New Test of Local Lorentz Invariance Using a ${}^{21}\mathrm{Ne}\mathrm{\text{−}}\mathrm{Rb}\mathrm{\text{−}}\mathbf{K}$ Comagnetometer

We develop a new comagnetometer using ${}^{21}\mathrm{Ne}$ atoms with nuclear spin $I=3/2$ and Rb atoms polarized by spin exchange with K atoms to search for tensor interactions that violate local Lorentz invariance. We frequently reverse the orientation of the experiment and search for signals at the first and second harmonics of the sidereal frequency. We constrain 4 of the 5 spatial Lorentz-violating coefficients $c_{jk}^n$ that parametrize anisotropy of the maximum attainable velocity of a neutron at a level of ${10}^{-29}$, improving previous limits by 2 to 4 orders of magnitude and placing the most stringent constraint on deviations from local Lorentz invariance.

Figure 1

The experimental apparatus is rotated around the local vertical. ${}^{21}\mathrm{Ne}$ spins are polarized down along $-\widehat{z}$ and the probe beam is directed horizontally along $-\widehat{x}$.

Figure 2

Magnetometer signal as a function of time. During the first 300 s the sensitive axis is alternated between North and South directions. After the transient from rotation decays, the background optical rotation is measured (dashed box). Then the $B_z$ field is set to the compensation point and the actual signal is measured (solid box). Rotations are paused after 13 reversals for calibration. In the following sequence the sensitive axis is alternated between East and West.

Figure 3

Long-term measurements of the N-S and E-W modulation amplitudes with a fit including first and second harmonics of the sidereal frequency.

Figure 4

Daily fits for the $c$ coefficients defined in Eq. (6). The final statistical uncertainty is increased to account for reduced ${\chi }^2=2$ to 4.