http://ieeexplore.ieee.org TOC Alert for Publication# 19 2013 May 16 62 6 C1 1395 71 62 6 C2 C2 143 62 6 1396 1396 23 62 6 1397 1399 101 62 6 1400 1406 800 62 6 1407 1412 623 $pm$ 0.03% at 95% confidence. Measurement results at 1780 and 2510 MHz are also reported.]]> 62 6 1413 1416 466 62 6 1417 1422 605 $10^{6}$ and in addition can be applied in wideband power measurements. Using our new method, the frequency response of cables and several other components is no longer a limiting factor for power measurements up to 1 MHz.]]> 62 6 1423 1427 431 62 6 1428 1435 1538 62 6 1436 1442 688 $muhbox{T}$ ) and 50–60 Hz, respectively. The system to generate a reference magnetic field is based on the Helmholtz coil formula. However, this well-known formula is precise only for dc magnetic fields. Therefore, we improve the formula for ac magnetic fields. The uncertainties of the reference magnetic field and magnetometer calibration are also improved. Validation is achieved by using two calibration systems based on different principles.]]> 62 6 1443 1449 435 $hbox{Mn}_{3}hbox{Ag}_{1 - x}hbox{Cu}_{x} hbox{N}$, around 23 $^{circ}hbox{C}$. The second-order term of the temperature coefficient $beta$ was found to show a rough correlation with the Cu content. For a Cu content $x = 0.4$ , the compound $hbox{Mn}_{3}hbox{Ag}_{0.6}hbox{Cu}_{0.4} hbox{N}$ shows the lowest temperature coefficients. The drift rate of the samples was also investigated. The drift rate was suppressed to about 1/30 by higher temperature annealing, and the lowest drift rate was 9.1 $(muOmega/Omega)$/year. We expect that this antiperovskite compound will be useful for precision resistors and improve the temperature coefficient of resistors by process optimization so that they can be used to produce standard resistors for example.]]> 62 6 1450 1453 712 $^{circ}hbox{C}$ –2000 $^{circ}hbox{C}$ with a 101-kPa Ar background gas lowers the rates of SiC decomposition and Si sublimation/diffusion and thus provides a means to control the rate of graphene layer development. We studied a wide range of growth temperatures and times and describe the resulting sample surface morphology changes and graphene layer structures. The experimental results are compared to a kinetic model based on two diffusion processes: Si vapor diffusion in the Ar-filled gap and atomic diffusion through graphitic surface layers.]]> 62 6 1454 1460 812 $hbox{SiO}_{2}$ substrate have been examined using scanning probe microscopy. Distinctive dot-like protrusions appearing nearly periodically on the edges of graphene were observed, and the density of the protrusions increased as the number of graphene layers increased. Imaging analysis revealed that the electrostatic properties of these protrusions are different from those of surrounding graphene. These findings are discussed and interpreted in terms of the local oxidation at the native graphene edges.]]> 62 6 1461 1466 718 62 6 1467 1472 1622 $mu hbox{s}$ (rise/fall time) impulse currents up to 50 kA peak value with a target uncertainty of 0.1% for the complete measurement system. The investigation method consists of using two current transducers based on different technologies that will detect simultaneously the same impulse current.]]> 62 6 1473 1478 548 62 6 1479 1485 689 62 6 1486 1490 447 $2 times 10^{-7}$ with a type A uncertainty of $3 times 10^{-7}$. A vacuum enclosure located in a dedicated laboratory with better thermal and vibrational isolation is ready to receive the present apparatus for further improvement.]]> 62 6 1491 1498 462 ${}^{28}$ Si material was published with an overall relative uncertainty of $3 times 10^{-8}$. The results were based on measurements of a very pure single crystal of highly enriched silicon-28. Where the result of crystal properties, molar mass, and lattice constant could be performed with satisfying uncertainty, the measurements on the spheres made of the same crystal were overshadowed by an unexpected effect; the surface of both silicon spheres were contaminated by small amounts of copper (Cu) and nickel (Ni). Although the contaminations were quantitatively measured and their impact into the different quantities was taken into account, the Avogadro group was worried about its negative influence. With the knowledge of various etching procedures and after due consideration, a special chemomechanical technique was applied to remove the metallic contaminations. Herein, we describe the decontamination process, the subsequent measurements, and their impact on the overall uncertainty of the Avogadro constant.]]> 62 6 1499 1505 1410 $(h)$. These effects were estimated, and corresponding corrections were applied to measurements of $h$ made at NRC between August and September of 2011. The effect due to the tilting of the beam required an estimated correction of $(0.238 pm 0.018) muhbox{J} cdot hbox{s}/(hbox{J} cdot hbox{s})$ , and the effect due to the tilting of the flat, which supports the beam, required an estimated correction of $(-0.636 pm 0.056) muhbox{J} cdot hbox{s}/(hbox{J} cdot hbox{s})$ . After the initial measurements were completed, we made several modifications to the balance which achieved a reduction of the two errors to negligible levels. A second set of measurements, made with the modified apparatus in May and June of 2012, shows very good agreement with the first measurements.]]> 62 6 1506 1511 535 $k$ with an electronic technique that measured the Johnson noise of a 100 $Omega$ resistor at the triple point of water and used a voltage waveform synthesized with a quantized voltage noise source (QVNS) as a reference. In this paper, we present measurements of $k$ using a 200 $Omega$ sense resistor and an appropriately modified QVNS circuit and waveform. Preliminary results show agreement with the previous value within the statistical uncertainty. An analysis is presented, where the largest source of uncertainty is identified, which is the frequency dependence in the constant term $a_{0}$ of the two-parameter fit.]]> 62 6 1512 1517 779 $k$, by comparing electrical and thermal noise power. A measurement with an integration period of 19 h and a bandwidth of 638 kHz results in a relative offset of $1 times 10^{-6}$, from the current Committee on Data for Science and Technology value of $k$, and a type A relative standard uncertainty of $17 times 10^{-6}$. Closely matched noise powers and transmission-line impedances were achieved, and consequently, the quadratic fitting parameters of the ratio spectrum show flat frequency responses with respect to the measurement bandwidth. This flat response produces a dramatically reduced systematic error compared to that of the National Institute of Standards and Technology measurement of $k$, in which the relative combined uncertainty was dominated by this error.]]> 62 6 1518 1523 779 62 6 1524 1530 1055 62 6 1537 1544 1059 $-$100-dBc/Hz flat phase noise standard at 10 MHz, which ensures traceability to the International System of Units. The flat phase noise standard signal is produced using a carrier combined with white noise. To ensure traceability, both the flat phase noise signal power and the power spectral density of the white noise are determined with a calibrated power meter and a noise standard, respectively. Sources of uncertainties in the system are discussed, and the overall expanded uncertainty is found to be 0.6 dB for a Fourier frequency of 1 Hz, 0.26 dB for 10 Hz, and 0.22 dB for 1, 10, and 100 kHz. We also report on the validity of a flat phase noise standard that consists of high-power white noise.]]> 62 6 1545 1549 434 62 6 1550 1555 936 $10^{-16}$-level fractional frequency stability of a “master laser” operated at 1.5 $muhbox{m}$ to a “slave laser” operated at 698 nm, using a femtosecond fiber comb as transfer oscillator. With the 698-nm laser, the $^{1}hbox{S}_{0} - {}^{3}hbox{P}_{0}$ clock transition of ${}^{87}hbox{Sr}$ was resolved with a Fourier-limited linewidth of 1.5 Hz (before: 10 Hz). Potential noise sources contributed by the frequency comb are discussed in detail.]]> 62 6 1556 1562 1078 62 6 1563 1567 671 62 6 1568 1573 848 62 6 1574 1580 354 $(k = 2)$ of the system is approximately 10 nV for voltages up to 20 V and 0.9 $muhbox{V/V}$ in the voltage range from 20 to 1000 V.]]> 62 6 1581 1586 1022 $10^{7}$ was achieved with the technique when measuring a 7 V RMS sine wave generated by a calibrator at 62.5 Hz.]]> 62 6 1587 1593 839 62 6 1594 1599 539 $mu{rm V/V}$. This performance is achieved because of the excellent differential linearity of the DVM used in the method. We have validated the overall accuracy of the method by directly comparing this approach with more traditional resistive divider techniques and have established that the expanded uncertainty is better than 0.2 $mu{rm V/V}$ up to 1000 V, limited only by the performance of the instruments used.]]> 62 6 1600 1607 295 62 6 1608 1614 890 62 6 1615 1620 320 $< 1 muhbox{V/V}$ above 10 Hz and $< 35 muhbox{V/V}$ at 1 Hz with lower uncertainties compared to the conventional method. The estimated overall uncertainty measured by using our system is 3.1 $muhbox{V/V}$ $(k = 1)$ at the frequency of 10 Hz and root-mean-square voltage of 3 V. Our measurement results above 10 Hz are in good agreement with the results obtained by the conventional method within a standard deviation of the mean. Our differential sampling measurement system using AC-PJVS is a useful tool for low-frequency ac voltage metrology.]]> 62 6 1621 1626 650 62 6 1627 1633 395 $-$110 dBc (carrier)] and amplitudes above the 1-V level. Fast and flexible code programming enables pulse-driven Josephson arrays to synthesize arbitrary waveforms. We have used the pulse-driven system to cancel the harmonic content from the binary array that can reach 1-V amplitudes. The quantum nature of the Josephson synthesizer ensures very stable amplitudes (better than $6 times 10^{-8}$) and no frequency drift. This performance, together with the ability to synthesize arbitrary waveforms, can make this synthesizer a relevant tool for characterizing precision electronics and ac voltage metrology.]]> 62 6 1634 1639 954 $-$6 parts in $10^{11}$ ( $-$0.6 nV), with a combined uncertainty of 7 parts in $10^{11}$ (0.7 nV).]]> 62 6 1640 1645 520 $1 times 10^{-6}$ ( $1sigma$ standard deviation) at 100 kHz. The temperature dependence of these devices has been found ten times smaller than those previously reported. In addition, MEMS-based references are theoretically independent of frequency beyond the mechanical resonant frequency. The frequency stability has been successfully tested between 40 and 300 kHz.]]> 62 6 1646 1651 820 $Omega$ shunt. The phase angle errors of resistive high current shunts and the inductive high current shunt have been measured. The level dependence in phase angle errors of three different types of high current shunts has been evaluated at currents from 1 to 100 A. The measurement results and standard uncertainties of the phase angle errors of high current shunts are reported.]]> 62 6 1652 1657 1151 62 6 1658 1663 1000 62 6 1664 1668 852 $5times 10^{-7}$ at a primary current ranging from 5 A to 2000 A.]]> 62 6 1669 1674 308 62 6 1675 1680 617 62 6 1681 1686 404 $mu{rm rad}$ and 300 nrad, respectively. The same digitizer has been used for the precise measure of the harmonic spurs up to 100 kHz. Instead, for the evaluation of the total harmonic distortion from 100 kHz up to 50 MHz, a wideband spectrum analyzer has been used. The algorithms employed allow the use of the synthesizer as a digital phase standard or as a polyharmonic arbitrary synthesizer with independent outputs.]]> 62 6 1687 1692 498 $mu{rm A/A}$ in magnitude and 0.8 $mu{rm rad}$ in phase for CT-under-test current ratios that do not differ by more than a factor of 5 from one of the ratios in the reference CT. The linearity of the bridge for input currents changing between 1% and 120% of nominal input current is better than 4 $mu{rm A/A}$ in magnitude and 0.5 $mu{rm rad}$ in phase.]]> 62 6 1693 1698 259 62 6 1699 1703 857 $muhbox{W/VA}$ at $k = 1$. Special high-accuracy current and voltage range extenders have been incorporated to extend the current and voltage ranges up to 200 A and 1200 V, respectively.]]> 62 6 1704 1709 302 62 6 1710 1715 430 62 6 1716 1722 937 $hbox{m}Omega$, 10 $hbox{m}Omega$, 1 $hbox{m}Omega$, and 100 $muOmega$, respectively. The measurement results of the three laboratories agree extremely well within the respective measurement uncertainties with the comparison reference value. Careful transport of the resistors was crucial for achieving this result. Still, some of the resistors showed steps in value at each transport which likely relates to the construction of the resistance elements of these resistors.]]> 62 6 1723 1728 520 $,times,$ 100 ${rm{M}}Omega$ network was developed to improve the traceability levels of dc resistance at 1 ${rm G}Omega$ level. Utilizing and revisiting this project, a Hamon 10$,times,$10 ${rm G}Omega$ network is developed to extend the capabilities of the Hamon scaling technique up to 100 ${rm G}Omega$. The novelty of the 10$,times,$10 ${rm G}Omega$ network is its improved guard system, and the improvement for INRIM is the extension of the range of the Hamon scaling method up to 100 ${rm G}Omega$ A description of this technique at INRIM, accurate construction details of the network and of its particular suitability for the modified Wheatstone bridge for high resistors calibration measurement method are given. The 1:100 transfer reliability test of the network gave satisfactory results. An uncertainty budget from 10 ${rm k}Omega$ to 100 ${rm G}Omega$ is worked out. Some measurement results on the network, both in parallel and in series configuration, are shown.]]> 62 6 1729 1735 710 $Hinfty$ controller for a two-terminal cryogenic current comparator is designed. To this end, a set of mathematical models covering the actual system is proposed. Simulation results compare the open- and closed-loop systems based on the proposed controller and the traditional integral control. According to these results, the new controller can significantly reduce the noise in the superconducting quantum interference device sensor.]]> 62 6 1736 1742 1065 $i = 2$ plateau at zero shield potential is not flat, and the linear frequency dependence of the QHR amounts to around $-0.15 ( muOmega/Omega)/hbox{kHz}$. Improved shield designs are proposed to reduce the frequency dependence.]]> 62 6 1743 1748 904 ${rm k}Omega$ to 100 ${rm T}Omega$. The resistance unit has been transferred from the quantum hall resistance (QHR) primary resistance standard to the resistance standard with a nominal value of 10 ${rm k}Omega$ by a cryogenic current comparator. For further transfer of the resistance unit from 10 ${rm k}Omega$ to 100 ${rm T}Omega$, Hamon transfer devices (Hamon networks) are used. For its complete realization five devices have been developed: 10–100–1000 ${rm k}Omega$, 1–10–100 ${rm M}Omega$, 0.1–1–10 ${rm G}Omega$ , 10–100–1000 ${rm G}Omega$, and 1–10–100 ${rm T}Omega$. The first two devices have a single insulation, and the next three have a double isolation. In the transfer devices with the double insulation, triax-type connectors are used. All transfer devices have individual temperature stabilization within ${pm}{rm 0.01}^{circ}{rm C}$. In the paper, the factors influencing device accuracy have been described. The insulation leakage is recognized, and the error caused by it is calculated. The guarding network for the new design of Hamon transfer devices is pres- nted. Voltage coefficients of resistance and settling times of resistors used in these devices have been determined and discussed. Uncertainties for the Hamon transfer devices ratios have been calculated. Calibration results have been presented. The developed system is currently tested in the Central Office of Measures in Poland.]]> 62 6 1749 1754 380 $hbox{k}Omega$ quantized Hall resistance (QHR) array device. This design realizes quantized resistance values that approach decade values with fewer Hall bars. In this design, the 10-$hbox{k}Omega$ QHR array device consists of only 16 Hall bars, and this number is about 16 times lower than that in the previous design, although the nominal value is identical to the previous one. The nominal value of this device shows a relative difference of only 0.034 $muOmega/Omega$, based on $R_{K - 90} (=25,812.807 Omega)$, from the exact value of 10 $hbox{k}Omega$. This fact might allow us to evaluate each Hall bar in the array device. The parameters of this 10-$hbox{k}Omega$ QHR array device were measured by a conventional QHR and cryogenic current comparator using the 100- $Omega$ standard resistor. According to the measurement, the value of the 10-$hbox{k}Omega$ QHR array device agrees with its nominal value within around one part in $10^{8}$. In the same manner, new combinations of Hall bars were designed for the 100-$Omega$ –1-$hbox{M}Omega$ range.]]> 62 6 1755 1759 884 ${rm M}Omega$ and up to and above 100 ${rm T}Omega$ is described. Whereas similar setups reported so far use only current null detection, the present system developed at VSL can use either voltage or current null detection. We compare the two null detection methods and find no significant difference between the methods over the resistance range between 10 ${rm M}Omega$ and 100 ${rm G}Omega$. Voltage null detection is more accurate for resistance values below 10 ${rm G}Omega$ at 10 V measurement voltage, whereas for resistance measurements above 100 ${rm G}Omega$ current null detection should be used. Careful automation and refinement of the measurement procedure lead to excellent results and uncertainties. For example, the measurement of the 1:100 ratio of the series and parallel connection of a 10$,times,$100 ${rm M}Omega$ Hamon device agrees within $(0.5pm 0.4)muOmega/Omega$ with the nominal Hamon ratio. The total measurement uncertainty at 1 ${rm G}Omega$ equals 1.7 $muOmega/Omega~(k=1)$, which has been confirmed in a recent international comparison.]]> 62 6 1760 1765 355 $LCR$ meter (usually in the 10-Hz range). A new impedance spectrometer based on multifrequency excitation and discrete Fourier transform analysis is presented here. The spectrometer defines the impedance under measurement as a two-port standard and measures impedance spectra covering 5 dec, reaching frequencies in the millihertz range. The instrument accuracy is verified by measurements on calibrated resistance standards. As an example of application, measurement results on pure-water samples and a comparison with measurements performed with an $LCR$ meter are reported.]]> 62 6 1766 1770 613 $10^{7}$ at kilohertz frequency, sufficient for calibration purposes, with the potential for further improvement.]]> 62 6 1771 1775 727 $Omega$ metal-foil resistor components. The current and frequency dependence of the resistance and phase angle have been characterized. Below a current of 5 mA, no current dependence is found within the uncertainty range, and no frequency dependence within the expanded measurement uncertainty range of around 2.5 $muOmega/Omega$ is found below a frequency of 2 kHz. From the phase-angle measurement results, the time constant of the resistor is estimated to be less than 3.7 ns.]]> 62 6 1776 1782 527 $hbox{m} Omega$ and 10 $Omega$ at frequencies from 40 Hz to 10 kHz. The possibility of using a digitizer to measure the complex voltage ratio is analyzed for both digitizer operating modes: the dual-channel mode with a simultaneous sampling and the single-channel mode with a sequential sampling in which an additional external multiplexer is used. The method of acquiring samples presented in this paper was used to investigate and compare the short-time stability and temperature influences on the metrological properties of the digitizer. This makes it possible to reduce the influence of the instability of the generator, being the signal source for the tested digitizer, on the research results. Moreover, presented in this paper are the frequency characteristics of the magnitude and phase of the complex voltage ratio, as well as the research results of the digitizer linearity and the phase measurement accuracy. All measurements discussed here were performed for both digitizer operating modes.]]> 62 6 1783 1788 871 62 6 1789 1794 783 ${mmb{mu}}{rm F}$, 100 ${mmb{mu}}{rm F}$, and 1 mF at frequencies of 100 Hz, 120 Hz and 1 kHz.]]> 62 6 1795 1800 366 62 6 1801 1806 836 $hbox{N}_{2}$ buffer gas, and the beam diameter were optimized. As a result, the average uncertainty in obtaining the Rabi frequency, whose accuracy is important for the measurement of microwave power, was reduced several-fold. We demonstrated the measurement of microwave power using the optimized atomic candle signal in our previous work. In this paper, we discuss the optimization process in detail and the utility of optimizing the atomic candle signal for the precise measurement of microwave power.]]> 62 6 1807 1813 390 62 6 1814 1820 695 62 6 1821 1827 831 62 6 1828 1833 645 62 6 1834 1839 769 62 6 1840 1846 281 62 6 1847 1852 634 $Omega$, which has a bifilar heater pattern to reduce the thermoelectric effects, the nominal current is 200 mA. The observed current level dependence of the ac–dc current transfer differences from 50 to 200 mA for the 5-$Omega$ MJTC was within 1 $muhbox{V/V}$ at frequency from 1 to 100 kHz. Compared with a conventional thermal converter, the wide range of available input current and small current level dependence of the new MJTCs may reduce uncertainties of ac–dc current transfer standards.]]> 62 6 1853 1858 476 ${-}{45%}$ is obtained upon application of elevated temperatures and a gas mixture containing ${rm NO}_{2}$ at a concentration of 10 parts per billion (10 ppb). The sensitivity response increased linearly with ${rm NO}_{2}$ concentration, reaching ${-}{60%}$ at a concentration of 250 ppb, followed by saturation at 1 part per million (ppm) level.]]> 62 6 1859 1864 3938 62 6 1865 1865 33 62 6 1866 1866 22 62 6 1867 1867 77 62 6 1868 1868 1155 62 6 C3 C3 103 62 6 C4 C4 133