Erratum for the Research Article “Observation of the Wigner-Huntington transition to metallic hydrogen” by R. P. Dias and I. F. Silvera

See allHide authors and affiliations

Science  18 Aug 2017:
Vol. 357, Issue 6352, eaao5843
DOI: 10.1126/science.aao5843

Correction to Dias and Silvera, Science 355, 715 (2017)

In the Research Article “Observation of the Wigner-Huntington transition to metallic hydrogen,” Fig. 3 and the first and fourth cell in the top row of Table 1 were corrected to fit the Drude model with only the long wavelength points. The authors have also corrected the reflectance values to include multiple light reflections, which were omitted in the original calculation. The conclusions of the Research Article are not affected by the change in data used to fit the model.

The revised caption for Fig. 3B is: A least-squares fit of the uncorrected reflectance to the two longest-wavelength data points for metallic hydrogen. The fit was carried out for reflectance at two temperatures, yielding the plasma frequencies and scattering times.

In the Research Article, the principle evidence that hydrogen had transformed to the metallic phase was a measurement of the reflectance as a function of wavelength. Metallic hydrogen (MH) (or metallic deuterium) has been previously observed in the high-temperature liquid phase in which the evidence of metallization was provided by reflectance. In the shock experiments, reflectance was often measured at a single wavelength (1, 2), whereas several wavelengths were used in static measurements (3). In measurements of the Wigner-Huntington (WH) transition to MH in a diamond anvil cell, four wavelengths were used (404, 732, 642.6, and 1550 nm). The Research Article used data from Vohra, who studied the optical density (OD) of stressed natural diamonds (4). Highly stressed diamonds can attenuate light in the blue and ultraviolet region of the spectrum. Ruoff et al. proposed that this was due to the closing of a bandgap in diamond (5). However, in a new study of stressed diamond (6), Gamboa et al. measured the bandgap of diamond and found that it opens, rather than closes, with increasing pressure. They suggested that optical attenuation is likely due to defects that are created in the stressed diamonds. Defects in high-purity diamond have been shown to dramatically change optical properties of diamond (7). If defects are the explanation of attenuation of light, the reflectance will depend on the quality of the diamond and the state of stress. Natural diamonds can have impurities, defects, and inclusions, whereas synthetics (which were used in the Research Article) may have a lower level of impurities and no inclusions. Stress can vary, depending on the mounting of diamonds [brilliant cut with the diamond table-mounted on a flat tungsten carbide (WC) backing plate, used by Vohra, or Boehler-Almax conics, mounted in conic holes of a WC, which was used in the Research Article]. Thus, the OD of stressed diamonds determined by Vohra may not have been appropriate to use in the Research Article.

The corrected reflectance data in Fig. 3 and Table 1 must include multiple reflectance of light between the sample and the diamond table (8); there was a computational error in this part of the original analysis. Because the long wavelength data in the Research Article is unaffected by Gamboa et al.’s new data, and there is uncertainty about how to correct the short wavelength region, to the authors have disposed of these points and refit the reflectance to a Drude model by using only the long wavelength (642.6 and 1550 nm) points. The fit is shown in the revised Fig. 3B. In Table 1, the updated data show that at temperature (T) = 5 K, the plasma frequency is 33.2 ± 3.5 eV (compared with the original 32.5 ± 2.1), which yields a carrier density of 81 ± 17 × 1022 electrons/cm3 (compared with the original 77 ± 11 × 1022 electrons/cm3.

References and Notes

Navigate This Article