1. Fiberglass Reinforcement of High Field Pulsed Magnets 2. Development of a High Pressure System for use at Low Temperatures and High Fields and David.

1. Fiberglass Reinforcement of High 1. Fiberglass Reinforcement of High Field Pulsed MagnetsField Pulsed Magnets

2. Development of a High 2. Development of a High Pressure System for use at Pressure System for use at Low Temperatures and High Low Temperatures and High

FieldsFields

and

David Barbee ’04

Advisor: Prof. Charles Agosta

OutlineOutline

A little about superconductivity and A little about superconductivity and pulsed field magnetspulsed field magnets

Getting fiberglass on a magnetsGetting fiberglass on a magnets

Creating high pressuresCreating high pressures

Pulsed Field MagnetsPulsed Field MagnetsWhy do we need pulsed magnets?Why do we need pulsed magnets?

To study the behavior of high field superconductors: the critical field (Hc2), vortex state, and the spin-magnetic field interaction. Studying these phenomena provides a better understanding of the mechanisms behind superconductivity (Ginzburg-Landau, BCS Theory.)

In the normal metallic state the motion of the electrons in aHigh magnetic field provides information about the effective electron mass, scattering time, and fermi surface via Haas von Alphen and Shubnikov de Haas oscillations.

Pulsed Field MagnetsPulsed Field Magnets

How do we make pulsed field magnets?How do we make pulsed field magnets?

By tightly winding several hundred turns (700) of copper wire around a delrin rod. The rod will eventually be removed to create the bore of the magnet.

How are these pulsed fields created?

By dumping the energy in a capacitor bank into the magnet, creating roughly 5,000 – 10,000 amps of current in the copper wire. The magnet can be treated as a solenoid to first approximation.

B = μ0NI/L = 4π · 10E-7 Tm/A · 700 turns · 5000 A / 0.1 m = 44 T

44 tesla! Won’t that produce some large lorentz forces?

A Few CalculationsA Few CalculationsCan copper wire alone handle the forces?Can copper wire alone handle the forces?

1.7 mm

2.8 mm

Properties of copper wire:

Tensile strength of Cu wire = 70 ksi (480 MPa) Area of Cu wire = 4.8E-6 m2 Breaking force of Cu wire = 2200 N

Hoop stress calculation for ½ loop:dF = I B R dθF = ∫ dF sin(θ) = IBR ∫ sin(θ) dθ F = 2 · 5,000 A · 40 T · 0.025 mMax Force on Cu wire = 10,000 N(8th Layer for most recent magnet)

Current

Fnet

Field

Copper wire alone will not work, extra reinforcement is necessary.

Reinforcement of MagnetsReinforcement of Magnets How do we “safely” make magnets?

- wires held in place to prevent rubbing,- reinforce the magnet with something of high tensile strengthto prevent hoops from stretching.

There are two methods of winding to fix the wires in place:

Our lab has already made and destroyed wet-wound magnets. The most dependable and highest field magnet in our lab is the Leuven magnet, which was made using the dry-wind technique.

1. Wet-winding – painting epoxy on every layer of the magnet as it is wound. Requires a great deal of external reinforcement.

2. Dry-winding – magnet is first wound, then epoxy impregnates the magnet. Allows for internal reinforcement via fiberglass.

We will make dry wound magnets because of the success of theLeuven magnet. We therefore need internal fiberglass reinforcement

The Wonders of FiberglassThe Wonders of Fiberglass

Maximum tensile strength of 4890 MPa (709 ksi)Maximum tensile strength of 4890 MPa (709 ksi)

Best case scenario, 0.7 mm (0.028”) of S-glass needed to reinforce the Best case scenario, 0.7 mm (0.028”) of S-glass needed to reinforce the

magnet at the greatest force (8magnet at the greatest force (8thth layer), layer), neglecting copper’s strength.neglecting copper’s strength.

1.7 mm2.8 mm

0.7 mm

Electrical resistivity of 9.0E12 Ω·cmfiberglass is a great insulator, reducing the chances of a short between wires.

A flywinder is needed to wind the fiberglass onto the magnet being wound because the magnet must be kept under constant tension while being wound.

How do we get it on the magnet?

Original

The FlywinderThe FlywinderCapable of quickly and accurately applying fiberglass from the spindle to a magnet as the magnet is being wound. The guiding posts are necessary to accurately positionand wind the fiberglass.

A single flywinder is capable of only applyingfiberglass to EVERY OTHER LAYER due to interference by the copper wire (see picture.)

When the flywinder is idle it is in the way ofwinding! Need to get the flywinder out of theway.

Fiberglass Addition vs. Rotational Rate

y = 1.98E-03x + 1.38E-04

0

0.005

0.01

0.015

0.02

0.025

4 6 8 10Rev/S

Ch

ang

e in

Dia

met

er

(in

ches

)

The Computer/ControllersThe Computer/ControllersTwo parameters needed for flywinder:1. Number of revolutions needed for theflywinder to traverse the magnet.2. The rate (rev/s) to turn the flywinder and apply fiberglass.

Program allows for 4 movements: windingfiberglass (both lateral and rotational motion)moving the flywinder forward and backward,and rotating only the flywinder.

Computer communicates with daisy chained controllers through a modem port and addresses each controller specifically and simultaneously through a 25-pin D-connector.

The controllers are fixed on a giant aluminumheatsink and powered by a 75 volt power supply

A linear dependence exists betweenthe rotational velocity and the changein diameter of fiberglass wound, meaning that we can specify the desired thickness of fiberglass betweenlayers.

Look at it go!Look at it go!

ResultsResults•Two magnets have been successfully wound and impgregnated.•The first was an overall test of the fiberglass/epoxy system and proved the system worked as advertised.•The second has been further reinforced with steel ribbon and pulsed at liquid nitrogen temperature. A maximum field of 35 T has been reached out of a theoretical 44 T.•We are now able to quickly and easily create our own magnets with customizable field characteristics.

Magnet 1 Magnet 2

Future Work for the FlywindingFuture Work for the FlywindingProjectProject

Development of another flywinder for winding fiberglass Development of another flywinder for winding fiberglass on every layer of copper.on every layer of copper.

– Another flywinder must first be built (5” ID bearing not necessary!)

– The new flywinder must be attached somehow to the coil winder’s horizontal threaded rod for lateral movement.

– The stepper motor controllers must be readdressed and the program modified to control at least one more motor.

– Extension rods must be implemented in the winding frame setup to allow for the new flywinder to keep out of the way when idle.

Superconductors at high pressureSuperconductors at high pressure

-Critical field is changed

-Effective mass, scattering time, fermi surface changed.

What happens as pressure is increased?

As the pressure increases two things happen:1. The interlayer spacing of the SC is decreased due to the compression of the lattice.

2. The electron density increases and affects the electron-electron interaction (destructive to superconductivity.)

Bringing SC’s to High PressureBringing SC’s to High Pressure How are high pressures reached?

• SC samples brought to high pressures via diamond anvil cells (DAC) or metal cells.

• DAC’s theoretically good to 25-50 kbar (our experiences, up to 5 kbar).Difficult to work with because pressure is difficult to set (in press) and maintain when not at STP.

• Metal cells will not work in our setup, eddy current heating.General rule to keep metal out of 50 T.

We want a system and nonconductive cell capable of changing pressure in situ and capable of reaching 2 kbar (~30 ksi).

What we got from DukeWhat we got from DukeDonated by Horst Meyer

Highlights:

• a frame with most components 40 yrsold and nonfunctioning

• system leaked like crazy

• overly elaborate for what we needed.

Front

• all unnecessary components removed from previous system: ~12 valves 5 valves 5 gauges 3 gauges removed ~5 feet of excess tubing removed 10 connecting blocks/ valves

• all O-rings in system changed: invalves, the pressure generator

• new 1/8” and 1/16” tubing used tomake connections

Back

How do we get to high pressure?How do we get to high pressure?

Treat as a thick walled vessel: Pburst = Tstr*log(OD/ID)

Typical tensile strength of S.S: 600 MPa (87 ksi)

What pressure is the tubing good to?

Theoretical Burst Pressures:1/8” OD, 0.020” – 70 ksi 1/16” OD, 0.020” – 40 ksi

Tank alone can give 2,500 psiHand Pump – 20,000 psiPressure Head – 20,000 psiGauges – 20,000 psiSafety Heads – 30,000 psi(Burst Valves)Pressure Generator – 30,000 psi

Our pressure limitations:

Low temperature should increasethe tensile strength allowing evengreater pressure.

Results thus farResults thus far

~ 4,000 PSI12,600 PSI

13,600 PSI (~0.8 kbar)

Results are favorable, we can already study several different superconductors at these pressures.

Cells cracked at a variety of pressures,but system did not leak.

Currently limited only by cell design.

k-(ET)2-Cu(NCS)2 has a predicted criticalpressure of 2.2 kbar, meaning thatthis range is already very useful tous.

The critical field is already reduced by a factor of 3 at 1.5 kbar

Quite a bit of work to be doneQuite a bit of work to be doneMore accurate measurement - Manganin wire gauge, strain gaugeMore accurate measurement - Manganin wire gauge, strain gauge

Electrical feed through for closed cell designElectrical feed through for closed cell design

Pressure experiments on superconductors – verifying/calibratingPressure experiments on superconductors – verifying/calibrating

More cell designMore cell design

Testing the system and cell at liquid helium temperaturesTesting the system and cell at liquid helium temperatures

Are the system and cell in hydrostatic equilibrium?

Acknowledgements:Acknowledgements:Vincent Ciarametaro, Luiz de Viveros, Geoff Esper,Vincent Ciarametaro, Luiz de Viveros, Geoff Esper,Katholike University – Leuven, BelguimKatholike University – Leuven, Belguim

Special Thanks to:Special Thanks to:Chuck Agosta, Joel Norton, Vin Ciarametaro, Mike Viotti,Chuck Agosta, Joel Norton, Vin Ciarametaro, Mike Viotti,

Catalin Martin, Izabela Mihut, and Charles GateteCatalin Martin, Izabela Mihut, and Charles Gatete

Funded by:National Science Foundation, Stanley Geschwind Memorial Fund