the material. (6) . ( 12b) * [12F(k. is defined as the ratio of the average demagnetizing field to the average magnetization of the entire sample [58]: (4) and po is the permeability of vacuum.)] .) and l2 -+ 0 (for N f ) . (7) + + Hd = H. A.12)E(k. in general. this x should be regarded as an effective x . (2b) If the applied field is along one of the principal axes. In this section we calculate exact values of N . is obtained when the solenoids have the same length. - NM.) + (a2 . which are commonly used in magnetic measurements. and the problem reduces to the self-inductance calculation. of the same thin solenoid and a coaxial single-turn loop of the same radius at its midplane is [62] L = (pOna/krn> [F(krn) . AND + + B' = B - poHa = po(Hd + M). (13a). (15) For N f . Formulas for Inductance When x = 0.

3116 1 0.8333 0./(po~a*).8513 0.3933 0.7126 0.007635 O. FOR x = 0" The definition of mutual inductance is ~~ Y NAO) Nf (0) N 1 . JULY 1991 The average demagnetizing field is (Hd) = *o/(pona2) .94 0.8860 0. as shown in Fig.6151 0.07991 0.6 0.40 0.30 0.00000200 0.' by Joseph [35].75 0. [60].E ~ r=O 2 i .3083 0.09835 0.7706 0.9389 0.5127 0.5 0. the volume magnetic pole density.6802 0 6 11 .3 1.9364 0.8586 0.50 0.07 0.9999 0.0008483 0.5 4 5 6 7 8 9 1 0 20 5 0 10 0 200 500 l0 o0 0.4321 0. $M ds = na2dM(z) + 2aaMr(z)dz = 0.'' surface.01889 0.9994 0.4842 0.0004299 0. U(Z) = -.001443 0. on the side surface.2572 0.6948 0. the data agree with those given by Brown [29].6158 0.6262 0.001 0.8261 0.009904 0. Ha.60 0.2187 0.9125 0.006749 0. equals 0 inside the cylinder.9124 0.02 0.08965 0. IV.8163 0.4745 0.04323 0.0001999 0. Results Values of N.8608 0. For Nf.05 0 89 0.5281 0.0754 1 0.00004999 O.71 0. proportional to V M.o 1. .6101 0. and Hd can also be written as scalar quantities. 21.38 0.4303 0. (13b) (14b) TABLE 1 EXACT FLUXMETRIC MAGNETOMETRIC AND DEMAGNETIZING Nf FACTORS AND N. is uniform in each cross section of the cylinder.6565 0.1819 0.09351 0.03461 0.21L. N is the demagnetizing factor for ellipsoids.(Z/O2'.04562 0.7505 0.2905 0.00843 8 0.OOOOO6601 L.9650 0.9993 0.6614 0.7281 0.8419 0. The final expression for the fluxmetric demagnetizing factor is Nf .5 3. 1.2592 0.1278 0.18 0.(i = 0.1096 0.2089 0.2488 0.2322 0.65 0. NO.5530 0.26 0...2861 0.m 0.01 0. n ) are constants.5563 0. by direct integration rather than inductance formulas.1 1.6456 0.12 0.04 00 .80 0. .Oo6061 0. Substituting a ( z ) = p o M r ( z ) gives.7450 0.06 0.6304 0. We further assume for this one-dimensional model that M. (20) .04799 0.3618 0.34 0.7845 0.9950 0..45 0.0 2.004927 0.08 0.8991 0. M r ( z ) is the radial component of M at the side surface.001245 0.5960 0.5b82 0.0209 1 0.4982 0.1555 0.7027 0. (16) Equations (15) and (16) have been derived. ONE-DIMENSIONAL MODELFOR LONGCYLINDERS A.3619 0. and (16) are given in Table I.7312 0.3349 0.7698 0.M.55 B = po(M + H) = poM(1 + l / x ) (17) at any point inside the cylinder.4516 0.00001 0o0 .4531 0.2 1..9546 0.8110 0.9402 0.70 0. and can be expressed by a scalar quantity as - M(z) = M. In Table I we also give N for ellipsoids of revolution with longitudinal axes 21 and transverse axes 2a calculated from well-known formulas [4].1886 0.5279 0.06728 0.the data agree with those obtained by Joseph [35] and by Chen and Li [38].5403 0.2664 0.09 0.01334 0. [31].05 8 52 0. n (18) where M2.4999 0.00000050 'Factors were calculated as functions of y using inductance formulas.(z) = r=O c M2.02842 0.(&)/A4 = 1 .7219 0.6778 0.10 0.03 0.1298 0. The material has constant susceptibility x. all poles are on the <.8905 0.1710 0.8733 0. For N. Since V * B = 0.6017 0.1501 0.00002363 O.8 2.20 0.7967 0.28 0.5662 0. 4.9638 0.(x = 0) and Nf(x = 0) as functions of y (= Z/a) computed using (6).3604 IEEE TRANSACTIONS ON MAGNETICS.24 0.3770 0.0004243 0.3944 0. VOL. the surface magnetic pole density n - 0.oooO1250 0. we have uniform surface magnetic pole densities: n U(* 1 ) = k p o ~ ( l = 5 PO ) .5438 0.5778 0.291 023 .3273 0.2186 0.0001248 0.7004 0.6432 0.36 0.041 9 1 0.5272 0. (15).002119 0. that is.4758 0.9998 098 .1087 0.16 0.9262 0.1986 0.8001 0.1351 0.3082 0.01 0.03653 0. which leads to 0. [59].6789 0.5801 0.8703 0.) For a section of cylinder of length dz at z . ( B .5947 0.9161 0.3480 0.9949 0.1736 0.32 0.8764 0.5005 0.4221 0.3333 0.4 16 .2429 0.7917 0.9845 0.41 0 1 0.9999 0.90 1 .8137 0.02029 0.1112 0.7553 0. = *o/z. 1.6352 0. the z component of M.22 0.4126 0.02865 0.004232 0.06544 0.3705 0.poa dM(z)/dz = -(poa/l) iM21(~/l)2r-'. 1.02382 0.O 3.8954 0. (19a) because $ M ds = V * M d v and V * M = 0.9694 0. Calculation o M f Assume that a cylinder of length 21 and diameter 2a is located in a uniform applied field Haalong the z axis.1941 0. For comparison.5604 0. (8).05110 0. C.4514 0. (19b) On the end planes of the cylinder.8367 0.14 0.3952 0.

For computation. r=O c [F. and where P +I z. Fig. * . * * * 9 121.1. the computed function M ( z ) oscillates. d 2 . we obtain Although the side surface magnetic pole density is a function of z only.(z) + F2(z) . + . for Long Cylinders ( 1 10) 7 In this model. and uniform a( k I) are. Similar data were presented in 1641. However. M2.(5). and x are given. n ) are set of n then obtained by solving these equations simultaneously. (27) (28) N. our assumptions of both uniform M. However. (30) (314 (3 1b) H.I II 'I CHEN et al. and get a * 1 linear equations. z. if n is too large. the demagnetizing field Hd is a function of z : H d ( z ) = Hdl(z) + Hd2(z). the susceptibility x . and ( M ) = I-' HdI({) = -a(2/h)-' -I {)2 fl(z>(z - s: M(z) dz = 1=O c M2.. respectively. As a result. and (18) and (21) into (25). At a point z = {on the z axis. and if we consider for only Nf. and M close to the ends is very small. and (25) as + + - Nf = -Hd(o)/MO = Ha/MO . Cylinder geometry and coordinate system.(z/l)21/XlM21 = -Ha.(I + {)[(l + { ) 2 + t12]-1/2 (24b) -I---- + i ( l . and Nfas functions of y when x = 0.( x -+ 0 0 ) obtained from this model are expected to be fairly accurate due to rather small end effects. and H substituting (22a) and (24a) into (21). which the middle part of the cylinder is more impoltant . * * . and F2 as z . (21) where Hdl and H d 2 are the demagnetizing fields produced by the side poles and the end poles. Rewriting the variable {in H d l .. which would lead to a nonuniform M. contradictory for constant x. and considering the H d . When x 03.I.H a ) . n (26) B.001 and lo9.. +I This is a general equation relating the expansion coefficients M2rof magnetization. if y (= l / a ) is large ( 2 lo). and Nffor x -+ 00... z = zO../(2i + 1).are unknown. n (29) [(z - + a2Ip3/*dz In principle. : DEMAGNETIZING FACTORS FOR CYLINDERS 3605 z X uy where F2(33 = . in a given Ha.. In our problem. where the ( ) brackets denote the volume average. n = 12 is a practical choice.. a. (i = 0 . can be obtained according to (4).l/x. F I I . The demagnetizing factors N for ellipsoids (dashed curve) are located between N.. We can choose n + 1 positions. we have z dependence of H and = (25) I Fig. 1. Hd is uniform in the cylinder ( H d = . the larger the number of terms in the expansion equation (26).. = . rc=! Table I1 and Fig. a uniform a +_ 1) and the side pole density would produce a nonuniform Hdzr the z component of the demagnetizing field. and the position z for a cylinder of length 21 and diameter 2a.( H d ) / ( M ) = H a / ( M ) . this one-dimensional model is a good approximation.{)2 + a2]-'/2 . zI. 1. 1. = 1.{ ) [ ( t . in fact. -'-?--'- -e IHa From (3). Nfand N. the more accurate are the results.. the applied field H. H. H ~ ( z . 2(a) and 2(b) give the calculated Nf as functions of y and 2 for -1 Ix I lo9. Calculation of Nfand N. values of N. (i = 0. = if/l2. 3 gives N .M(z)/x ) -Ha.l/X. and the n 1 coefficients M..

316 1.17 "The row for x = 0 is comparable to data for Nf(0) Table I.669 2.83 16. IO5.27 15.315 5.50 16.390 5.390 5.83 21.333 1.28 15. lo4. I 0.11 20.94 59.352 2.642 4.196 1. 300.44 10.931 1. (a) For 10 5 y < 200.944 5.026 2.250 1.16 36.616 1.384 5. AS FUNCTION y OF 20 AND x CALCULATED THE ONE-DIMENSIONAL USING MODEL' 50 Y 10 100 200 500 1000 2000 N/ X -1 -0.36 47. The last row gives A .23 13.000 2.695 9. 2.414 2.239 4.294 4.191 2.135 5.143 5.776 4.945 1.14 11.108 5.130 1.210 1.58 21.018 5.162 1.332 1.296 1.60 37.186 8.011 5. 3 x lo3. and .500 1.132 11.27 15.286 1.004 2.007 2.889 7. 300.252 1. and .664 6.965 4.237 1.238 5.980 2.377 5.77 19.808 3.058 2.3 X IO4.796 4. NO.87 11.30 (IO-)) IO-^) 1.972 1. = calculated by in '() Templeton and Arrott [45].8 -0.09 12.22 12.369 1.999 2.391 5.53 11. the curves from top to bottom are for x = OD.48 37. 0.24 12.267 1.14 14.71 38.800 10.250 1.20 (IO-') (10-7 1. VOL.256 1.289 1.77 20. .255 5.255 1.21 12.302 1.24 12.63 37.600 4.99 so 100 200 500 IO3 2 5 2 5 2 5 X X X X X X IO3 IO3 IO4 IO4 IO4 IO5 IO5 IO5 IO6 IO' IO9 rn 1.027 2.106 2.917 1.11 (IO-') 4.28 15. 30.275 8.494 8.811 11. 4.168 5.900 1.16 11.067 5.77 57.28 21.089 2.95 12.73 53.1.071 2.363 5.865 4.391 1.23 12.474 1.040 2.90 11.187 5.265 1.221 6.28 15. 100.391 5.98 (IO-') 1. I" 10 100 Y (a) Y (b) Fig.000 5.12 34.305 1.199 1.223 13.27 15.260 1.33 37.277 1.25 15.776 5.967 2.255 1.438 5.497 9.684 6.299 1. lo3. 100. 0.02 11.268 1.448 7.377 11. 10.211 5.21 14.10 11.308 1.570 11.047 5.59 (IO-') 4. the curves from top to bottom are for x = OD.642 9.03 15.5 1 2 5 10 20 (IO-)) 3.083 2.141 2.2 0.216 1.1 .661 4. 3.734 2.565 3.91 26.19 15.288 5.000 5.388 5.963 5.010 5. 21.26 15.4 0 0.210 2.292 1.215 1.620 7.43 60.62 37.021 5.779 10.803 5.016 2.328 1.011 2.237 1.950 5. Calculated Nffrom the one-dimensional model.78 36.68 15.639 6.23 12.3606 IEEE TRANSACTIONS ON MAGNETICS.82 60.21 15.28 15.68 15.037 5.322 5.100 5. (b) For 10 5 y < 1000.293 1.55 37.88 37.231 7.260 1. 1000.005 2.47 18.16 12.248 1.15 15.395 1.275 1.401 1.126 5.10 14. 1.12 22.049 2.391 5.082 5.32 31.688 3.247 5.394 5.237 4.28 15.63 37.852 4.24 12.17 11. JULY 1991 TABLE I1 N.040 5.44 21.252 1.

A method of obtaining demagnetizing fields in bodies of arbitrary shape was presented by Vallabh Sharma [36] using rectangular volume elements. where n is the unit vector outward normal to the surface of the ring. the maximum difference is above 30% 1231. / p . and z . and N.4% smaller than our result [20]. given in (32). we assume that the cylinder consists of material with constant x.6% f o r y = 10.(x = O ) . -5. especially near the comers.r:) on the end planes of the cylinder and into rings of area 2na(z2 . so the demagnetizing fields are completely specified by the surface pole density.zl) on the side surface of the cylinder. and results for x our data agree within 1. + + and use the resulting distribution to obtain the magnetometric and fluxmetric demagnetizing factors. + H d ) . and -8. The dashed ) curve is N for ellipsoids. [21]. and we are therefore able to use elements in the shape of rings about the central axis. and MJis the magnetization at the jth ring. and Della Torre [67]. Both of these studies were interested primarily in the field and magnetization distributions rather than the demagnetizing factors. A method that involves dividing the volume into uniformly magnetized elements was used by Brug and Wolf [57] for the case of thin disks that undergo magnetic phase transitions.. y( . and0. O. A calculation of demagnetizing factors for a short cylinder (y < 10) must take this variation into account.(x = 0 . For x -+ 0 0 . Templeton and Arrott’s [45] 00 (Table 11) are the most accurate.: DEMAGNETIZING FACTORS FOR CYLINDERS 3607 10 100 1000 Y Fig. Templeton and Arrott [45] used the principle that the magnetic potential is 0 at each point on a grid inside a body with infinite susceptibility. A method of calculating these factors is given in the Appendix. The interactions between the surface poles at each ring is specified using the equation M = x(H. The solid curves from top to bottom are N. and 1000 [22]. 20. J (324 where. N ! is a scalar factor that relates the surface pole on the jth ring to the r component of the demagnetizing field that it produces at the ith ring. Volume elements have also been used by Soinski [69] for rectangular and ring-shaped samples. where the magnetization sharply diverges for susceptibilities far from 0. . . N.. 0. Compared with the exact results for x = 0 in Table I.. The method of dividing the surface of a magnetized body into a set of interacting elements of uniform pole density has been used before for the calculation of the magnetic fields of rectangular bodies. varies throughout the cylinder in both the radial and axial directions.for 10 5 y < 1000. -3. The surface is divided into rings of area 7~ ( r . The following method solves for the pole distribution at the surface of a cylinder for an arbitrary value of the susceptibility.3%. but for their smallest x (12.1 % . Wurschmidt’s result for y = 50 and x 00 is 3.0% f o r y = 10. and z2 are the side-surface ring limits.4%.CHEN et al. --* m). rl and r2 are the inner and outer radii of the end-surface ring.2%. the data in Table I1 have errors o f 0 . 7 % .O%. Kadar. 3. They calculated the demagnetizing factors of cylinders and bars and later extended this to the case of a material that saturates [70]. We solve the set of equations that describes the interaction between these rings of surface poles v. the maximum deviation of the results of Stablein and Schlechtweg from our results is 3 % . 100.x Earlier results for y > 10 exist for Nf but not for N. 50. in general. Ruehli and Ellis [65] assumed a constant susceptibility. 20. we divide the surface of the cylinder into a set of nonoverlapping elements of uniform pole density. They used a demagnetizing matrix that was derived by Hegedus. The cylinder diameter is 2a. The z and r components of the demagnetizing field at a point i that results from the surface poles uJ at each ring are given by HLz = - TWO-DIMENSIONAL MODELFOR SHORT CYLINDERS A . is written in terms of the magnetization at each ring using u j / p o = n * MJ = M i . As in the one-dimensional model. for example.5%. The demagnetizing field Hd . With the cylindrical symmetry there is no azimuthal dependence of the pole distribution.56). . To obtain the distribution of poles on the surface for a specified susceptibility. N. c NY a . a n d y L 50.(x -t a). The resulting equations are for the magnetization at each point. and Normann and Mende [66] used a field dependent magnetization with the assumption that the volume pole distribution is negligible. Neumann and Warmuth’s results deviate from our results by 1. and N. [68] for interacting volume elements in cylindrical geometries. Calculation of Surface Pole Density The magnetization.

For the end planes. NO. The adjustable parameters almost entirely to the finite number of elements with which a. (34a) H. we refer to them as the surface and the volume A r = {ao a l [ ( l . For y to the distance r between the pole and the point at which = 10.0081227 Nfl 0. The second method. .1 / 3 (1 + (35) Since the first method involves surface flux calculaWe let the width of the rings A r be inversely proportional tions. Because only the normal component of M is needed at each point. We have $ M ds = 0 for a material with constant susceptibility.23196 0. (38) where SI is the surface consisting of the top half of the cylinder. * 1 ) * [ ( 1 . A total of 148 rings was used. 27.23258 0. uo(r.23225 0. We have.. over the cylinder surface above the cross section. and the volume-averaged M. sj is the area of thejth ring.(1 r / ~ ) . 20. This error is mainly due to the division of After obtaining the surface pole distribution. is calculated from a weighted average over these cross sections. For the case of infinite susceptibility. (rings on the side surface). two-dimensional model. using (37a). Therefore.1. based on the normal component of Hd is produced by poles on the the two equations ring itself with significant contribution from poles on adNf.0001 10000 Y 0. The A similar equation determines the width of the rings on source of the disagreement is a systematic error that is due the side surface of the cylinder. = x .1 1 10 0. In the end-plane rings and 132 side rings. there will be a set of n simultaneous linear equations.82582 0.. and a I are determined by the total number of rings used the surface pole density is calculated.). are averaged over the midplane for Nf and over the volume for N. with i denoting an interior point of the cylinder.. two ways to calculate the demagnetizing factors.00099941 where we have constrained the applied field to be in the axial direction.78460 0.".m = . For the first method.015670 0.3608 IEEE TRANSACTIONS ON MAGNETICS.. the pole density near the corner at the endplane is approximated as [45] where Hd and M. the average M.. Table I11 gives some examples of ~ the results obtained from both methods. For both large susceptibilities and susceptibilities near .l +~ C~N Y M J . JULY 1991 TABLE 111 Nf VALUES CALCULATEDSURFACE VOLUME BY AND METHODS AND N. VOL. The number of rings required to adequately specify the surface pole density depends on how rapidly the pole distribution varies. These at equations are solved for M. To reduce the total number of rings. from (37a). We use (32a). and the sum is over the rings on the top end-plane and on the side surface above the midplane. (34c) With the surface of the cylinder divided into n rings. The differences are even larger for N.22912 0. the calculations were repeated with 8 the field is considered. a series of cross sections corresponding to each side-surface ring is considered.) USING 74 SIDE RINGS (NjA X -1 Nfl 0. 4.r / a ) . with an r P 2dependence. In the case of N f ."M$ J (rings on the bottom end-plane). respectively. / ( M .. 0. the division of the surface of the cylinder into rings will produce a discretization error because Hd is calculated in the center of a region of Uniform B.' . Calculation o N .r / ~ ) .064998 0. in the calculation. N f .x . especially for large y or large x. (34b) o = x . the average is taken over the midplane for Nf and over the volume for N. at the center of each ring on this surface. while the second involves interior field calculato uo near the comer of the cylinder. = -x-'M.73312 0. (37a) jacent rings. and Nj f pole density. multiplied by the cross-sectional area S equals the surface integration of M. Again. there are the side surface. and 50. ) . the cross section is taken to be the midplane.' M i + c NYML J (rings on the top end-plane). using (37b).016904 Nf..I .( H d ) / < M . each ring using matrix inversion [7 11.. > = H . methods.82568 0.' / ~ ] } . We see from this (36) table that the results do not agree. + J N. the poles on the same + + . 37 on The field produced by a magnetic pole is very sensitive each end-plane and 74 on the side of the cylinder. which gives tions. To calculate N.7846 1 0. is calculated from ( M. 0. for each transverse cross section of the cylinder.0042004 0.. the equations can be rewritten as H. requires a calculation of ( Hd ) . the density of the rings is made roughly proportional to the pole density. the pole distribution diverges sharply near the corner of the cylinder.23195 0.0049267 Nf.018951 Nf$ 0.

We assume that the error in the surface pole density calculation is not too large so that Hdu/Hd Mzs/Mz. Fig. the error is 0.22% for N. the curves from top to bottom are for x = 03. and . The error becomes larger as y increases because the size of the rings on the side surface also increases. (404 0 Solving (40a) for N gives the interpolation/extrapolation equation N -0. corresponding to the cylinder end.. The normalized pole density is expressed as M .. and 0. Variation of normalized M. 5. Exceptions are when y = 20 and x 2 100 for the 74-side-ring calculation. The deviations of the one-dimensional from the two-dimensional results are -1. giving a demagnetizing factor error of < l %. 100. The entries in Table I11 provide data for illustrative computations.x). they are odd functions of z.1) and N .' = . -0.X). we calculate five values each of N m ( .0001. The side poles are narrowly distributed for x < 0.. From (37). / I M. respectively. 0 . (39c) (394 + 1). we have --$ N = M z s' - x -I . and when y = 50 and x 2 100 or y = 10 and x = 10 000 for the 132-side-ring calculation. The curves shown are for the top half of the cylinder with z > 0. On the right half.. From Taylor's data. the results are compared with the values calculated by Templeton and Arrott [45]. On the left half of the figure. on the surface of a cylinder with y = 1 for selected values of susceptibility. can correct M. for all Nf and N. They are presented in Tables IV and V and Figs. The results of Nf for y 1 10 can be compared with our one-dimensional calculation (Table 11).1 .6%. and H d . and this should give a relatively large error. the corrected quantities are written as N . all the curves have the actual sign of the pole density when H . and 7. and write ( M . can be compared with N . and Hdr. ( r = 0) I. The largest deviation is 0. 6.I U I CHEN et a l . 4 . 7 % .and Hd using we (39b) and (39d) and compare them with M z s and Hdt. : DEMAGNETIZING FACTORS FOR CYLINDERS 3609 N. Such deviations are due mainly to the approximations made in the one-dimensional model. The values for y = 20 and 50 are from the 132-side-ring calculation. For N f ( 0 3 ) . the curves are for x = 00. giving a small error. The relative effect of the end poles increases and the error decreases as y decreases. 1.5.8%.( .2% f o r y = 10 and x = 10 000. Therefore. when x 00. 4 gives the surface pole distribution of a cylinder with y = 1 for x = -1. 1. in the worst case..2%. Nf for x = 10 000 and y = 20 is 55. the error will be small or zero.1 . 0. . [41] as shown in Section VI-A. ( m ) for 0. the side poles are broadly distributed with a larger density. 0. the discretization error is less than l o % . As can be seen from the figure. 10.x ( H i . -0.2%. Ha is arbitrarily set equal to 1. For y = 20 and 50 the maximum deviation reduces to 1. compared to 53. Thus (39e) is rather accurate in most cases and the interpolation/extrapolation approach is valid.98% at y = 10.. and . For the surface and volume methods. 1 . respectively. 0. for y > 1 when.5. The greatest difference between the two-dimensional results and Taylor's method is 0.5 = Nr)(1 + N s x ) / ( 1 + N. N-' = -x(Hdl + l). With the corrected Nf. surface do not contribute to the normal component of H d . corresponding to the cylinder side. M. For simplicity. except for the pole at the point being considered.85 x lop4 for 132 side rings. .5. we write the demagnetizing factors as Ns and N. calculated with the two-dimensional model. This systematic error is correctable if it is not too large. (394 We use (40b) and the results (up to 5 significant digits) from the two methods to obtain the final Nf and N. The results agree with the exact self-inductance calculations in Table I to four significant figures except for N .. (00). A comparison can be made with less-general published results obtained for specific values of x . respectively. 1..25 Iy I4 (Table VI). For the cases of small or zero side-pole density (low or zero x ) . and 10 000. -1.17% for N . (40b) 1 r/a 0 Fig.. 1. ) and ( H d ) as M.12%.calculated from (39a) and (39c). is positive. (39e) Thus we have from (39) NJl + N x ) / [ N ( l + Nvx)l = (1 + N x ) / ( 1 + N. The two-dimensional calculation for N . We find that. 0. A region of uniform pole density adjacent to the point does not produce a discretization error.9%. which gives a smaller discretization error for these cases.46 X lop4 when computed with 74 side rings.. As an example of the worst case. -0. and -4.1 ) and 0..2% and 1.1. data deduced from the work of Taylor [40]. Values for y I10 were obtained with the 74side-ring calculation.

02 0. the magnetization is uniform but the demagnetizing field is not.5 3 4 5 7 10 20 50 9781 9576 9383 9201 9028 8704 8265 7060 6106 5309 4626 3513 2319 1199 689.1 9695 9464 9259 907 1 8896 8575 8149 7029 6200 5550 5025 4226 3410 2580 2076 1737 1494 1167 958.9 104.1 327.3 79.039 9586 9268 8990 8739 8508 8094 7558 6214 5262 4539 3968 3125 2304 1520 1078 801.21 1.2 154.2 82.27 12.946 9690 9453 9242 9048 8865 8528 8073 6840 5879 5093 4436 3399 2325 131 1 808. In cylinders.1 386.7 199.1 51.7 752.3 0.999 9628 9345 9097 8872 8664 8285 7788 6494 5535 4780 4169 3242 2320 1442 965.7 663.6 185.5 0.9 -0.8 157.110 9562 9225 8929 8663 8419 7984 7425 6054 5 109 4405 3858 3060 2291 1556 1138 871.8 176.44 1.4 9677 9434 9220 9026 8845 8516 8082 6947 6108 545 1 492 1 41 14 3295 2469 1973 1643 1407 1094 894.7 508.01 0.70 9642 9373 9139 8927 8733 8380 7920 6741 5880 521 1 4674 3862 3047 2238 1761 1449 1229 941.7 323.61 2.4 0.6 560.0 705. In these three cases.41 13.81 45.9 99.2 654.6 683.4 66.2 0.1 262.5 -0.1 233.984 N.4 689.5 505.70 9628 9347 9103 8882 8680 8315 7841 6637 5768 5096 4558 3749 2942 2 144 1675 1370 1156 897.2 0. a cylinder has N. 27.5 120. Moreover.07 0.0001 0.2 671.2 143.5 2 2. 9639 9365 9125 8906 8704 8334 7846 6566 5605 4843 4222 3273 2322 1418 935.8 762.2 84.18 9589 9275 9000 8754 8529 8127 7609 6332 5440 4768 4239 3455 2682 1921 1475 1184 981. and -1 f In ellipsoids.1 42.6 580.974 9655 9393 9163 8952 8756 8398 7922 6659 5697 4926 4292 3314 2325 1385 894.8 721.96 9544 9191 8881 8603 8349 7879 7321 5929 4990 4302 3773 3009 2278 1581 1182 927.84 12.1 137.27 12. For x + 03. Nmy.7 119.0 85.337 9548 9195 8888 8612 8359 7909 7335 5945 5006 4316 3784 3016 2280 1578 1177 920.7 293.0 616.3 0.4 89.60 9574 9247 8960 8703 8469 8050 7515 6207 5306 4640 4118 3349 2593 1850 141 1 1124 922.68 42.45 1.2 356.47 11.7 1 1.8 745.20 2.937 9728 9513 9316 9132 8957 863 1 8188 6974 6016 5223 4549 3466 2323 1246 738.1 390.8 114.9 46.3 422. For x = . VOL.9 80. For x = 0.25 19.7 1 1.1 244.3 1 -1 -0. the demagnetizing field is uniform (and exactly opposite to the applied field) but the magnetization is not.6 531.2 461..and NmZ.5 60.07 0..38 11.1 0.0 201.6 97.5 440.59 1.77 9608 93 10 9052 8819 8605 822 1 7726 6485 5604 4930 4395 3596 2804 2023 1567 1270 1063 796.67 47.0 303.0 513..8 55.6 266. Nf.03 0.1 390.961 9669 9418 9196 8993 8803 8454 7988 6738 5776 4999 4353 335 1 2325 1354 857.1 0. DISCUSSION A .4 199.5 3 4 5 7 10 20 50 9780 9576 9385 9205 9035 872 1 8300 7200 6393 5766 5260 4488 3692 2860 2339 1979 1717 1358 1123 834.9 -0.9 286.9 309.83 1.5 390.7 552.3 0.2 505.5 336.8 634.7 564.5 505.8 328.3 72.1 198.3 0.2 210.53 965 1 9390 9162 8955 8765 8420 7968 6803 5948 5282 4745 3933 3116 2301 1818 1501 1277 983.4 258.5 2 2.5 -0.3610 IEEE TRANSACTIONS ON MAGNETICS.05 9577 925 1 8967 8711 8478 8062 7529 6224 5327 4657 4134 3363 2604 1858 1418 1131 930.0 707.0 522.6 91. o r x = 0.3 1 3 10 100 10000 0.7 3 10 100 10000 Y 0.6 400.4 479.43 14.4 0.1 447.78 9665 9413 9 192 8993 8808 847 1 8029 6881 6035 5373 4839 4029 3210 2388 1898 1574 1345 1040 848.51 2. the magnetization and the demagnetizing field are both nonuniform except in two cases.0001 0.1 255.85 TABLE V N . demagnetizing factors can be calculated more accurately by introducing electromagnetic scalar potentials.7 448. and N.3 543.8 221.75 41.5 53.0 601.6 264.9 226.05 49. AS FUNCTION x AND OF X CALCULATED USING THE TWO-DIMENSIONAL MODEL -1 -0.6 152.8 128.8 251.05 0.0 0 .6 93. there are the some simple relations among N.01 0.05 0.04 0.03 1.3 106.17 12. For x = 0.4 619.03 0.6 449. the nonuniform magnetization and nonuniform demagnetizing field in the cylinder combine to exactly cancel the applied field so that the flux density B = 0.4 9730 9517 9323 9 143 8974 8660 8240 7135 6318 5681 5167 4382 3576 2744 2229 1878 1623 1277 1053 779.5 165.6 77. 4.012 9610 9313 9052 8817 8599 8206 7693 6378 542 1 4679 4084 3193 2315 1477 1013 731. JULY 1991 N.57 44.28 11.1 350.5 316.5 616.5 71. magnetometric demagnetizing factors along the three principal orthogonal axes.5 188.5 210.6 179.4 334.04 0. Nmy.0 391.4 487. a uniform applied field produces a magnetization and a demagnetizing field that are both uniform.6 548.3 144.3 584. that are .0 212.02 0.. X AS TABLE IV FUNCTION x A N D y CALCULATED THE TWO-DIMENSIONAL OF USING MODEL -0..1 798.4 411.6 468.5 0.69 2. NO.91 VI.8 139.5 655.1 93.7 -0.0 241.1.

rather than N. with N. for 0.. However. it is greater than 1. (42) 0.1 5 x < 01 from the two-dimensional model.6. Equations (41) and (42) are valid for N . the lower boundaries of the y range with the same indicated errors extend to 2 and 7 for Nfand N. the sum is less than 1.12(log 7) = N ( y ) [1 - 3)1..2 according to the more general theorem of Brown and Morrish [72]-[74]. Therefore we can relate the longitudinal CY/ and transverse CY.. As examples from Table I. based on Table V . 10. ( x -.. e a ) (ab) (UC) (ad) = Nm(03>-’. and 1.I II I CHEN er al. “conducting” means “without field penetration. 1. based on Table IV. The curves are for x = .2i 0. 6. Taylor calculated the anisotropic electric (CY) magand 0 netic ( )polarizabilities of conducting cylinders [40].25 (log 0. CY/. 0.1 . while for x < 0.8 0. (41) 0. such as a cube.: DEMAGNETIZING FACTORS FOR CYLINDERS 3611 1.0 0.1 5 x < 01 from the two-dimensional model.5 % . y) = N ( y ) [ l + 0.6 z c 0. and the longitudinal and transverse P. The curve labels refer to the left side of the figure.o N~(x Nm ( X 03. the sum is always less than 1.4 0. respectively. (43a) (43b) + 0. Nffor 0. A cube is also equivalent to a sphere according to the theorem [75]-[77].11-’9 = WA-1) 0. 0.01 Fig.1 1 10 Y Fig. Including the results for y < 10 in Tables IV and V. .. 0.30.6 z E 0. + Nmy + N.80.01 0. A cylinder with y = 0.” so electrically E = 0 and magnetically B = 0. these two shape-isotropic geometries can be used as alternatives to spheres.3333) [36]. In Taylor’s terminology. 00) and N . cylinders with y = 0.lV. For a conducting solid of revolution with volume Vo. I)]. respectively.22.01 I y 5 20 and . We can fit the Nf(03) and N .5.N m J . values of N f ( x are rather close to N for ellipsoids. only when x = 0. = Nmy = i ( 1 . = 1.o equivalent to those of a possible ellipsoid of revolution. N. the transverse magnetometric demagnetizing factors are N ./(EOVO/O) = CYff/(%~O) Nrnz(W)-’. 1 .90. and OD. (43a) and (43b) fit the model results with maximum deviations of 1 % and 1.0 0. for weakly magnetic and saturated ferromagnetic materials. If Nf is considered. 5 . and 00. In the range 10 Iy I 100.11-13 where eo is the permittivity of vacuum.. For x > 0. the theorem of Brown and Monish on the equivalence of a body of arbitrary geometry (including a cylinder) and a possible ellipsoid cannot lead to an a priori estimate of the value of N .6 are equivalent to ellipsoids with y = 0.5. ( x = 0) except for a body with center symmetry. 00. magnetic polarizabilities to N . Experimentally. Equations (44a) and (44b) are in accordance with the analogy between . For cylinders. The curves from top to bottom are for x = . -0. 10. 0.8 0.1 . 2 and 3 by simple equations: + + + 1. o..01 5 y 5 20 and . and 1.9065 is equivalent to a sphere ( N = 0. electric polarizabilities to Nm(03). 03) For a given y.P. (03) versus y data in Table I1 and Figs. = [Nrnz(-l) .1 1 10 POP///VO Y cC.4 where the demagnetizing factor N ( y ) for ellipsoids can be found in Table I. (.1). -0. [4 13.

01 0.2371 0.’s deviate from 1 by about 1 1 %. Nmz(-1).1361 0 1.5258 0.1030 1.3154 0.2590 0. x) > Nf(y.1 z “ 0..2341 0. 27.:forx~mandX= -1. (b) When both y and x are fixed.3692 0. for different values of x and y. (c) N. the ratio N.... In fluxmetric ferromagnetic measurements.05 0.3612 IEEE TRANSACTIONS ON MAGNETICS.(@) = Nmy(@)= .(dashed curves) from the two-dimensional model.5712 0. From this table we see that the sum of the N.25 0.46. and this could help one estimate N. For y = 10. .1409 0.Nmz(-l)]. the end effect is weaker at the midplane. [41]. x). (f) The minimum x for Nf(x) > O. POPI[ We give some general rules for the variation of Nf and N .15 0. the minimum values are 200.9302 1 N.5 0.5 E(-l) 1 0. Nf increases when y > 1 and decreases when y < 1.4596 0..0553 1 “The data are obtained from the electric and magnetic polarizabilities of conducting cylinders calculated by Taylor 1401.. both Nf and N . N.1) 9 N.(m) 0 E(m) 1 0. N. this rule tells us that can be used for dM/dH larger than a minimum value that depends on y. TableVIlists Nmz(oo).. as Functions of x and y y 0 N. E (m) and E ( . (d) With increasing x . and (d) give the following relation for Nf(x) and N .3669 0.8895 0. (a) For a given x . When y increases from 1 to 10 to 1000. TABLE VI B.0815 1.(O)/N’(O) increases from 1.. the deviation should be less. 1 I y I 10. electrostatic shielding in a conductor and magnetostatic shielding in the same body but with x -+ 03. (c).6764 0.Nm(@).10 0.06635 0 0. and 1000. decrease with increasing y.4319 0. 4.. Nmx(-I) FORSHORTCYLINDERS~ and y I 20.14 to 1.. (b) Linear scale.. ( .??:(-I) 1 N.3829 0.(m) 1 N.mO).(m). and lo6.00 0.1036 1. 1411.4237 0. 0. N..01 0.4111 0.(-l).(m) Nm.25 0. Rules (b).8948 0. there is a region around y = 1 where Nf is insensitive to x.1114 1.9067 0. (x) when y > 1: Nrn(-l) > Nrn(0) > Nm(03) > Nf(@) > Nf(0)> Nf(-l). NO. When y increases..2928 0.99Nf(m) increases with increasing y. with y and x based on the tables and figures. The General Rules for Nf and N.1 ) 0 0. VOL. N . N.5 1 2 4 m 0. There is a relation between and at.20 z zSE 0.1 1 10 1 2 3 4 5 6 7 8 9 10 Y Y I (b) (a) Fig. decreases with increasing x at any y. That is. JULY 1991 1 0. andNm(-1) for short cylinders based on the CY and p data given by Taylor. (a) Logarithmic scale. ( .. N. (03) / N f ( m ) is smaller and increases from 1.” although the demagnetizing fields depend on both end and side surface poles./Nfincreases with increasing y and decreasing x .(y. 5 X lo4. (454 and (45b) which leads to an exact relation between N.+N. (solid curves) and N.1618 0. 100.. 7.01 curve labels refer to the right side of the figure.8853 0.1) are thesumsN.37 to 8 to 800. (46) = -CYY.32 to 1. N. This is because the demagnetizing effect is an “end effect. the effect of cylinder ends on the midplane and the entire volume decreases.2136 0.[I .. (e) For y > 1 . For other values of x .+N.x(m).

5. and (c) are for x = . and 8(f) on finer scales.o Fig. and lo5 computed from the one-dimensional model with n = 10. x Dependence of N.0 - 0.0 3.0 -0. and lo5. 0 W N s ! 1. the signs for M ( z ) and ~ ( zare opposite to the signs shown ) for M ( i ) / M ( O ) and a ( z ) / a ( + Z ) in Fig. 30.98BE x =0. To have N f ( x ) almost equal to Nf(03). M ( 0 ) and a( 1 ) are positive for x > 0 and negative for x < 0. 8(e). H d . Graphs (d). x w) we can see in Fig. Bloomberg and Arrott [78] derived M ( z ) / M ( O ) . where N is the local demagnetizing factor. + E. (b).oooz z - -0. respectively. respectively. which makes H very small (0for x -. M .001 h0.. 8(d).001. : DEMAGNETIZING FACTORS FOR CYLINDERS 40 3.I II I CHEN er al. and f 0 For x > 0.0 1 . M becomes H d ( H d = .5 0. :fi 0 0. we have to know the details of the position dependence of M . 00) and M ( z ) + -1- 7--lr--- - -- . C. with increasing y. 8(a). Hd(Z)/Ha. (e). and a.0000 J-o.0 1 h X 0. which feeds back to the input. 9 9 N f ( w ) increases. Since H / H a = (1 + N x ) . and gain of the amplifier are Ha. Therefore.the minimum x to satisfy N f ( x ) > 0 . The curves in graphs (a). Therefore. The result is that M becomes smaller and nonuniform. the demagnetizing effect resembles the response of an amplifier with negative feedback.001.0 n U 0 (a) 5 2 .001 0. H d .0006 CI Y (c) < W U + b 1. output. M ( z ) / M ( O ) [graphs (a) and (d)].1 .5. The input.9eoE v - SI 0. and the origin of rule (f) may be understood as follows.0002 a o. with sign opposite to that of Ha. H d ( z ) / H a [graphs (b) and (e)]. 8 .5 and -1. N f ( w ) decreases with increasing y. and x .0 1 z"i: 1. 30. M ( z ) and &(z) are even functions of z . When Ha is positive. -0. and (f).oooo . After operation.0 n 0.0002 I 0.001 -1 1 I e v fl N 1. to be small everywhere. Our curves for x = 0. There is perfect feedback when x 03. in order to have N f ( x ) almost equal to N f ( m ) . -0. Also. and o ( z ) / o ( + l )[graphs (c) and (f)] as functions of z / l for a cylinder with y = 20. where N is the local demagnetizing factor). for 1 Ix < w and y = 100.' . 8(b). 8(b) that H d ( z )is a constant equal to -Ha in the entire cylinder..N M .001 are replotted in Fig. Minimum x for N f ( x ) > 0 .0002 1. while H becomes smaller than Ha.(1 + N x ) must be sufficiently large for H I H . 9 9 N f ( m ) D.001 on an expanded scale. 0. n x =0. To understand the other rules. and a(z)/a(+Z) as functions of z for y = 20 and x = -1. for x = -0. 0. + We have explained rules (a).0004 0./Nf For x = lo5 (that is.o 1 -0. Position Dependence o M . Actually Nf is the smallest local N . a smaller N f ( w ) must be balanced by a larger x to fulfill the condition of a sufficiently large (1 + N x ) . Fig. 8(a) and 8(c). From Table I1 we can deduce this minimum x to be k / N f ( w ) with 15 < k < 17. but a ( z ) iS an odd function of z . and ( f ) are for x = 0. (b).using a similar approach. and H = 0 everywhere. and 8(c) gives the curves of M ( z ) / M ( O ) .

is the largest. This increases the volume-averaged M to more than 0. fluxmetric measurements should be made with long samples. as can be seen from Fig. NO. the end-pole density has a sign opposite to that of the side-pole density nearby. and Nf takes its largest value when x = .. is 0. The average Hd is equal to . continue the trend without sudden change.. or Nf on susceptibility measurements. replacing N by Nf.. we examine the influence of the error in N. x ( .1 . (48) On the right-hand side of (48). This makes Nf a little lower than for x = 0. When x decreases to 30. Thus Nf is its largest for the highest x . which has a greater contribution from the poles in the central region than from the ends. < 1 . 4 . 8(a) and 8(b). We consider a cylinder consisting of material with constant x in a longitudinal applied field H a . F.. both M ( z ) and H d ( z ) remain nearly constant in Fig. 3. and less than 10% of the N f . depending on whether N . which gives rise to a larger end-pole density and a larger N f . M ( z ) is approximately parabolic as shown in Fig. Therefore Nf is the smallest. This is because. where the average is taken over the whole volume or the midplane. and 0 .. so that the demagnetizing effect of the end poles is partially compensated by the effect of the side poles.... ) at the midplane. H d ( z )and the absolute value of M ( z ) remain constant until z > 0. / N f largest when x = .. which can be obtained from the results of this work.. if we expand the scales [Fig.. the end pole density increases continuously with decreasing x regardless of its sign. calculation to the x determination. The reason is that. 0. When y < 1. we can see that the variation of M ( z ) and H d ( z ) continues the trend of decreasing x from lo5 to 30. When x is decreasing. N.5. 4. The second case is a magnetometric measurement for x = .. 8(c)]. and Nf can reach 1 % in general.67M(O). while the magnetic pole density in the central region becomes gradually lower. when I x( = 0. Therefore.. thus the end poles are further enhanced. To know if this accuracy is good enough for the purpose of magnetic measurements. y aged N. when x < 0. x Dependence of Nf G. In the first case.1 since Nf.. and Nf should be its smallest.81.. An external susceptibility xe can be defined as ( M ) / H a . and the modification of the field by such a small Hd ( z ) causes an invisible change in M in Fig.. VOL.. JULY 1991 finite. 8(c) and 8(f). Nf decreases with decreasing x . so the volume-aver- x = xe(1 - Nf. We can explain this as follows. When x = 0.. and Hd (0). To . the end poles are the main contributors to the demagnetizing field Hd at the midplane. Nf and N. Although all the poles are distributed on the ends when x = 0. [27]. I x ( and a l are small. To reduce the error due to the large a l ..H a . a(z) for z close to 1 increases with decreasing x and its sign is opposite to that of a ( + 1 ) [Fig.5 and . high-x materials are considered. The side poles produce a field at the center directed opposite to the field produced by the end poles. the variation of u(z) is progressively greater in the end regions. 21. In fact. They greatly increase the value of Hd in the end regions. For a given ( M . / N f even larger. the accuracy of N . can be large and still result in a small trqnsmission error.001. From (47).1 . 8(b)]. 8(a) and 8(b). increases with decreasing x even for negative x. At this point. we obtain To understand the susceptibility dependence of Nf for > 1. the end-pole density is smallest when x -+ 00 since. c y 1 .1 . the first term is the transmission error from the erroneous Nf.and then they suddenly increase. For the midplane and the entire volume. the error in Nf.. A smaller positive x repels the poles to the end regions. the variation of M ( z ) is gradual for small ? and sudden in the end regions.rn/N~ml + ( x / x A IAxelxel. in this case. so &(I) is even smaller than H. / N f is larger. 8(d) and 8(e)]. is transmitted to the final x result. The variation of the extremely small H d ( z )cannot be seen in Fig. But because Hd ( z ) becomes rather z dependent and its absolute value at z = 1 is much larger than at z = 0 [Fig. From (3). We examine three typical cases.. This means that the magnetic poles are the most uniformly distributed on the cylinder. respectively. we have a l < 0.rnxl lANf. N . the average M is M ( 0 ) and approximately 0.1. the poles are the most uniformly distributed on the entire surface. When x is small but negative. the susceptibility dependence of Nf is the opposite. both M ( z ) and H d ( z ) change their signs. As x decreases to 0. For example. Error Transmission in Susceptibility Measurements From the above analysis we see that. only a small part of the error in N. In the third case. with signs opposite to those of the end poles. This error equals the error in Nf..3614 IEEE TRANSACTIONS ON MAGNETICS. have a different effect on N. 8(a).1 in Fig. 0 6 for y = 1. 8(a) and as already pointed out by other authors [2 11... and Nf decreases with increasing y. / N f at x = . The side poles close to the ends. although a(x) remains 0 in a large central region for x < 0. The situation for x < 0 can be seen from the curves for x = -0. all the poles are at the ends. Nf continues to decrease when x becomes negative. This makes N ../Nf should be close to 1. The reason is that M ( z ) is very small compared to Ha. For oblate cylinders..1 . As a consequence.67M(O). . obtained from Table V .37. we focus on the magnetic pole distribution shown in Fig.and M by ( M )..mxc)-l* (47) This equation is accurate under the above assumptions and definitions. at present. multiplied by a factor a1 = I N f . since Nf < N. However.17. and 10. This makes N . Finally. and the measurement error of x e can be derived as I A X / X I = INf. Further discussion of this in terms of pole distribution is given in the next section. . a(z) is 0. Therefore. error is reflected in the final x result. the relative error in x caused by the calculation error of Nf. we have the largest value of N .. a(z) varies with z almost linearly on the cylinder surface except for the regions close to the ends. the pole density on the ends is not the largest. or Nf is considered. For these two cases. For the largest x . 8(b).

we have a l + a2 = 1 . and an averaged differential susceptibility in the sample. we have a2 = 2 from (47). weakly magnetic materials. . In an ideal type-I superconductor. : DEMAGNETIZING FACTORS FOR CYLINDERS 3615 ensure that aI < 1 based on Table I1 and Fig. y should be 12. no matter how large the xe measurement error is. respectively. In has very large y and x.01) only very small demagnetizing corrections are needed.. and H = H. where the subscript m denotes the maximum value at the endpoints of a symmetric magnetization loop. = M . + 03. when y is small so that N. These magnetic moments allow us to ascribe values of M . We accounted for a 0. I. We can regard x as the normal susceptibility x. from (47). one uses Nf. .. large negative values of x arise in ac magnetic measurements of normal conductors and both ac and dc measurements of superconductors. to determine x accurately. our values of N. our 0. But a knowledge of x is required to select the appropriate Nf value. / H . These include paramagnets. respectively. The resultant x is an effective susceptibility xeff. Nf is estimated based on the measured xe using Table I1 or Fig. and 700. From a magnetic measurement point of view. We have xeff= x. For the high-x case.2% accuracy for x = -1 is required for superconductor calibration standards.at x = 0 is more than adequate.. and lo5.. a 1% accuracy may be sufficient for many uses but is inadequate for others. For these cases.(x) values except in the initial state and when approaching saturation. [64]. so the error in x is almost the same as the xe measurement error. or when the sample ) . our Nf. For other x values. which ensures that this error will arise. depending on the measurement conditions. even in these cases. respectively. this treatment involves some error. ( x = . when H . Zijlstra [58] has suggested a simple method to obtain a reasonably accurate loop in which the demagnetizing corrections for M and H are performed using Nf(0)and Nf(03). our results can still be used satisfactorily for long magnetically soft materials over a wide field range. even if the requirements of [79] are followed [38].1) can be used.. our calculation accuracy for Nf. For measurements of semihard ferromagnets with intrinsic coercivity H. lo4. Thus there is an equivalence between these cylinders and a normal perfectly diamagnetic cylinder. to obtain accurate results of x for high-x materials.1) and Nf( x = . and H . An iterative process may be used. such materials can also be measured by fluxmetric methods. some caution is required. x = -1 because B and the permeability po( 1 + x) both equal 0 at every point in the material. with the uncertainty based on the measured scatter in xe. so the unknowns can be calculated simultaneously. 2. 2.001 from (47). Then a better estimate of Nf is made. + 1 and a1+ 1. nonzero finite effective x (corresponding to the differential susceptibilities at H = H . and the number of calculated points is sufficient for accurate interpolation over a wide range of values. the higher is the required calculation accuracy of Nf. values of N. 58. other cases. values are satisfactory. With our results. A source of error in fluxmetric measurements is if the sample diameter is less than that of the measurement coil. fluxmetric measurements on weakly magnetic materials require many coil turns. and iron-powder cores and ordinary ferromagnets in the initial and saturation states. the higher the I X I of the material..4% volume decrease of both the standard sphere and the sample cylinder upon cooling to 4 K.I II I CHEN er a l . For materials with linear or nearly linear magnetization curves. The susceptometer was calibrated using the known demagnetizing factor and dipole field of a superconducting niobium sphere [80]. and deduced a value of N. and x to these materials.Its value is between x. For magnetometric measurements of weakly magnetic materials (x < 0.. a2 is very close to 1 since x = xe.361 0. the resultant xeffmay be larger or smaller than xn. This is interesting because. Furthermore. We give an example below. For example. Since the differential susceptibility is field dependent in ferromagnets. However.. For our second case. and x is calculated from xe and Nfusing (47). 2. --$ 0 or H . and Tables IV and V. However. leading to a double transmission error to x from the xe measuring error. . equal to 0. In summary. For the first case where cyI is very small. The error in x due to this effect can be as large as 30%..033) below the critical temperature.) can be used more properly. However. where bulk magnetic moments have their source in eddy currents and shielding supercurrents. The corresponding factor is a2 x/xe. and our results can be used for demagnetization corrections of their magnetic measurements. the requirement for calculation accuracy will be determined by the particular purpose of the magnetic measurement. H . the x error is mainly due to the error in N. As a consequence.. Application to Materials with x > 0 Most materials have x > 0. as mentioned above. spin glasses. is close to H . The second term in (48) is the transmission error due to the measurement error of xe.. However. The known xe and y and the unknown x and Nfare related by (47) and Table I1 or Fig.. We have verified this experimentally with a magnetometric low-field ac susceptibility measurement of a niobium cylinder (y = 1. A large demagnetizing effect would occur in the measurement of M because the measured flux linkage is contributed not only by the M of the sample but also by the Hd within the coil volume produced by the sample’s poles. if a1 = 1 . ( x = 0) are more than adequate for experimental work. The magnetization curves of ferromagnetic materials are nonlinear. H . The same applies to a normal conductor in an ac field when the skin depth is negligible compared to its dimensions. Remarks Concerning x < 0 and Nonuniform x For normal diamagnetic materials with uniform x. as recommended by at least one measurement standard [79]. lo3. for x = lo2... ( x.. including the use of N. both accurate Nf and accurate xe are strongly required if y cannot be made very large. 200. and use Nf(x). and it is difficult to assign to them specific x and Nf. To extract an unknown x from fluxmetric measurements of xe on samples with known y.

our Nf. who calculated the field produced by a disk of charge. The results of this work have to be used more carefully for materials that do not have constant susceptibility. = exp ( . G(z. or a finite height./<2lz. VOL. Otherwise.[a -a iom exp exp ( ..l>l[ r2 Iom exp (-xlz.h ~ ~ ~ ( ) J ~ ( X a ) Jd~ ( h r ...1). 4) dr d4 = ( u r 2 / 2 ~ 0 ) exp (-A som I zj - zi I) ('45) h-'J1(hr2)Jo(hri) d h . and asrepresent the potential due to a ring on the end-plane and side surface. 4 ) dz d4 - where F ( r . and the effective x should be close to the differential susceptibility at each point.I)Jl(hr2)JO(X~. and a complex susceptibility with a negative real part can be defined [81]. (A61 where Jo and J I are Bessel functions and 2a is the diameter of the cylinder. different from our model assumption of uniform susceptibility. where a portion of the magnetization comes from penetrated supercurrents that follow the critical-state model [82]. In iom a+./ar. 4 ) represent the integrands in (Al) and (A2). Since most cases of magnetic measurements involve nonlinear magnetization curves.1 may be used in the limit of small skin depth.[r2 . 4) and G(z.values for an effective x < 0 can be used. l ~ J l ~ ~ d ~ ] . Thus we see that cylindrical superconducting standards for magnetic measurements for use at low fields and temperatures can be made using accurate values of N . 4) dr d4 - j:' F(r. ) dX 1. r = -.. an effective susceptibility should be found. the demagnetizing correction using the factors calculated for constant x should be made cautiously.. For this. as on the end-plane of a cylinder.3616 IEEE TRANSACTIONS ON MAGNETICS. y must be large enough so that only a small correction is needed.. For example.f.zi)2 dz.demagnetizing effect is most important. M ( H ) increases when H > H. A similar case arises in hard superconductors. ~ J ('47) ~ ~ . (-43) (A4) +s = G(z. 4) dz d4. ~ - ~ I ~d h . l ~(A81 (~ rj[(Zj - Zj)2 + rj + r: (All and.. ~ hl o ~ N : = ( P o / 4 a+.f4. (. . The magnetic scalar potential at a point (rir z. The method follows that of Gray [83].~ l z . + a2 + r: (A3 = . . Results are presented both for the case where a ring has a finite width. F(r. The parameters rl and r2 are the inner and outer radii of the ring of poles on the end surface./% = fS. for rings on the side surface.3622. Normal conducting cylinders in ac fields have M ( H ) loops. Similar caveats apply to the intermediate state of superconductors.rl som exp (-A lz. This causes a discontinuity in the value of N . d4. as on the side surface of a cylinder. Similar expressions where the limits of integration are r l and z1 are obtained from the previous expressions by replacing r2 with rl and z2 with z I . Nf. APPENDIX The demagnetizing field produced by a ring-shaped distribution of magnetic poles is calculated below. 27. = f S r . In the mixed state. The pole density is taken to be uniform over the width or height of the ring. The interaction constants NY and N:' are defined for rings on the ends as N : = (Lco/u) a@. N:' = (PO/U> fiz. as . which is positive.2rjrj cos 4]-'12drj d4. 4) dr d4. = rz. a deep understanding of the magnetization process and the . where (P. NO. The definite integral S F ( r .2~ia COS S": Sz* [(zj . and zI and z2 are the limits in the z direction of the ring on the side surface. 4 ) dr d4 has been evaluated by Gray [83].~ l z . this susceptibility is due to eddy currents constrained by the skin effect. respectively. to obtain good results. ~ 1 (A101 o ~ ~ ~ ~ . with x = .) produced by a pole density U on the j t h ring is given by 9 = (U/4Tpo) . 4. above H.f ./az. The limits of integration are changed to give +e = p sp F ( r . JULY 1991 Our two-dimensional calculations give 0.) d h - rl I exp ~ .. In these cases. and a proper demagnetizing correction should take account of this effect. I)JlW2)Jl O r . we evaluate the other integrals with the results. l ) J l ( ~ ~ l ) J ld h ]~.. the M ( H ) curve of an ideal type-I1 superconductor in fields below the lower critical field H. is linear. ) h 1 . Using a similar method.1. However.for x = . ('49) +$ = (ua/4?rp0) .

vol. 219-276. Ewing. 470-471. 1-6. and T. REFERENCES [I] G. a useless concept. Soc.” Proc.” Ann. translator./ 1 zj I when needed to account for the sign reversal that occurs when zj < zi. 23-43. 1934. . Warburton. 140-159. Oxford: Clarendon. Theorie des Enrmagnerisierungsfakrors und der Scherung von MagnefisierunRskurven. London. 359-369. 1896. [20] J . L. Warmuth. 1901. 1896. [9] Lord Rayleigh (J. 5 . 26. W. A. Evans and A. 1885. vol. Kos and J. pp. Sweden: Wezata-Melins Aktiebolag. 31. 1925. [I81 G. Thomson (Lord Kelvin). (AI 1) We have arbitrarily set zi = 0 for ease of notation and have included the factor z. 48. Atkinson. pp. B. 5 . steel. Aug. ser. Strutt). [8] J . A. M a g .. Soc. First published 1873.. [13] C. “Mean intensity of magnetization of soft iron bars of various lengths in a uniform magnetic field. 15.p)’I2(1 . =fir -fir ACKNOWLEDGMENT dX = [z2/(21z201 a iom exp (-hlz2l)J.” in Encyclopaedia Brirannica. “Ueber die Entmagnetisirungsfactoren kreiscylindrischer Stabe (Demagnetizing factors of cylindrical rods). 6 ) is related to the Heuman lambda function and is expressed in terms of the elliptic integrals of the third kind. Rowland. A Treafise on Elecfricifyand Magnefism. 175-183. London. 1892. [6] H. Foster. 1886.-Apr. f&. “Uber die rechnerische Auswertung ballisticher Entmagnetisierungsfaktoren kreiszylindrischer Stabe (Mathematical evaluation of the ballistic demagnetization factor of circular cylindrical rods).(Xa)J. 1883. M a g . 43. [I21 H. Rev. Clerk Maxwell. pp. 2. 1929. Nov. London: Macmillan. “Tables of demagnetizing factors for iron rods..” Wiss.. Acud. “On the magnetic character of the armourplated ships of the Royal Navy. 6 .” Am. Phys. B. vol. - [z1/(21z101 a ’jomexp(-Xlzl()J.(1 . 165-169. vol. London: Longmans. and S . apparent hysteresis. L. “Experimental researches in magnetism. Chrystal.I CHEN et al. vol. Smith. “The magnetic pole. pp. Thompson and E. Li.” Phil. The parameters p . Phys. Moss. Mann. 1873. [I71 P. [ 1 I ] A. [2] W. 21. 66-72. Feb. M a g . Arrott. Dec. [I91 S. W. Sept.” Proc. Wiirschmidt. Nottingham.) dX. 263-323. Preece. pp. 321-340. VerOf. Physik. 1888. “Demagnetization factors for cylinders. 450-456. The elliptic integrals are evaluated numerically using the procedure given in [MI.” Phil.” Z.” Phil. vol. 155. Mar. 1865. factors f. pp. 8. W. 176. “1. R. 1151 C. Dec. 4. B. and the maximum of magnetism of iron.. “Magnetism. 1954. (161 F. Wiirschmidt. pp. Rev. London. . 1872. W.” Phil.. E. ser. vol. Reprinted Goteborg. pp.-Z. NoguCs for assistance in computer programming. pp. 304313. Benedicks. Trans. The integrals in theffactors are given in [84] and reduce to where a2 = r2 and b2 = ri. 2. 1958. P. .: DEMAGNETIZING FACTORS FOR CYLINDERS 3617 N: = (po/u) a+. 52. 1828. [5] H. and sin2P = z2/[(a .k 2 ) sin2P]. 22. “On magnetic permeability. Soc. pp. Rowland. (A15b) where a2 = ri and b2 is the radius of the cylinder. A. 185-256. pp. A(cy. Aug. p .k 2 / p ) ’ k I ( a . F. J. k = sin a.: Wheelhouse. “Demagnetizing factors. [I41 C. and + b2(2na2)Y1Mcy2. Trans.” Phys. vol. pp. pp. vol. 66-73. Phys. Amer. pp. vol. and F ( k ) and E ( k ) are the complete elliptic integrals of the first and second kind.(Xa)Jl(Xr. ser./&-. P) = (1 . . “Magnetische Anfangspermeabilitat.P2))z2/lz2l (a2 < b2). U. Teacher (Am. 4 .-On the energy of magnetized iron. vol. 1907. 1910. and 6 are specified by p = k 2 / [ 1 . Braunschweig: Sammlung Vieweg. Shtrikman for helpful comments. 1909. and nickel.(Xr. “On the magnetic permeability and maximum of magnetism of nickel and cobalt. vol. pp. Shuddemagen.” School Sei.. Petersen for help in preparing the figures. 11. and on the effect on the compass of particular arrangements of iron in a ship. du Bois. Neumann and K. 309-315. Green.. Roy. [4] J. W. [2 I ] J . 1970. 128-164. cy. B. ser.) We thank A. Aug. The f&.b)2 + z 2 ] . 7r/2). [22] H. 1923. vol. vol. 622-633.” Phys.. Jul. Physik. 726-740. and behavior under vibration). An Essay on the Application of Mathematical Analysis to rhe Theories of Electriciry and Mugnerism. vol. The Magnetic Circuit in Theory and Practice. vol. 3. NOV. pp. 46. pp. 7 . Mag. S .” Phil. 12. [IO] D. scheinbare Remanenz und Verhalten bei Erschiitterungen (Magnetic initial permeability. [3] F. A ( a . Sept. 25-35. 1932. “The demagnetizing factors for cylindrical iron rods. Reprinf of Papers on Elecfrosfaficsand Magnerism. Siemens-Konzern. ser. pp. A. For each case k = 4ab/[(a + b)2 + z2]. Arts and Sei. “An experimental method for the determination of the ballistic demagnetization factor. 3rd ed. and f i r are given by the same expressions except r2 is replaced by rl and z2 is replaced by zI . vol. “On the self-demagnetizing factor of bar magnets. pp.?. 1874. Roy. Shuddemagen. Rev. 9th ed. vol. 523-640. [7] C.K. Dec. pp.).” Phil. Notes on magnetism.’’ Phil. Mag. Reprinted New York: Dover. pp. Tanakadate. J . Green.

1965. [66] N. vol. Smythe. [72] W.” J. vol. 1960. pp. and A. Phys. vol. 1945. Phys. 2382-2387. [55] R. “Electric polarizability of a short right circular conducting cylinder. [29] W. 1977. Kraus. pp. pp. 2 . B. “Magnetization distributions in bulk magnetic material. Jan. pp. and E. pp. vol. MAG-14. Templeton.. Archer and E. 21. vol. I. 1987. MAG-23. “A diagonal sum rule concerning demagnetization tensors in composite bodies. 135143. vol.. “The magnetostatic energy of coaxial cylinders and coils.3618 IEEE TRANSACTIONS ON MAGNETICS.” Proc. Heinrich. A. S . Appl. Stoner. vol. 4. Osborn. 1984.” Phil. S. Arrott. Leeds Phil. (271 R. Nut. vol. [32] G. 1761 E. A . “Ballistic demagnetizing factor in uniformly magnetized cylinders. 1963. vol.” J. pp. Magn. 209-21 1. 31-38. [59] J. “Formulas and tables for the calculation of mutual and self-inductance. 488-492. 1967. 512-519. Inst. [74] R. 33. and T.” J. New York: Van Nostrand.” IEEE Trans. Lit. vol. pp. Phys. Jr. 54. Sinica. vol. [61] E. [30] F. Phys. E.” Am. 478-482. Magn. M. (251 K. 52. Phys. Res. pp.” Brit. 49-53. Press. Heinrich. no. Morrish. 242.. Mende. “Theoretical aspects of demagnetization tensors. pp. “Rapid computation of magnetic anomalies and demagnetization effects caused by bodies of arbitrary shape. “Micromagnetics of curling configurations in magnetically soft cylinders. 4685-4694.” J. 1973. Appl. Smythe. Ackers. NOV. B. Appl. Wolf.” IEEE Trans. 199210. “Uber den ballistischen Entmagnetisierungsfaktor zylindrischer Stabe (Ballistic demagnetization factor of cylindrical rods). 24. 3. Jul. L. 39. 89-109. 761-779. W. 1986. MAG-19. (451 T. 2387-2389. [38] D.. “Numerical calculations of the fielddependent magnetization of short rectangular prisms.” IEEE Trans. 1241 K. Aug. Templeton. 739-744. vol. Jan. May 19821. 1198-1201. pp. R. Jun. [37] M. R. B . 2650-2652. 1491 A. pp. Mag. Appl. Rhodes and G. J . 19. NOV. “Charged right circular cylinder. Standards. 36. Chen. 542-551. pp. 55. pp. Oct. 1990. Warmuth. pp. equation (73). Amsterdam: North-Holland. Eng. Normann and H. vol. Rosa and F. Teukolsky. Sat0 and Y. 1117-1125.” Phys. W.. Physical Basis of Magnetic Measurements. [64] D. Phys. [39] D. vol. Della Torre. [41] T. Okoshi. Soc. 6840-6843.” J. Klem. Jul.” J . vol. Vetterling.” J . Bozorth. and W. Magnetostatic Principles in Ferromagnetism.” IEEE Trans. NOV.” Bull. 66. Beijing: China Mechanical Industry. 1945. 30. “New analytical expressions for flux distribution and demagnetizing factor of cylindrical core. Heinrich. MAG-15. “Demagnetizing fields in the de Haas-van Alphen effect. pp. 1968. JULY 1991 1231 F .. pp. 16. 1939. Phys. pp. 1986. Kadar.’’ J. vol. Experimental Methods in Magnetism. vol.. 142-162. 1936. E . 15. Bur. 51-56. 235-242. pp. M. Thin discs. T. 54. “Numerical calculation of magnetic fields in the vicinity of a magnetic body. Aharoni. vol. “Zur Darstellung des ballistischen Entmagnetisierungsfaktors zylindrischer Stabe (Ballistic demagnetization factor for cylindrical rods). 102. 1-8. 46-47. 1960. L. Its Appl. 1960. Taylor. [56] L. 7.. “On the error of measurement of feebly magnetic material in regard to demagnetizing field. R e v . Math. T. Ah. “Die Bestimmung des ballistischen Entmagnetisierungsfaktors mit dem magnetischen Spannungsmesser an Staben von quadratischem Querschnitt (Determination of ballistic demagnetization factor for specimens of square section). 4364-4373. Flannery. Appl. 630-646. pp. pp. Brown. MAG-9. H. 39. “Partially saturated ferromagnetic cylinders. Instrum.. Bureau Standards. Chen. pp.” IEEE Trans. “Magnetic polarizability of a short right circular conducting cylinder. Geophys.” J . 1935. 3094-3097. J. Jan.” Archiv Elektrotechnik.” J. vol. May-Jun. Magn. 2. [43] W. [69] M. 50. F. . Oct. Hegedus. [48] A. [67] C. W. “A note on the demagnetizing energy of a uniformly magnetized cube. “Phenomenology of ferromagnetism: I. P. 1937. E. Chapin. 1331 R. Jul. pp. Bureau Standards. J. pp. “Magnetostatic energy of a saturating cylinder. 69-72. 77-85. Guancial. Sol. pp.-X. Amsterdam: North-Holland. 747-763.” J. and A. Jr. 37. Aharoni. 95. Appl. Appl. 4639-4643. Aharoni. E. Rosa and F. pp. 1979. 1990. vol. Ruehli and D. Iwata. 1912. Mar.” Phys. pp. Inductance Calculations. Dec. [44] W. pp. New York: McGraw-Hill. 1211. 704-710. Develop. J. and E. “Demagnetizing factors of the general ellipsoid. 1954. T. C. [42] W. Kadar. Res. Magn. “Charged right circular cylinder.. 124-130. Jr.” Bull. and R. 1966. Arrott. vol. vol. A. vol. pp. 1956. Phys. 8. Jul. pp. Joseph and E. “Demagnetizing fields in magnetic measurements. Mar. “Single-domain particles: New uses of old theorems. Res. pp. B. H. [53] T. pp. pp. Magnetostatic Principles in Ferromagnetism. [35] R. Grover.” J. pp. E. Nov. 1912.. 1981. vol. vol. NO. 803-821. Dec. [60] E. May 1977. “The demagnetizing factors for ellipsoids. 1983. Aug. Grover. IEEE. 1985. Phys. “Magnetostatics of rods and bars of ideally soft ferromagnetic materials. 25. 1983. 27. pp. H.” Proc. pp. “Simple and approximate expressions of demagnetizing factors of uniformly magnetized rectangular rod and cylinder. Ferromagnetism. Arrott. “Demagnetizing field in nonellipsoidal bodies. pp. L. vol. Feb. pp. “Demagnetization matrices for cylindrical bodies. 15-19. vol.. “Uher den Entmagnetisierungsfaktor zylindrischer Stabe (Demagnetizing factor for cylindrical rods). 67. 1962. 1989. Appl. 295-303.. S. Warmuth. 13.” Phys. Brown. pp. F. ’ 1946. 983-985.-Dec. 6 . Sept. vol. Appl. Bureau Standards. Phys. “Demagnetizing factors of rods. Cambridge.” IEEE Trans. Cohen. Rev. U. “An exact formula for the mutual inductance of coaxial solenoids. Schlomann. “The demagnetization tensor of a cylinder. B. May 1973. pp. R. [51] W.” 2. Rowlands. pp. Feb.-X. 1579-1593. Ishii. equation (54).. “Demagnetization effect of rectangular and ring-shaped samples made of electrical sheets placed in a stationary magnetic field. G. Nut. “Boundary integral equations analysis of induction devices with rotational symmetry. 102A.-X. A. Sept. 276-277. Stablein and H. Appl. [34] J. Amsterdam: North-Holland. 31. 1621 E. P.” Electr. 845-849.. Lee and J. Phys. Magn.. Arrott. Burke. Meas.. and A. K. Della Torre. Appl.” J.” IEEE Trans. vol. 13. 187-192. Jr. KaczCr and Z. ser. vol.. Oct. vol. vol. C. Smythe.. “Effect of a cavity on a singledomain magnetic particle. 1989. vol. 14.. 1966. “Magnetostatic energy of a ferromagnetic cylinder. Appl. Warmuth. 35.. Appl. pp. pp. Oct. 351-357. Phys. 1461 A. J. pp. Soinski. pp. Sept. M. Bur. 23.” J. 1-237.. pp. Magn. May 1907. 1978. F. 279-291. “Cylindrical demagnetization matrix. Moskowitz and E. 1681 G.. Feb. 64B. Magn. Static and Dynamic Electricity. M. Effects of magnetostatics on susceptibility. Crabtree. Bozorth and D. 36. vol. Ballistic and Bridge Methods of Magnetic Measuremenls of Materials. (281 R. [73] W. W.” Pure Appl. 1957. May 1976. 1966. Standards. M. pp. Della Torre. Yamamoto and H. 917-920.1979. Brown. Fawzi. L. Brug and W .. 1982 [Denki Gakkai Ronbunshi. Joseph. B. . Stat. VOL. 64B. 1979. 217-224. Phys. Mar.” J. I. pp. vol. “Formulas and tables for the calculation of mutual and self-inductance.” Acta Metall. Dec. : Cambridge University Press. pp. Vallabh Sharma. Taylor. pp. [75] P. Chen.” J. M. H. and A.. [31] W. vol. 1631 L. 2189-2191. 1971. Templeton and A. Templeton.. Aharoni. vol.” Archiv Elektrotechnik.” Czech.. 1471 A. Aug. Phys. I. vol. Zijlstra. pp. Sept. vol. W.. 36. B.. “Demagnetizing factors of rods and tubes computed from analog measurements. p. [71] W.” IEEE Trans. 1701 T. Rev. S . 1951. vol. Phys. Princeton: Van Nostrand. Aharoni. Beijing: China Metrology. 57. Hegedus. 3rd ed. Jan. vol. [77] T.” Archiv Elektrotechnik. S. May 1965.” Bull. vol. vol. vol.-Z.. T. “Point singularities and magnetization reversal in ideally soft ferromagnetic cylinders. 36-44... MAG-2. 8. [54] Y. Jpn. Numerical Recipes. 1228-1235. F. May 1985. pp. Li. 2966-2967. [57] J . vol. 105. Grover. Brown.K. p. Moskowitz.” IBMJ. Schlechtweg.” Appl. 255-262.. pp. 28. Dec. pp. 1521 T. 1966.-Sept. Della Torre. [40] T. Chen and B. [58] H. 191-210. [65] A. A. vol. (261 K. Ellis. K. F. Appl.. Physik. vol. May 1942. Phys. vol. 1962. 1962. 1983. Arrott. 1-237.” Phys. Jun. and P. Dec. “Demagnetising energies of uniformly magnetised rectangular blocks. “Tabulation of magnetometric demagnetization factors for regular polygonal cylinders. 64. pp. Dec. 320-326. (361 P. [50] A. Phys. 33. S. Yamada.

Instrum. deo Continuous Media. CO. 2nd ed. vol. and at the National Institute of Standards statics of an iron single crystal whisker. versitat Autbnoma de Barcelona. [80] R.D. Mag. vol.. A247. B. D. He received the [81] L.. 1975. His areas of research are ac losses and coupling effects in at Beijing University from 1958 to 1964. and Technology in Boulder. Magnetic Measurements of Materials (China Metrology. vol. and the magnetics of magnetoresistive recording heads. 1919. pp.. 36. 30-36. J . Beijing.S. and W. “Magnetization of high-field superconductors. Cambridge. Aug. Annual Mechanical Industry. Sci.S.S. His research has focused on noise [83] A.S. 529-551. and L.04. ~ / 2 ) in 1951.D. 03. He has pub[79] American Society for Testing and Materials. Academia Sinica. U. Press. He is currently on leave from Hewlett-Packard Laboratories in Palo Alto.” Phil. degree from low. He studied physics Boulder. Goldfarb and J.: Pergamon. Yetterling. pp.” Rev. U.I I1 I CHEN et al. Roy. “Standard test methods lished two books entitled Physical Basis of Magnetic Measurements (China for permeability of feebly magnetic materials. Bean. Philadelphia: ASTM. Sneddon. Beijing. 761-764. cryogenic magnetic properties the Institute of Physics. and I. vol. M. May 1984. China. 201-214. PP. Goldfarb (M’79-SM’86) was born in Mexico City.” Can.and high-temperature superconductors. Soc.K. Colorado. 1985) and Ballistic and Bridge Methods of Book of ASTM Standards. [84] G. in 1981. Colorado. 53. 1984. N. P. Phys. degrees in physics from Colorado State University. n . Goldfarb is a member of the American Physical Society and the measurements. E. 183-190. 38. Brug (M’87) was born in Cody. H. Mexico. Electrodynamics B. Boulder. Minervini. He is a member of the Superconductor and Magnetic Measurements Group at the National Institute of Standards and Technology (NIST) in Du-Xing Chen was born in Shanghai. Spain. [82] C. P. A. Trans. Eason. pp. Jan. from 1985 to 1989.” Phil. and magnetic degree from the Central Iron and Steel Research Institute. Teukolsky. Ph. B. Bloomberg and A. vol. WY. James A. Apr. M. magnetic materials.: Cambridge University Press.D. 1955. Dr. as an ASEE postdoctoral fellow at the National Institute of Standards Mod. II(*/2. in 1985. pp. P. f gree in applied physics from Yale University. instrument development. S. Numerical Recipes. CA.. S.p . of concentrated spin glasses and weakly magnetic alloys. in 1954. Gray. pp. received the M. measurement of magnetic properties at expression in terms of Bessel functions and elliptic integrals. Ronald B. and Technology. “On certain integrals of Lipschitz-Hankel type involving products of Bessel functions. vol. His areas of research are magnetic Dr. in 1941. pp. . and superconductor theory and meaAmerican Society for Testing and Materials.A. Lifshitz. 1990). Brug is a member of the American Physical Society. “Notes on electric and magnetic field constants and their mechanisms in magnetic thin films. Landau. He received the B. B. Oxford. Aug. Pitaevskii. and with k = sin 01 and n = . 31-39. Stockholm. He was a visiting scientist at the Royal Institute of Technology.” Rev. London. 5 5 . Arrott. V. 205-207. degree in physics from Montana State University and the Ph. “Calibration of ac susceptometer for cylindrical specimens. [85] W.K. small length scales. degree in materials science from Rice University and the M. 1990. k) corresponds to our ~ ( L Y p . S. and received the Ph. 1986. during 1988 and 1990. “Micromagnetics and magnetosurements.” A342-84. Flannery. From 1979 to 1981 he was a National Research He is a visiting Professor in the Electromagnetism Group at the UniCouncil postdoctoral research associate at NIST. pp. Phys. Noble. degree in electrical engineering and the . 1964. Beijing. T .: DEMAGNETIZING FACTORS FOR CYLINDERS 3619 [78] D. 1454-1471.