Project Narrative

Cover Page

Company Name & Address: Particle Beam Lasers, Inc.

18925 Dearborn Street
Northridge, CA 91324-2807

Principal Investigator: Robert J. Weggel

Project Title: Development of Open-Midplane Dipole Magnets

For Muon Accelerators

DOE Grant Number: DE-SC0004494

Topic No: 64 Advanced Concepts and Technology for High

Energy Accelerators

Subtopic: (b) Technology for Muon Colliders and Muon Beams

1
7. Degree to which Phase I has Demonstrated Technical Feasibility

A primary objective of Phase I was to develop a conceptual and preliminary design of high-field
open-midplane dipoles appropriate for a muon accelerator or collider and to confirm that there
were no “show-stoppers” that would preclude a Phase II. The preliminary design had: a) good
field quality (~0.01%); b) magnetically-supported inboard coils; c) an unobstructed channel to an
energy-deposition warm absorber far from any coils; d) acceptable stresses and deformations at a
central field of at least 10 T; and e) the potential for substantially higher fields with HTS and the
stress-management techniques proposed for Phase II.
Phase I also predicted the energy deposition—both energy density and integrated power—for a
variety of coil and absorber geometries. Phase II would have continued these energy-deposition
simulations in order to refine parameters such as gap width and absorber location to reduce the
heat load on the coils.
An R&D plan for Phase II was developed. This work included a conceptual design and structural
analysis of the coils, support structure and hardware that would have been needed to build and test
a proof-of-principle test magnet in Phase II.
The work performed in Phase I is summarized in the following sections.

7.1. Design of Open-Midplane Dipole: Equations for Field, Force and Field Homogeneity

To generate designs with optimized combinations of central field B 0, field homogeneity ∆B/B0,
peak-field ratio Bmax/B0 and conductor volume or cost, while guaranteeing that the vertical
magnetic force Fy on each inboard coil will attract it away from the magnet midplane, analytic
equations may be preferable to finite-element methods (FEM) to compute field, force and field
homogeneity. For a bar of infinite length, rectangular-cross section and carrying a uniform current
density J in the z direction, the vertical field By is [20]:

−μ0 J (u−x )du dv

B y=
2π
∬ ¿¿
¿

where cB = μ0J/4𝜋, and ui and vj are shorthand for ai−x; and bj−y, the horizontal and vertical
distances, respectively, from a corner [ai, bj] of the bar cross section to the field point [x, y]. Bx is
of the same form, with ui and vj interchanged.
This SBIR provided the motivation to derive corresponding equations for the horizontal and
vertical components of force, Fx and Fy, between two parallel bars of conductor, of current density
J and J ' , and to incorporate the formulas into computer programs. The equation for each
component of force has sixteen terms. For Fy they are of the form:
2 2
F y =c F {(3 v −u )u ln (r ¿)+ 2 v ¿¿
where cF ≡ cB J ' /3 and r = u2+v2. u and v are shorthand for ui,m and vj,n, the horizontal and vertical
distances from one corner [ai, bj] of the first bar to a corner [am, bn] of the other bar; i, j, m and n
each run from 1 to 2. The equation for the horizontal force Fx is similar, with u and v interchanged.
The field along the x or y axis of a dipole with mirror symmetry about the planes x=0 and y=0
may be expanded in a power series of distance from the center point [x=0, y=0]; because of the

2
mirror symmetry, the expansion will include only even-order terms—e.g., B y = ∑ B y . With the
(n )

n even
shorthand of the previous equations, u ≡ ai−x, v ≡ bj−y, r ≡ u +v , and c B now written as C, the 2 2

field-inhomogeneity coefficients have the form:

2
v(2) d By
B ≡ 2
=C
dy 2r
B =−C v ( 3 u −v ) r /12
(4 ) 2 2 −3

B(6 )=C v ( 5 u−10u 2 v 2+ v 4 ) r −5 /30

B(8 )=−C v ( 7 u 6−35 u4 v 2+21 u2 v 4 −v 6 ) r −7 /56
10 ¿ 2 2 6 4 2 2 4 6 −9
B ¿=C (3 u −v )(3 u −27 u v +33 u v −v )r /90
B(12)=−C ( 11u10−165 u8 v 2 + 462u6 v 4−330 u 4 v 6−55 u2 v 8 −u10 ) r −11 /132
By evaluating these equations, an optimization routine such as Excel’s “Solver” can iteratively
adjust the conductor placement and other parameters of a dipole magnet in order to achieve: 1) a
desired central field; 2) zero inboard force on its most-inboard conductor; and 3) field-
inhomogeneity coefficients of desired magnitude, typically zero up to order N—thereby achieving
a field homogeneity termed “N th order”. The iterative procedure can succeed in finding a solution
even when starting from initial parameters that are quite far from satisfying any of the above
constraints. This is particularly true if the order of field homogeneity is modest. For systems with
field homogeneity of high order, a fruitful starting point is a magnet which satisfies constraints #1
and #3, by the technique to be described below.
To reduce the number of magnet parameters to manageable size and visualize more easily the
effect of conductor placement, the magnet employs conductors that are not bars but wires. The
field contribution By and first seven even-order terms in the y-axis field expansion B y =∑ B for
(n)

n
a wire at [x, y] carrying a current I are:

(0 ) x μ0 I 2 2
B ≡ B y =−C ; C= ;r ≡ x + y
r 2π
x ( x 2−3 y 2 )
2
(2) d By
B ≡ 2
=−C 3
dy r
(4 )
B =C x (x −10 x y +5 y 4 )/r 5
4 2 2

(6 ) 6 4 2 2 4 6 7
B =−C x (x −21 x y + 35 x y −7 y )/r
B(8 )=C x (x 2−3 y 2)(x 6 −33 x 4 y 2 +27 x 2 y 4−3 y 6 )/r 9
(10) 10 8 2 6 4 4 6 2 8 10 11
B =−C x (x −55 x y +330 x y −462 x y +165 x y −11 y )/r
B(12)=C x( x12−78 x 10 y 2 +715 x 8 y 2 −1716 x 6 y 4 +1287 x 4 y 6 −286 x2 y 8 +13 y 10)/r 13
(14) 2 2 4 2 2 4 8 2 6 4 4 6 2 8 15
B =C x ( x −3 y )(x −10 x y +5 y )(x −92 x y +134 x y −28 x y + y )/r

3
 Fig. 8 plots B0 through B(12), evaluated with C = 1 and x = 1. Note that each curve B(2n) is quasi-
sinusoidal, of decreasing frequency, with n zeroes, not including the one at y = ∞. In order to plot
values for y>1, without allocating an inordinate fraction of the graph to do so, the abscissa υ,
which is identical to y when υ<1, has been distorted, when υ>1, to υ ≡ 2−y −1, so that y ≡ (2−υ)−1;
for example, υ = 1.8 (the right-hand limit of the graph) corresponds to y = (2–1.8) −1 = (0.2)−1 = 5.
To improve the readability of the graph when y >> 1, each function B(2n) has been multiplied by
(1+y2)n. 2 n
Field Derivatives, Multiplied by (1+y ) , at [0, 0] from Wires at [1, y]; Iz = 2.5 MA

1.0

0.8

0
Field: B [T]
0.6

0.4

0.2
(12)
B
(y)

(10) (6) (4) (2) 2

B B B B [T/m ]
0
(2n)

(8)
B
By

-0.2

-0.4

-0.6

-0.8

-1.0
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
-1
y = (2-abscissa) if abscissa 1
Bob Weggel 5/16/2011
Fig. 9: Field derivatives B(2n), multiplied by (1+y2)n, from a wire at [1, ±y].

A magnet of 4th order field homogeneity requires B(2) = 0; to accomplish this requires only a
single wire in each quadrant, each with location given by the zero crossing of the black curve: y/x
= ±1/√3 ≈ 0.577. The gray curve, evaluated at y = 0.577, reveals that the field generated by the
magnet is 75% that were the wires at y = 0 instead of y/x = ±1/√3.

To design a magnet of, say, 12th order field homogeneity, one can locate wires at the five zero
crossings of the B(10) curve (turquoise): y = 0.1438, 0.4567, 0.8665, 1.5560 and 3.4057. No wire in
the set will generate a 10th derivative of field, whatever its current. Lower-order derivatives will
arise from each wire individually, but the set of wires as a whole can be made to have zero
derivatives of all orders 2, 4, 6 and 8 by solving a set of five linear equations, with coefficients
calculated from the equations above, and plotted as the black, red, magenta and blue curves of Fig.
9. In this example the resulting system is quite inefficient: the currents that solve the set of linear
equations are, respectively, 1.0, 1.18, 1.72, 3.35 and 12.34; conductors #4 and #5 are inefficiently

4
far from the origin. Obtaining an appealing solution therefore calls for the optimizer program to
penalize inefficient usage of conductor. In this example, the optimizer program was able to zero
out conductors #4 and # 5 completely, resulting in a 12th-order magnet with only three conductors
per quadrant: y = [±0.16939, ±0.56473, ±1.31460] with currents, respectively, of [1, 1.38946,
4.20104]. This solution guided the input values for the program which optimizes magnets with
conductors in the form of bars instead of wires and which simultaneously guarantees that the force
on the most-inboard bars be away from the magnet midplane.

For muon colliders, cos(θ) dipoles are expensive because the bore needs to be large to
accommodate shielding to protect the conductor from radiation from the decaying muons. Open-
midplane dipole designs banish windings from the path of this radiation. The design concept
proposed here—an outgrowth of R&D for an LHC luminosity upgrade [10, 11]—banishes
structure, too, from the midplane. Support for the windings closest to the midplane is via
magnetic attraction from outboard windings [Fig. 8].

Figs. 10 through 12 show the conductor cross section and selected field-homogeneity contour
lines from 10 ppm to 1000 ppm for magnets with field homogeneity of, respectively, 4 th, 6th, and
8th order.

4
100 ppm

80

60
3
y [mm]

40

20
2

0
0 1 2 3 4 5 6 7

x [mm]

5
 nd
Contours of ppm Field Homogeneity of OMD2 15mm
1.1

1.0

0.9

0.8

1000 ppm
0.7

0.6

y [cm]
0.5 500

0.4

200
0.3

100
0.2
50
0.1 20
10 ppm

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

x [cm] Bob Weggel 4/16/2011

Fig. 8: Simple (two bars per quadrant) OMD of 30-mm -gap. Left: 1st-quadrant windings cross
section & field magnitude B ≡ (Bx2+By2)½ (color & contours). B0 ≡ B(0, 0) = 10 T at 200 A/mm2;
Bmax/B0 is only 107%. The muon beam is at [0, 0]. The lobed end of the keyhole accommodates a
radiation absorber. Right: Contours of field homogeneity; red curve is ∆B/B0 = 1x10-4.

th
Contours of ppm Field Homogeneity of OMD4 15mm
2.0

1.8

1.6

1000 ppm
1.4

1.2
500
y [cm]

1.0
200

0.8
100

0.6 50

20
0.4
10 ppm

0.2

0
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
x [cm] Bob Weggel 4/14/2011

The magnet midplane can be truly open, because the inboard bar of conductor experiences a
vertical Lorentz force that is upward—not only in total but on the left and right halves separately,
to preclude tipping toward the midplane; the horizontal force is 1,356 kN/m. For the outboard bar
the force components are Fy = −3,650 kN/m and Fx = 4,194 kN/m.
FEM computations confirm that support structure of sufficient cross section can limit stresses
and deformations to acceptable levels with a central field of 10 T [Fig. 9]. The von Mises stress to
the right of the keyhole is benign, being compressive. The average tension in the web between the
inboard and outboard bars is only ~150 MPa at 10 T; the predicted maximum deformation δ max is
less than 0.27 mm. One goal of Phase II will be to minimize stresses and deformations by
techniques such as coil partitioning, to increase the feasibility of fields as high as 20 T.

6
Fig. 9 Stress and strain in OMD of Fig. 8 with support structure xmax = 40 cm; ymax = 20 cm. Left:
Von Mises stress, σvM. To the right of the keyhole the primary stress is compressive, with a
maximum von Mises stress σvM of 246 MPa. The average tension in the web between the two coils
is ~150 MPa. Right: Predicted total deformation, magnified twentyfold.

The open-midplane geometry is amenable to countless variants. For example, Fig. 10 shows a
magnet with three conductor bars per quadrant, with field homogeneity of so-called “4 th order”—
i.e., zero 2nd-order coefficients ∂2B/dx2 and ∂2B/dy2. Its region of 0.01% homogeneity is four times
larger in area than in Fig. 8.

16

14

12

200 ppm

10

100
y [mm]

8
40

10

0
0 2 4 6 8 10 12 14

x [mm]
2 2 2 2
Fig. 10: OMD magnet with three bars per quadrant and ∂ B/dx = ∂ B/dy = 0; B0 = 10 T at 200
A/mm2. As in Fig. 9, the field ratio Bmax/B0 is only 107%. Left: Field magnitude (color &
contours) & direction (arrows). Right: Contours of field homogeneity ∆B/B0 in parts per million.

7
 Fig. 11: OMD of Fig. 10. Left: Contours of von Mises stress, σvM; average σvM is ~180 MPa in
the web at [x = 0; 3.6 cm < y < 6.6 cm]. Right: Total deformation, amplified twentyfold.

The stresses in the web between the windings range up to 180 MPa (26 ksi), even discounting
localized stress concentrations; deformations range up to 0.37 mm. Doubling the field to 20 T
would quadruple these values. A challenge in pursuing the design of a very-high-field OMD
magnet is to limit stresses and deformations to avoid mechanical failure, magnet quenching, and
the degradation of field quality. Phase II proposes to address these concerns.

th
Contours of ppm Field Homogeneity of OMD6 15mm
5.5

5.0

4.5

4.0

1000 ppm
3.5

500
3.0

y [cm]
200
2.5
100

50
2.0
20
10 ppm
1.5

1.0

0.5

0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
x [cm] Bob Weggel 4/12/2011

th
Contours of ppm Field Homogeneity of OMD6 15mm
5.5

5.0

4.5

4.0

3.5
1000 ppm

3.0 500
y [cm]

200
2.5
100
50
2.0
20
10 ppm
1.5

1.0

0.5

0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

x [cm] Bob Weggel 4/17/2011

8
 th
Contours of ppm Field Homogeneity of OMD8 15mm
6.0

5.5

5.0

1000 ppm
4.5

500
4.0

200
3.5
100

y [cm]
3.0 50

20
2.5
10 ppm

2.0

1.5

1.0

0.5

0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

x [cm] Bob Weggel 4/13/2011

Dipoles are capable—in theory at least—of field homogeneity adequate for magnetic resonance
imaging. Figs. 12 & 13 show the conductor-placement in dipole magnets (modeled as infinitely
long) with field homogeneity of 1 ppm (part per million) throughout a cross section more than 30
cm in diameter, the standard for thoracic MRI magnets. The magnet of Fig. 13 is of “12 th order”;
i.e., the leading term in the polynomial expansion of its field is proportional to the 12 th power of
distance from the origin.

22

20

18

16
10 ppm
5
14
2
1
12
y [cm]

0.5
0.2
10
0.1 ppm
8

0
0 3 6 9 12 15 18 21

x [cm]

Fig. 12: Dipole magnet with midplane gap and field homogeneity appropriate for MRI. Left: 1 st-
quadrant coil placement and field magnitude. Distance between inboard faces of inboard coil = 50
cm. B0 = 2 T. Right: Contours of field homogeneity, from 0.1 to 10 parts per million.

9
 th
Contours of ppm Field Homogeneity of OMD12 15mm
13

12

11

10
1000 ppm
9 500
200
8
100
50
7

y [cm]
20
10 ppm
6

0
0 1 2 3 4 5 6 7 8 9 10 11 12
x [cm]
Bob Weggel 4/18/2011

21

18

10 ppm
5
15
2
1
0.5
12
0.2
y [cm]

0.1 ppm

0
0 3 6 9 12 15 18

x [cm]

Fig. 13: Compact dipole magnet (no significant midplane gap) with MRI-quality field
homogeneity. Left: 1st-quadrant coil placement and field magnitude. B0 = 2 T. Right: Contours
of field homogeneity.

Conductor MA-m, Bmax/B0 and Mean Radius of Homogeneity Cylinder

20
/B [%]; and mean radius of homogeneity [cm]

10
Bmax/B0

2 MA-m

1000 ppm
1
0

500
max
B

.5 200
Conductor [MA-m]; 

100 ppm

50
.2
20

10 ppm
0.1
2 4 6 8 10 12

Order of field homogeneity

Bob Weggel 5/18/2011

7.2. Predictions of Energy Deposition and Consequent Temperature Rise

7.2.1. Energy-Deposition Predictions

10
 In a 1.5 TeV center-of-mass muon collider storage ring, muons decay into electrons (and into
two neutrinos) at a rate of 5x10 9decays/s per meter. About 1/3 of the muon energy is carried by the
electrons, which are deflected toward the inside of the ring by the dipole magnetic field. The
radiation (energetic synchrotron photons and electromagnetic showers) is ~200 W/m per
circulating beam, directed mostly outward in the horizontal plane of the storage ring. The energy
deposition must not exceed the quench tolerance of the superconducting coils. To predict the
energy deposition we use the code MARS15 [21].
Our simulations assume either one unidirectional beam or two counter-circulating muon beams
of 750 GeV, with 2x1012 muons per bunch at a rep rate of 15 Hz. Absorbing tungsten rods are
place in the mid-plane to intercept the bulk of the radiation. Figure 14a shows the result for a
unidirectional muon beam traversing an open-midplane dipole of 6-m length and 15-mm half-gap.
For this example, the peak power density on the inboard coil (nearest the midplane) is 0.13 mW/g
on the right (inward) side of the bend and 0.05 mW/g on the left (outward) side. For the outboard
coil the respective peak power densities are 0.14 mW/g and 0.07 mW/g. These values are within
the nominal quench limit of 1.6 mW/g [22].
Note that the tungsten absorber on the inward bend side has a slot in its left side (as in Fig. 2b),
to reduce backscattering from the absorber. To eliminate backscattering completely it may be
possible to remove the right-hand absorber—the one that backscatters more radiation—by
completely opening the magnet midplane on its right side, as in Fig. 14b. Preliminary stress
predictions suggest that such a design is indeed feasible.

Fig. 14a & b: Left: Energy deposition from a unidirectional muon beam at the downstream end of
a 6-m-long open-midplane dipole with half-gap of 15 mm. Right: OMD magnet with structure of
“C” shape, without the right-hand absorber, to eliminate its backscattering of radiation onto nearby
conductors; maximum σvM to left of keyhole = 353 MPa.

We study the energy deposition from the muon beam in the muon collider on the open-midplane
dipole for a + − collider of 1.51.5 TeV. Fig. 15 shows the MARS model of open-midplane
dipoles with a) two coils per quadrant (similar to Fig. 9) and b) three coils per quadrant (similar to
Fig. 10). This work follows the work of N. Mokhov and S. Striganov from 1996 for a non-open-
midplane dipole for a + − collider of 2 TeV on 2 TeV [23].
Our Phase I calculations using MARS [21] give heating estimates similar to Mokhov and
Striganov. The major backgrounds come from the decay of − into electrons—or + into positrons
—and other particles. Figure 16 shows the simulated positron energy spectrum, which is

11
consistent with the results of Mokhov and Striganov.

Fig. 15: MARS model of cross section of 6-meter-long open-midplane dipoles and sagitta orbit.
Left: Two coils per quadrant. Right: Three coils per quadrant. The red blocks are
superconducting coils; the arrows indicate the direction of the magnetic field.

Fig. 16: Positron energy spectrum from decaying muons (50,000 events).

12
 Positrons/electrons from muon decay have a mean energy of ~250 GeV (~1/3 that of the muons).
Generated at a rate of 5x109/s per meter of ring, they travel toward the inside of the ring and
radiate energetic synchrotron photons in the plane of the ring. The positrons/electrons shower to
produce not only electrons and photons but also—eventually, and to a much lesser extent—
neutrons and other charged and neutral hadrons and even muons, which create high background
and radiation levels both in the superconducting coils and in the storage ring. Each muon beam
generates~200 W/m of heat. Figure 17 shows the energy deposition near the beam exit of the
dipole magnet.

Fig. 17: Energy deposition in dipoles of Fig. 15 at downstream end, where it is expected to be
greatest. Left: Two coils per quadrant. Right: Three coils per quadrant.

In Figure 3 we see the energy deposition predicted by Mokhov, et al. for an open-midplane
dipole for the LHC. Mokhov and Striganov studied the attenuation of azimuthally-averaged
energy deposition density in the first superconducting cable shell as a function of the tungsten
liner thickness for a cos(θ) dipole and confirmed that thicker liners are better. Similarly, we have
calculated the energy deposition for open-midplane dipoles with half-gaps of 15 mm, 30 mm, 50
mm and 75 mm (Figs. 18 and 19).

13
 Fig. 18: Predicted energy deposition. Left: Half-gap = 15 mm. Right: Half-gap =30 mm.

Fig. 19: Predicted energy deposition. Left: Half-gap = 50 mm. Right: Half-gap = 75 mm.

Table 3 lists the peak power density in the inboard and outboard coil in each quadrant.
Increasing the gap tends to reduce the maximum energy deposition density, but half-gaps of 50
mm and 75 mm are worse than 30 mm because their tungsten absorbers are too close to the coils
and therefore backscatter radiation onto them.

Table 3: Peak Power Density [mW/g] vs. Gap of OMD for Unidirectional Muon Beam

Half-gap Inboard coil Inboard coil Outboard coil Outboard coil

height in Q1/Q4 in Q2/Q3 in Q1/Q4 in Q2/Q3
15 mm 0.06 0.018 0.115 0.105
30 mm 0.009 0.012 0.0028 0.008
50 mm 0.04 0.021 0.0355 0.001

14
 75 mm 0.0175 0.011 0.0065 0.0002

7.2.2. Temperature Rise in Open-Midplane Dipoles from Steady-State Energy Deposition

Equations derived and evaluated for Phase I reveal that at least some of the power-dissipation
densities of the previous section are within range of conduction cooling through the stainless steel
(Sst) structure surrounding the superconducting bars. The equations model the winding pack and
it’s surrounding Sst as concentric annuli centered on the muon beam. Heat flows radially through
each annulus, of thermal conductivity k [W/cm·K], from its inner radius r i to its outer radius ro.
The power deposition can be a surface heat flux w s or a power density wv that may be uniform or
non-uniform, decreasing inversely as the 1st, 2nd, 3rd or 4th power of the radius.
For a surface heat flux density, the temperature rise is ∆T = c ln(r), where c = ws ri / k, and r is
the normalized outer radius ro/ri. For a volumetric power density, the equations are of the form ∆T
= c Fn(r). For a uniform power density, wv = constant, F0 = [r2 – 2 ln(r) – 1] / 4. The remaining
equations are F1 = r ln(r) – 1; F2 = ln2(r) / 2; F3 = ln(r) + 1/r – 1; and F4 = [2 ln(r) + 1/r2 – 1] / 4.
Table 4 presents the results for the temperature rise in the Sst from power deposited in the Sst
itself. To obtain the total temperature rise in the Sst, one needs to add the contribution from the
surface heat flux density ws at its inner surface from the heat flowing into the Sst from the winding
pack. To estimate the total temperature rise in the winding pack one can model it as another
concentric annulus of inner radius riʹ, outer radius roʹ = ri and thermal conductivity kʹ. This
contribution to temperature rise is likely to be small, because of the high thermal conductivity of
the copper stabilizer that accompanies the superconductor.

Table 4: Power-Deposition Density for 1 K ∆T in OMD’s Cooled at Outside of Sst

half-gap, ymin cm 1.500 3.000 5.000 7.500

inboard ymax cm 2.780 4.646 7.113 9.793
inboard xmin cm 3.073 4.690 6.935 9.247
inboard xmax cm 10.36 17.46 27.33 37.90
center of dump cm 21.58 28.66 35.26 48.90
left edge of dump cm 19.24 25.11 30.20 42.60
angle to corner degrees 147.3 139.8 135.3 130.0
core cross section cm2 44.5 115 238 473
radius of core cm 7.53 12.09 17.41 24.54
xsteel cm 20.0 25.0 30.0 37.5
ysteel cm 40.0 50.0 60.0 75.0
Asteel + Acore cm2 800 1250 1800 2813
outer radius cm 31.9 39.9 47.9 59.8
∆r of annulus cm 24.4 27.8 30.5 35.3

15
 radius ratio -- 4.24 3.30 2.75 2.44
304 SSt ck W/cm·K 0.003 0.003 0.003 0.003
wv @ i.r. mW/cm3 1.00 1.00 1.00 1.00
∆T with unif. Wv K 8.84 7.56 6.58 6.47
∆T if wv~1/r K 4.51 4.46 4.28 4.47
∆T if wv~1/r2 K 2.62 2.87 2.97 3.25
∆T if wv~1/r3 K 1.71 2.00 2.18 2.47
∆T if wv~1/r4 K 1.22 1.49 1.68 1.94
1K wv if wv = c mW/cm3 0.11 0.13 0.15 0.15
1K wv if wv~1/r mW/cm3 0.22 0.22 0.23 0.22
1K wv if wv~1/r2 mW/cm3 0.38 0.35 0.34 0.31
1K wv if wv~1/r3 mW/cm3 0.59 0.50 0.46 0.41
1K wv if wv~1/r4 mW/cm3 0.82 0.67 0.60 0.51

Table 4 shows that the stainless steel of the open-midplane dipole designs of the previous section
will tolerate a power deposition density of ~0.1 to 1.0 mW/cm 3 (~0.015 to 0.15 mW/g) with an
allowed temperature rise of 1 K and the Sst cooled only at its outside. For the four magnet designs,
the permissible power deposition density values range is 0.11-0.15 mW/cm 3 if the energy
deposition is uniform and 0.51-0.82 mW/cm3 if the power dissipation is localized as (ri/r)4.
With some difficulty, one can incorporate either copper conduction paths or helium cooling
channels into the support structure, to increase the permissible energy deposition density to that
permitted by conduction cooling at the external surfaces of the conductor bars.

We now examine the energy deposition profile of a single circulating beam through a set of three
contiguous dipoles each of 6 m length. The dipoles have a 30 cm layer of tungsten following each
dipole. Each tungsten layer has a 2.6 cm aperture located at the position where the beam intersects
the plane hence allowing the tungsten layers to act as collimators.

Fig. XX Three 6 m long dipoles with 30 cm thick tungsten layers at each exit end. The muon
beam direction is left to right.

16
The cross-section of the dipoles is shown in Fig. XY. The mid-plane of the dipole has a total gap
of 6 cm with the upper and lower portions of the gap each lined with 1 cm low-Z material (for
thermal insulation), 5 mm of tungsten, and 2 mm of stainless steel thus giving a total clear gap of
2.6 cm for the muon beam and the generated decay products and radiation.

Fig XY. Cross-section of each dipole showing various layers: thermal insulation (yellow),
tungsten absorber (orange), and stainless steel. The cold mass consists of iron (blue) and
superconducting coils (green).

For this simulation, muon beam decay was confined to the interior volume of the initial (1st) dipole
for a total released energy of 1200W. The simulations yielded energy depositions in the 1st, 2nd,
and 3rd dipole cold masses of 1.1, 3.7, and 0.05W respectively for a total of ~5W or 0.4% of the
radiating power from the decay of the muon beam.

7.3 Design Studies for Proof-of-Principle Open-Midplane Dipole

The following magnetic and mechanical models develop a preliminary design of a proof-of-
principle (PoP) open-midplane dipole whose design is to be refined and then built and tested in

17
Phase II. It is a truly-open-midplane dipole, devoid of material that would backscatter radiation on
its way from the beam pipe to a warm absorber beyond the coils.
For economy this novel open-midplane dipole structure is to use coils which are available from
other programs or at least can be made with tooling from these programs. This restricts the design;
however, we were able to find solutions. For Nb 3Sn coils, the leading candidates are designs from
LBL and/or BNL. For HTS coils, we propose to use the coils that are being built for the Facility
for Rare Isotope Beams (FRIB).
For the proposed Nb3Sn PoP open-midplane dipole we considered open-midplane gaps (coil-to-
coil separation between the inboard faces of the inboard coils) of 10 mm, 20 mm and 30 mm. In
all cases we were able to find coil parameters that guarantee that the outboard coil attract the
inboard coil away from the midplane. Thus, magnet designs of large gap are viable. However, the
gap of 10 mm (Figs. 20 & 21) gives the best field homogeneity and the highest central field, 9.7 T,
and therefore is the leading candidate for the proof-of-principle magnet. The details of the coil
geometry will be described in this section, with more details in Phase II.
The FRIB coil (Fig. 22), of high-temperature superconductor, is to generate 1.4 T at 50 K and 5
T at 4 K.

Fig 20. Dimensions of Nb3Sn coil with coil-to-coil gap of 10 mm and free gap of 4 mm.

18
Fig. 21: Nb3Sn open-midplane dipole with coil-to-coil gap of 10 mm. Top left: Field magnitude,
B (color). Top right: B(x). Bottom left: B(y). Bottom right: B(z).

Fig 22. FRIB coil of HTS. B0 ≈ 1.4 T at 50 K and ≈ 5 T at 4 K.

7.4. Summary of Phase I Accomplishments

19
 Phase I has advanced the feasibility of open-midplane dipoles for accelerator and storage rings
of muon accelerators and colliders. First-order magnetic and structural designs and analytic
techniques have been developed to advance the design process. Preliminary energy-deposition
predictions—to be refined greatly in Phase II—show promise of adequately limiting the energy
deposition in the superconducting coils. The SBIR has generated a candidate design to fabricate
and test, for the first time, a proof-of-principle dipole of a truly-open-midplane dipole.

9) Phase I Work Plan

 Develop parameters of the Open-Midplane Design

o Basic lattice and overall machine design
o Specify preliminary field quality requirements
o Magnet aperture
o Clear gap (no material)
o Magnet length

 Develop magnetic design

o Coil to coil gap
o Conductor requirements
o Pure HTS vs. hybrid design
o Conductor choices
o Preliminary cost of various conductors

 Mechanical design
o Stress/deflection calculations
o Preliminary mechanical design concept

 Energy deposition estimates

o This work will play a major role in determining the open midplane gap

References
1. http://map.fnal.gov/organization/MAP-Approval.PDF
2. Y.A. Alexahin, et al., “Muon Collider Interaction Region Design”, IPAC10.
3. First workshop on Muon Colliders (Napa, CA, 1992), Nucl. Inst. Methods, Vol. A350, pp. 24-56
(1994)
P. Chen & K. MacDonald, Summary of the Physics Opportunities Working Group, AIP Conference
Proceedings, 279, Advanced Accelerator Concepts, 853 (1993)

20
 Mini-Workshop on mu+ mu- Colliders, Particle Physics and Design, Napa CA, Nucl. Inst. Met., A350
(1994), ed. D. Cline
Physics Potential and Development of mu+ mu- Colliders, 2nd Workshop, Sausalito, CA, ed. D. Cline,
AIP Press, Woodbury, NY (1995)
Ninth Advanced ICFA Beam Dynamics Workshop, ed. J. Gallardo, AIP Press (1996)
Symposium on Physics Potential and Development of mu+ mu- Colliders, San Francisco, CA, Dec.
1995, Supplement to Nucl. Phys. B, ed. D. Cline and D. Sanders.
4. M. Alsharo, et al., “Recent Progress in Neutrino Factory and Muon Collider Research within the Muon
Collaboration”, Phys. Rev. ST Accel. Beams 6, 081001 (2003)
5. M. Green and E. Willen, “Superconducting Dipoles and Quadrupoles for a 2 TeV Muon Collider”,
IEEE Trans. on Appl. Supercond., Vol. 7, No. 2, June 1997.
6. P. McIntyre and D. Gross, “m+ m- Collider Dipole Concept”, Informal Seminar, April 2, 1998.
7. B. Parker, et al., "Magnets for a Muon Storage Ring", PAC’01 (2001).
8. P. Snopok, M. Berz and C. Johnstone, “A New Lattice Design for a 1.5 TeV CoM Muon Collider
Consistent with the Tevatron Tunnel,” PAC’07, Albuquerque, NM (2007).
9. N.V. Mokhov, et al., “Energy Deposition Limits in a Nb3Sn Separation Dipole in Front of the LHC
High-Luminosity Inner Triple”, PAC’03, Portland, USA, May 2003, http://www.jacow.org.
10. R. Gupta, et al., "Optimization of Open Midplane Dipole Design for LHC IR Upgrade," PAC’05,
Knoxville, TN, USA (2005).
11. R. Gupta, et al., “Open-Midplane Dipole Design for LHC IR Upgrade”, MT18, Morioka City, Japan
(2003).
12. R. Gupta and W. Sampson, "Medium and Low Field HTS Magnets for Particle Accelerator and Beam
Lines", ASC, Chicago, August 2008.
13. G. Greene, R. Gupta and W. Sampson, "The Effect of Proton Irradiation on the Critical Current of
Commercially Produced YBCO Conductors", ASC, Chicago, August 2008.
14. R. Gupta, et al., “Status of High Temperature Superconductor Magnet R&D at BNL,” MT18, Morioka
City, Japan (2003).
15. Principles of Magnetic Resonance, 3rd ed., Springer, New York (1996)
16. Becker, ed., High Resolution NMR Theory and Application, Academic Press, New York (1980)
17. D. Bobela and Taylor “Nuclear Magnetic Resonance Studies of Tellurium and Antimony Bonding in
Crystal Sb2Te3 GeTe and Ge2Sb2Te5”, Jap. J. of Appl. Physics 47, 10, 2008, pp. 8162-8165.
18. Fyfe, Solid State NMR for Chemists, CFC Press, Guelph, Ontario (1983)
19. Mehring, M. “High Resolution NMR of Solids”, Springer, Heidelberg (1981)
20. D.B. Montgomery, assisted by R. Weggel, Solenoid Magnet Design, 2nd printing, pp. 260-271, Krieger
(1980)
21. N. V. Mokhov, http://www-ap.fnal.gov/MARS/
22. N.V. Mokhov, et al., “Energy Deposition Limits in a Nb3Sn Separation Dipole in Front of the LHC
High-Luminosity Inner Triple”, PAC, Portland, USA, May 2003, http://www.jacow.org.
23. N. V. Mokhov and S. I. Striganov, “Simulation of Backgrounds in Detectors and Energy Deposition in
Superconducting Magnets at μ+ μ− Colliders”, 9th Advanced ICFA Beam Dynamics Workshop, ed. J. C.

21
 Gallardo, AIP Conf. Proc. 372, 1996.
24. D.R. Dietderich, et al., “Correlation Between Strand Stability and Magnetic Performance”, IEEE Trans.
Appl. Supercond., Vol. 15, no. 2, pp. 1524-1528.
25. R. R. Hafalia, et al., “An Approach for Faster High Field Magnet Technology Development”, 2004
Applied Superconductivity Conference.

22