= 1962 rocket discovery of cosmic X-ray sources
In 1962, <Riccardo Giacconi> and colleagues flew three <Geiger counters> to search for lunar <X-ray fluorescence>. Different <mica> thicknesses tested penetration; lampblack blocked <ultraviolet>. Each counter sat in an <anticoincidence detector> using a <scintillation counter> to reject <cosmic rays>. There was no mechanical <collimator>. Rotation scanned the sky; optical sensors supplied orientation.<Evidence for x Rays From Sources Outside the Solar System by Riccardo Giacconi and others>[pp. 439–440, apparatus and readout]{ref}
Counts were sorted by <azimuth>; a faulty counter and tumbling descent data were excluded. A directional peak lay above a diffuse component. Window, altitude and angular responses tested <X-rays> against <charged particles>; the <Moon>'s nondetection checked optical leakage. The authors argued for an origin outside the <Solar system>, while residual <cosmic ray> background was not directly measured and source size, position and <spectrum> remained assumption-dependent.<Evidence for x Rays From Sources Outside the Solar System by Riccardo Giacconi and others>[pp. 440–443, Figs. 1–4 and concluding limitations]{ref}
\Image[/x-ray-rocket-scan-response.png]
{title=Rotation encodes direction; a teaching collimator narrows an angular response}
{description=Functional schematic and invented response curves. The 1962 rocket had no mechanical collimator; the separate collimator panel illustrates a possible angular filter. Curves are model expectations, not flight data.}
A simple original <forward problem> makes the distinction precise. Let $\phi$ be the scan <azimuth>, let $g(\phi)$ be a dimensionless directional response with maximum one, let $F$ be a source's band-integrated <photon> arrival rate per unit area, let $A_0$ be its on-axis <X-ray effective area>, and let $b$ be a constant accepted background rate. The expected rate is
$$
r(\phi)=F A_0 g(\phi)+b.
$$
For a broad illustrative response $g(\phi)=\max(0,\cos\phi)$, the full width at half maximum is $120^\circ$: $\cos\phi=1/2$ at $\phi=\pm60^\circ$. For a one-dimensional ideal <collimator> model $g(\phi)=\max(0,1-|\phi|/\theta_c)$, the full width at half maximum is $\theta_c$. Choosing $\theta_c=20^\circ$ narrows the response sixfold without focusing any additional <photons>. These are teaching response functions, not reconstructions of the rocket's window transmission or geometry. A narrower scan peak can reflect the instrument response even when the source is unresolved.
Back to article page
