stephen benson u.s. particle accelerator school january 17, 2011 the fel as a diagnostic* * this...
Post on 19-Dec-2015
218 views
TRANSCRIPT
![Page 1: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/1.jpg)
Stephen Benson
U.S. Particle Accelerator SchoolJanuary 17, 2011
The FEL as a Diagnostic*
* This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150, the Air Force Office of Scientific Research, DOE Basic Energy Sciences, the Office of Naval Research, and the Joint Technology Office.
![Page 2: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/2.jpg)
Outline
• FEL tutorial• Dependence of FEL gain on:
• Beam steering• Emittance• Bunch length and energy spread
• Some examples• Conclusions
![Page 3: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/3.jpg)
Relativistic Hertzian Dipole
![Page 4: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/4.jpg)
FEL MECHANISM
![Page 5: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/5.jpg)
How an FEL works
![Page 6: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/6.jpg)
The Synchronous Condition, cont.
• If put in mathematical terms, the synchronous condition is:
λWvz
=λr
c−vz
This can be rewritten by using the formula for the electron velocity:
βx =vxc
=Kγ
and βz = 1−1+K 2
γ2
⎛
⎝ ⎜
⎞
⎠ ⎟
⎡
⎣ ⎢
⎤
⎦ ⎥ 1/2
So we we now have:
€
λr =λ W (1+ K 2 )(1+ β z )βzγ
2≅
λ W (1+ K 2 )2γ 2
Where K=0.934 T-1cm-1 lWBrms is the wiggler parameter, and g is the electron relativistic energy normalized to the electron rest mass.
![Page 7: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/7.jpg)
The Pendulum Equation• Let us define a pondermotive phase z
ζ =(kr +kW)z−ωt
For an electron at the resonant energy the pondermotive phase is constant and the electron sees a force proportional to sin(z). The equation of motion of the electron motion with respect to this pondermotive potential is that of a pendulum:
d2ζdτ2 =
dνdτ
=asin(ζ +φ)
Where we define the dimensionless field amplitude a as
a=2πNeKFLγmc2
Eeiφ
F =J 0
K 2
2(1+K 2)
⎛
⎝ ⎜
⎞
⎠ ⎟ −J 1
K 2
2(1+K 2)
⎛
⎝ ⎜
⎞
⎠ ⎟
where
![Page 8: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/8.jpg)
Motion in the Pendulum Phase Space
• The height of the separatrix is proportional to a1/2.• The vertical position is proportional to the energy
detuning.• The maximum small signal gain occurs outside the
separatrix.• Saturation occurs in the separatrix.
![Page 9: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/9.jpg)
The Field Equation
• As the electrons bunch they radiate coherently. The radiated field adds to the original bunching field. Depending on the initial electron velocity it can add in phase, in quadrature, or out of phase. This defines a new equation:
j =4N(πeKFL)2ρ
γ3mc2
dadτ
=−j e−iζ
ζ0,ν0
Where we define the normalized current as
Where r is the electron density. For low gain (j<1), the gain is just
G=0.135j . The pendulum and field equations must be solved self-
consistently and are applicable in low and high field and low and high
gain.
![Page 10: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/10.jpg)
Spontaneous radiation and Gain
• The spectrum of the spontaneous radiation is just:
• The width is inversely proportional to the number of wiggler periods and the height is proportional to the square of the number of wiggler periods.
• The gain is proportional to the slope of the spontaneous radiation so it is proportional to the cube of the number of periods.
€
P(ν ) = P0
sin2(ν )ν 2
⎛
⎝ ⎜
⎞
⎠ ⎟
![Page 11: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/11.jpg)
Exercise
Remember that the pondermotive phase is
And that
Show that
And
Thus show that gain occurs for energies lower than the resonant energy and for wavelengths longer than the resonant wavelength.
ζ =(kr +kW)z−ωt
€
ν =dζdτ
where τ =z
Nλ W
€
ν ≅4πNΔγγ
€
ν =−2πNΔλλ
![Page 12: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/12.jpg)
Gain
The gain spectrum is given by:
Note that this is for small signal only. The curve broadens and reduces in amplitude for large signal.
Also note that the peak gain does not occur for the resonant wavelength. As we will soon see it is even further from resonance for a typical device.
€
G0 =d
dνsin2(v)
v2
⎛
⎝ ⎜
⎞
⎠ ⎟⋅
j2π
![Page 13: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/13.jpg)
Now Reality Raises its Ugly Head
• So far, I have assumed a spherical cow:• A monochromatic plane wave• A mono-energetic electron beam• A beam with infinite transverse extent with electrons
traveling along parallel paths.• A perfect wiggler.
• Now we have to transition to the real world where beams have transverse extent, bandwidth, energy spread, emittance, and wigglers and beam transport magnets have focusing and steering errors.
• For this course we will assume perfect mirrors since this is an electron diagnostics course.
![Page 14: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/14.jpg)
Reality #1: Real Optical Modes
A realistic optical cavity contains a superposition of Gauss-Hermite or Gauss Laguerre eigenmodes, the lowest of which is defined as:
With the waist size and radius of curvature defined by
And the Guoy phase shift given by
€
E(r,z) = E0
w0
w(z)exp
−r2
w2(z)
⎛
⎝ ⎜
⎞
⎠ ⎟exp −ikz − ik
r2
2R(z)+ iζ (z)
⎛
⎝ ⎜
⎞
⎠ ⎟
€
w(z) = w0 1+zzR
⎛
⎝ ⎜
⎞
⎠ ⎟
2
and R(z) = z 1+zR
z
⎛ ⎝ ⎜
⎞ ⎠ ⎟2 ⎡
⎣ ⎢
⎤
⎦ ⎥ with zR =
πw02
λ
€
ζ(z) = arctanzzR
⎛
⎝ ⎜
⎞
⎠ ⎟
![Page 15: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/15.jpg)
Higher order modes
Higher order modes multiply the basic Gaussian mode by a set of Hermite or Laguerre polynomials.The Guoy phase shift for a TEMn0 L-G mode is 2n+1 times that of the fundamental mode.
The minimum mode volume for a Gaussian mode occurs for
€
zR =L12
![Page 16: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/16.jpg)
Reality #2: Electron beam size
Ideally you want all the electrons to sit where the electric field is largest, i.e. on axis. The wiggler focuses in one direction and the electrons drift freely in the other.
Unfocused direction:
Just like the optical mode. Can fit electron beam inside the optical mode if and if we set
Focused direction
Wiggler field usually has the form:
The stronger field away from the axis introduces a wavelength shift and provides a linear restoring force. Electrons oscillate with a period:
€
ρ =ρ0 exp −r2 /(2σ (z)2 )[ ] and σ 2(z) =βx0ε nx
γ1+
zβ x 0
⎛
⎝ ⎜
⎞
⎠ ⎟
2 ⎛
⎝ ⎜ ⎜
⎞
⎠ ⎟ ⎟
€
By (z,y) = B0 cos(kW z)cosh(kW y)
€
λβ =2πβm =γλ W
Krms
€
εn < γλ 4π
€
β0 = zR
![Page 17: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/17.jpg)
Electron beam size (cont.)
• If the matched betatron period is less than the wiggler length (strong focusing case) the rms beam size scallops as it goes down the wiggler. In the transverse space the electrons travel on circular trajectories. The increase in the average phase space volume leads to an increase in the effective emittance
€
εeff =ε2
β0
βm
+βm
β0
⎡
⎣ ⎢
⎤
⎦ ⎥
![Page 18: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/18.jpg)
Weak focusing case
• For a matched betatron period much longer than the wiggler it might be better to focus the beam in the wiggler center to improve the filling factor.
Focused envelope
Unfocused envelope
Matched envelope
![Page 19: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/19.jpg)
Filling factor
• In general you would like to have as many of the electrons as possible in the part of the optical mode with the largest field. You can define an effective filling factor
€
η f =4λ r
2.974λα
1+α 2 7dx
Ae(x)+1+α 2x 2
0
1
⌠
⌡ ⎮
Ae(z) =4πεN
γλzR
β x x( )β y x( ) , α =L
2zR
, x =2zL
β x (x) = β x 0 1+zR
β x 0
⎛
⎝ ⎜
⎞
⎠ ⎟
2
α 2x 2 ⎛
⎝ ⎜ ⎜
⎞
⎠ ⎟ ⎟, β y (x) =
γλ W
2πK
The 1+a2/7 term is due to the Guoy phase shift. This factor is unity for a filamentary beam, N>>1, and zR=L/3.
![Page 20: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/20.jpg)
Reality #3: Angular effects
• The wavelength is shifted by changing K, which changes the longitudinal velocity, Changing the angle of the electron beam does the same thing. In fact
• This leads to a detuning
• Note that the detuning is unidirectional so an angular spread leads to an asymmetric distribution in n
€
λr (θ ) =λ W
2γ 21+ K 2 + γθ( )
2
( )
€
Δν =−2πNγ 2θ 2
(1+ K 2 )
![Page 21: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/21.jpg)
Reality #4: Energy spread
• Real beams have energy spread. This leads to a distribution in n0. The distribution is as symmetric as the energy distribution. The spread in no is
• We can define a inhomogeneous gain reduction factor for Gaussian distributions as
€
σ ν =4πNσ γ
γ
€
ηI =2π
0.85dα
α (1− α )sinψ e−α 2σ 2 2
(1+ ax2α 2 )(1+ ay
2α 2 )4
0
1
⌠
⌡ ⎮
ψ = αν −12
tan−1(axα )−12
tan−1(ayα )
σ = 4π Nσ γ
γ, ax =
2π εN Lβ x 0γλ R
, ay =2π εN Lβ y 0γλ R
€
ν opt ≅ 2.606e−0.526σ +1.07σ + (2 +σ 5)ax
ax + 3+
ay
ay + 3
⎡
⎣ ⎢
⎤
⎦ ⎥
![Page 22: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/22.jpg)
Longitudinal match
• For a given longitudinal emittance you can have a longer or shorter bunch as long as the energy spread is adjusted to keep the emittance constant. In general you want a very short bunch but:• If the energy spread gets too large the gain reduction factor starts
going down fast.• A short bunch might be short compared to the slippage length Nl.
The gain vs. bunch length goes as:
• But all this assumes an upright bunch. If you have a skewed longitudinal distribution the energy spread will be large but the bunch length will be larger as well so the peak current won’t be as high as it could be. One therefore expects the gain to vary quadratically around the bunch length minimum.
€
ημ = 1+NW λ3σ z
⎡
⎣ ⎢
⎤
⎦ ⎥
−1
![Page 23: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/23.jpg)
Offsets instead of spreads
• Everything so far has concentrated on distributions. What if the whole beam moves?• Angular shift: Want the angular shift to be less than the mode
divergence so we want
• As an example for 400 nm and 90 cm zR we want the angle error to be less than 380 µrad.
• Position shift: Want the transverse position to be smaller than half the waist size (mode jumps if you exceed this) so
• Energy shifts just lead to wavelength shifts, which is bad for users but not for gain and power.
€
θe <λ
πzR
€
Δx <zRλ4π
or for previous parameters Δx <170μm
![Page 24: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/24.jpg)
What about Jitter?
• An important parameter for jitter is the resonator critical frequency
• For example, for frt=4.678 MHz and Q=10, fres = 74.5 kHz.
• For jitter at frequencies much less than fres the FEL generally follows the changes so the jitter looks like a DC change at any one time.
• For jitter at frequencies higher than fres the FEL integrates over the motion so the effect is to blur the beam out (e.g. a phase jitter would effectively increase the bunch length).
• For jitter near fres the FEL can react strongly to the jitter. The effects can be larger than the DC shift would indicate.
€
fres =frt
2πQ where Q =
1Γ
![Page 25: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/25.jpg)
Three Dimensional Simulations
![Page 26: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/26.jpg)
FEL MECHANISM
![Page 27: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/27.jpg)
UV Demo Beamline Layout
E = 135 MeV67 pC pulses @ 4.68 MHz
(>20 μJ/pulse in 250–700 nm UV-VIS)
(UV beamline and commissioning funded by AFOSR and BES. Wiggler on loan from Cornell U.
![Page 28: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/28.jpg)
Cornell Undulator A Prototype
![Page 29: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/29.jpg)
UV Wiggler trajectories in Cornell Wiggler
![Page 30: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/30.jpg)
Very High Gain Seen at 400 nm
![Page 31: Stephen Benson U.S. Particle Accelerator School January 17, 2011 The FEL as a Diagnostic* * This work was supported by U.S. DOE Contract No. DE-AC05-84-ER40150,](https://reader035.vdocuments.us/reader035/viewer/2022062407/56649d385503460f94a11a33/html5/thumbnails/31.jpg)
Conclusions
• For optimum laser performance three things are required
1. The electron beam should be as small as possible compared to the optical mode.
2. The detuning spread must be less than about p/4 from all effects.
3. The beam must be stable. The parameters should vary by no more than about 1/10 of their spreads.
If any of these things are not true the FEL will let you know!