arXiv:nucl-ex/0504024v1 25 Apr 2005
A high power liquid hydrogen target for the
Mainz A4 parity violation experiment
Technische Universit¨ at M¨ unchen, D-85748 M¨ unchen, Germany
E. Schilling , S. Baunack , L. Capozza , J. Diefenbach ,
K. Grimm , T. Hammel , D. von Harrach , Y. Imai ,
E. M. Kabuß , R. Kothe , J. H. Lee , A. Lopes Ginja ,
F. E. Maas∗∗, A. Sanchez Lorente, G. Stephan , C. Weinrich
Institut f¨ ur Kernphysik, Universit¨ at Mainz, D-55099, Mainz, Germany,
We present a new powerful liquid hydrogen target developed for the precise study
of parity violating electron scattering on hydrogen and deuterium. This target has
been designed to have minimal target density fluctuations under the heat load of
a 20µA CW 854.3 MeV electron beam without rastering the electron beam. The
target cell has a wide aperture for scattered electrons and is axially symmetric
around the beam axis. The construction is optimized to intensify heat exchange by
a transverse turbulent mixing in the hydrogen stream, which is directed along the
electron beam. The target is constructed as a closed loop circulating system cooled
by a helium refrigerator. It is operated by a tangential mechanical pump with an
optional natural convection mode. The cooling system supports up to 250 watts of
the beam heating removal. Deeply subcooled liquid hydrogen is used for keeping
the in-beam temperature below the boiling point. The target density fluctuations
are found to be at the level 10−3at a beam current of 20 µA.
Key words: Charge conjugation, parity, time reversal, and other discrete
symmetries, Particle sources and targets, Polarized and other targets
PACS: 11.30.Er, 29.25.-t, 29.25.Pj
∗Associated Member of the St. Petersburg Nuclear Physics Institute, Russia
∗∗Corresponding author. E-Mail: email@example.com
Preprint submitted to Elsevier Science8 February 2008
Cryogenic liquid hydrogen targets have become common experimental equip-
ment in and particle physics experiments. The most powerful new targets have
been requested for electron scattering parity violation experiments, which are
able to register the small effects of weak interaction at the level 10−6. In these
targets heating can reach a few hundred watts, but target density fluctua-
tions should not become larger than 10−3to reach high statistical accuracy
of measurements. One source of target density fluctuations is the heating of
the liquid hydrogen by the energy deposit of the electron beam. Overheating
can cause boiling. Therefore, the task of in-beam heat removal is to be solved.
Recently, Experiments at the MIT-Bates accelerator [1,2], at the TJNAF ,
and at SLAC  have reported on their powerful targets. In this paper we
present the target design and realization for the A4-experiment at the MAMI
accelerator in Mainz [5,6,7,8,9].
The aim of the A4-experiment is a precise measurement of the parity violating
asymmetry in the elastic scattering of right and left handed electrons on an
unpolarized proton target. The expected value of the asymmetry is on the or-
der of 10−6. The experiment is carried out at the MAMI facility in Mainz with
a 80% polarized 854.3 Mev electron beam. With modern GaAs photocathodes
high currents of polarized electron beam combined with online intensity and
polarization monitoring have been realized . A large calorimeter with 1022
crystals to handle a total of 100 MHz counting rate has been developed. An
innovative feature of this detector is a use of new, high resolution and very
fast PbF2 Cerenkov radiators . The 10 cm liquid hydrogen target pro-
vides a luminosity of 5 × 1037cm−2s−1at 20 µA electron beam current. This
corresponds to about 100 watts of heat absorbed in the liquid hydrogen and
in the aluminum windows. The target cell is included in a closed loop with
a hydrogen-helium heat exchanger, where the liquid hydrogen circulation is
forced by a mechanical tangential pump. This is a common scheme for the
high power cryogenic system in which a high mass flow and a high heat trans-
fer rate can be achieved. We outline the characteristic features of this target
- The target has a 140◦wide aperture for scattered electrons in forward (op-
tionally in backward) direction. It is axially symmetric with respect to the
incident beam axis.
- The hydrogen moves along the beam. Therefore a special flow geometry was
applied to intensify a turbulent transverse mixing in the hydrogen stream.
The turbulence causes a mass exchange and a heat removal from the heated
region transverse to the beam direction.
- The helium cooling system was designed in a way to maintain deeply sub-
cooled (down to freezing point) liquid hydrogen in order to enlarge the head
room for the beam heating without boiling.
- A powerful cryogenic pump was used. It produces up to 0.3 bar overpressure
and can provide a mass flow of several hundred grams per second of liquid
hydrogen. This would cover the needs of a few kilowatt cryogenic system,
provided the corresponding helium refrigerator power is provided.
- The circulation contour was designed in a way to apply the largest available
pressure decline very locally at the target cell right to the region of beam
heating. To this end a low flow resistance hydrogen-helium heat exchanger
- The target works well even without running the mechanical pump up to
120 watts heat load since an intensive natural convection driven circulation
takes place in the loop. This regime is useful for applications when the
density fluctuations do not play an important role.
- The target gas system is simple and reliable. It contains only 5 m3of hydro-
gen at 2.2 bar of absolute pressure. All safety requirements, conventional for
the accelerator environment, are fulfilled and the target it inherently safe
in case of a sudden loss of cooling.
In the following sections we explain our design approach, give a description of
the target and present the results obtained.
2 Local beam heating and luminosity fluctuations
In this section we aim at an estimate of the luminosity fluctuations caused by
temperature variations in the hydrogen stream. We try to reveal the essential
parameters for the effective suppression of the fluctuations considering a prim-
itive model of turbulence, since the exact treatment of fluid dynamics in the
turbulent regime is difficult and would require detailed numerical simulations.
The numerical estimates are based on an electron beam with 854.3 MeV en-
ergy and 35 µA. It deposits a heating power of 160 W into the liquid hydrogen
target. We assume a hydrogen flow in a tube coaxial with the electron beam.
The diameter of the tube is 1.2 cm.
2.1 Estimation of local beam heating
Following Prandtl’s concept of turbulent mixing length  we have the space
scale of turbulent local fluctuations as
lt=κ y (1)
where κ = 0.4 is an empiric constant and y is the distance from a parallel wall
to the flow direction. This corresponds to half of a tube diameter D. In our
case the resulting turbulent mixing length lt= 2.4 mm, which is of order or
larger than the possible electron beam diameter d in mm. We can also apply
the concept of relaxation time ttof the turbulent transfer, that defines the
time scale of turbulent fluctuations:
with the dynamic velocity v∗= v
and the Reynolds number Re. In our case the relaxation time is in the region
of several milliseconds. For example, for hydrogen of T=16 K and v = 2 m/s
the relaxation time results in tt= 12.5 ms. In this simple model any existing
local fluctuation, for example temperature fluctuation ∆Tf, will exponentially
dissipate with time t as
ζ/2, the friction coefficient ζ = 0.316/
The heating due to the electron beam occurs through Bethe-Bloch energy
deposit at a constant specific volume energy deposition rate qv. Fluctuations
of temperature arises from cold fluid entering the beam and dissipate as the
heated mass is replaced by cold mass from the region outside the beam. For
a static case, the temperature rise due to the beam heating is
where Cpdenotes the specific liquid hydrogen heat capacity and ρ the hydrogen
density. Combining this with Eq. 3, we get a time structure of the temperature
Cpρt e−(t/tt). (5)
This function has a maximum at t = tt. This gives the amplitude of tempera-
ture fluctuations as
which drops with the beam size (due to qv) and with flow velocity as shown
at Fig. 1. An efficient heat transfer from the beam area takes place when both
d > 2 mm and v > 2 m/s.
2.2Luminosity fluctuations without boiling
Even without boiling, luminosity fluctuations ∆L/L can originate from target
density fluctuations ∆ρ/ρ arising from temperature fluctuations. The scale of
these fluctuations is given by the value of ∆Tmax. An accomplishable value, is
∆Tmax≈ 1 K, as one can see from Fig. 1. The corresponding relative luminosity
variation amounts to
for 16 K hydrogen. This estimate is for a single turbulent fluctuation with size
of order lt. In order to get an estimate of the overall luminosity fluctuations
over the whole target length, we average the single fluctuations over the the
beam volume containing a number n of single fluctuations contributing to
beam scattering during the integration time window ti= 20 ms. Since lt> d,
the number of simultaneous fluctuations penetrated by the beam is l/lt, where
l is the target length of 10 cm. This value has to be integrated over ti, yielding
With the assumptions as stated above, the total luminosity fluctuations result
Fig. 1. Beam heating for 12 mm tube.
L)single/√n ≈ 1.7 · 10−3. (9)
This estimate for the luminosity fluctuations is of the same order as required
by the experiment. It can be much larger if local boiling takes place.
2.3 Boiling at the aluminum windows
The electron beam deposits energy not only in the hydrogen volume, but also
in the thin aluminum entrance and exit windows. This heating amount to
1.9 W for 0.1 mm of aluminum thickness which corresponds to a specific heat
flow from the beam spot at the aluminum window of qf= 60W/cm2for beam
diameter of d = 2 mm. A more detailed analysis shows that the heat is not
transferred directly to the liquid, but is spread over the aluminum window
due to the aluminum heat conductivity. Nevertheless, the specific heat flow
from the aluminum to the liquid hydrogen does not get below qf≈ 5 W/cm2
to 10 W/cm2. This exceeds the first boiling crisis value (∼ 1.5 W/cm2) 
and causes unavoidable boiling around the beam spot in the hydrogen volume
right at the aluminum window. It should be noted that at such a large value
of qf the boiling regime is not stable in time. The wall temperature is also
unstable and its average value is determined mainly by the heat spreading
over the wall trough the bulk aluminum and can, in our case, reach a few
ten degrees. This process does not depend much on beam size or on aluminum
wall thickness. With increasing flow velocity its contribution to the luminosity
fluctuations will be proportionally less.
2.4Design criteria summary
We summarize the target efficiency parameters as they follow from our esti-
mations which is a delicate problem down to the level of fluctuations of 10−3.
We need a beam size of a few millimeters in diameter as an alternative to
the method of beam rastering. The temperature of liquid hydrogen should be
close to the freezing point in order to give several degrees of headroom to the
boiling temperature. The cross section of the tube (nozzle), which guides the
hydrogen stream along the beam, should be minimized. This minimum has to
provide the necessary mass flow (a mass flow of 20 g/s gives a 1 degree rise of
the bulk hydrogen temperature at 160 W heating power). The flow velocity
has to be maximized. A limiting value is defined by the pressure drop in the
nozzle. They are evaluated as a local pressure decline
∆P = ξρv2
with the coefficient ξ ∼ 2. This is a central item for the design of the heat
exchanger cooling system since it requires a careful optimization with respect
to the flow resistance, which has to be small compared with the value given
by formula 10 for the remaining parts in the hydrogen system.
3 Experimental setup
Fig. 2. The experimental setup of the Mainz A4 parity experiment.
Fig. 2 shows the experimental setup. The electron beam enters from the left
and hits the high power liquid hydrogen target in the vacuum vessel of the
scattering chamber. The scattering chamber is surrounded by the large ac-
ceptance (0.64 sr) multichannel PbF2-detector that registers electrons in the
electron scattering angle range of 30◦< θe< 40◦. The luminosity is mea-
sured under small scattering angles of 4.4◦< θe< 10◦. The target cryostat is
placed on an elevator system. The large elevator bellow allows a 20 cm expan-
sion along the vertical axis in order to move the target in and out of the beam
line. An adjustable incline of the axis under any azimuthal angle is possible,
so that the target cell can be aligned.
4 The target cell
The axial symmetry of the detector which is required by the physics of the A4
parity experiment is also reflected in the liquid hydrogen target cell design.
Fig. 3 shows the target cell design. The target consists of aluminum parts
Fig. 3. Target cell design of the liquid hydrogen target. Left: view from the side.
Right: view from top.
joined with an indium sealing, so that the target can be reassembled many
times. The target thickness is limited to 10 cm for forward scattering experi-
ments by the variation of the scattered electron energy. The liquid hydrogen
fluid enters the toroid shaped collector. The hydrogen flow is directed into
a nozzle of 12 mm diameter along the beam path, where the maximum flow
velocity and the maximum degree of turbulence is reached. The fluid leaves
the target cell through the exit collector. The aluminum entrance window has
a thickness of 75 µm. The large parabolic aluminum cap which contains the
hydrogen has a regular wall thickness of about 250 µm, but at the place of the
beam exit it is thinned to 100 µm. The nozzle walls are 200 µm thick, so that
a cone of up to 66◦has a low thickness of material in the way of the scattered
electrons. This construction has been tested up to 6 bar pressure in the cell
against vacuum. The target cell is fixed through a thermally isolating bridge
to the bottom cryostat flange in order to minimize mechanical vibrations and
The heating by the electron beam is concentrated in the small volume of the
electron beam path. As the flow is directed along the beam axis, the only effi-
cient mechanism of heat removal from this overheated volume is the turbulent
transverse mixing, which is more efficient at higher velocity. This mechanism
was considered in section 2. An even smaller nozzle diameter would enhance
the degree of turbulence to such an extend, that the mass flow along the beam
would drop too much. The nozzle cross section must be large enough to avoid
additional background arising from scattered electrons at small scattering an-
5The target cooling system
A schematic of the target cooling system is shown in Fig. 4. The system
consists of a closed liquid hydrogen circulation loop and a cold helium gas
cooling loop including a 400 watt Linde helium refrigerator. For the heat
transfer from hydrogen to helium we use a counterflow heat exchanger, placed
at the top of the hydrogen loop. The hydrogen target cell is located at the
lowest point in the loop. The liquid hydrogen circulates clockwise driven by
a tangential pump from the target through the pump to the heat exchanger
top. After cooling in the heat exchanger it flows again to the target. In order
to stabilize the total heat load of the refrigerator a temperature controlled
electrical heater is added in the loop right behind the target. This provides a
stable operation at variable electron beam currents and provides a heat load
even in case the beam is turned off. The loop hight is about 2 m. For safety
reasons, the hydrogen loop is always connected to an external gas tank of 5
m3at room temperature. The hydrogen system in total contains, including
the gas tank, an amount of hydrogen corresponding to 2.2 bar of absolute
pressure at room temperature. When the helium temperature gets below the
hydrogen boiling point, the heat exchanger works as a liquifier until the loop
is completely filled with liquid hydrogen. The loop volume of liquid hydrogen
is about 5 l, therefore the pressure in the system drops to 1.2 bar. When the
liquid covers the heat exchanger completely, it will be cooled down further to
helium temperature since the condensation process stops. When there is no
cooling the liquid is naturally evaporated filling the buffer gas tank back to the
2.2 bar. There is an additional electrical heater in the helium loop, in order to
maintain the temperature level of incoming helium at an appropriate level. In
the helium-helium heat exchanger incoming cold helium gas is heated by warm
Helium, coming from the helium-hydrogen heat exchanger. The additional
helium-helium heat exchanger, controlled by a bypass valve, allows further
regulation of the temperature at the working point without changing the load
on the helium refrigerator. This is especially useful when deuterium is used
Fig. 4. Schematic of the target cooling system.
instead of hydrogen. The system has eight temperature sensors. It is controlled
by a computer with an interface card having analogue to digital converters
(ADC) and digital to analogue converters (DAC). Two modes of operation
are possible. A soft mode, where the control system regulates only the power
of the helium heater to keep the incoming helium temperature T2right above
the hydrogen freezing point (∼15 K). It is convenient for automatic target
cool-down and for operation in natural convection mode. When operating
with the electron beam on the target, the control system regulates both the
power of the hydrogen heater (T8 is kept stable) and the incoming helium
The hydrogen loop has been optimized to provide the conditions for a hydrogen
circulation based on natural convection driven by the density differences of
cool and hot hydrogen liquid. For many target applications this regime is
quite sufficient. It is known , that optimized natural convection can provide
significant heat removal, up to a few kilowatts of power. For the optimization
high flow resistance parts in the loop have been removed such as pipe bends,
sharp changes of cross sections and so on. As a result of this approach about
2/3 of the total pressure loss along the whole loop comes from the target cell.
For natural convection the difference of static pressure ∆Pstbetween cold (on
the right side) and warm (on the left) loop branch should be compensated
by the dynamic pressure loss ∆P over the whole circulation loop. The static
pressure difference can be calculated from the density difference ∆ρ
∆Pst= g ∆ρ L (11)
where g = 9.81 m/s2and L ≈ 2 m is the loop height. The maximum of ∆ρ is
achieved when the hydrogen temperature in the cold branch is close to freezing
temperature and the warm branch is close to boiling temperature. From this
follows that ∆Pstis of order 1 mbar and 2/3 of that have to be applied in the
nozzle. This corresponds to a local pressure loss caused mainly by narrowing
of the stream before and its expansion after the nozzle. It is proportional to
the square of the flow velocity according to formula 10. With a 12 mm nozzle
diameter we achieve a flow velocity of v = 0.9 m/s and a hydrogen mass
flow of G ≈ 7.5 g/s. These values provide a stable operation of the target in
natural convection mode, but they are not sufficient to suppress luminosity
fluctuations down to the required level of 10−3. Operation with a mechanical
tangential pump gives a significant improvement since the possible pressure
drop increases due to a pump about 300 times reaching 0.3 bar. A sufficiently
high velocity of 15 m/s, as estimated from
and a mass flow of 130 g/s is possible in this case.
5.3Construction design of the hydrogen loop system
The liquid hydrogen loop and target cooling system is incorporated in one
cryostat with all inputs going through a top flange. Fig. 5 shows a photograph
of the inner part of the cooling system without the super isolation. The pho-
tograph corresponds to the elements in the diagram of Fig. 4. The target cell
at the bottom of the picture is connected on the left side to the hydrogen tur-
bine and on the right side to the main hydrogen-helium heat exchanger located
right under the top flange. This heat exchanger consists of 78 tube-in-tube el-
ements with total heat transfer surface about 1.4 m2. The helium-helium heat
exchanger in the center is smaller. It contains 32 short tube-in-tube elements.
The total height of the loop is about 2 m, the top flange diameter 0.6 m. The
hydrogen target cell is made from aluminum, all other parts are made from
stainless steel. Joints are made using conflat flanges with copper gaskets unless
welded. The system is fixed to the top flange by the cold valve body, attached
to the helium-helium heat exchanger top collector. All thermal length changes
are compensated by bellows. As a hydrogen pump (on the left of the main
heat exchanger) we use the Barber-Nichols tangential pump BNHeP-15-000
giving a closed pressure difference of 0.3 bar. It has originally been designed
for high flow rate (up to 600 g/s at 100 Hz) transfer of liquid helium. We run
the pump at a rotation speed of 70 Hz. We estimate the heating of the hydro-
gen by the pump as ∼ 20 W corresponding to about half of the total parasitic
heat load of the system. The additional liquid hydrogen volume required by
the tangential pump corresponds to about 1 l. The turbine can be substituted
by a short piece of tube if only the natural convection mode is needed. A 400
watt Linde helium refrigerator is available for target cooling. It produces a
flow of cold helium gas at a temperature of 7-11 K. Helium transfer lines of
18 m length deliver the coolant to the experimental hall. The helium mass
flow is about 6-8 g/s. Though the nominal refrigerator power is rather high,
it cannot be utilized for the hydrogen system because of the poor mass flow.
That puts currently limit on the cooling power of below 250 W, though the
cooling system design allows to use a helium mass flow of up to 30 g/s pro-
viding a possible cooling power of up to 1000 W. The heat exchanger design
allows freezing of the liquid in it without safety hazard. This is important as
the input helium temperature can be safely reduced to 12 K and below.
6 The gas system
Fig. 6 presents an operational flow diagram of the complete hydrogen target
system. The cryogenic part is located in the vacuum container formed by the
target cryostat united with the scattering vacuum chamber and the attached
beam guide. The cold hydrogen loop is connected to the 5 m3hydrogen gas
buffer tank through the gas handling stand. For the standard operation of
the cold target the valves AV2, AV3 and BV1 have to be open providing a
free transfer of hydrogen between the cold part and the buffer tank. This
connection is ensured by a one-way release valve EV1. Access to the hydrogen
Fig. 5. Photograph of the hydrogen circulation and cooling loop system.
system with vacuum pump and various gases for its preparation and filling is
achieved from an access valve AV1. The safety valve EV2 prevents overfilling
of the system by venting the hydrogen to the exhaust line on the building
roof. The vent tube is kept filled with nitrogen gas that is trapped at its far
end by 100 mbar release valve EV5 in order to keep the line always free of
oxygen from air. In the case of a target crash the beam guide will be closed by
the fast shutter valve VV5 preventing hydrogen from entering the high power
microwave system of the accelerator. In the case of hydrogen escape to the
experimental hall two hydrogen sensors control a special venting procedure of
the experimental hall.
Fig. 6. Target operation scheme.
7 Target control
The target control system includes the readout of all the operation parameters
(16 input channels), the control of two feedback loops for regulating the helium
and the hydrogen heater power, and the control of the cold helium bypass
valve. The scheme is realized with a computer interface card DAP 2400e/5
board of Microstar Laboratories. The board is controlled by DIGIS software
on a single PC. The helium refrigerator has an independent automatic control
and readout. The temperature sensors (see Fig. 4) are standard silicon diode
sensors, except T6, which is a carbon glass resistor. This type of sensor has
a high radiation resistance, which is necessary for the a temperature sensor
located near to a target cell.
8 Test results
8.1 Tests with the electrical heater in the hydrogen circuit
In order to commission the hydrogen target cooling system we operated the
system after liquifying the hydrogen gas without the electron beam, only using
the electrical heater in the hydrogen loop. All target cooling system parameters
like the temperatures, the heating power of both heaters, the valve positions
etc. (as displayed in Fig. 4) have been measured during the tests. The efficiency
of the hydrogen heater is 100 % i.e. all the electrical power is transferred to
the hydrogen liquid. It simulates the global heat load of the electron beam.
For a high heat load of 250 W the highest temperature in the system T8
Fig. 7. Target performance.
behind the target rises 19.7 K which is still about 1.3 K below the boiling
point of hydrogen. The target entrance temperature T5 for this high heat
load is 18 K corresponding to 3 K subcooled hydrogen. For working conditions
corresponding to a more realistic heat load of 100 W and the tangential pump
rotation frequency to 50 Hz, the maximum system temperature T8 settles at
17.3 K. The target entrance temperature T5 stabilizes at 16 K corresponding
to 5 K subcooled hydrogen. These first tests show rather quantitatively, that
the target can be operated at the working conditions of the A4 experiment.
In order to determine the efficiency of the hydrogen-helium heat exchanger we
have studied the temperature variation in the system for different tangential
pump rotation speeds (corresponding to different values of the hydrogen mass
flow) as a function of the electrical heater power. The four plots in Fig. 7 show
the results for T2, T3, T5, T7, and T8 for natural convection (lowest mass
flow rate), tangential pump speed of 7.5 Hz, 15 Hz, and 30 Hz. T2 has been
stabilized to 14 K for all four pump speeds. T2 corresponds to the helium inlet
temperature of the hydrogen-helium counterflow heat exchanger and T3 corre-
sponds to the helium outlet temperature (see Fig. 4) whereas T8 corresponds
to the hydrogen inlet temperature and T5 to the hydrogen outlet temperature.
The difference ∆THe=T3-T2 is proportional to the cooling power at constant
helium mass flow. One sees that for all four mass flow values the helium cool-
ing power ( which is proportional to ∆THe) varies linearly with the electrical
heater power. Only for the case of natural convection, the difference shows
nonlinear behavior due to the fact, that the mass flow for natural convection
is a function of the temperature. The difference of T8-T7 is proportional to
the heat deposition coming from the tangential pump. When the heater is off,
∆THe≈ 0.5 degree that corresponds to about 15 W of parasitic heating. This
parasitic heating is sufficiently small and does not distinctly depend on the
pump speed. The outlet helium temperature T3 linearly rises as the heating
power approaches the hydrogen boiling temperature (∼ 21 K) at about 200
W. The Helium is heated by hydrogen, therefore the inlet hydrogen temper-
ature T8 is always higher than the outlet helium temperature T3. The data
for 30 Hz pump speed show a very efficient heat transfer since T8≈T3. The
efficiency becomes worse as the speed reduces. In case of natural convection
the difference T8-T3 arrives at about 3 degree being almost independent from
the heater power. The maximal hydrogen temperature depends on the outlet
helium temperature which is, in its turn, defined by the helium mass flow. The
temperature T5, at which hydrogen comes into the target cell, is limited by
T8. This test with the heater demonstrates that the target cooling loop per-
forms like expected from our estimates based on simple fluid dynamics rules.
The overall performance of the target cooling loop is currently limited only
by the available helium refrigerator.
8.2 Tests of the target cooling system with the MAMI 85 MeV electron beam.
8.2.1 Influence of Hydrogen Mass Flow and Temperature on the Target Den-
We have studied the target density fluctuations in the liquid hydrogen tar-
get by hitting the hydrogen target cell with the MAMI 854.3 MeV electron
beam  at different target cooling system conditions. For these tests we have
measured the luminosity L(which is the product of electron beam current and
target density) with a system of water Cherenkov detectors detecting electrons
from Møller scattering at electron scattering angles 4.4◦< θe< 10◦. We
have measured the electron beam current simultaneously. During the tests,
the cross section of the electron beam has been kept small with an approxi-
mate size of 3× 104µm2. In Fig. 8 we show on overlay of three histograms.
The peak labeled “Without turbine” shows the distribution of the luminosity
signal for the case where the tangential pump was not rotating and the loop
was circulating in natural convection mode with low mass flow. The measured
normalized RMS for this case amounts to σ = δL/L = 0.0092. The peak la-
beled “PIMO monitor” shows a histogram of the beam current monitor signal.
The normalized RMS of the beam current I for this measurement has been
Fig. 8. Target density fluctuations as measured in 20 ms samples with a small cross
section 20 µA electron beam at different pump speeds (corresponding to different
hydrogen mass flow).
σ = δI/I = 0.0004 which is a factor 23 smaller than the luminosity fluctua-
tions, showing that the observed luminosity fluctuations arise to a high extend
from target density fluctuations only. The peak labeled “Turbine 30 Hz” shows
the luminosity distribution for the case when the tangential pump was run-
ning at 30 Hz giving a higher hydrogen mass flow. The effect of the turbulent
mixing in the hydrogen target cell leading to a higher effective heat transfer
can be clearly seen, since the normalized RMS of the luminosity distribution
with high mass flow is σ = δL/L = 0.0026 corresponding to a factor of 3.5
decrease in the width of the target density distribution. We have made more
tests varying the pump rotation speed and the beam current. Fig. 9 shows
some of the results. One recognizes that the normalized RMS of the target
luminosity varies from σ = δL/L = 0.032 at 30 µA and low mass flow (nat-
ural convection mode) to a value of σ = δL/L = 0.0023 at high hydrogen
mass flow and 20 µA electron beam current corresponding to an improvement
of a factor of 14 due to the turbulent mixing of cold and warm hydrogen in
the target cell caused by the higher mass flow. We studied the target density
fluctuations at 20 µA in more detail since the working point of the A4 parity
violation experiment at the MAMI electron accelerator in Mainz corresponds
to an electron beam current of 20 µA. The results of those measurements are
shown in Fig. 10. As expected, the target density fluctuations diminishes with
the hydrogen mass flow (tangential pump speed) due to higher degree of tur-
bulent mixing in the target cell. The observed decrease of the fluctuations with
the hydrogen temperature is also expected since the temperature difference of
the hydrogen to the boiling point temperature is larger corresponding to more
8.2.2Influence of Electron Beam Cross Section and Exact Beam Position
As estimated in Sec. 2 the size of the cross section of the electron beam should
have an essential influence on the target density fluctuations. Fig. 11 shows
the the measured absolute fluctuations over a running period of about 1000
hours of 854.3 MeV electron beam as a function of the run number. There
are two regions in Fig. 11, which are separated by the vertical dashed line at
run 6600. For the runs up to run 6600, we tried to widen the electron beam
diameter or electron beam cross section. The change of the electron beam
diameter on the target position is not straightforward, since besides the beam
diameter on the target other beam parameters like divergence, energy stability,
helicity correlated position and angle fluctuations and also the beam tune in
the accelerator have to be optimized at the same time. Over the running time
of a few months, we were able to continuously increase the beam diameter of
the electron beam on the target by studying the beam tune in the electron
accelerator and in the beam optics of the focusing and bending magnets in the
beam line to our accelerator hall. The effect of continuously widening the beam
cross section can be be seen in Fig. 11 between run 2800 and run 6600 where
one recognizes a steady decrease of the measured absolute fluctuations which is
caused by the increase of the beam diameter over the time of the runs. In order
Fig. 9. Target density fluctuations at various conditions of tangential pump rotation
speed and electron beam current.
Fig. 10. Fluctuations at 20 µA.
Fig. 11. Investigations of target density fluctuations.
to verify this, the measured absolute fluctuations are plotted as a function of
the measured area of the elliptical beam cross section in Fig. 12. On can clearly
recognize a correlation between the measured fluctuations and the size of the
beam area. A further reduction of the target density fluctuations was possible
after we found, that there is a cool spot on the entrance or exit window of
the target cell. When we move the electron beam 1 mm vertically out of the
center of the target cell, the measured fluctuations decrease by an additional
factor of 4-8 as can be seen in Fig. 11 after run 6800. This can be explained
tentatively by the nonsymmetric flow of the hydrogen inside the target cell,
since right at the axis of symmetry all transversal directions of movement are
equivalent to each other. This leads to a suppression of transversal movement
and forms dead zones at windows. The asymmetric inlet in the target cell
causes a rotation of the hydrogen stream around the axis of symmetry, that
is a transversal velocity component. When the beam is moved from the axis
of symmetry, we see the effect of this rotation on heat removal. The resulting
low level of residual target density fluctuations can best be illustrated by the
two plots in Fig. 13. The upper plot shows the luminosity monitor signal
of the water Cherenkov monitor at small scattering angles downstream the
hydrogen target as a function of time in units of 20 ms integration gates. The
lower plot shows the signal of the electron beam current RF-cavity monitor in
Fig. 12. Measured fluctuations of the hydrogen target luminosity as a function of
the electron beam surface.
Fig. 13. Comparison of the signals of luminosity monitor (downstream the hydrogen
target) and beam current monitor (upstream of hydrogen target).
Fig. 14. Correlation analysis of the measured PIMO and LUMO data from Fig. 13.
The upper plot shows a scatter plot. The linear correlation comes from the electron
beam current fluctuations. For the middle plot, a straight line has been fitted to
the averaged data of the upper plot in order to disentangle correlated beam current
fluctuations from uncorrelated intrinsic monitor fluctuations. The lower plot shows
the histogram of the differences of the data (upper plot) and the straight line (middle
plot) (see text).
the accelerator beam line 3 m upstream the hydrogen target. The luminosity
signal follows even in details the electron beam current signal which means
that the contribution of target density fluctuations to the observed luminosity
fluctuations is very small. A quantitative result on the luminosity fluctuations
in the 20 ms of the integration time window can be yielded from a correlation
analysis of the PIMO and LUMO signals of Fig. 13. which is shown in Fig. 14.
The upper graph shows a scatter plot of the data from Fig. 13. The luminosity
monitor # 5 is for this plot normalized to its mean value as well as the current
monitor signal PIMO, which is normalized to its mean value too. The linear
correlation arises from electron beam current variations during this run. The
width of the band of points in the scatter plot is a measure for the individual
fluctuations of the monitors either due to their intrinsic noise or due to residual
target density fluctuations. A linear fit to the averaged data (middle plot) is
used to disentangle the correlation due to beam current fluctuations. For each
point in the upper plot the closest difference of the data points to the straight
line fit is calculated. These histogrammed differences are shown in the lower
plot. The width of this histogram ∆PL is the independent (quadratic) sum
of the three contributions: a) the intrinsic noise of the current monitor PIMO
(∆P/P), b) the intrinsic noise of the luminosity monitor LUMO ∆L5/L5, and
c) the residual target density fluctuations (∆ρ/ρ). The observed RMS of the
histogram of the normalized differences (lower plot) in Fig. 14 is ∆PL = 4.56×
10−4. If the intrinsic noise of both PIMO and LUMO monitor would be zero,
∆PL would directly correspond to the residual target density fluctuations.
The intrinsic noise of the luminosity monitor can be determined by applying
the same method as in Fig. 14 to the pair of luminosity monitors # 5 and
# 6 to ∆L5/L5= 1.25 × 10−4. The intrinsic PIMO noise can be determined
similarly by combining PIMO # 27 with PIMO # 08 in the beam line further
upstream to ∆P/P = 1.97 × 10−4. The residual target density fluctuations
within the integration time window of 20 ms can be calculated from
∆PL2− (∆L5/L5)2− (∆P/P) (13)
to ∆ρ/ρ = 3.92×10−4corresponding to ∆ρ/ρ = 3.20×10−6in the data taking
run time of five minutes.
We have designed a new cryogenic liquid hydrogen target system for the A4-
Experiment at the MAMI accelerator in Mainz. The special requirements
posed by the smallness of the measured parity violating cross section asym-
metry of order 10−6has put special demands on the design and construction
of the target cooling system. We have chosen the new approach which uses a
high turbulent flow in the target cell in order to enhance the transverse mixing
of the cold target liquid and thus enhancing the transverse heat transfer. This
new approach allows in combinations with an enlarged beam spot slightly off
center on the target to abandon rastering of the narrow beam spot on the tar-
get cell. In tests with and without electron beam we have been able to reduce
the target density fluctuations from a factor 23 larger then the electron beam
fluctuations to negligibly lower than the electron beam fluctuations. We plan
on further parity violation measurements including the use of deuterium and
a cell with 20 cm length at backward angles. The stability and the low level
of target density fluctuations of the liquid hydrogen target as it has been pre-
sented in this paper has shown to fulfill the high demands of a parity violation
This work has been supported by the Deutsche Forschungsgemeinschaft in the
framework of the SFB 201 and the SPP 1034. We are grateful to V. Lobashev
for useful discussions and participation in a very early state of the experiment.
 E. J. Beise et al., A high power liquid hydrogen target for parity violation
experiments, Nucl. Instr. and Meth. in Phys. Res. A 378 (1996) 383.
 D. T. Spayde, et al., The strange quark contribution to the proton’s magnetic
moment, Phys.Lett. B583 583 (2004) 79.
 K. A. Aniol et al., Parity violating electroweak asymmetry in polrized e−-p
scattering, Phys. Rev. C 69 (2004) 065501.
 P. L. Anthony et al., Observation of parity nonconservation in møller scattering,
Phys. Rev. Lett. 92 (2004) 181602.
 F. E. Maas et al., Measurement of strange quark contributions to the nucleon’s
form factors at q2=0.230 (gev/c)2, Phys. Rev. Lett. 93 (2004) 022002.
 F. E. Maas et al., Measurement of the transverse beam spin asymmetry in elastic
electron proton scattering and the inelastic contribution to the imaginary part
of the two-photon exchange amplitude, Phys. Rev. Lett. 94 (2005) 082001.
 F. E. Maas et al., Evidence for strange-quark contributions to the nucleon’s
form factors at q2=0.108(gev/c), Phys. Rev. Lett. (2005) in print.
 F. E. Maas et al., Parity violating electron scattering at the mami facility in
mainz: The strangeness contribution to the form-factors of the nucleon, Eur.
Phys. J. A 17 (2003) 339.
 F. E. Maas et al., Proc. of the ICATPP-7, World Scientific, 2002, p. 758.
 K. Aulenbacher et al., The mami source of polarised electrons, Nucl.Ins.Meth.A
391 (1997) 498.
 P. Achenbach et al., measurements and simulations of cerenkov light in lead
fluoride crystals, Nucl. Instrum. Meth. A 465 (2001) 318.
 W.Frost, T.H.Moulden, Handbook of Turbulence, Plenum press, 1977.
 J. W.G.Fastowski, A.E:Rowinski, Kryotechnik, Akademie-Verlag Berlin, 1970.
 A. I.S.Altarev, V.A.Mityukhljaiev, A.A.Zakharov, Cold and ultracold neutron
sources in gatchina, russia, Journal of Neutron Research 1 (4) (1993) 71.
 H. Euteneuer et al., Beam measurements and operating experience at mami, Download full-text
Proc of the EPAC (1994) 506.
 T. Hammel et al, A luminosity monitor for the a4 parity violation experiment
at mami, to be submitted to Nucl. Instr. and Meth. A .