Proposed new determination of the gravitational constantGand tests of Newtonian gravitation

Abstract
The first "constant of nature" to be identified, Newton's constant of universal gravitation G, is presently the least accurately known. The currently accepted value (6.672 59±0.000 85) × 1011 m3 kg1 s2 has an uncertainty of 128 parts per million (ppm), whereas most other fundamental constants are known to less than 1 ppm. Moreover, the inverse-square law and the equivalence principle are not well validated at distances of the order of meters. We propose measurements within an orbiting satellite which would improve the accuracy of G by two orders of magnitude and also place new upper limits on the field-strength parameter α of any Yukawa-type force, assuming a null result. Preliminary analysis indicates that a test of the time variation of G may also be possible. Our proposed tests would place new limits on α=α5(q5μ)1(q5μ)2 for characteristic lengths Λ between 30 cm and 30 m and for Λ>1000 km. In terms of the mass mb of a vector boson presumed to mediate such a Yukawa-type force, the proposed experiment would place new limits on α for 7×109 eV<mbc2<7×107 eV and for mbc2<2×1013  eV. Two distinct tests of the inverse-square law, one employing interactions at intermediate distances and having a peak sensitivity if Λ is a few meters (i.e., mbc2107 eV), and the other employing interactions at longer distances and having a peak sensitivity for ΛREarth (mbc23×1014eV), would both place limits of 105 to 106 on α. These interactions also provide tests of the equivalence principle (Eötvös' experiment). The intermediate-distance interaction would test the equivalence principle to 5 parts in