Skip to main content

Development of a 3He magnetometer for a neutron electric dipole moment experiment


We have developed a highly sensitive 3He magnetometer for the accurate measurement of the magnetic field in an experiment searching for an electric dipole moment of the neutron. By measuring the Larmor frequency of nuclear spin polarized 3He atoms a sensitivity on the femto-Tesla scale can be achieved. A 3He/Cs-test facility was established at the Institute of Physics of the Johannes Gutenberg University in Mainz to investigate the readout of 3He free induction decay with a lamp-pumped Cs magnetometer. For this we designed and built an ultra-compact and transportable polarizer unit which polarizes 3He gas up to 55% by metastability exchange optical pumping. The polarized 3He was successfully transfered from the polarizer into a glass cell mounted in a magnetic shield and the 3He free induction decay was detected by a lamp-pumped Cs magnetometer.

PACS numbers

07.55.Ge Magnetometers for magnetic field measurements; 13.40 Electric and magnetic moments; 14.20 Protons and neutrons


In 1950, Purcell and Ramsey [1] were the first to point out that the existence of an electric dipole moment (EDM) in an elementary particle with spin would violate the parity symmetry P. Although there were strong arguments favoring the view that elementary particles and their interactions conserve parity, Ramsey and Purcell called for an experimental search for EDMs. Shortly after the discovery of P violation in 1957 [2]-[5] in weak interactions, Smith, Purcell, and Ramsey [6] reported the first (null) result of an experimental search for a neutron EDM. Besides P, an EDM also violates time reversal invariance, T, and the discovery of CP-violation in the neutral K meson’s decay [7] in 1964 (being equivalent to T-violation when assuming CPT conservation) removed objections against the possible existence of an EDM in elementary particles. However, besides the recent observation of CP-violation in the B meson system [8],[9], no other experimental evidence for CP- or T-violation has been revealed in elementary particle physics experiments nor in experiments on atoms or molecules up to date. Among all elementary particles, neutrons provide the best possibilities for the EDM search, since they are electrically neutral, may have a long-lived spin coherence, and can be stored for a long time in an environment which can hold a strong electric field.

The current upper limit of the neutron electric dipole moment (nEDM) [10], dn ≤ 2.9·10−26 e·cm (90% c.l.), represents one of the most precise measurements in physics and continues to challenge our understanding of fundamental physics, in particular the so-termed "‘Strong CP-Problem"’ and "‘SUSY CP-Problem"’, see, e.g. [11],[12]. Today, various efforts to improve the sensitivity to an nEDM are underway [13]-[17]. The nEDM collaborationa at the Paul Scherrer Institute (PSI) in Villigen, Switzerland is currently developing a next generation EDM setup which will use ultracold neutrons (UCN) delivered by a new spallation-based UCN source with a solid deuterium moderator/converter combination [18]. The nEDM experiment at PSI uses the conventional approach of storing UCN in a trap at room temperature. Additional improvements of the experiment are aimed at a sensitivity gain of more than one order of magnitude. The gain in measurement sensitivity, sets new challenges to the monitoring of systematic effects, in particular effects associated with magnetic field changes that are correlated with the switching of the electric field and which may mimic a false nEDM signal.

In this paper, we describe the technical realization of a 3He magnetometer system that shall be implemented into the PSI n2EDM experiment.

3He magnetometer for the n2EDM experiment

The planned n2EDM experiment at PSI [19],[20] is an attempt to increase the sensitivity of the neutron electric dipole measurement by more than one order of magnitude. The parallel electric and magnetic field configurations will be in the vertical plane, as shown in the left part of Figure 1. The spin precession chamber, i.e., the neutron storage volume will have a cylindrical double chamber geometry similar to the one used in [21] in order to minimize systematic effects. Further improvements are planned by the implementation of a new and effective magnetic shield, a better control of the magnetic field and its gradients as well as an improved magnetometry system.

Figure 1
figure 1

Projected layout of the spectrometer chamber (left) and UCPU of the n2EDM experiment. (1) coil system, (2) CsOPM, (3) 3He magnetometer vessel, (4) neutron guides, (5) storage chamber shutter, (6) double neutron spin precession chamber surrounded by electrode system, (7) valve, (8) multi-layer magnetic shielding, (9) non-magnetic piston compressor, (10) polarized 3He storage cell, (11) volume for metastable optical pumping, (12) transfer line and (13) 3He reservoir.

Precise knowledge of the magnetic field and its vertical gradient inside the neutron precession chamber during Ramsey cycles is of crucial importance for controlling several systematic effects. One approach described here, is to infer the magnetic field amplitude from the Larmor precession frequency of nuclear spin polarized 3He, detected by its free induction decay (FID) [22].

The 3He magnetometer will be placed in the electric field free region near the UCN storage chamber to avoid geometrical phase dependent systematic effects on the 3He nuclei proper [23]. Since the helium magnetometer is based on the free precession of nuclear spins, it is not prone to many systematic frequency shifts that are inherent to phase-locked optically pumped alkali vapour magnetometers and additionally gains in overall sensitivity due to the long 3He coherence times. The proposed 3He sandwich configuration, Figure 1, enables the control of vertical gradients of the magnetic holding field with high precision, whose knowledge and control is of great importance for tracing tiny geometrical phase effects in the nEDM spectrometer.

A magnetometric sensitivity in the femto-Tesla range has been demonstrated using SQUID (superconducting quantum interference device) magnetometers for the readout of the precessing 3He spins [24]. However, the required cryogenic infrastructure is a severe obstacle for operating these magnetometers under the experimental conditions of an EDM experiment (vacuum, high voltage, stabilized temperature, exclusive use of nonmagnetic materials). As shown for the first time by Cohen-Tannoudji et al. [25] in 1967 optically pumped alkali (Rb) magnetometers are a promising alternative for the readout of the 3He precession signal. Based on the expertise with optically pumped Cs magnetometers (CsOPMs) gained in the ongoing phase of the PSI-nEDM experiment, we plan to deploy a CsOPM-based, rather than a SQUID-based detection of the 3He FID signal. The CsOPM are developed by the Fribourg Atomic Physics group (FRAP) [26]-[28], member of the nEDM collaboration at PSI. The polarizer unit necessary for spin polarization of the 3He gas was built at the Johannes Gutenberg University (JoGu), Mainz, and is described in detail below.

During nEDM measurement, 3He at a pressure of 1 mbar will be polarized in an ultra-compact and transportable polarizer unit (UCPU) outside the n2EDM chamber in a field of 1 Gauss. The gas is then compressed to a pressure of ≈ 100 mbar and stored in low-relaxation glass cell. The whole UCPU is enclosed in a mu-metal cylinder to provide shielding from environmental fields and to allow generation of a homogeneous magnetic field by an internal cylindrical coil. Upon request, the compressed and polarized 3He in the storage cell will be transferred into the magnetometer vessels through a polytetrafluoroethylene (PTFE) tube surrounded by coils producing a suitable magnetic holding and guiding field on the order of 1 to 200 μ T.

The magnetometer vessels will be two flat cylindrical glass cells mounted inside the n2EDM vacuum tank, above and below (top, bottom) the neutron spin precession chamber, respectively (Figure 1). Since both magnetometer cells are traversed by almost the same magnetic flux as the UCN chambers, the average value of frequency measurements ωHe=(ωt,He+ωb,He)/2 yields a best guess for the neutron spin precession frequency ωn=(γn/γHe)ωHe, while the frequency difference determines the magnetic field gradient |B|/ z=(ωt,Heωb,He)/(γHe·Δ z) to a high precision. Here, Δ z is the distance between the centers of gravity of the upper and lower magnetometer vessel, ωt,He and ωb,He are the free spin precession frequencies of the 3He in the top and bottom magnetometer vessels, respectively, and γHe is the 3He gyromagnetic ratio.

The 3He FID is started by applying a π/2 spin-flip pulse. The long coherence time of the 3He FID signal, which strongly depends on the absolute magnetic field gradients ( 20 pT/cm), the magnetometer cell material, the size of the magnetometer vessels as well as the 3He pressure (pmbar)[29], reaches the order of one hour under the typical operating conditions of the EDM apparatus. Assuming a typical Ramsey cycle time for the ultracold neutrons of ≈ 200 s, it is thus possible to run 2–3 Ramey cycles with a single 3He filling without significant decay of the 3He FID signal amplitude. After these cycles the partly relaxed 3He will be pumped out, eventually recovered [30] and replaced by freshly polarized gas that has been prepared by the UCPU during the Ramsey cycles. This procedure will require the UCPU to deliver freshly polarized 3He gas every 10 to 15 min in order to ensure a high signal to noise ratio (SNR) of the recorded spin precession signal. A quasi-continuous operation of the 3He magnetometer will thus be possible.

Results and discussion

3He/Cs magnetometer test facility

A dedicated test facility was installed at the JoGu in order to construct and test the proposed 3He/Cs magnetometer system. As main part of this facility, a cylindrical four layer mu-metal shield (MS) [31] with magnetic shielding factors of 100 (longitudinal) and 1000 (transverse) was adapted to the requirements imposed by the 3He FID studies. In order to lower existing magnetic field gradients and to record the 3He FID signal, the innermost mu-metal cylinder (610 mm diameter, 1300 mm length) was equipped with a magnetic coil system (solenoid and cosine-theta coil) providing magnetic fields parallel and perpendicular to the cylinder axis. The coils were wound on a double-walled cardboard cylinder which was then covered with five layers of Metglas [32]. With these arrangements, magnetic field gradients of 150 to 350 pT/cm could be achieved in the longitudinal direction. For the transverse field gradients, values in range of 350 to 450 pT/cm were determined. Within this magnetic environment two basic types of experiments have been performed:

- FID detection of in-situ polarized 3He in a small cell, and

- FID detection of externally polarized and transferred 3He.

The setup for in-situ polarization of 3He is presented in Figure 2. The mechanical support of the components was produced from non-magnetic materials, e.g., acrylic glass and polyethylene. The optical axis was aligned to the cylinder axis of the solenoidal coil. A beam of linearly polarized light from a 5 W ytterbium doped 1083 nm fiber laser was sent via an optical fiber into the magnetic shield. A beam expander coupled to the fiber (Figure 2, item 1) widened the incoming laser beam to a diameter of 30 mm. The light was circularly polarized by a polarizing beam splitter cube (PBS) with an attached λ/4-plate, item 2 of Figure 2. The laser beam can be blocked after the λ/4-plate by means of a compressed-air driven shutter (item 3 of Figure 2). The cylindrical 3He precession cell, made from borosilicate glass and filled with 3He at a pressure of 1 mbar is positioned in the center of the setup. A capacitively coupled AC high voltage produced by a Tesla transformer (item 4 of Figure 2) driven at 1.7 MHz is used to ignite and sustain a discharge inside the 3He cell during the metastability exchange optical pumping (MEOP) [33]. The optical pumping beam is retro-reflected by a dichroic optical mirror (item 7 of Figure 2) for a second passage through the cell. Fluorescense light of 668 nm transmitted by the dichroic mirror is analyzed by an optical polarimeter (OPN), item 8 of Figure 2. The degree of circular polarisation of the fluorescense light measured by the OPN is proportional to the degree of spin polarisation of the 3He gas and can hence be used to monitor the 3He polarisation build-up [34]. A discharge lamp pumped cesium magnetometer (Figure 2, item 6) operated in the self-oscillating Mx-mode [26],[35] is positioned close to the 3He cell. The output of this type of magnetometer is a continuous sinusoidal voltage signal of constant amplitude, whose frequency ωCs is related to the modulus of the magnetic field via the Cs gyromagnetic ratio

ω Cs = γ Cs B .
Figure 2
figure 2

Inner setup of the 3 He/Cs test setup consisting of a beam expander (1), beam splitter cube (2), shutter (3), Tesla transformer coil (4), closed glass cell filled with 3 He (5), Cs-magnetometer (6), dichroic mirror (7) and optical polarimeter (8).

For the field (≈ 1 μ T) applied in the present study, ωCs/2π is ≈ 3.5 kHz. When the solenoid coil is powered and a discharge is ignited inside the cell, the 3He pump laser produces nuclear spin polarization along the direction B ̂ sol of the solenoid axis. The degree of polarization reaches approximately 65% after two minutes of optical pumping. There are two different ways to initiate the 3He FID after the discharge and the laser have been switched off:

- The magnetic field can be switched rapidly from the longitudinal solenoid (along ŷ-axis) to the transverse cosine-theta coil (field-flip) along z ̂ -axis, inducing a non-adiabatic field rotation when the switching time is much shorter than the 3He precession period

1 t switch ν He .

- A pulse of resonant rf radiation can be applied with the cosine-theta coil to rotate the 3He spin polarization by 90° (π/2 spin-flip pulse).

Please note that after a field flip the 3He polarization will precess around B ̂ cos in the x ̂ ŷ plane, while a spin flip results in a precession around B ̂ sol in the x ̂ z ̂ plane. The macroscopic magnetization m that is associated with the spin-polarisation of the 3He atoms produces - outside of a spherical cell containing the 3He gas - a magnetic dipole field that is given by

B dipole (r)= μ 0 4 π 3 r m · r m · r 2 r 5 .

When the magnetization precesses around the applied magnetic field B0 it thus gives rise to a small alternating magnetic field bHe(t)=Bdipole cosωHet in the vicinity of the 3He cell which is vectorially added to B0. The magnetic field at the position of a nearby CsOPM is thus modulated at the 3He’s larmor frequency

ω He = γ He | B 0 + b He (t)| γ He | B 0 + B ̂ 0 · b He (t)|.

As a consequence the CsOPMs output signal becomes frequency modulated (FM) by the 3He FID, the FM showing up as sidebands in the Fourier transform of the CsOPM’s oscillatory signal. Eq. 4 shows that the Cs magnetometer reading is only sensitive (to first order in |bHe|/|B0|) to the projection of bHe onto B0. Inspection of Eq. 3 shows that for a spherical cell in the FID-geometry of Figure 3 this projection is maximal when the Cs magnetometer is located on a cone with a half-opening angle of 45° with respect to B0. Numerical calculations have shown that the 45° condition is also a good approximation for the (60 mm long, 60 mm diameter) cylindrical cell used in the actual experiments. Consequently, the CsOPM was positioned such that the center of the Cs sensor cell was roughly on the 45° cone to maximize the measurement signal. It is equally important to minimize the distance between the CsOPM and the 3He cell since the dipole field drops with 1/r3, as can be seen from Eq. 3. The minimum achievable distance imposed by the sizes of the 3He and Cs cells was d=70 mm. A typical result of a 3He FID measurement is shown as the Fourier transform of the Cs magnetometer’s oscillating signal in Figure 4. The displayed FFT spectrum was calculated from a 10 s long chunk of data. The characteristic carrier and sideband components of the FM signal are clearly visible. The sidebands resulting from the 3He FID are offset by 32.8 Hz with respect to the carrier, each one having an amplitude AHe = 0.1 mV with a carrier amplitude of ACs = 2.2 V. One can show that [36] the amplitude bHe of the oscillatory field produced by the precessing 3He nuclear spins and the characteristic parameters of the Fourier spectrum is given by

b He = 4 π ν He γ Cs A He A Cs .
Figure 3
figure 3

Alternating dipole field produced by the precessing magnetization m of spin polarized3He atoms around the main magnetic field B 0 . The sensor cell of the CsOPM must be positioned as close as possible in a 45° arrangement to maximize the CsOPM signal.

Figure 4
figure 4

Discrete Fourier spectrum obtained from a 10 second long data sample from the lamp-pumped CsOPM. Peak of carrier frequency at 3513 Hz, sidebands from the 3He FID and 50 Hz sidebands are visible. 3He was polarized in situ and precession occured around the cosine-theta field direction. The output signal of the CsOPM was digitized at a sampling rate of 50 kHz with a resolution of 18 bits.

For the measurement presented in Figure 4, Eq. 5 yields bHe = 8.5 pT which represents a SNR of about 34 in a bandwidth of 1 Hz. We inferred the T 2 -time from a measurement of the time dependent decay of the amplitude AHe. For different measurements in both the Bsol and Bcos field configurations spin coherence times of up to one hour were obtained. The feasibility of the new test setup for further investigations of the functionality of an ultra-compact 3He polarizer unit (UCPU) is successfully demonstrated.


Ultra compact polarizer unit

As mentioned above the main component of the future 3He magnetometer system, the 3He polarizer unit (UCPU) should be able to deliver periodically i.e., every 10-15 min, 16 mbar·l of spin-polarized 3He gas, corresponding to a 3He pressure of about 1 mbar in the two magnetometer vessels that have a total volume of 16 l. The UCPU is an ultra-compact transportable 3He polarizer following several generations of polarizer facilities which were developed at JoGu for medical application respectively fundamental research [37]-[45]. It has been specifically designed to fulfill the needs of the planned n2EDM experiment at PSI.

As shown in Figures 5 and 6, the working principle of the polarizer can be described as follows. 3He from a high pressure reservoir is fed into the gas system of the polarizer. The pressure inside the system is controlled by a Barocel pressure sensor (item B100 in Figure 5) from the company Edwards Vacuum and a feedback-driven electrically adjustable valve from the company MKS. After purification by means of a getter-based purifier from the company SAES (impurities < ppm), the 3He gas is filled via a specific nonmagnetic valve (item EV in Figure 5) into the optical pumping cells OPZ1 and OPZ2. The EV is a custom-made all-metal construction that is actuated by compressed air. The optical pumping cells consist of two connected borosilicate glass tubes (length 1.1 m, diameter 54 mm) with vacuum tight antireflection-coated (λ =1083 nm) fused silica windows at the ends. The vacuum grease Apiezon H is used as a sealing agent between the fused silica windows and the glass tubes. Specific cover flanges allow differential pumping to prevent capillary cracks in the vacuum grease induced by outer air pressure [36]. The total optical pumping volume is approximately 5 l. The pressure inside the pumping cells is controlled by a second Barocel pressure sensor (item B10 in Figure 5) to ensure efficient optical pumping. Thin stripes of cooper foil at top and bottom of both cells serving as electrodes connected to the power-amplified output of a frequency generator for sustaining a weak discharge in the 3He gas, necessary for MEOP. The light beams enter both pumping cells through a polarizer- λ/4 plate combination and are back-reflected by dichroic mirrors after the cells’ exit windows. A tunable fiber laser from the company Keopsys with a maximal output power of 5 W (linewidth ≈ 2 GHz) is used for optical pumping. The laser is tuned to the C8-lineb of the 3He line spectrum and its frequency stability can be monitored by measuring the fluorescence light from a small auxilliary 3He reference cell. The linearly polarized laser beam is widened by a beam expander to a diameter of 30 mm to illuminate the whole cross-section of the pumping cells. The light becomes circularly polarized after passing the PBS and λ/4 plate and is then sent through the pumping cells. The mirrors at the far ends of the cells have a high reflectivity in the 1083 nm range but high transmission for the 668 nm fluorescence light, providing a low-loss double-passage of the 1083 nm light through each cell while permitting to monitor the degree of 3He spin polarization with an optical polarimeter (OPN) by polarization analysis of the 668 nm fluorescence light. The longitudinal relaxation time T1 was measured by using the OPN to observe the decay of polarization after switching off the laser. With a burning discharge we find T1,bright=1.5 min while without gas discharge the measurement yields T1,dark=8 min. When the nuclear spin polarization of 3He in the optical pumping cells has reached the steady state value of typically 50-55%, a valve (item V1 in Figure 5) is opened to suck out part of the polarized gas into the empty compressor volume of V = 3 l. The piston compressor then pumps the gas into a small storage cell of 0.33 liter volume. The storage cell is made of a special glass mixture GE180 to minimize 3He polarization relaxation during wall collisions [46]. The compressor is built entirely from non-magnetic parts and driven entirely by compressed air. The compressor volume is closed by a head with six vacuum tight valves (items V1-V6 in Figure 5), which open and close the connections between the compressor volume, the optical pumping volume, the storage cell and the vacuum system. All valves are operated by custom-modified compressed air driven actuators from the company Festo. Great care was taken to isolate the 3He gas system from the compressed-air line. In places where compressed air could possibly leak into the 3He circuit, small separation volumes connected to a vacuum pump were introduced to remove any possible leakage flow i.e., item PL3 in Figure 5. All relevant volumes of the system are connected by four tubes (items PL1-PL4 in Figure 5) to two turbo molecular pumps (items TP1 and TP2 in Figure 5) and can be evacuated. During movement of the piston we measure a residual gas pressure as low as 10−6 mbar in the compression volume. The UCPU is mounted inside a solenoid coil (length 1.8 m, diameter 0.52 m) which produces a magnetic field of 250 μ T at a coil current of 200 mA. In order to minimize stray fields and to improve the homogeneity of the magnetic field the coil is enclosed in a cylindrical single layer mu-metal shield. Additionally, a correction coil is mounted at each of the flat end-caps of the shield to further homogenize the field inside. All components inside the shielding, except the pressure sensors, are manufactured from non-magnetic materials (e.g. titanium, brass, aluminum) to avoid any degradation of the field homogeneity. As a result, the relative magnetic field gradients in the area of the optical pumping volume and the storage cell are below ( Bz/ z)/Bz<3.8· 10−4 cm−1, where z is the direction along the cylinder axis. The automatic operation of the UCPU is established by a software package written under the LABVIEW environment. The system can work autonomously and no user interaction is required under normal conditions. In various tests the polarizer unit has successfully demonstrated its working performance by periodically delivering polarized and compressed 3He. In a cyclic mode of operation the degree of spin polarization in the optical pumping volume reaches up to 52%, as shown in Figure 7. The losses of polarization occuring during compression into the storage cell are below 1% and thus negligible. We have also demonstrated that the required 16 mbar ·l of polarized 3He for the planned magnetometer setup can be produced within an optical pumping time of less than seven minutes (within 9 cycles of 45 s).

Figure 5
figure 5

Schematic sketch of the UCPU. Details, see text. Solenoid and magnetic shielding are not shown.

Figure 6
figure 6

Ultra-compact polarizer (height ≈2 m, width of mounting plate 70 × 70 cm2), gas system and vacuum pumps. Right: Optical pumping cells with ignited discharge.

Figure 7
figure 7

Time course of polarization observed with OPN demonstrating the cyclic operation of the UCPU. The maximum polarization of ≈ 52% is reached seven times in sequence. The sharp peaks are artefacts that result from changes in discharge color and brightness when polarized gas is transferred into the compression volume or when non-polarized gas is filled into the optical pumping volume.

Polarization maintaining transfer

Once the polarized 3He is prepared by the UCPU it is delivered to the installed magnetometer vessels using polarization-stabilizing magnetic guiding fields. As shown in Figure 1, the distance between the polarizer unit and the n2EDM setup is on the order of several meters and may contain a few bended sections. In order to investigate and, if necessary, to optimize polarization losses during transfer, we have installed a demonstration transfer line system consisting of a 6 m long PTFE tube with an inner diameter of 6 mm, as shown in Figure 8. This transfer line connects an outlet valve of the compressor head (see Figure 5, item V5) with a spherical glass cell (GE180, volume Vmag=0.09 l) inside the test setup (see Figure 9, item 3). After injection of the gas the cell is shut off by a separate compressed air driven nonmagnetic valve (see Figure 9, item 2). The main challenge in the polarized gas transfer was to provide a suitable holding field along the transfer line which has two 90° bends. In order to minimize polarization losses by gradient-induced spin-flips, the gradients along the transition from the 250 μ T field inside the UCPU’s shield to the 1 μ T field of the cosine-theta coil in the test facility had to be minimized. Along the vertical parts of the line, above the UCPU and at the transition through the MS, solenoid coils (Figure 8, (1) and (3)) provided these fields. The relative gradients ( Bz/ z)/Bz were found to be 1−6·10−2 cm−1 for the first solenoid and 3−5·10−2 cm−1 for the second respectively, where z ̂ is the solenoid axis. For the horizontal part of the transfer line Figure 8 (2), a dedicated arrangement of four wires running along each side of the tube was used to generate the required holding field. Field simulations were done to ensure small gradients in the x ̂ and ŷ direction. Measurements of the relative field gradient along the horizontal part of the transfer line yielded ( Bz/ x)/Bz≤3·10−3 cm−1. To quantify the transport losses, 16 mbar·l of 3He with a polarization of about 52% were prepared by the UCPU and filled via the transfer line into the evacuated magnetometer cell. The filling process took less than 0.5 s and yielded an inner cell pressure of approximately 30 mbar. The 3He FID was initiated by a non-adiabatic field switch and the FID was measured as described in Sec.3He/Cs magnetometer test facility. The 10 s FFT spectrum of the CsOPM signal shows a 3He amplitude of AHe=18 mV with a carrier amplitude of ACs=2,2 mV (see Figure 10). Using Eq. 5 we infer that the magnetic field seen by the CsOPM is 157 pT, which corresponds to a SNR of 21 per 1 mbar and 1 Hz bandwidth. Assuming no transfer losses we can estimate the expected field strength from Eq. 3. The magnetic moment m inside the magnetometer cell is given by

m= μ He PN= 2.127 μ K P N A p V R T .
Figure 8
figure 8

CAD drawing of the transfer setup: The storage cell of UCPU (shown without solenoid coil and shielding) at the left is connected via a transfer line (red) to a magnetometer vessel inside the MS on the right. (1) - (3) Coils to provide polarization conserving guiding fields along the transfer line.

Figure 9
figure 9

Interior of the MS coil system for transfer measurement. (1) PTFE transfer tube, (2) non-magnetic pressed air driven valve, (3) 3He magnetometer cell and (4) lamp pumped Cs-magnetometer.

Figure 10
figure 10

Discrete Fourier spectrum of lamp pumped Cs-magnetometer signal with 30 mbar magnetometer cell pressure and 52% initial polarization. Cs-carrier frequency shifted to zero. Sidebands at ± 32 Hz resulting from 3He FID and 50 Hz perturbations are visible. The small satellite peaks result from vibrations of the setup during measurement.

With a polarization P of 52%, a pressure p of 30 mbar and a volume V of 0.09 l one obtains m=3.66·10−7 Am2. Assuming that at t=0 the magnetic moments are oriented in the x ̂ -direction and knowing that the centers of the 3He cell and the CsOPM cell are placed in the x ̂ ŷ plane of the MS coil system on the 45° cone (Figure 3) at a distance r of 70(2)mm the expected 3He magnetic field seen by the CsOPM is bHe=160(14) pT. The difference between the prediction, assuming a lossless transfer, and the measured value is below 10%. From this we expect the transfer losses to be 2.0(9)%.


A test facility for the recording of 3He FID by a lamp-pumped CsOPM was installed at University of Mainz. A compact, transportable 3He polarizer was built and its functionality demonstrated. It was shown that the device is capable of delivering the required amounts of polarized gas for the n2EDM experiment with 52% of spin polarization during cyclic operation. For the first time the transfer of polarized 3He gas from a cyclically working polarizer into a multilayer mu-metal shield, even through a transfer line having two 90° bends, with negligible depolarization was successfully demonstrated. Thus, together with 3He magnetometers made of flat cylindrical cells [22], all essential 3He magnetometer parts (see Figure 1) were built and tested. Further improvements of the magnetometer largely depend on the magnetic field conditions in the next generation nEDM spectrometer (longer 3He transverse relaxation times, lower magnetic field noise, higher SNR in the 3He precession detection). Besides this the construction of a recovery and recycling system for used 3He gas will help to drive the whole magnetometer setup in a more economical way [30]. A detailed analysis of the achievable sensitivity and possible systematic effects in the n2EDM spectrometer with a combined 3He/Cs-magnetometer is ongoing and will be published elsewhere.


a list of n2EDM members, see also

b C8: 2 3S1, F=1/2 → 2 3P0, F=1/2, λ=1083,06 nm.


  1. Purcell EM, Ramsey NF: On the possibility of electric dipole moments for elementary particles and nuclei. Phys Rev 1950, 78: 807. 10.1103/PhysRev.78.807

    Article  ADS  Google Scholar 

  2. Lee TD, Yang CN: Question of parity conservation in weak interactions. Phys Rev 1956, 104: 254. 10.1103/PhysRev.104.254

    Article  ADS  Google Scholar 

  3. Wu CS, Ambler E, Hayward RW, Hoppes DD, Hudson RP: Experimental test of parity conservation in beta decay. Phys Rev 1957, 105: 1413. 10.1103/PhysRev.105.1413

    Article  ADS  Google Scholar 

  4. Garwin RW, Lederman LM, Weinrich M: Observation of the failure of conservation of parity and charge conjugation in meson decays: the magnetic moment of the free muon. Phys Rev 1957, 105: 1415. 10.1103/PhysRev.105.1415

    Article  ADS  Google Scholar 

  5. Friedman JI, Telegdi VL: Nuclear emulsion evidence for parity nonconservation in the decay chain π+ μ+ e+. Phys Rev 1957, 106: 1290. 10.1103/PhysRev.106.1290

    Article  ADS  Google Scholar 

  6. Smith JH, Purcell EM, Ramsey NF: Experimental limit to the electric dipole moment of the neutron. Phys Rev 1957, 108: 120. 10.1103/PhysRev.108.120

    Article  ADS  Google Scholar 

  7. Christenson JH, Cronin JW, Fitch VL, Turley R: Fundamental aspects weak interaction. Phys Rev Lett 1964, 13: 138. 10.1103/PhysRevLett.13.138

    Article  ADS  Google Scholar 

  8. Aubert B, Boutigny D, Gaillard DJM, Hicheur A, Karyotakis Y, Lees JP, Robbe P, Tisserand V, Zghiche A, Palano A, Pompili A, Chen GP, Chen JC, Qi ND, Rong G, Wang P, Zhu YS, Eigen G, Stugu B, Abrams GS, Borgland AW, Breon AB, Brown DN, Button-Shafer J, Cahn RN, Gill MS, Gritsan AV, Groysman Y, Jacobsen RG, Kadel RW, et al.: Search for T and CP violation in B0−= B0 mixing with inclusive Dilepton events . Phys Rev Lett 2002, 88: 231801. 10.1103/PhysRevLett.88.231801

    Article  ADS  Google Scholar 

  9. Abe K, Abe K, Abe T, Adachi I, Aihara H, Akatsu M, Asano Y, Aso T, Aulchenko V, Aushev T, Bakich AM, Banas E, Bay A, Behera PK, Bizjak I, Bondar A, Bozek A, Bracko M, Brodzicka J, Browder TE, Casey BCK, Chang P, Chao Y, Chen KF, Cheon BG, Chistov R, Choi Y, Choi YK, Danilov M, Dong LY, et al.: Improved measurement of mixing-induced CP violation in the neutral B meson system. Phys Rev D 2002, 66: 071102(R). 10.1103/PhysRevD.66.071102

    Article  ADS  Google Scholar 

  10. Baker CA, Doyle DD, Geltenbort P, Green K, van der Grinten MDG, Harris PG, Iaydjiev P, Ivanov SN, May DJR, Pendlebury JM, Richardson JD, Shiers D, Smith KF: Improved experimental limit on the electric dipole moment of the neutron. Phys Rev Lett 2006, 97: 131801. 10.1103/PhysRevLett.97.131801

    Article  ADS  Google Scholar 

  11. Khriplovich IB, Lamoureux SK: CP Violation Without Strangeness . Springer, Berlin, Heidelberg; 1997.

    Book  Google Scholar 

  12. Pospelov M, Ritz A: Electric dipole moments as probes of new physics. Ann Phys 2005, 318: 119. 10.1016/j.aop.2005.04.002

    Article  ADS  MATH  Google Scholar 

  13. Matsuta K, Masuda Y, Hatanaka K, Kawasaki S, Matsumiya R, Mihara M, Jeong SC, Watanabe Y, Adachi T, Nishimura D, Asahi K, Martin JW, Konaka A, Miller A, Bidinosti C, Dawson T, Lee L, Davis CA, Ramsay WD, van Oers WTH, Korkmaz E, Buckman L: An EDM measurement with a new comagnetometer and a high density UCN source. AIP Conf Proc 2013, 1560: 152. 10.1063/1.4826742

    Article  ADS  Google Scholar 

  14. Serebrov AP, Kolomenskiy EA, Pirozhkov AN, Krasnoshekova IA, Vasiliev AV, Polyushkin AO, Lasakov MS, Fomin AK, Shoka IV, Solovey VA, Zherebtsov OM, Geltenbort P, Zimmer O, Ivanov SN, Alexandrov EB, Dmitriev SP, Dovator NA: New measurements of neutron electric dipole moment. arXiv 2013,1310.5588. [nucl-ex].

  15. Zenner J: The search for the neutron electric dipole moment. PhD dissertation. University of Mainz; 2013.

    Google Scholar 

  16. Altarev I, Beck DH, Chesnevskaya S, Chupp T, Feldmeier W, Fierlinger P, Frei A, Gutsmiedl E, Kuchler F, Link P, Lins T, Marino M, McAndrew J, Paul S, Petzoldt G, Pichlmaier A, Stoepler R, Stuiber S, Taubenheim B: A next generation measurement of the electric dipole moment of the neutron at the FRM II. Nuovo Cimento-C 2012,35(4):122.

    Google Scholar 

  17. Ito TM: Plans for a neutron EDM experiment at SNS. J Phys: Conference Series 2007,69(1):012037.

    ADS  Google Scholar 

  18. Anghel A, Atchison F, Blau B, van den Brandt B, Daum M, Doelling R, Dubs R, Duperrex PA, Fuchs A, George D, Gültl L, Hautle P, Heidenreich G, Heinrich F, Henneck R, Heule S, Hofmann T, Kasprzak M, Kirch K, Knecht A, Konter JA, Korhonen T, Kuzniak M, Lauss B, Mezger A, Mtchedlishvili A, Petzoldt G, Pichlmaier A, Reggiani D, Reiser R, et al.: The PSI ultra-cold neutron source. Nucl Instrum Methods Phys Res A 2009, 611: 133. 10.1016/j.nima.2009.07.077

    Article  ADS  Google Scholar 

  19. Altarev I, Bison G, Ban G, Bodek K, Burghoff M, Cvijovic M, Daum M, Fierlinger P, Gutsmiedl E, Hampel G, Heil W, Henneck R, Horras M, Khomutov N, Kirch K, Kistryn S, Knappe-Grüneberge S, Knecht A, Knowles P, Kozela A, Kratz JV, Kuchler F, Kuzniak M, Lauer T, Lauss B, Lefort T, Mtchedlishvili A, Naviliat-Cuncic O, Paul S, Pazgalev AS, et al.: Towards a new measurement of the neutron electric dipole moment. Nucl Instrum Methods Phys Res A 2009,611(2–3):133–136. 10.1016/j.nima.2009.07.046

    Article  ADS  Google Scholar 

  20. Altarev I, Bison G, Ban G, Bodek K, Burghoff M, Chowdhuri Z, Daum M, Düsing C, Fertl M, Fierlinger P, Franke B, Grab C, Gutsmiedl E, Hampel G, Heil W, Henneck R, Horras M, Khomutov N, Kirch K, Kistryn S, Knappe-Grüneberg S, Knecht A, Knowles P, Kozela A, Kraft A, Kuchler F, Kratz JV, Lauer T, Lauss B, Lefort T, et al.: An improved measurement of the electric dipole moment of the neutron. Nucl Phys A 2010,844(1–4):47C. 10.1016/j.nuclphysa.2010.05.012

    Article  ADS  Google Scholar 

  21. Altarev IS, Borisov Yu V, Brandin AB, Egorov AI, Ezhov VF, Ivanov SN, Lobashov VM, Nazarenko VA, Porsev GD, Ryabov VL, Serebrov AP, Taldaevet RR: A search for the electric dipole moment of the neutron using ultracold neutrons. Nucl Phys A 1980,341(2):269. 10.1016/0375-9474(80)90313-9

    Article  ADS  Google Scholar 

  22. Borisov Y, Heil W, Leduc M, Lobashev V, Otten EW, Sobolev Yu: Feasibility study of a3He-magnetometer for neutron electric dipole moment experiments . Nucl Instrum Methods Phys Res A 2000,440(3):483. 10.1016/S0168-9002(99)01024-4

    Article  ADS  Google Scholar 

  23. Pendlebury JM, Heil W, Sobolev Yu, Harris PG, Richardson JD, Baskin RJ, Doyle DD, Geltenbort P, Green K, van der Grinten MGD, Iaydjiev PS, Ivanov SN, May DJR, Smith KF: Geometric-phase-induced false electric dipole moment signals for particles in traps. Phys Rev A 2004, 70: 032102. 10.1103/PhysRevA.70.032102

    Article  ADS  Google Scholar 

  24. Gemmel C, Heil W, Karpuk S, Lenz K, Ludwig Ch, Sobolev Yu, Tullney K, Burghoff M, Kilian W, Knappe-Grüneberg S, Müller W, Schnabel A, Seifert F, Trahms L, Bael~er St: Ultra-sensitive magnetometry based on free precession of nuclear spins. Eur Phys J D 2010,57(3):303–320. 10.1140/epjd/e2010-00044-5

    Article  ADS  Google Scholar 

  25. Cohen-Tanudji C, DuPont-Roc J, Haroche S, Laloë F: Detection Of the static magnetic field produced by the oriented nuclei of optically pumped3He gas . Phys Rev Lett 1969,22(15):758–760. 10.1103/PhysRevLett.22.758

    Article  ADS  Google Scholar 

  26. Groeger S: Laser-pumped cesium magnetometers for the PSI-nEDM experiment. PhD. dissertation. University of Fribourg; 2005.

    Google Scholar 

  27. Groeger S, Bison G, Schenker J-L, Wynands R, Weis A: A high-sensitivity laser-pumped M x magnetometer . Eur Phys J D 2006, 38: 239. 10.1140/epjd/e2006-00037-y

    Article  ADS  Google Scholar 

  28. Knowles P, Bison G, Castagna N, Hofer A, Mtchedlishvili A, Pazgalev A, Weis A: Laser-driven Cs magnetometer arrays for magnetic field measurement and control. Nucl Instrum Methods Phys Res A 2009,611(2–3):306. 10.1016/j.nima.2009.07.079

    Article  ADS  Google Scholar 

  29. Cates GD, Schaefer SR, Happer W: Relaxation of spins due to field inhomogeneities in gaseous samples at low magnetic fields and low pressures. Phys Rev A 1988,37(8):2877–2885. 10.1103/PhysRevA.37.2877

    Article  ADS  Google Scholar 

  30. Zimmer St: Optimierung und Charakterisierung eines3He Magnetometers fuer nEDM-Experimente. Diploma thesis. University of Mainz; 2013.

    Google Scholar 

  31. Brys T, Czekaj S, Daum M, Fierlinger P, George D, Henneck R, Kasprzak M, Kirch K, Kuzniak M, Kuehne G, Pichlmaier A, Siodmok A, Szelc A, Tanner L, Assmann C, Bechstein S, Drung D, Schurig Th, Ciofi C, Neri B: Magnetic field stabilization for magnetically shielded volumes by external field coils. Nucl Instrum Methods Phys Res 2005,554(1–3):527–539. 10.1016/j.nima.2005.08.040

    Article  ADS  Google Scholar 

  32. Kubo S, Sasada I, Harada K: Effective shielding for low?level magnetic fields. J Appl Phys 1988,64(10):5696–5698. 10.1063/1.342278

    Article  ADS  Google Scholar 

  33. Happer W: Optical pumping. Rev Modern Phys 1972, 44: 169–250. 10.1103/RevModPhys.44.169

    Article  ADS  Google Scholar 

  34. Hasse J: Charakterisierung und Optimierung eines3He-Kompressors. Diploma thesis. University of Mainz; 2000. Hasse J: Charakterisierung und Optimierung eines3He-Kompressors. Diploma thesis. University of Mainz; 2000.

  35. Groeger S, Pazgalev AS, Weis A: Comparison of discharge lamp and laser pumped cesium magnetometers. Appl Phys B 2005,80(6):645. 10.1007/s00340-005-1773-x

    Article  ADS  Google Scholar 

  36. Kraft A: Aufbau und Inbetriebnahme eine hochsensitiven3He-Magnetometers fuer ein zukuenftiges Experiment zu Bestimmung eines elektrischen Dipolmoments des freien Neutrons. PhD. dissertation. University of Mainz; 2012.

    Google Scholar 

  37. Becker J, Heil W, Krug B, Leduc M, Meyerhoff M, Nacher PJ, Otten EW, Prokscha Th, Schearer LD, Surkau R: Study of mechanical compression of spin-polarized3He gas . Nucl Instrum Methods Phys Res Sect A 1994,346(1–2):45–51. 10.1016/0168-9002(94)90687-4

    Article  ADS  Google Scholar 

  38. Ebert M, Grossmann T, Heil W, Otten EW, Surkau R, Thelen M, Leduc M, Bachert P, Knopp MV, Schad LR: Nuclear magnetic resonance imaging with hyperpolarised helium-3. The Lancet 1996,347(9011):1297–1299. 10.1016/S0140-6736(96)90940-X

    Article  Google Scholar 

  39. Surkau R, Becker J, Ebert M, Grossmann T, Heil W, Hofmann D, Humblot H, Leduc M, Otten EW, Rohe D, Siemensmeyer K, Steiner M, Tasset F, Trautmann N: Realization of a broad band neutron spin filter with compressed, polarized3He gas . Nucl Instrum Methods Phys Res A 1997,384(2–3):444–450. 10.1016/S0168-9002(96)00923-0

    Article  ADS  Google Scholar 

  40. Deninger AJ, Eberle B, Ebert M, Grossmann T, Heil W, Kauczor H, Lauer L, Markstaller K, Otten E, Schmiedeskamp J, Schreiber W, Surkau R, Thelen M, Weiler N: Quantification of regional intrapulmonary oxygen partial pressure evolution during apnea by3He MRI . J Magn Reson 1999,141(2):207–216. 10.1006/jmre.1999.1902

    Article  ADS  Google Scholar 

  41. Ebert M: Entwicklung eines leistungsstarken Polarisators und Kompressors fuer3He fuer medizinische MR Tomographie. Ph. D. dissertation. University of Mainz; 2000.

    Google Scholar 

  42. Schmiedeskamp J: Weiterentwicklung einer Produktionsanlage und der Speicherungs- bzw. Transportkonzepte fuer hochpolarisiertes3He, Anwendungen in Kernspintomographie und physikalischer Grundlagenforschung. PhD. dissertation. University of Mainz; 2004.

    Google Scholar 

  43. Thien F, Friese M, Cowin G, Maillet D, Wang D, Galloway G, Brereton I, Robinson PJ, Heil W, Thompson B: Feasibility of functional magnetic resonance lung imaging in Australia with long distance transport of hyperpolarized helium from Germany. Respirology 2008,13(4):599–602. 10.1111/j.1440-1843.2008.01262.x

    Article  Google Scholar 

  44. Krimmer J, Distler M, Heil W, Karpuk S, Kiselev D, Salhi Z, Otten EW: A highly polarized3He target for the electron beam at MAMI . Nucl Instrum Methods Phys Res A 2009,611(1):18–24. 10.1016/j.nima.2009.09.064

    Article  ADS  Google Scholar 

  45. Tullney K, Allmendinger F, Burghoff M, Heil W, Karpuk S, Kilian W, Knappe-Grüneberg S, Müller W, Schmidt U, Schnabel A, Seifert F, Sobolev Yu, Trahms L: Constraints on spin-dependent short-range interaction between nucleons. Phys Rev Lett 2013,111(1–5):100801. 10.1103/PhysRevLett.111.100801

    Article  ADS  Google Scholar 

  46. Schmiedeskamp J, Heil W, Otten EW, Kremer RK, Simon A, Zimmer J: Paramagnetic relaxation of spin polarized3He at bare glass surfaces . Eur Phys J D 2006,38(3):427–438. 10.1140/epjd/e2006-00050-2

    Article  ADS  Google Scholar 

Download references


The described work benefited from the excellent support of the mechanical workshops of the Institute of Physics and the Institute of Nuclear Chemistry of the University of Mainz, especially L. Funke and glassblower R. Jera. We also thank Yu. V. Borisov, Konstantinov Nuclear Physics Institute, St. Petersburg, who did some preliminary studies including prototyping of the flat magnetometer vessels and 3He polarimetry. The research was enabled by the loan of the mu-metal test shield by K. Kirch and B. Lauss, PSI, Switzerland. This work was supported by the DFG under the contract number HE 2308/10-1. We wish to thank M. J. Kraft, Wiesbaden, Germany providing support in the creation of the illustrations.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Andreas Kraft.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

AK designed and built the apparatus especially the UCPU, carried out the experiments and drafted the manuscript. DN constructed essential parts of the 3He/Cs magnetometer test facility. HCK, TL, DN, AP and YS participated in the experiments. HCK, MD and TL participated in manuscript preparation. WH supervised the described work. WH and AW both gave scientific advises and helped draft the manuscript. All authors read and approved the final manuscript.

Authors’ original submitted files for images

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (, which permits use, duplication, adaptation, distribution, and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Kraft, A., Koch, HC., Daum, M. et al. Development of a 3He magnetometer for a neutron electric dipole moment experiment. EPJ Techn Instrum 1, 8 (2014).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: