Review report · by Lena Hartmann
2026-09-30
Candidate issues, not verified.
Checked the conversion from integrated optical density to atom number against the cross-section and the pixel size, and the fit residuals by eye. Nothing to raise.
Example team · Fictional people and results, shown on the real Library screens. The two papers and the checks of them are real. What am I looking at?
Shared in Quantum Science in Space · Result
Published 2026-09-29 · Last reviewed 2026-09-30 · Reviewed · 0 candidate issues
Only the members of Quantum Science in Space can see it. It does not appear in the public Library, in search for anyone else, or in another claim's “cited by”.
After 1.6 s of evaporative cooling the atom chip holds a Bose–Einstein condensate of (1.19 ± 0.06) × 10⁵ ⁸⁷Rb atoms, with a condensed fraction of 0.74 ± 0.02, measured by absorption imaging 20 ms after release.
Holds when: Ground test setup; imaging calibrated to 4.0 µm per pixel, to 5 %; the atom number assumes resonant imaging with the ideal cross-section 3λ²/2π.
Evidence 1 · Reasoning · markdown
The image is summed along one axis and fitted with the two-part profile of a partly condensed cloud: an inverted parabola for the condensate, a Gaussian for the thermal atoms. Each part's area, divided by the resonant cross-section, is an atom number. The condensed fraction follows from the two.
Evidence 2 · Code · python
The fit, as it ran on the camera frame.
import numpy as npfrom scipy.optimize import curve_fitPIXEL = 4.0e-6 # m per pixel, object planeSIGMA0 = 3 * 780.241e-9**2 / (2*np.pi) # resonant cross-section, m^2def bimodal(x, a_bec, r_tf, a_th, s_th, x0, off): u = np.clip(1 - ((x - x0) / r_tf)**2, 0, None) return a_bec*u**2 + a_th*np.exp(-(x - x0)**2 / (2*s_th**2)) + offod = np.load("run-0929-17/od_tof20ms.npy") # optical density imageprofile = od.sum(axis=0) # integrate along zx = np.arange(profile.size, dtype=float)p, cov = curve_fit(bimodal, x, profile, p0=[60, 18, 10, 28, x.size/2, 0])err = np.sqrt(np.diag(cov))def atoms(area): # integrated OD -> atom number return area * PIXEL**2 / SIGMA0n_bec = atoms(p[0] * p[1] * 16/15) # ∫(1-u²)² du = 16/15 r_tfn_th = atoms(p[2] * p[3] * np.sqrt(2*np.pi))print(f"N_BEC = {n_bec:.3e} N_th = {n_th:.3e} " f"fraction = {n_bec/(n_bec+n_th):.3f}")print(f"relative uncertainty of N_BEC: fit {np.hypot(err[0]/p[0], err[1]/p[1]):.4f}, " f"with 5 % calibration {np.hypot(np.hypot(err[0]/p[0], err[1]/p[1]), 0.05):.3f}")Evidence 3 · Parameters · yaml
evaporation: 1.6 s, final RF 3.62 MHztime_of_flight: 20 msimaging: resonant, σ⁺, 50 µspixel_size: 4.0 µm (object plane)camera_frame: run-0929-17/od_tof20ms.npyEvidence 4 · Output
What the fit printed.
fit: a_bec = 113.22 ± 0.29, r_tf = 17.97 ± 0.03 px, a_th = 11.56 ± 0.43, s_th = 26.49 ± 1.61 px N_BEC = 1.194e+05 N_th = 4.225e+04 fraction = 0.739 relative uncertainty of N_BEC: fit 0.0030, with 5 % calibration 0.050
Review report · by Lena Hartmann
2026-09-30
Candidate issues, not verified.
Checked the conversion from integrated optical density to atom number against the cross-section and the pixel size, and the fit residuals by eye. Nothing to raise.
Jonas Weber (2026). A Bose–Einstein condensate of 1.2 × 10⁵ rubidium atoms on the chip. Berry Science Library. 2c629f0c-c247-58bc-a757-6ac5a9d9fb37