Euclid preparation: V. Predicted yield of redshift 7

may be selected from Euclid $OYJH$ photometry alone, but selection over the

redshift interval $7

expected in 2024. It will be possible to determine the bright-end slope of the

QLF, $7

We provide predictions of the yield of $7<z<9$ quasars from the Euclid wide

survey, updating the calculation presented in the Euclid Red Book (Laureijs et

al. 2011) in several ways. We account for revisions to the Euclid near-infrared

filter wavelengths; we adopt steeper rates of decline of the quasar luminosity

function (QLF; $Phi$) with redshift, $Phipropto10^{k(z-6)}$, $k=-0.72$,

consistent with Jiang et al. (2016), and a further steeper rate of decline,

$k=-0.92$; we use better models of the contaminating populations (MLT dwarfs

and compact early-type galaxies); and we use an improved Bayesian selection

method, compared to the colour cuts used for the Red Book calculation, allowing

the identification of fainter quasars, down to $J_{AB}sim23$. Quasars at $z>8$

may be selected from Euclid $OYJH$ photometry alone, but selection over the

redshift interval $7<z<8$ is greatly improved by the addition of $z$-band data

from, e.g., Pan-STARRS and LSST. We calculate predicted quasar yields for the

assumed values of the rate of decline of the QLF beyond $z=6$. For the case

that the decline of the QLF accelerates beyond $z=6$, with $k=-0.92$, Euclid

should nevertheless find over 100 quasars with $7.0<z<7.5$, and $sim25$

quasars beyond the current record of $z=7.5$, including $sim8$ beyond $z=8.0$.

The first Euclid quasars at $z>7.5$ should be found in the DR1 data release,

expected in 2024. It will be possible to determine the bright-end slope of the

QLF, $7<z<8$, $M_{1450}<-25$, using 8m class telescopes to confirm candidates,

but follow-up with JWST or E-ELT will be required to measure the faint-end

slope. Contamination of the candidate lists is predicted to be modest even at

$J_{AB}sim23$. The precision with which $k$ can be determined over $7<z<8$

depends on the value of $k$, but assuming $k=-0.72$ it can be measured to a 1

sigma uncertainty of 0.07.

http://arxiv.org/icons/sfx.gif