We used the following prime diagnostics. For the electron beam we used two e-beam Beam Profile Monitors (BPM’s), one located downstream and below, following the bend of a 90-deg dipole. This dipole is normally used as a part of a high-resolution spectrometer for the STELLA experiment (located upstream of the Thomson chamber). For the x-rays we used an x-ray detector placed at the beam line exit behind a 250 um Be window. For the CO₂ laser we used IR camera that images the CO2 laser beam with the x8 optical magnification. The location of the e-beam and x-ray diagnostics is shown on Figure 1. The interaction cell is equipped with the capillary discharge setup and a parabolic mirror to focus the laser beam.

Below we present rough results from the run spiced with just minor speculation on a possible nature of the observed effects. Processing and thorough explanation of these results are still ahead of us.

 Tests of the high-power CO₂ laser propagation through the capillary discharge

 We show below sample images that illustrate the plasma-channel's effect on the CO₂ laser beam propagation and preliminary results on the high laser energy transmission through the channel. Figure 2 represents a set of typical images that we usually observe at the alignment stage before energizing the capillary discharge. Fig. 2a shows the laser image at its focal spot; 2b the distribution at the output plane of the “cold” capillary at a relatively low (<1 J) laser energy when optical plasma does not affect the distribution; 2c is the same as 2b but with 5 J into the capillary.

Images 2b and 2c are obtained with X100 times attenuation to allow comparison with the image at the focal spot.

 When the laser energy has been increased to 5 J, it caused bright white light flashes from the capillary. The laser distribution changed, as shown in Fig. 2c. We can see an extra concentration of the intensity in the axial region. The total energy transmission drops by 30%. There is a possibility that the leading portion of the laser pulse produces plasma at the wall that contracts the tail of the pulse (partial self-guiding). The box size in Fig. 2a is 1x1 mm that allows estimating the size of the laser spot. The focal spot shown in Fig. 2a as measured by our Spiricon image analyzer is 85 microns FWHM. All the laser images are on the same scale.

Figure 3 shows the effect of the capillary discharge on the laser beam at various input laser energies. The images in Fig.3 are obtained at the optimized discharge conditions: ~17 kV, close to the first “zero” current. The images in Fig. 3 show laser channeling, with the image observed as before at the output of the capillary discharge tube. Fig. 3a is at 30 mJ input, 3b at 1 J,  3c at 5 J and 3d is at similar conditions as 3b but image taken with a higher attenuation to avoid saturation.

 The unsaturated images on Fig. 3a and 3d have 100-120 um FWHM quite comparable to the free-space focus. The high energy beam shows more complex structure (possibly filamentation). However, the main energy is still concentrated in a single core. Energy transmission through the plasma channel at up to 1 J input was similar or higher than through the “cold” capillary. At 5 J, the transmission dropped by 50%. This could be explained by partial clipping of the laser pulse in the plasma due to additional optical ionization.

                Due to aperture limitations on the optics internal and external to the interaction cell, just a potion of the transmitted laser beam can be collected on the diagnostics. This affects the accuracy of the absolute laser transmission measurements through the channel. We trust, however, that the above quoted percentages are close to the absolute transmission numbers. 

E-beam manipulation by the capillary discharge plasma

 Figure 4 shows typical BPM images obtained without the capillary discharge. Fig. 4a shows the energy spectrometer’s BPM image, where beam energy is along the X axis (higher energy to the left) and Y is the horizontal profile of e-beam [~1.5% x 4 mm full screen]; Fig. 4b shows the straight-ahead BPM image [4mm x 4 mm full screen]

 To test the effect of the plasma on the e-beam we introduced the electron beam at a variable delay after the capillary discharge current terminates as is shown in the following figures.

 Figure 5 shows the oscilloscope image of a discharge event, showing the trigger discharge, main discharge and the electron beam at one particular setting.

 Figure 6 shows the beam images on the straight-ahead beam profile monitor (top images) and on the energy spectrometer BPM (bottom) at various delay intervals: 1.2, 1.8, 3.7 and 5 microseconds from left to right.

Counter-interaction of the CO2 laser and e-beam in a plasma channel

In order to observe the interaction of the e-beam with the laser in the capillary we set the synchronization to the “zero” of the discharge current. This position corresponds to a minimal effect of the discharge magnetic field on the e-beam and is optimal for the laser channeling. 

When all three elements (plasma, laser and electron beam) are present at these conditions, we observe a significant smearing of the e-beam spot on the straight-ahead BPM and a strong image perturbation on the spectrometer BPM. The BPM images are shown in Figure 7 for laser off (left) and laser on (both with plasma on). With laser on we observed a strong x-ray signal. Although a part of this signal may be due to the Thomson scattering effect, a significant contribution is due to enhanced inverse bremsstrahlung. This may indicate that the e-beam is strongly manipulated by the laser beam in the plasma and is partially lost inside the capillary or on the mirror. This is confirmed by a reduction of the Faraday cup signal. The x-ray signal disappears immediately when the e-beam is delayed after the laser. This rules out a possible explanation that the laser affects the e-beam via generating higher plasma density. We look into a possibility that the laser beam ponderomotively expels plasma electrons from the e-beam propagation path that may result in the strong e-beam focusing.

Back to Top

The objective of this run was to observe more Electro-Optic (EO) signals on the streak camera. We have realigned the optics to let the laser beam path select a location on the EO crystal different than that of the August 1 run, see Figure 1.

1. Before the beam was turned on, the minimum dc light level (destructive interference) from the dark port was as low as 10 mV (into a high impedance), while the maximum light level (constructive interference) was 750 mV. The spatial beam profiles of the reference, EO, and the combined EO+ref are good, see Figure 2. However, like all previous experiments, the beam profile of the destructive interference degraded with increasing exposure time to the electron beam field, also see Figure 2. The dc light level increased to  ~50 mV at the end of the run. The beam charge was ~0.4 nC.
2. Before the run, Don and Triveni were trained by Vitaly to become ATF duty operators. We now have our own duty operators, who are officially allowed to manipulate the electron beam parameters.
3. The HeNe laser was turned on at 9:30 am. Our beam run started on August 14 Wednesday at about 11:30 am when the laser reached a quiet period of ~ 1 hour, see Figure 3. The electron beam stability was similar to previous runs. The beam profile jittered by about one beam diameter (~ 1-2 mm) horizontally (momentum variation) from shot to shot.
4. EO signal was obtained immediately after the electron beam was turned on. We then sent the EO signal onto the streak camera. Several sets of measurement were taken, each was an average of 32 shots, see Figure 4 – data set #1. The streak camera time scale was at 3 ns window. On the next 2 sets of streak camera data, Figure 4- data sets #2 & #3, we decreased the time window to 1.5 ns. Each data trace was still an average of 32 shots. The EO signal was clearly observed at the expected time location.
5. It appeared that the timing jitter of either the EO signal or the streak camera trigger became important. It is possible that signal averaging might wipe out the exact time structure of the EO signal. Therefore, we reduced the number of EO signal averages to 8 in the afternoon beam run.
6. Although the laser did not reach a satisfactory level of quietness, a series of measurements were made at this lower averaging number. Figure 4 – data set #4 shows most of this data and its overall average. It was observed that more ‘spike-like’ signals appeared at the expected EO signal time location. A few ‘representative’ signal traces are plotted in Figure 3. They appeared to have a timing jitter of  ~400 ps, and if each can be consider as an EO signal it has a width of ~40 ps. This is our first indication of a fast EO signal.
7. We have a total of 55 min. actual beam time on the EO crystal – 17 min. in the morning run and 38 min. in the afternoon run. There is no doubt that EO signals were observed on the streak camera, the ~40 ps EO signal might still be ambiguous. Many factors can contribute to its broadening/narrowing. We know that we have to check and minimize our timing jitter, see Figure 5 (exp.setup) and Figure 6 (Streak camera timing jitter), and record single shot EO signal – no averaging. The photocathode/image-intensity-unit of the streak camera is degrading and it has a non-uniform responsivity, we have to pick a location on the photocathode where it gives a ‘good’ response, see Figure 7(Streak camera photocathode response - pulse), Figure 8(Streak camera photocathode response – light) and Figure 9(Streak camera photocathode response – dark).
8. Some incremental improvements: in this run our EO signal was only ~1 mV on the oscilloscope, previously we observed 4 mV in single shot, therefore the EO signal can be improved. Also, the beam charge can be increased from 0.4 to 0.6 nC.
9. Substantial improvements: we wasted most of our time waiting for the laser to quiet down, we should replace the laser with a low noise HeNe laser, see Figure 10(Noise of HeNe lasers – Ours and Peter’s). Also we may need a scheme to replace/switch samples in the vacuum after it has been ‘invasively’altered by the electron beam field.
10. We have made progress on this run with very minimal use of beam time. We’ll have a meeting at a later time to discuss and plan for the next beam run – on November.

Back to Top

We continued to do beam studies for transverse phase space tomography over two weeks in July and August. The results are summarized below:

1. XJ Wang and his collaborators observed the transverse emittance is strongly dependent on the RF gun’s phase. It is interesting to investigate the real transverse phase space distribution to gain more information. Transverse phase spaces vs. various RF gun phases are studied. See Figure 1. Further analysis is under way.
2. Different initial laser distributions can produce electron beams with different performance. Their emittances had been measured in the previous beam studies. This time we compared their transverse phase spaces for initial regular laser and hollow laser beam. Their phase spaces vs. different charges is studied. See Figure 2. We are trying to convert (x-x’, y-y’) phase spaces into (r-pr, theta-pthea) and hope to find interesting physics. It is obviously interesting to see whether the phase spaces of theta-ptheta are changed with bunch charges or not. In addition, note that the y-y’ phase space is slightly smaller than x-x’ with the increase of the bunch charge. It is repeatable for any data. This may imply some external factors (e.g., misalignment, wakefield) to affect the beam. Further analysis is under way. 

Back to Top

As mentioned in the last newsletter, the STELLA-II experiment is awaiting the upgrade of the ATF CO2 laser to provide the higher peak power needed to drive the tapered undulator. We have also discovered that the laser beam inside the buncher is annular in shape, which accounts for the below-normal energy modulation we are observing. We have determined that the primary cause for this annular beam appears to be due to the optics inside the TW CO2 amplifier. Apparently these optics are somehow distorting the beam, perhaps through diffraction effects, and causing it to evolve into an annular profile as it propagates to the Experimental Hall.

We attempted to counteract this effect by placing a partially closed iris downstream of the TW amplifier along the STELLA laser beam transport line. Diffraction from the inside edges of the iris can be made to produce a laser spot in the center of the annulus at the buncher (i.e., Spot of Arago) with the sacrifice of some laser power. Unfortunately, this did not improve the modulation significantly. We also attempted to steer the e-beam through the ring of the annulus, but the ring is too large for our steering coil capabilities.

Thus, it was concluded that the best solution is to change the optics inside the TW amplifier in order to eliminate the annulus. This will be done during the current shutdown.

On a different matter, we also completed a set of spectrometer resolution measurements for both the wide-energy-acceptance, variable-angle spectrometer and the high-resolution, 90-degree spectrometer. This information is important for the model, which includes this resolution factor as part of its predictions, and for answering the question whether the current spectrometers can be used when the energy gain becomes very large (e.g., >20%) once the laser is upgraded. Based on our modeling analysis we determined for the higher laser power experiments that we need a spectrometer with a total energy acceptance of ~25-30% and a resolution of sigma = 1% or smaller.

The variable-angle spectrometer when set to 4-degrees has an energy acceptance of over 25%. Our measurements show that its resolution is sigma = 0.28%. This is well below our 1% requirement and means the 4-degree spectrometer has the capability to satisfy our needs. However, the camera detector at the output of the spectrometer does not have a wide enough view to see the entire spectrum. It may be possible to change the camera lens system to increase its field of view or to use two separate cameras to view different portions of the spectrum. This modification of the camera system will be investigated later.

During the current shutdown we will be performing further characterization studies of the STELLA-II laser beam transport/focusing system and adding improvements to the beamline system.

Back to Top

Last Modified: December 3, 2007
Please forward all questions about this site to: Vitaly Yakimenko