Past exam of the mathematics course of the University of Cambridge / 2018 / iii / Paper 338 (Optical and Infrared Astronomical Telescopes and Instruments)

Use the gnomonic projection associated with an undistorted focal plane. Write $\Delta =\alpha -A$ and rotate the celestial Cartesian coordinate system so that the pointing meridian has longitude zero. The stellar direction and an orthonormal basis adapted to the optical axis areHere $\mathbf{e}$ points east and $\mathbf{n}$ points north. Intersect the ray $t\mathbf{s}$ with the plane $\mathbf{x}\cdot \mathbf{b}=f$. It gives $t=f/(\mathbf{s}\cdot \mathbf{b})$, so the detector coordinates areThese equations also avoid spurious singularities caused by writing individual tangents or cotangents.

To put this gnomonic projection in the desired form, let $\rho =({\cos }^2\delta {\cos }^2\Delta +{\sin }^2\delta {)}^{1/2}$ and choose $q$ locally byThus $\cot q=\cot \delta \cos \Delta$. The denominator becomes $\rho \cos (q-D)$, the numerator for $\eta$ becomes $\rho \sin (q-D)$, and $\cos \delta \sin \Delta =\rho \cos q\tan \Delta$. ConsequentlyUse the local meridian chart $\cos \Delta >0$, with $q\in [-\pi /2,\pi /2]$ chosen continuously near $D$ and $q=D$ at the image centre. A visible gnomonic projection requires $\mathbf{s}\cdot \mathbf{b}>0$, but this front-hemisphere condition alone does not select that meridian chart. A narrow field near a celestial pole can cross the opposite meridian; for such fields use the Cartesian expressions above and distinguish the oriented great circle parameter from ordinary declination. In particular, the small-field limit is $\xi \simeq f\cos D(\alpha -A)$ and $\eta \simeq f(\delta -D)$, with angles in radians. The factor $\cos D$ is the shrinking angular distance per unit right ascension near the pole.

The great circle containing the pointing and celestial pole is the intersection of the unit sphere with the meridian plane $y=0$. The orthogonal projection of $\mathbf{s}$ onto that plane, normalized to unit length, isWithin the local meridian chart $\cos \Delta >0$, its declination is therefore $q$. Outside that chart, the same construction gives an oriented great circle parameter; the foot can lie on the opposite right-ascension half of the meridian, and $q$ need not be its ordinary declination.

To verify the right angle geometrically, the tangent to the great circle from $L$ toward $S$ is proportional to $\mathbf{s}-(\mathbf{s}\cdot \mathbf{\ell })\mathbf{\ell }$. Since $\mathbf{s}\cdot \mathbf{\ell }=\rho$, that tangent has only a $y$ component. The tangent from $L$ toward the pole, $\mathbf{P}-(\mathbf{P}\cdot \mathbf{\ell })\mathbf{\ell }$, lies in $y=0$. Their orthogonality proves the claimed spherical right angle. Thus within the local meridian chart, $q$ is the declination of the perpendicular foot on the pointing meridian.

There are antipodal perpendicular feet on the complete great circle. On the local meridian chart, the telescope field selects the foot near the pointing; the cotangent equation alone determines $q$ only modulo $\pi$. At a coincident foot or pole, the corresponding angle is understood by continuity rather than by a nonzero tangent vector.

A Newtonian reflector uses a concave primary generated by a parabola, followed by a flat secondary inclined at $45^{\circ }$ to the optical axis. Parallel marginal rays converge toward the primary focus; the secondary intercepts that converging beam and folds it sideways to an accessible focal plane.

The Newtonian reflector needs only one powered mirror, so it is relatively simple to fabricate and inexpensive. A parabolic primary has no on-axis spherical aberration. Its principal wide-field limitation is coma, accompanied by field curvature and off-axis astigmatism. The diagonal and its supports obstruct the entrance pupil and introduce diffraction; the tube is long compared with a folded two-powered-mirror design, and heavy instruments at the side focus can be awkward to support. These comments also answer the unheaded design-comparison clause on the next PDF page.

In a classical Cassegrain reflector, a concave parabolic primary sends light toward an intermediate focus, but a convex hyperbolic secondary intercepts it before it reaches that focus. The secondary returns the beam through a central hole in the primary to a focal plane behind the primary. The two relevant foci of the secondary’s hyperbola are the primary’s would-be focus and the final focus.

The secondary magnifies the effective focal length, giving a compact tube and convenient rear-mounted instruments. The classical conics correct on-axis spherical aberration, but not the off-axis coma, astigmatism or field curvature. There is secondary obscuration, diffraction from its supports, and sensitivity to mirror alignment. A long effective focal length is useful for a small angular image scale per detector pixel, but yields a small field for a fixed detector size.

The Ritchey–Chrétien reflector follows the same folded path as a Cassegrain reflector, but both the concave primary and convex secondary are hyperbolic. Their conic constants and separation are chosen to cancel third-order spherical aberration and coma. The focal plane is behind the perforated primary.

The absence of third-order coma gives a much more useful wide field than a classical Cassegrain reflector, making this design attractive for research imaging. It retains astigmatism and field curvature, so a large flat detector generally needs corrective optics; higher-order optical aberrations are not all removed. Both aspheric mirrors are more demanding to manufacture and align, and the usual secondary obstruction remains. The sketch illustrates the beam routing and mirror types; its conics are not an optimized aplanatic prescription.

A classical Schmidt telescope combines a concave spherical primary with a thin aspheric corrector plate at its centre of curvature. The plate is also near the aperture stop. It supplies just the extra ray bending needed to cancel the mirror’s spherical aberration; rays then return toward an internal focus roughly halfway between the plate and primary. The native best-focus surface is curved, as indicated in red.

The spherical primary is easy to manufacture, and the stop at its centre of curvature gives a nearly symmetric optical geometry that suppresses coma and astigmatism over a large field. Fast Schmidt cameras are consequently excellent survey instruments. The costs are a long enclosed optical assembly, a large precision corrector, an internal camera that obstructs the beam, and a curved focal surface. A flat detector needs field flattening. The transmissive corrector introduces wavelength-dependent effects and transmission losses, unlike an entirely reflecting system.

A Gregorian telescope has a concave parabolic primary and a concave ellipsoidal secondary. Unlike the Cassegrain reflector, the secondary is beyond the primary focus: the rays cross that focus before reaching it. The two foci of the secondary’s ellipse are the intermediate focus and final rear focus, so the reflected rays return through the primary hole to the detector.

The real intermediate focus permits a field stop to reject unwanted light, and the two image inversions give an erect image. The rear focus is convenient and the effective focal length can be large. The secondary beyond prime focus makes the tube longer than the corresponding Cassegrain reflector, often with a larger secondary obstruction. The classical design retains off-axis coma, astigmatism and field curvature, and needs careful alignment.

An ordinary passive primary must keep its optical figure while the direction of gravity relative to the mirror changes as the telescope tracks. A thin unsupported disk bends enough to introduce wavefront errors, degrading the point spread function. Increasing thickness strongly raises its resistance to bending.

For an isotropic elastic plate of thickness $t$, Young's modulus $E_Y$ and Poisson's ratio $\nu$, the flexural rigidity is $\mathcal{B}=E_Y{t}^3/[12(1-{\nu }^2)]$. Under its own weight the load per area scales as $\rho g_{0}t$. With support spans of order the diameter $D$, plate bending therefore givesNumerical coefficients depend on the support and boundary conditions. A larger $t$ greatly reduces gravitational sag, even though it adds weight. Near normal incidence a mirror displacement $h$ changes optical path length by approximately $2h$, so surface errors must be much smaller than the observing wavelength. Thickness provides passive stiffness needed to preserve the mirror figure.

The difficulty is an unfavorable engineering scaling, rather than a fundamental diameter cutoff. For the same passive support concept and fixed allowed sag, the elastic plate estimate gives $t\propto D^2$. Consequentlywhere $\kappa$ is the thermal diffusivity. Thus increasing diameter makes the casting, annealing, transport and support progressively harder, while its long thermal diffusion time prevents the mirror following the changing ambient temperature. Temperature gradients deform the surface; a warm mirror also creates local turbulence that worsens astronomical seeing. A thinner solid disk has a shorter thermal response but is too flexible without better support.

The quoted four-metre scale is a practical rule of thumb for the simple thick passive design, not an absolute physical impossibility. The official BTA description records a 6.05-metre, 42-tonne primary already operating well before this examination. Its existence rules out interpreting the stated number as a hard bound.

Three complementary approaches replace the simple thick solid disk:

“Operational” and “planned” here refer to the 2018 examination date. The next generation included the 39-metre Extremely Large Telescope with 798 segments, the thirty-metre Thirty Meter Telescope with 492 segments, and the Giant Magellan Telescope using seven 8.4-metre primary mirrors. These designs extend segmentation or lightweight casting with active control, rather than simply making a thicker slab. The specifications are documented by the 2017 ELT announcement, TMT optics design and GMT primary-mirror description. Adaptive optics is also needed to exploit the large apertures against atmospheric turbulence; it corrects faster disturbances than active optics.

The telescope forms the sky image at the entrance slit. A collimator makes the transmitted beam parallel, a reflection diffraction grating disperses it, and a camera’s optical lens focuses each wavelength to a different detector position.

With a uniformly illuminated slit, ideal imaging and negligible slit diffraction, each monochromatic slit image has an approximately top-hat intensity profile of physical width $p$. A small wavelength separation $\Delta \lambda$ displaces two such images by $q\Delta \lambda$, using the positive magnitude of grating dispersion. In the geometrical, slit-limited convention, resolution occurs when the displacement is of order the slit-image width. HenceHere $p/q$ is the slit’s apparent wavelength extent. A finite diffraction grating broadens the sharp edges; the top-hat sketch and the following identities presume slit-limited resolving power of a grating, rather than the regime where the grating’s own diffraction determines the line width. A precise resolution criterion for nonideal profiles can change order-unity factors.

Let the telescope and collimator have focal lengths $f_t,f_c$, and let $w=f_t\theta$ be the physical slit width. The collimated beam diameter in the dispersion direction is $A_c=f_{c}D/f_t$, so $w{A}_c/f_c=\theta D$.

The diffraction grating changes both the angular width and the beam diameter. At fixed wavelength, differentiating the grating equation gives $|d\beta /d\alpha |=\cos \alpha /\cos \beta$. Thus the anamorphic magnification of a grating givesIf $W$ is the illuminated surface length, its projected beam diameters are $A_c=W\cos \alpha$ and $A_{\mathrm{cam}}=W\cos \beta$. Multiplication cancels the anamorphic factors:ThereforeThis is a one-dimensional optical-invariant relation: a grating cannot independently magnify the slit and shrink the corresponding beam without compensating angular changes. The calculation uses local paraxial imaging about each instrument’s chief ray and an unclipped beam.

From the slit profile, the spectral resolving power obeys $R=\lambda q/p$. Substitute into the preceding optical-invariant relation:The slit-image width cancels. Although $q$ is a detector length per unit wavelength, $\lambda q$ is a detector length, so this expression for $L$ has units of length, as does $R\theta D$.

Use the signed-angle convention in which the reflection grating equation is $m\lambda =d(\sin \alpha +\sin \beta )$, with positive diffraction order $m$. For fixed incidence,The second identity is the local focal-plane scale, measured about the camera axis aligned with the central diffracted ray. Since $A_{\mathrm{cam}}=W\cos \beta$, the projected beam size cancels the cosine in the grating dispersion:Writing $N=W/d$ for the number of illuminated grooves also gives $L=mN\lambda$. This connects the geometrical instrument invariant to the phase span of the illuminated diffraction grating.

Substitute the grating equation into the preceding result:Moving across the illuminated grating by one groove spacing changes the incident-plus-outgoing optical path length by $d(\sin \alpha +\sin \beta )=m\lambda$. Across length $W$, the total path difference is therefore $L$. Thus $L$ is the optical path difference between contributions from the two ends of the illuminated grating.

The number of coherent phase cycles across it is $L/\lambda =mN$, which is the intrinsic diffraction-limited spectral resolving power of the diffraction grating under the usual first-minimum criterion. In the slit-limited regime, $R\theta D=L$ also states how much sky angle can be accepted at a given resolution and aperture. Holding slit angle and resolution fixed while increasing telescope diameter requires a larger optical path span. The geometry gives $|L|\le 2W$; large incidence and diffraction angles increase resolution per unit grating length, though grazing beams become impractical.

The slit equations do not imply unlimited resolution when $\theta$ tends to zero. Once $\theta D$ is comparable to $\lambda$, finite-aperture diffraction matters and the actual resolving power is bounded by about $mN$.

An integral field spectrograph obtains spectra throughout a two-dimensional field instead of along one slit alone. Its reduced spectral data cube is $I(x,y,\lambda )$: two coordinates locate a spatial sampling element and the third labels wavelength. One slice at fixed wavelength is an image; one column at fixed spatial position is an optical spectrum. The third axis is spectral, not a third spatial direction.

Six ways to obtain such a cube illustrate the distinction between field reformatting and scanning:

The first four are simultaneous spatially multiplexed grating arrangements. The last two deliver equivalent cube coordinates by scanning; they are imaging spectrometers rather than simultaneous grating integral-field units, and variability during the scan can corrupt the cube. Detector packing, sampling, calibration and throughput determine the practical tradeoffs.

In angular differential imaging, the instrument is operated in pupil tracking: its pupil and associated quasi-static speckle patterns remain nearly fixed on the detector, while the sky rotates with the parallactic angle. A reference stellar point spread function is formed from other exposures, preferably excluding frames where a companion remains at almost the same location. Subtract the reference from each exposure, then derotate the residuals into the sky frame and combine them. A companion adds coherently after derotation; the stellar residuals are reduced. This is the technique described in the original angular-differential-imaging analysis. At angular separation $\rho$, the approximate displacement is $\rho \Delta \psi$, so useful diversity normally requires $\rho |\Delta \psi |\gtrsim \lambda /D$.

In simultaneous spectral differential imaging, acquire nearby spectral-band images simultaneously, for example with a beam splitter and filters or an integral field spectrograph. Stellar speckle patterns approximately move radially in proportion to wavelength. Rescale each image by ${\lambda }_0/\lambda$ and normalize the stellar flux before subtracting bands. The speckles then approximately align, while a companion at a fixed sky position moves in the rescaled coordinates. A companion absorption band can also distinguish its spectrum from the star: methane bands are useful for cool companions, as in the TRIDENT instrument description, but methane is not a universal companion property. Positional diversity scales as $\rho |\Delta \lambda |/\lambda$.

Both methods reduce structured stellar residuals, rather than removing all noise. In angular differential imaging, changing atmospheric turbulence, imperfect adaptive optics, flexure and thermal drift change the point spread function between frames. The reference then fails to represent the instantaneous stellar field. Small sky rotation makes close companions contaminate their own reference, causing differential-imaging self-subtraction. Extended disks are especially susceptible; subtraction can alter shape as well as total flux. More images help independent photon shot noise, but do not necessarily average away correlated residuals.

Simultaneous spectral differential imaging avoids the time delay, but different channels have non-common-path wavefront errors. Chromatic optical aberrations, wavelength-dependent amplitude errors and out-of-pupil propagation prevent a perfect radial rescaling of speckle patterns. Filter throughput, detector calibration, image registration and atmospheric dispersion also leave subtraction residuals. Nearby bands align the stellar field better but give less positional diversity; wider separation gives more displacement but larger chromatic mismatch. A smooth-spectrum companion, or one too close for appreciable rescaled displacement, can undergo severe differential-imaging self-subtraction.

Artificial-companion injection through the complete processing pipeline and forward modelling can calibrate lost throughput and photometric or astrometric biases. They do not guarantee that every correlated residual is a real source. Independent epochs or spectral evidence remain valuable.

First include the mean-gain calculation preceding the numbered clauses. At each electron-multiplying CCD stage one electron contributes an expected $1+p$ electrons to the next stage. If $N_j$ is the charge after stage $j$, conditional expectation gives $\mathbb{E}[N_{j+1}\mid N_j]=(1+p)N_j$. Iteration from $n$ input electrons yieldsFor small $p$, $g\simeq e^{rp}$ if corrections of order $r{p}^2$ are negligible.

In the high-gain approximation the output from one input electron has an exponential distribution of scale $g$. Independent input electrons produce independent cascades, and their charges add. By convolution of independent random variables, assuming the formula for $n$ inputs,Together with the single-electron base case this provesIt is a gamma distribution with shape $n$ and scale $g$, so it is normalized and has expected value $ng$ and variance $n{g}^2$. The formula is a continuous approximation to discrete output charge, not an exact integer-valued branching distribution; for $n=0$ there is a point mass at zero instead.

Let $N$ be the input photoelectron count and $\mu =\eta Q$ its mean, with quantum efficiency $\eta$ and mean incident photon count $Q$. Independent arrivals give a Poisson distribution, so $\mathrm{Var}N=\mathbb{E}N=\mu$. Conditional on $N$, the preceding gamma distribution gives $\mathbb{E}[X\mid N]=gN$ and $\mathrm{Var}(X\mid N)=g^{2}N$. The law of total variance therefore yieldsOne contribution is the ordinary photon shot noise; the other is multiplication excess noise. Thus, neglecting read noise and backgrounds,This is an excess-noise factor $\sqrt{2}$ in analog operation: actual photon conversion efficiency has not been halved. At low occupancy, thresholding each pixel as zero or one event can avoid most multiplication noise, but high arrival rates produce coincident events that cannot be counted separately. That is why the high-rate result concerns charge measurement rather than ideal binary photon counting.

Treat $Q$ and $B$ as mean detected counts, or as photon counts with unit quantum efficiency. Let the independent patch measurements be $X\sim \mathrm{Poisson}(Q+B)$ and $Y\sim \mathrm{Poisson}(fB)$. For the specified unweighted background subtraction $\hat{Q}=X-Y$,The first error is a fixed bias of an estimator; the last expression is the sum of independent photon shot noise variances. Since $Q,B$ are already patch totals, no extra factor of the pixel count $n$ is needed, and read noise is neglected.

To obtain the systematic-error ceiling requested in the following clause, define the accuracy measure $Z$ using total root-mean-square error relative to the true source count. By the bias-variance decomposition of mean squared error,For $f\simeq 1$, the shot-noise term is approximately $Q+2B$, but the mismatch term must be retained.

There is a terminology qualification: the usual variance-based signal-to-noise ratio in photon counting is $Q/\sqrt{Q+(1+f)B}$ and does not include a fixed bias as noise. The printed next-part limit requires the root-mean-square accuracy convention above. A deterministic background mismatch contributes $b^2$ to mean squared error, not to the statistical variance.

Put $Q=R_{Q}T$ and $B=R_{B}T$ in the root-mean-square accuracy measure:For a fixed uncorrected $f\ne 1$ and $R_B>0$, the squared bias eventually dominates the shot-noise terms. ThereforeThe PDF omits the absolute value. Its expression is the positive accuracy ratio only if $f<1$; for $f>1$ the residual background changes sign, but an error magnitude and a signal-to-noise ratio in photon counting remain nonnegative. The printed signed formula can instead be read as source divided by signed bias.

For exact background matching $f=1$, there is no systematic-error signal-to-noise ceiling: $Z=R_Q\sqrt{T/(R_Q+2R_B)}$ grows without bound in this idealized model. With $R_B=0$ the same conclusion holds. Limits in which $f$ itself approaches one with exposure time are different from the fixed-mismatch limit used here.

Increasing exposure reduces fractional photon shot noise as $T^{-1/2}$, but a fixed background mismatch grows in proportion to the signal. The fractional systematic error is $|1-f|R_B/R_Q$, independent of exposure. The systematic-error signal-to-noise ceiling can therefore be poor for a faint source in a bright sky even when random fluctuations are tiny.

Better background subtraction requires matching sky location and time, correcting detector response, dithering or modelling spatial background variations. If $f$ were known exactly, one could instead use $X-Y/f$, which is unbiased and has variance $Q+B+B/f$. The ceiling arises from an uncorrected or unknown mismatch in the prescribed subtraction, not an unavoidable property of measuring two patches.

To reach an accuracy ratio $Z_{\ast }$ at all, a necessary condition in the fixed-mismatch model isEquality allows only an asymptotic approach; finite exposure adds positive random error. More exposure improves precision, but only better calibration removes the fixed background bias.

An H2RG detector is a hybrid infrared detector: a photosensitive array is bonded, pixel by pixel through indium contacts, to a separate silicon readout circuit. In infrared use the absorber is normally mercury cadmium telluride, whose composition sets the absorption cutoff. The H2RG readout architecture can also be used with other absorbing materials, so the readout name does not fix the detector’s wavelength response. The manufacturer’s H2RG specification gives a $2048\times 2048$ grid and $18\mu \mathrm{m}$ pixel pitch, together with reference-pixel and guide-window capabilities.

Each illuminated pixel contains a photodiode connected to its own readout node. Absorbed photons generate carriers; their accumulated charge changes the node voltage, with $|\Delta V|=|\Delta Q|/C$ for node capacitance $C$. A reset establishes the baseline; row and column addressing and multiplexed output amplifiers sample the pixel voltage. This is not the serial transfer of charge packets through adjacent pixels used by a charge-coupled device.

Nondestructive detector readout allows repeated measurements during an integration. Correlated double sampling removes the reset baseline, Fowler sampling differences groups of reads at the beginning and end, and up-the-ramp sampling estimates the signal from its slope and can flag cosmic-ray steps or saturation. The “R” denotes detector reference pixels, which monitor electronic offsets; “G” denotes a programmable guide window that can be read rapidly between full-array operations. Reference pixels do not measure sky background. Cooling suppresses dark current and thermal background. Read noise, nonlinearity, persistence and interpixel coupling still require calibration.