1: \documentclass[12pt,preprint]{aastex}
2: %%\documentclass[manuscript]{aastex}
3: %% \documentclass[preprint2]{aastex}
4: %% \documentclass[preprint2,longabstract]{aastex}
5: %
6: \newcommand \ergs {\ifmmode {\rm ergs\,s}^{-1} \else ergs s$^{-1}$\fi}
7: \newcommand \ergcms {\ifmmode {\rm ergs\,cm}^{-2}\,{\rm s}^{-1}
8: \else ergs\,cm$^{-2}$\,s$^{-1}$\fi}
9: %
10: % Solar units
11: %
12: \def\Msun{\ifmmode M_{\odot} \else $M_{\odot}$\fi}
13: \def\Lsun{\ifmmode L_{\odot} \else $L_{\odot}$\fi}
14:
15: \slugcomment{Accepted for publication in ApJ main journal}
16:
17: \shorttitle{Star formation in the hosts of high-z QSOs}
18: \shortauthors{Lutz et al.}
19:
20: \begin{document}
21:
22: \title{Star formation in the hosts of high-z QSOs: Evidence from Spitzer
23: PAH detections}
24:
25: \author{D. Lutz\altaffilmark{1},
26: E. Sturm\altaffilmark{1}, L.J. Tacconi\altaffilmark{1}, E. Valiante\altaffilmark
27: {1}, M. Schweitzer\altaffilmark{1}
28: H. Netzer\altaffilmark{2},
29: R. Maiolino\altaffilmark{3},
30: P. Andreani\altaffilmark{4,5},
31: O. Shemmer\altaffilmark{6},
32: S. Veilleux\altaffilmark{7}}
33: \altaffiltext{1}{Max-Planck-Institut f\"ur extraterrestrische Physik,
34: Postfach 1312, 85741 Garching, Germany \email{lutz@mpe.mpg.de,
35: sturm@mpe.mpg.de, linda@mpe.mpg.de,
36: valiante@mpe.mpg.de, schweitzer@mpe.mpg.de}}
37: \altaffiltext{2}{School of Physics and Astronomy and the Wise Observatory,
38: The Raymond and Beverly Sackler Faculty of Exact Sciences,
39: Tel-Aviv University, Tel-Aviv 69978, Israel\email{netzer@wise1.tau.ac.il}}
40: \altaffiltext{3}{INAF, Osservatorio Astronomico di Roma, via di Frascati 13,
41: 00040 Monte Porzio Catone, Italy\email{maiolino@ao-roma.inaf.it}}
42: \altaffiltext{4}{ESO, Karl-Schwarzschildstra\ss{}e 2, 85748 Garching, Germany
43: \email{pandrean@eso.org}}
44: \altaffiltext{5}{INAF, Osservatorio Astronomico di Trieste, via Tiepolo 11,
45: 34143 Trieste, Italy}
46: \altaffiltext{6}{Department of Astronomy and Astrophysics, 525 Davey
47: Laboratory, Pennsylvania State University,
48: University Park, PA 16802, USA\email{ohad@astro.psu.edu}}
49: \altaffiltext{7}{Department of Astronomy, University of Maryland,
50: College Park, MD 20742-2421, USA\email{veilleux@astro.umd.edu}}
51:
52: \begin{abstract}
53: We present Spitzer rest-frame mid-infrared spectroscopy of twelve z$\sim$2
54: mm-bright type 1 QSOs, selected from unlensed and lensed QSO samples and
55: covering a range of AGN optical luminosities L$_{5100}=10^{45}$ to
56: 10$^{47}$ erg s$^{-1}$.
57: On top of the AGN continuum, we detect PAH emission from luminous star
58: formation in nine objects individually as well as in the composite spectrum
59: for the full sample. PAH luminosity and rest frame far-infrared luminosity
60: correlate and
61: extend the similar correlation for lower luminosity local QSOs. This provides
62: strong evidence for intense star formation in the hosts of these mm-bright
63: QSOs, sometimes exceeding 1000\Msun yr$^{-1}$ and dominating their rest frame
64: far-infrared emission. The PAH-based limit on star formation rates is lower
65: for luminous z$\sim$2 QSOs that are not preselected for their mm emission.
66: Partly dependent on systematic changes
67: of the AGN dust covering factor and the dust spectral energy distribution of
68: the AGN proper, the spectral energy distributions of mm-faint high-z QSOs
69: may be AGN
70: dominated out to rest frame far-infrared wavelengths. Towards the most
71: luminous high-z QSOs, there is a flattening of the
72: relation between star formation and AGN luminosity that is observed for
73: lower redshift QSOs. No QSO in our sample has a PAH-measured
74: star formation rate in excess of 3000 \Msun yr$^{-1}$.
75:
76: \end{abstract}
77:
78: \keywords{galaxies: active, galaxies: starburst, infrared: galaxies}
79:
80:
81: \section{Introduction}
82: Studies of the evolution of galaxies and active galactic nuclei (AGN) are
83: drawing an
84: increasingly detailed picture of their global evolution from redshifts
85: above 6 to the present. The cosmic star forming density as well as the number
86: density of AGN rise and fall over cosmic time, with a maximum in the `quasar
87: epoch' at redshifts around 2. Less is known about the connection of star
88: formation and nuclear activity in individual high redshift objects, and
89: the controlling physical processes. A related basic observational goal is
90: to quantify both AGN luminosity and star formation in an as wide as possible
91: set of galaxy populations, a task that is posing considerable technical
92: difficulties for some types of objects.
93:
94: A particularly interesting population are the luminous AGN that are
95: traced by optically selected QSOs.
96: Evolutionary connections between intense starburst events and QSOs have
97: been proposed
98: early on \citep[e.g.][]{sanders88,norman88} and naturally suggest that the
99: luminous AGN phase coincides with or follows after a period of
100: intense star formation. While still limited in spatial resolution and requiring
101: simplifying assumptions, increasingly advanced models of these processes
102: now become available. They combine galaxy merging, gas transport, star
103: formation, accretion and feedback
104: processes \citep[e.g.][]{springel05}. In turn, such scenarios are part
105: of recent global modes of galaxy and merger evolution
106: \citep[e.g.][]{granato04,hopkins06}.
107:
108: Indications have been obtained for intense star formation events in the host
109: of high redshift QSOs. Perhaps the most important is the detection of
110: submm or mm continuum (rest frame far-infrared) from dust in part
111: of the high-z QSO population. If heated by star formation, this
112: emission indicates
113: high star formation rates similar to those in submillimeter galaxies
114: \citep[e.g.][]{omont03,priddey03,barvainis02}. Large molecular gas reservoirs
115: sufficient to sustain intense star formation have been detected
116: in CO \citep[see summaries in][]{solomon05,greve05},
117: and in some of the brightest systems also in tracers of dense molecular gas
118: \citep{barvainis97,solomon03,carilli05,riechers06,gracia06,guelin07}.
119: Finally, the [CII] 157$\mu$m fine structure line
120: was detected in a few high-z quasars \citep{maiolino05,iono06} at a ratio
121: to the rest frame far-infrared emission similar to the ratio in local
122: ULIRGs, consistent with massive star formation. While these observations
123: indicate a plausible picture of massive star formation in some QSO hosts, they
124: need independent support, in particular to better discriminate between star
125: formation and AGN heating of the observed rest frame far-infrared continuum.
126: We have hence initiated a program using the mid-infrared `PAH' emission
127: features as tracers of star formation in the hosts of high-z QSOs.
128:
129: Because of their presence in star forming environments with a wide range of
130: physical conditions but absence close to powerful AGN, the main mid-infrared
131: PAH emission features at 6.2, 7.7, 8.6, and 11.3$\mu$m have been used
132: extensively as extragalactic star formation indicators. The calibration
133: factor between PAH luminosity and star formation rate
134: depends on average physical conditions in the interstellar medium of the
135: galaxy probed. This is indicated both
136: in the relatively weaker PAH in galaxies with intense interstellar
137: radiation fields \citep[as measured by the large grain dust temperature from
138: their SEDs,][]{dale01}, and in the sublinear increase of PAH emission
139: with P$\alpha$
140: hydrogen recombination line emission, that is observed in spatially resolved
141: studies of nearby
142: galaxies \citep{calzetti07}. For the specific case of star formation
143: indicators for QSO hosts, the associated uncertainty is outweighed by several
144: advantages. PAH emission from starbursts is very luminous and thus
145: less easily outshone by the powerful AGN. Unlike other popular star
146: formation indicators like H$\alpha$ or mid-infrared 24$\mu$m
147: continuum, it will not be quickly overwhelmed by strong AGN emission in the
148: very same tracers. It is also less sensitive than e.g. the optical
149: [OII] 3727\AA\ line
150: to the significant internal obscuration that is typical for starburst
151: regions.
152:
153: In \citet{schweitzer06} and \citet{netzer07} we have applied these techniques
154: to local z$\leq$0.3 Palomar-Green (PG) QSOs, detecting PAH emission in many
155: objects and concluding that likely most of their
156: $\sim 10^{10}$--10$^{12}$\Lsun\ far-infrared luminosity is due to star
157: formation. Comparing average spectra and SEDs of groups of QSOs that are
158: differing in their level of far-infrared emission, we were able to
159: derive an infrared SED for the
160: pure AGN, after subtraction of host star formation.
161:
162: Several spectroscopic studies have shown that Spitzer-IRS spectroscopy with
163: modest integration times of at most a few hours per source can successfully
164: detect the PAH emission of the most luminous z$\sim$2.5 star forming events,
165: which approach infrared luminosities of 10$^{13}$\Lsun\ and star
166: formation rates around 1000\Msun /yr. Such detections have been obtained for
167: submillimeter
168: galaxies (SMGs) \citep{lutz05,valiante07,menendez07,pope08} as well as for
169: a subset of bright Spitzer 24$\mu$m selected galaxies
170: \citep[e.g.][]{weedman06,yan07}. Substantially
171: longer integrations exceeding 10 hours in an IRS spectral order have been
172: demonstrated to be able to detect yet fainter PAH emission at similar
173: redshifts with a peak of the rest frame 7.7$\mu$m feature of 0.4mJy or less
174: \citep{teplitz07}. Probing PAH emission to these levels would be highly
175: interesting for
176: a complete sample of luminous z$\sim$2.5 QSOs but expensive in observing
177: time. Systematic difficulties arise in detecting PAH features on a strong AGN
178: continuum that
179: additionally has structure due to silicate emission. The silicate emission
180: is strong in regions of the spectrum
181: where the main PAH complexes are weak, and thus tends to wash out the
182: PAH-induced structure, as demonstrated for some local PG QSO spectra in
183: \citet{schweitzer06}.
184:
185: In this paper, we hence focus as an important first step on
186: measuring PAH emission and inferred star
187: formation in a modest size sample of twelve z=1.8 to z=2.8 QSOs that all
188: have detected
189: (sub)mm emission. If this mm emission is due to host star formation,
190: its associated PAH emission will be detectable with IRS in modest integration
191: times by analogy to the SMGs at similar redshift. In \citet{lutz07} we have
192: presented first results for the brightest object in our sample, the Cloverleaf
193: (H1413+117). Here we present the full mm-bright sample and discuss in
194: section 4.2 implications for the
195: high redshift QSO 1 population as a whole.
196:
197: Throughout the paper we adopt an $\Omega_m =0.3$, $\Omega_\Lambda =0.7$
198: and $H_0=70$ km\,s$^{-1}$\,Mpc$^{-1}$ cosmology. When converting IR
199: luminosities to star formation rates we adopt for consistency
200: with much of the previous work the \citet{kennicutt98} value
201: which is based on a 0.1-100\Msun\ Salpeter IMF. For a more realistic Chabrier
202: or Kroupa IMF star formation rates would be a factor $\sim$1.6 smaller. Our
203: sample includes lensed sources. For these,
204: we present fluxes or flux densities as observed, but luminosities, star
205: formation rates and
206: masses corrected for the adopted lensing magnification.
207:
208: \section{Sample and Observations}
209:
210: We have selected from the literature z$\sim$2 Type 1 radio quiet QSOs with
211: robust ($>5\sigma$) and consequently bright submm or mm dust continuum
212: detections. The
213: sample includes lensed as well as unlensed QSOs and is summarized in
214: Table~\ref{tab:sample}. Sources have been chosen from the surveys of (sub)mm
215: emission in z$\sim$2 QSOs by \citet{omont03} and \citet{priddey03}, the lensed
216: QSO submm survey of \citet{barvainis02}, and one target each from
217: \citet{guilloteau99} and \citet{knudsen03}. One of the targets
218: (SMMJ04315+10277, \citet{knudsen03}) was first reported in a submm survey
219: and may be
220: intermediate in properties between optically selected QSOs and typical submm
221: selected
222: galaxies. The other targets are selected from the mentioned studies of
223: optically selected QSOs solely
224: on the basis of a bright and robust (sub)mm detection and will provide a view
225: of the rest frame far-infrared bright part of the general z$\sim$2 unobscured
226: QSO population.
227: Our unlensed targets have AGN rest frame optical luminosities
228: $L_{5100}= \nu L_\nu(5100\AA)$ in the range 10$^{46}$ to 10$^{47}$\ergs, some
229: of the lensed QSOs are intrinsically an order of magnitude less luminous
230: and overlap with the highest luminosity local sources studied with
231: similar methods by \citet{schweitzer06} and \citet{netzer07}.
232:
233: For targets in \citet{omont03} the optical luminosities were adopted from
234: the rest frame blue absolute magnitudes in this reference, converting
235: to our adopted cosmology. In the other cases we used published near-infrared
236: photometry \citep[][and 2MASS point source catalog]{priddey03}, and the
237: assumption of a $f_\nu\propto\nu^{-0.44}$ rest frame optical/UV continuum
238: \citep{vandenberk01}.
239:
240: We obtained low resolution (R$\sim 60-120$) mid-infrared spectra of the QSOs
241: using the Spitzer infrared spectrograph IRS \citep{houck04} during the period
242: July 2006 - October 2007. The IRS LL slit is wide enough to include
243: all images of the lensed sources. For each target, 30 cycles of
244: 120 sec integration time per nod position were taken in the
245: LL1 (19.5 to 38.0 $\mu$m) module and 15
246: cycles in the LL2 (14.0 to 21.3 $\mu$m) module, leading to effective on-source
247: integration times of 2 and 1 hours, respectively. We start from the pipeline
248: processed basic calibrated data (version 14.4 to 16.1 for different sources),
249: subtracting individually data for the same data collection
250: event within the data collection sequences of the two nod positions.
251: We use own deglitching routines identifying outliers from the default bad
252: pixel masks, by identifying points discrepant by more than a multiple of
253: the local noise from their spectral neighborhood, and by visual inspection
254: of the 2D frames, finally replacing them with values representative of
255: their spectral neighborhood. We coadded the images for each cycle,
256: checking pixel per pixel for outlying values in the stack, and subtracted
257: residual background derived from a first order spatial fit to source-free
258: regions. SMART \citep{higdon04} and SPICE (optimum extraction,
259: for the weak continuum source SMMJ04315 only) were used for
260: extraction. We compared the independent spectra from the two nod positions
261: before averaging, finding no mismatches that would be indicating
262: mispointings or other serious problems with the data. Before combining
263: the segments of the spectrum observed in the
264: different modules into the final spectrum, the segment observed in the LL2
265: module was multiplied with an object-dependent scaling factor (within a few \%
266: of 1) in order to provide best match in the overlapping wavelength range.
267:
268: \section{Results}
269: Figure~\ref{fig:allhighz} shows the spectra of all our targets, and
270: Fig.~\ref{fig:averagehighz} a composite derived from all sample objects.
271: The main spectral features are the ones also seen in the spectra of local PG
272: QSOs \citep[e.g.][]{siebenmorgen05,hao05,schweitzer06,hao07,netzer07}: A
273: strong AGN continuum
274: with superposed silicate emission at $\lambda\gtrsim 9\mu$m, and aromatic
275: `PAH' emission features at 6.2, 7.7, 8.6$\mu$m at varying strength
276: relative to the local continuum. An unusual individual spectrum is the one
277: of SBS 1408+567 with convex continuum, a smooth broad emission feature
278: around 8$\mu$m, but no unambiguous identification of PAH emission. This broad
279: feature is weaker relative to the continuum but has similarity with the
280: 8$\mu$m feature in the mid-infrared spectrum of the well known Type 1 AGN
281: and ULIRG
282: Mrk 231 \citep{rigopoulou99,spoon02,weedman05,armus07} which may originate
283: from a combination of silicate emission and absorption. PAH 7.7$\mu$m may
284: also contribute to some extent to this feature.
285:
286: In this paper, we focus on analysis of the PAH emission and its implications
287: for star formation in the QSO hosts. Given the S/N of the spectra and the
288: variations in slope and shapes of continuum and silicate emission
289: in AGN, we do not attempt to derive independent fluxes for the various PAH
290: features. Informed by the spectra of local QSOs and supported by the average
291: spectrum of Fig.~\ref{fig:averagehighz}, we quantify the overall
292: level of PAH emission by decomposing
293: the QSO spectra into a scaled and resampled starburst template
294: \citep[based on the ISO-SWS spectrum of M82,][]{sturm00} at the
295: nominal redshift of the QSO, and the underlying AGN emission which
296: consists of a dust continuum and silicate emission that is turning up
297: beyond $\sim$8$\mu$m. Results of this decomposition are shown in
298: Figure~\ref{fig:allhighzdecomp}.
299: The combination of AGN continuum and silicate emission varies
300: smoothly with wavelength over the range of interest. The first signature
301: of PAH in such spectra that is already visible at modest S/N then is the
302: strongest 7.7$\mu$m feature and its short wavelength steep flank, while its
303: longer wavelength side is often less pronounced, due to the combined presence
304: of 8.6$\mu$m PAH and the upturn of silicate emission. The strength of the
305: starburst component in the decomposition is adjusted to produce a smooth
306: residual AGN emission at the locations of the 6.2 and 7.7$\mu$m features, in
307: particular to remove the steep short wavelength flank of the 7.7$\mu$m feature.
308: Given the S/N of the data we do not attempt to derive PAH feature ratios by
309: multi-feature fits, but obtain from the simple template a more robust
310: characterisation of the overall level of PAH emission which we represent
311: by the flux of the brightest (7.7$\mu$m) feature. These PAH 7.7$\mu$m fluxes
312: (Table~\ref{tab:results}) are measured in the scaled starburst template
313: resulting from the decomposition, by fitting Lorentzian
314: features for the PAHs plus an underlying residual continuum.
315: PAH 7.7$\mu$m fluxes listed in Table~\ref{tab:results} can thus be
316: directly compared to the results for local QSOs presented by
317: \citet{schweitzer06}.
318:
319: We list in Table~\ref{tab:results} the measured noise per resolution
320: element over the rest frame 6.5--8$\mu$m region of the actual spectra,
321: defined as RMS around a linear fit to the data after subtraction of the
322: scaled PAH template and its associated feature structure. By quoting
323: the noise measured in this way from the spectra, we implicitly include
324: possible noise contributions for stronger continua, that may arise for
325: example due to more difficult outlier identification and fringing.
326: This noise may be conservatively overestimated if the continuum is slightly
327: curved over this range. Noise $\sim$0.2mJy per spectral sample as obtained
328: for some of our weaker continuum targets is in good agreement with the
329: nominal IRS sensitivity for our observing parameters, as implemented
330: in the Spitzer Science Center SPEC-PET tool.
331:
332: For the 9 sources for which we report PAH detections in
333: Table~\ref{tab:results}, the peak height of PAH emission in the starburst
334: component obtained from the decomposition is required to be at least 3 times
335: the measured local noise. These significances of 3--13 for the PAH emission
336: effectively reflect the significance of the short wavelength
337: flank of the 7.7$\mu$m feature. At longer wavelengths, systematic variations
338: in shape of underlying continuum and in particular the strength and detailed
339: shape of the onset of AGN silicate emission at $\geq 8\mu$m start to
340: become more and more limiting, preventing reliable use of these wavelengths
341: to verify or improve the formal significance. Several
342: of the targets (SMMJ04315+102177, KUV0806+4037, HS1002+4400, LBQS1230+1627B)
343: show PAH with peak height that is for the individual sources just above
344: the peak/noise=3 detection limit. In combination with the possibility
345: that the mentioned uncertainties on AGN continuum and silicate shape affect
346: to a lesser extent also the short flank of the 7.7$\mu$m PAH, we estimate
347: that individual PAH fluxes for these weak PAH detections can be uncertain
348: at the 30--50\%\ level. We caution that they should not be used for
349: analyses going significantly beyond the modest accuracy requirements of
350: the luminosity analyses presented below.
351:
352: Upper limits are estimated analogously, for the PAH template scaled to
353: a level where the features would remain visibly detectable on top of the
354: continuum and the silicate feature. The IRS spectra of the sources with
355: PAH nondetections are still consistent
356: with the presence of PAH emission at peak feature fluxes similar to the
357: sources with detections.
358: Again, the main limitation is systematic, due to the strong and structured underlying
359: continuum and silicate emission in these targets. In the spectrum of
360: LBQS 0018-0220, for example, inspection of Fig.~\ref{fig:allhighz} shows that
361: the 7-8$\mu$m region of the spectrum, observed with good S/N, may
362: represent either PAH emission partly filling in the region before an
363: upturn of silicate emission, or be due to variations in shape/slope of this
364: feature, which is strong for LBQS 0018-0220, and the underlying continuum.
365: Such variations are observed in local AGN \citep[e.g.][]{hao05,schweitzer06,netzer07}.
366: We list in Table~\ref{tab:results} PAH upper limits considering this possiblity
367: rather than only reflecting the detector noise.
368:
369: We will compare below the PAH emission of high-z QSOs with the rest frame
370: far-infrared emission. Partly, this is possible by simply using the rest
371: frame $\sim 250\mu$m continuum that can be obtained from the published
372: (sub)mm photometry
373: with minimal extrapolation. For other purposes and for comparison with
374: the local QSO population we need to extrapolate to the rest frame far-infrared
375: peak which is constrained by direct observations for very few high redshift
376: QSOs only. We do so by assuming that the detected (sub)mm emission reflects
377: a rest frame T=40K, emissivity index $\beta$=1.5 greybody spectrum. These
378: parameters are in reasonable agreement with the dust temperature
379: $\sim$36K suggested
380: for the submillimeter galaxy population by \citet{chapman05} and with
381: 40-60K
382: suggested by \cite{beelen06} for high-z QSOs, from the combination of
383: 350$\mu$m photometry with longer wavelength observations.
384: For three QSOs overlapping with our sample (KUV 0806+4037, HS 1002+4400, SBS 1408+567), \cite{beelen06} find 32K, 38K,
385: and 35K respectively, and the FIR-warmer QSOs in their small sample may include
386: objects with a smaller ratio of host FIR/submm to AGN emission. Similarly, our
387: simple extrapolation to rest frame 60$\mu$m for H1413+117 (the Cloverleaf) is
388: in good agreement with the observations of \citet{benford99} and
389: \citet{roro00}. At redshift z$\sim$2.4, changing the adopted temperature
390: by $\pm$5K will change the estimated rest frame 60$\mu$m flux by $+$60\% and
391: $-$45\%, respectively. This is the dominant source of uncertainty in the estimated
392: rest frame 60$\mu$m emission, since the original (sub)mm measurements for our IRS sample are
393: S/N$>$5 by sample definition. Using the T=40K $\beta$=1.5 assumption we extrapolate
394: to the rest frame FIR peak of this
395: component for all our sources, which we express as $\nu L_\nu (60\mu$m) for
396: consistency with the local QSO work of \citet{schweitzer06}. These 60$\mu$m
397: luminosities are listed in Table~\ref{tab:results}. For sources with
398: published independent 850$\mu$m and
399: 1.2mm photometry we have averaged results from the two possible extrapolations.
400:
401:
402: \section{Discussion}
403: \subsection{Consistent evidence for luminous starbursts from PAH and rest
404: frame far-infrared}
405:
406: Comparison of the PAH detections and limits to the rest frame far-infrared
407: emission of the sample QSOs, and comparison to other populations is the key
408: to testing the hypothesis that both PAH and far-infrared trace intense
409: starbursts in the hosts.
410: We do such comparisons first in directly observed quantities, and then using
411: PAH luminosities in combination with FIR luminosities derived with the
412: greybody extrapolation described above.
413:
414: Fig.~\ref{fig:pah222} shows a comparison of the PAH 7.7$\mu$m peak flux
415: density with the flux at rest frame wavelength 222$\mu$m. The latter
416: corresponds to
417: an observed wavelength of 850$\mu$m for z$\sim$2.8, the specific redshift
418: chosen for consistency with previous work on SMGs \citep{lutz05,valiante07}.
419: This is a simple diagram based on directly observed quantities,
420: and has minimal sensitivity to SED assumptions for the extrapolation from the
421: observed submm/mm point to rest frame 222$\mu$m.
422: The distribution of the ratio of
423: continuum-subtracted PAH peak flux density and rest frame FIR continuum
424: $S_{PAH 7.7\mu m}/S_{222\mu m}$ is
425: very similar for our mm-bright QSOs, for the submillimeter galaxy sample
426: of \citet{valiante07}, and for local ULIRGs. The median
427: log($S_{PAH 7.7\mu m}/S_{222\mu m}$) is -1.14, -1.32 and -1.26 for the QSOs,
428: SMGs, and ULIRGs, respectively, and the distribution for the QSOs can be
429: drawn at $>$25\% probability from the same distribution as either the SMGs
430: or the ULIRGs, according to a two-sided Kolomogorov-Smirnov test.
431:
432: Moving from observed quantities to PAH and far-infrared luminosities,
433: Fig.~\ref{fig:pahtofir} shows that the mm-bright high-z QSOs extend
434: the relation between these luminosities found for local PG QSOs by
435: \citet{schweitzer06} by an order of
436: magnitude towards higher luminosities, and that the relation for the QSO
437: population crosses the locus for star-forming ULIRGs. The PAH and FIR data
438: of the mm bright QSOs agree with a scenario where these two tracers are
439: jointly originating in
440: starbursts in the hosts. These starbursts reach SMG-like luminosities of
441: $\sim 10^{13}\Lsun$ in the most luminous (unlensed) targets of our sample.
442:
443: Individual PAH-based star formation rates can be derived for our sample and
444: are listed in Table~\ref{tab:results}. Concerning the calibration
445: factor applied, the similarity of the PAH/FIR ratio in our sample objects and
446: ULIRGs permits
447: us to use the ratio L(PAH)/L(FIR)$\sim$ 0.013 derived for the starburst-like
448: ULIRGs in
449: \citet{schweitzer06}, in combination with the calibration of FIR-based star
450: formation rates by \citet{kennicutt98}. Because of the
451: changes of PAH/FIR with galaxy properties discussed above, it is worth
452: recalling that the resulting
453: $SFR\ [\Msun\ year^{-1}]=3.5\times 10^{-42}L_{PAH}\ [erg\ s^{-1}]$ assumes
454: ULIRG-like physical conditions in the interstellar medium of the galaxy
455: observed. PAH based star formation rates for our high-z sample
456: range from about 200 to 3000 \Msun\//yr. As was found for
457: local QSOs and their far-infrared emission \citep{schweitzer06}, use of
458: the PAH diagnostic provides strong support for massive star forming
459: events in the mm bright QSOs, at yet an order of magnitude larger
460: luminosities.
461:
462: Molecular gas CO emission has previously been searched for in six of our
463: targets, detecting five (Table~\ref{tab:results}, all gas masses based on
464: the conversion factor 0.8 \Msun/(K km s$^{-1}$ pc$^2$) of \citet{downes98}
465: and our adopted cosmology). Another target (HE 1104-1805)
466: has been searched for CO \citep{barvainis02b}, but the nondetection may not be
467: meaningful given the narrow bandwidth of this CO search, which may miss the
468: target CO line for plausible uncertainties of the optical input redshift.
469: Consistent with their intense star
470: formation, the QSO hosts are gas-rich systems (molecular gas mass of a few
471: 10$^{10}$\Msun). Despite the large gas masses, gas exhaustion times
472: M(H$_2$)/SFR are short and range from 10$^7$ to $8\times 10^{7}$ years, where
473: measured.
474:
475: \subsection{From mm-bright QSO1s to the unbiased QSO1 population}
476:
477: The previous section has discussed properties of the mm-bright
478: part of the z$\sim$2 luminous QSO1 population only. As an indication of
479: the significance of this selection, we note that we actually
480: observed 6 mm-bright QSOs out of the combined samples of \cite{omont03} and
481: \citet{priddey03}, which in total contain 83 distinct objects. We discuss
482: the properties of our objects in the context to these two samples, the
483: 1.7$<$z$<$3 objects among the lensed sample of \citet{barvainis02},
484: and the local
485: PG QSOs, in order to elucidate implications for the complete and unbiased
486: population of luminous z$\sim$2 QSO 1s.
487:
488: Figure~\ref{fig:pahto5100} shows the QSO PAH luminosity as a function of
489: AGN rest frame optical luminosity. The mm bright QSOs clearly extend to
490: higher luminosity the relation for local PG QSOs already discussed by
491: \citet{netzer07}, with a possible saturation (flattening of the relation)
492: at the largest
493: AGN luminosities. An extension to individual PAH detections or significant
494: limits for the full population including mm-faint QSOs would be extremely
495: difficult even with Spitzer capabilities. Substantial insight can already be
496: gained, nevertheless, by adding to Fig.~\ref{fig:pahto5100} the limit on PAH
497: emission set by \citet{maiolino07} for the
498: {\em average} of a similar redshift QSO1 sample that
499: is not mm preselected. Unlike their average spectrum, the individual spectra
500: from the \citet{maiolino07} sample are of insufficient S/N to put a
501: limit on PAH emission
502: that is meaningful for our analysis. Compared to mm bright QSOs from our
503: sample of similar AGN luminosities, the limit of PAH emission for the
504: not mm-biased
505: average spectrum is clearly lower, by a factor $\sim$3 compared to our
506: PAH detections. The corresponding average SFR limit is
507: $\lesssim$700$\Msun\ yr^{-1}$.
508:
509: In an analogous way, Fig.~\ref{fig:firto5100} compares the QSO far-infrared
510: luminosity, for which a larger number of individual measurements is available,
511: to the AGN rest frame optical luminosity. This diagram also reminds
512: that the z$\sim$2 mm-bright QSOs, having L$_{FIR}\sim$L$_{5100}$ i.e.
513: L$_{FIR}\sim$0.1 L$_{Bol}$, are far from being extreme infrared excess objects
514: and resemble in this property the average local QSO. In relation to the
515: AGN luminosity, the strongest IR emission
516: for the high-z QSOs in this diagram is for SMMJ04315+1027, which was
517: originally detected at submm wavelengths. In this diagram, the bulk of the
518: SMG population would extend from the location of SMMJ04315+1027 to the left,
519: because of the typically modest luminosity of the AGN they are often hosting
520: \citep[e.g.][]{alexander05,valiante07}.
521: As concerns the relation of
522: mm bright QSOs to the total QSO population, we obtain a similar result as from
523: the PAH data. Again, the mm-bright QSO studied here extend the relation
524: for local QSOs and represent the far-infrared bright
525: end of a distribution in L$_{FIR}$/L$_{5100}$ among the z$\sim$2 QSOs.
526: Considering only the L$_{5100}>10^{46}$erg s$^{-1}$ QSOs, the mm bright
527: QSOs have median L$_{FIR}$=10$^{46.6}$erg\,s$^{-1}$, a factor of 4 larger
528: than the median
529: for the full population. The specific value of this factor should be
530: considered tentative only, since the distribution for the full population
531: is based on (sub)mm photometric points of typically low individual
532: significance.
533:
534: The PAH spectroscopy presented for mm bright high-z QSOs here, and for
535: local QSOs in \citet{schweitzer06} and \citet{netzer07}, has provided
536: strong support for
537: a star formation origin of the rest frame far-infrared emission in these
538: systems. This conclusion has been derived for QSOs with
539: L$_{FIR}\sim$L$_{5100}$, however, and needs to be critically reviewed for
540: the even more AGN dominated objects at the mm-faint end of the
541: high-z QSO population. In \citet{netzer07}, we have presented a
542: host-subtracted intrinsic
543: AGN SED for local PG QSOs which is very broadly characterised by
544: $\nu L_\nu$(5100\AA\/)$\sim \nu L_\nu$(6$\mu$m)$\sim 5\times \nu L_\nu$(60$\mu$m),
545: with considerable scatter between individual objects. This intrinsic SED drops
546: significantly towards the rest frame far-infrared, but in combination with
547: the properties of the high-z QSOs (Fig.~\ref{fig:firto5100}) would
548: imply that the mm-faint end of the z$\sim$2 QSO population
549: in fact may contain almost `pure' AGN without strong host star formation.
550: The AGN-heated dust could already provide the observed weak rest frame
551: far-infrared
552: emission. A caveat to be noted here originates from recent
553: evidence for systematic changes
554: in the mid-infrared to optical SEDs of luminous high-z QSOs compared
555: to their local low-luminosity analogs \citep{maiolino07,treister08}, and from
556: the possible detection of high-z AGN without hot dust \citep{jiang06}.
557: We defer a detailed analysis of the AGN-heated dust in our sample to a future
558: paper, but note that the larger PAH equivalent width of our
559: average spectrum compared to the local PG average
560: (Fig.~\ref{fig:averagehighz}) is in support of these results,
561: given that it is combined with similar
562: or lower ratio of PAH and 5100\AA\ continuum in our sample compared to locally
563: (Fig.~\ref{fig:pahto5100}).
564: These results
565: suggest a lower mid-infrared AGN continuum luminosity in relation
566: to the optical AGN luminosity, and thus smaller AGN dust
567: covering factors. They imply changes of the intrinsic AGN SED for luminous
568: high-z QSOs compared to the \citet{netzer07} intrinsic SED for local QSOs,
569: likely weakening the AGN-related 60$\mu$m emission along with the weakening
570: of the rest frame mid-infrared AGN dust emission. This decrease of AGN
571: rest frame 60$\mu$m emission would reduce again the part of
572: the total far-infrared continuum directly ascribed to the AGN and perhaps
573: suggest a noticeable host contribution even in mm-faint high-z QSOs.
574: Firmly establishing the nature of the weak
575: rest frame far-infrared emission at the mm-faint end of the z$\sim$2 QSO
576: population will hence require a more complete SED characterisation over
577: the full rest frame infrared range.
578:
579: \subsection{A flattening relation of starburst and AGN luminosity}
580:
581: For local QSOs with L$_{5100}\sim 10^{44}$ to 10$^{46}$erg\,s$^{-1}$,
582: we have found clear correlations of the star formation indicators PAH and
583: far-infrared continuum with AGN optical luminosity \citep{netzer07}, that
584: are approximately linear over this luminosity range. The high-z QSOs extend
585: the AGN luminosity range by another order of magnitude. The relation of PAH
586: (star formation) and AGN luminosity is flattening at
587: L$_{5100}> 10^{46}$erg\,s$^{-1}$ (Fig.~\ref{fig:pahto5100}). This is not yet
588: unambiguous from the comparison of our mm-bright sample alone with the
589: local PG QSOs. There may be a slight flattening, and the slope of a
590: linear fit in
591: Log(L$_{PAH}$) vs. log(L$_{5100}$) is 1.01 for the PAH detections among
592: the local PG QSOs, versus 0.92 adding the high-z PAH detections to the
593: sample, but these changes are not yet significant. The effect becomes clear
594: by inclusion of the PAH upper limit which has been derived
595: for the sample of \citet{maiolino07} that is not selected to be mm-bright
596: (Fig.~\ref{fig:pahto5100}). In agreement with mm continuum
597: based studies, our PAH spectroscopy
598: does not find evidence for star formation rates above
599: 3000\Msun yr$^{-1}$ (starburst luminosities much larger than 10$^{13}\Lsun$).
600: This saturation occurs at luminosities similar to the upper end of the
601: luminosity range spanned
602: by the submm galaxy population, a limit that is often related to the star
603: formation rate in a maximal starburst that is converting much of its gas
604: into stars over just a few dynamical times \citep[e.g.][]{tacconi06}.
605: In support of this view, a slower than linear increase (or saturation at
606: high AGN luminosities) is also inferred for the relation of cold gas
607: mass and AGN luminosity, from a synopsis of CO measurements for
608: high-z/high-L QSOs \citep{maiolino07b}.
609:
610: Local PG QSOs are found in a region of the fundamental plane for elliptical
611: galaxies and spheroids that is hosting
612: moderate mass ellipticals with corresponding black hole
613: masses $\sim10^8$\Msun\/, and accreting
614: efficiently at an Eddington ratio $\sim$0.25 \citep{dasyra07}. Relative to the
615: host star formation rate, their AGN growth rates are more than an order of
616: magnitude larger than needed to explain the locally observed BH mass to
617: bulge mass relation under the assumption of coeval growth \citep{netzer07}.
618: The increased L(AGN)/L(SF) for the high-z QSOs and in particular the mm-faint
619: ones places them in a yet more pronounced phase of AGN growth proceeding much
620: more rapidly than host growth.
621: The failure up to now to identify matching extreme $>>10^{14}\Lsun$ star
622: forming events also in other galaxies and surveys indicates this is not
623: just due to a time delay between AGN and host growth,
624: but must reflect intrinsically different durations of the AGN growth and host
625: growth phases.
626:
627: \section{Conclusions}
628:
629: In the context of quantitatively constraining the coexistence of star
630: formation and AGN activity in different classes of high redshift galaxies, we
631: have used Spitzer-IRS mid-infrared spectroscopy to detect PAH emission
632: in the spectra of nine out of twelve mm-bright z$\sim$2 radio quiet QSOs,
633: with optical luminosities L$_{5100}=10^{45}$ to
634: 10$^{47}$ erg s$^{-1}$.
635: The detections provide strong support for the presence of intense star
636: formation up to $\sim$3000\Msun yr$^{-1}$ in their hosts, that is dominating
637: their rest frame far-infrared emission. The typical luminous
638: z$\sim$2 QSO that is not preselected for mm emission has lower star
639: formation rate $<$700\Msun yr$^{-1}$ and the
640: mm-faint part of this population may include objects that are AGN dominated
641: out to rest-frame far-infrared wavelengths, due to luminous AGN but at best
642: modest level of star formation. Toward the high luminosity and redshift QSOs,
643: there is a flattening in the relation between starburst and AGN luminosity
644: that is observed for local PG QSOs, and no object in our sample has a
645: PAH-measured star formation rate above 3000\Msun yr$^{-1}$.
646:
647: \acknowledgments
648: This work is based on observations made with the
649: \textit{Spitzer Space Telescope},
650: which is operated by the Jet Propulsion Laboratory, California Institute of
651: Technology, under a contract with NASA. Support for this work was provided by
652: NASA under contracts 1287653 and 1287740 (S.V.,O.S.).
653:
654: %up to 8 authors, then first et al.
655: \begin{thebibliography}{999}
656: \bibitem[Alexander et al.(2005)]{alexander05} Alexander, D.M., Bauer, F.E.,
657: Chapman, S.C., Smail, I., Blain, A.W., Brandt, W.N., Ivison, R.J. 2005,
658: \apj, 632, 736
659: \bibitem[Armus et al.(2007)]{armus07} Armus, L., et al. 2007, \apj, 656, 148
660: \bibitem[Barvainis et al.(1997)]{barvainis97} Barvainis, R., Maloney, P.,
661: Antonucci, R., Alloin, D. 1997, \apj, 484, 695
662: \bibitem[Barvainis \& Ivison(2002)]{barvainis02} Barvainis, R.,
663: Ivison, R. 2002, \apj, 571, 712
664: \bibitem[Barvainis et al.(2002)]{barvainis02b} Barvainis, R., Alloin, D.,
665: Bremer, M. 2002, \aap, 385, 399
666: \bibitem[Beelen et al.(2006)]{beelen06} Beelen, A., Cox, P., Benford, D.J.,
667: Dowell, C.D., Kov\'acs, A., Bertoldi, F., Omont, A., Carilli, C.L.
668: 2006, \apj, 642, 694
669: \bibitem[Benford(1999)]{benford99} Benford, D. 1999, PhD Thesis, California
670: Institute of Technology
671: \bibitem[Calzetti et al.(2007)]{calzetti07} Calzetti, D., et al. 2007,
672: \apj, 666, 870
673: \bibitem[Carilli et al.(2005)]{carilli05} Carilli, C.L., et al. 2005,
674: \apj, 618, 586
675: \bibitem[Chapman et al.(2005)]{chapman05} Chapman, S.C., Blain, A.W.,
676: Smail, I, Ivison, R.J. 2005, \apj, 622, 772
677: \bibitem[Dale et al.(2001)]{dale01} Dale, D.A., Helou, G., Contursi, A.,
678: Silbermann, N.A., Kolhatkar, S. 2001, \apj, 549, 215
679: \bibitem[Dasyra et al.(2007)]{dasyra07} Dasyra, K.M., et al. 2007,
680: \apj, 657, 102
681: \bibitem[Downes \& Solomon(1998)]{downes98} Downes, D., Solomon, P. 1998,
682: ApJ, 507, 615
683: \bibitem[Grac\'ia-Carpio et al.(2006)]{gracia06} Grac\'ia-Carpio, J.,
684: Garc\'ia-Burillo, S., Planesas, P., Colina, L. 2006, \apj, 640, L135
685: \bibitem[Granato et al.(2004)]{granato04} Granato, G.L., de Zotti, G.,
686: Silva, L., Bressan, A., Danese, L. 2004, \apj, 600, 580
687: \bibitem[Greve et al.(2005)]{greve05} Greve, T.R., et al. 2005, \mnras,
688: 359, 1165
689: \bibitem[Gu\'elin et al.(2007)]{guelin07} Gu\'elin, M., et al. 2007,
690: \aap, 462, L45
691: \bibitem[Guilloteau et al.(1999)]{guilloteau99} Guilloteau, S., Omont, A.,
692: Cox, P., McMahon, R.G., Petitjean, P. 1999, \aap, 349, 363
693: \bibitem[Hainline et al.(2004)]{hainline04} Hainline, L.J., Scoville, N.Z.,
694: Yun, M.S., Hawkins, D.W., Frayer, D.T., Isaak, K.G. 2004, \apj 609, 61
695: \bibitem[Hao et al.(2005)]{hao05} Hao, L., et al. 2005, \apj, 625, L75
696: \bibitem[Hao et al.(2007)]{hao07} Hao, L., et al. 2007, \apj, 655, L77
697: \bibitem[Higdon et al.(2004)]{higdon04} Higdon, S.J.U., et al. 2004,
698: \pasp, 116, 975
699: \bibitem[Hopkins et al.(2006)]{hopkins06} Hopkins, P.F., Hernquist, L.,
700: Cox, T.J., Di Matteo, T., Robertson, B., Springel, V. 2006, \apjs, 163, 1
701: \bibitem[Houck et al.(2004)]{houck04} Houck, J.R., et al. 2004, \apjs, 154, 18
702: \bibitem[Iono et al.(2006)]{iono06} Iono, D., et al. 2006, \apj, 645, L97
703: \bibitem[Jiang et al.(2006)]{jiang06} Jiang, L., et al. 2006, \aj, 132, 2127
704: \bibitem[Kennicutt(1998)]{kennicutt98} Kennicutt, R.C. 1998, \araa, 36, 189
705: \bibitem[Knudsen et al.(2003)]{knudsen03} Knudsen, K.K., van der Werf, P.P.,
706: Jaffe, W. 2003 \aap, 411, 343
707: \bibitem[Lutz et al.(2005)]{lutz05} Lutz, D., et al. 2005, \apj, 625, L83
708: \bibitem[Lutz et al.(2007)]{lutz07} Lutz, D., et al. 2007, \apj, 661, L25
709: \bibitem[Maiolino et al.(2005)]{maiolino05} Maiolino, R., et al. 2005
710: \aap, 440, L51
711: \bibitem[Maiolino et al.(2007)]{maiolino07} Maiolino, R., Shemmer, O.,
712: Imanishi, M., Netzer, H., Oliva, E., Lutz, D., Sturm, E. 2007, \aap, 468, 979
713: \bibitem[Maiolino et al.(2007b)]{maiolino07b} Maiolino, R., et al. 2007,
714: \aap, 472, L33
715: \bibitem[Men\'endez-Delmestre et al.(2007)]{menendez07}
716: Men\'endez-Delmestre, K., et al. 2007 \apj, 655, L65
717: \bibitem[Netzer et al.(2007)]{netzer07} Netzer, H., et al. 2007,
718: \apj, 666, 806
719: \bibitem[Norman \& Scoville(1988)]{norman88} Norman, C., Scoville, N.Z., 1988,
720: \apj, 332, 124
721: \bibitem[Omont et al.(2003)]{omont03} Omont, A., Beelen, A., Bertoldi, F.,
722: Cox, P., Carilli, C.L., Priddey, R.S., McMahon, R.G., Isaak, K.G. 2003,
723: \aap, 398, 857
724: \bibitem[Pope et al.(2008)]{pope08} Pope, A., et al. 2008, ApJ, 675, 1171
725: \bibitem[Priddey et al.(2003)]{priddey03} Priddey, R.S., Isaak, K.G.,
726: McMahon, R.G., Omont, A. 2003, \mnras, 339, 1183
727: \bibitem[Riechers et al.(2006)]{riechers06} Riechers, D.A., Walter, F.,
728: Carilli, C.L., Weiss, A., Bertoldi, F., Menten, K.M., Knudsen, K.K., Cox, P.
729: 2006, \apj, 645, L13
730: \bibitem[Rigopoulou et al.(1999)]{rigopoulou99} Rigopoulou, D., Spoon, H.W.W.,
731: Genzel, R., Lutz, D., Moorwood, A.F.M., Tran, Q.D. 1999, \aj, 118, 2625
732: \bibitem[Rowan-Robinson(2000)]{roro00} Rowan-Robinson, M. 2000, \mnras,
733: 316, 885
734: \bibitem[Sanders et al.(1988)]{sanders88} Sanders, D.B., Soifer, B.T.,
735: Elias, J.H., Madore, B.F., Matthews, K., Neugebauer, G., Scoville, N.Z.
736: 1988, \apj, 325, 74
737: \bibitem[Schweitzer et al.(2006)]{schweitzer06} Schweitzer, M., et al.
738: 2006, \apj, 649, 79
739: \bibitem[Siebenmorgen et al.(2005)]{siebenmorgen05} Siebenmorgen, R., Haas, M.,
740: Kr\"ugel, E., Schulz, B. 2005, \aap, 436, L5
741: \bibitem[Solomon et al.(2003)]{solomon03} Solomon, P., Vanden Bout, P.,
742: Carilli, C., Gu\'elin, M. 2003, \nat, 426, 636
743: \bibitem[Solomon \& Vanden Bout(2005)]{solomon05} Solomon, P.M.,
744: Vanden Bout, P.A. 2005, \araa, 43, 677
745: \bibitem[Springel et al.(2005)]{springel05} Springel, V., Di Matteo, T.,
746: Hernquist, L. 2005, \mnras, 361, 776
747: \bibitem[Spoon et al.(2002)]{spoon02} Spoon, H.W.W., Keane, V.,
748: Tielens, A.G.G.M., Lutz, D., Moorwood, A.F.M., Laurent, O. 2002, \aap, 385,
749: 1022
750: \bibitem[Sturm et al.(2000)]{sturm00} Sturm, E., Lutz, D., Tran, D.,
751: Feuchtgruber, H., Genzel, R., Kunze, D., Moorwood, A.F.M., Thornley, M.D.
752: 2000, \aap, 358, 481
753: \bibitem[Tacconi et al.(2006)]{tacconi06} Tacconi, L.J., et al. 2006, \apj,
754: 640, 228
755: \bibitem[Teplitz et al.(2007)]{teplitz07} Teplitz, H.I., et al. 2007, \apj,
756: 659, 941
757: \bibitem[Treister et al.(2008)]{treister08} Treister, E., Krolik, J.H.,
758: Dullemond, C. 2008, \apj, in press (arXiv 0801.3849)
759: \bibitem[Valiante et al.(2007)]{valiante07} Valiante, E., et al. 2007,
760: \apj, 660, 1060
761: \bibitem[Vanden Berk et al.(2001)]{vandenberk01} Vanden Berk D.E., et al. 2001,
762: \aj, 122, 549
763: \bibitem[Venturini \& Solomon(2003)]{venturini03} Venturini, S., Solomon, P.M.
764: 2003, \apj, 590, 740
765: \bibitem[Weedman et al.(2005)]{weedman05} Weedman, D.W., et al. 2005, \apj,
766: 633, 706
767: \bibitem[Weedman et al.(2006)]{weedman06} Weedman, D.W., Le Floc'h, E.,
768: Higdon, S.J.U., Higdon, L.J., Houck, J.R. 2006, \apj, 638, 613
769: \bibitem[Wei\ss\ et al.(2003)]{weiss03} Wei\ss\/, A., Henkel, C., Downes, D.,
770: Walter, F. 2003, \aap, 409, L41
771: \bibitem[Yan et al.(2007)]{yan07} Yan, L., et al. 2007, \apj, 658, 778
772: \end{thebibliography}
773:
774: \clearpage
775:
776: \begin{deluxetable}{llrrrlc}
777: \tablecolumns{7}
778: \tablecaption{Sample}
779: \tablehead{
780: \colhead{Name} &
781: \colhead{z} &
782: \colhead{RA} &
783: \colhead{DEC} &
784: \colhead{$\mu_{L}$}&
785: \colhead{Ref.}&
786: \colhead{log($L_{5100}$)}\\
787: \colhead{}&
788: \colhead{}&
789: \colhead{J2000.0}&
790: \colhead{J2000.0}&
791: \colhead{}&
792: \colhead{}&
793: \colhead{erg/s}
794: }
795: \startdata
796: LBQS 0018-0220 &2.596& 0:21:27.30&$-$2:03:33.0& 1.0& &46.82\\
797: HE 0230-2130 &2.162& 2:32:33.10&$-$21:17:26.0&14.5&B02&45.29\\
798: SMMJ04315+10277 &2.837& 4:13:27.26& +10:27:40.5& 1.3&K03&45.70\\
799: KUV 0806+4037 &1.795& 8:12:00.54& +40:28:13.0& 1.0& &46.30\\
800: RXJ0911.4+0551 &2.793& 9:11:27.61& +5:50:54.1&20.0&B02&45.32\\
801: HS 1002+4400 &2.097&10:05:17.43& +43:46:09.3& 1.0& &46.84\\
802: HE 1104-1805 &2.305&11:06:33.40&$-$18:21:23.0&10.8&B02&45.70\\
803: LBQS 1230+1627B &2.735&12:33:10.40& +16:10:52.0& 1.0& &46.89\\
804: SBS 1408+567 &2.583&14:09:55.57& +56:28:26.5& 1.0& &46.89\\
805: H1413+117 &2.558&14:15:46.27& +11:29:43.4&11.0&V03&46.11\\
806: HS1611+4719 &2.35 &16:12:39.90& +47:11:57.0& 1.0& &46.61\\
807: J164914.9+530316&2.26 &16:49:14.90& +53:03:16.0& 1.0& &46.89\\
808: \enddata
809: \label{tab:sample}
810: \tablerefs{\ Lensing magnifications $\mu_{L}$ are adopted from: B02: \citet{barvainis02},
811: K03: \citet{knudsen03}, V03: \citet{venturini03}. Throughout the paper,
812: we present fluxes or flux densities as observed, but luminosities, masses, and
813: star formation rates corrected for these adopted magnifications.}
814: \end{deluxetable}
815:
816: \clearpage
817:
818: \begin{deluxetable}{lrrrrrrr}
819: \tabletypesize{\scriptsize}
820: \tablecolumns{7}
821: \tablecaption{Results and host starburst properties}
822: \tablehead{
823: \colhead{Name} &
824: \colhead{Peak (PAH7.7)} &
825: \colhead{Noise\tablenotemark{a}} &
826: \colhead{F (PAH7.7)} &
827: \colhead{L (PAH7.7)} &
828: \colhead{$\nu L_\nu (60\mu m)$}&
829: \colhead{SFR (PAH)}&
830: \colhead{M (H$_2$)}\\
831: \colhead{}&
832: \colhead{mJy}&
833: \colhead{mJy}&
834: \colhead{$10^{-21}$W\,cm$^{-2}$}&
835: \colhead{$10^{44}$erg s$^{-1}$}&
836: \colhead{$10^{44}$erg s$^{-1}$}&
837: \colhead{\Msun yr$^{-1}$}&
838: \colhead{10$^{10}$\Msun}
839: }
840: \startdata
841: LBQS 0018-0220 &$<$1.9&0.27&$<$2.3&$<$12.0&543&$<$4200&$<$3.5\tablenotemark{b}\\
842: HE 0230-2130 & 1.7&0.21& 2.3& 0.6& 48& 197&\\
843: SMMJ04315+10277 & 0.8&0.20& 0.8& 4.8&615& 1680&13.3\tablenotemark{b}\\
844: KUV 0806+4037 & 1.1&0.30& 1.7& 3.7&392& 1300&\\
845: RXJ0911.4+0551 & 2.7&0.21& 3.0& 1.0& 43& 340&0.5\tablenotemark{b}\\
846: HS 1002+4400 & 1.1&0.35& 1.6& 5.0&337& 1760&\\
847: HE 1104-1805 & 2.8&0.36& 3.8& 1.4& 43& 496&\\
848: LBQS 1230+1627B & 1.2&0.28& 1.4& 8.1&609& 2840&2.4\tablenotemark{c}\\
849: SBS 1408+567 &$<$3.0&0.32&$<$3.6&$<$19.6&888&$<$6840&6.6\tablenotemark{b}\\
850: H1413+117 & 5.1&0.75& 6.1& 2.9&147& 1020&4.0\tablenotemark{d}\\
851: HS1611+4719 & 1.0&0.19& 1.2& 5.3&394& 1840&\\
852: J164914.9+530316&$<$1.5&0.24&$<$2.0&$<$ 7.9&399&$<$2780&\\
853: \enddata
854: \tablenotetext{a}{Measured noise per spectral sample in the rest frame 6.5--8$\mu$m range}
855: \tablenotetext{b}{\citet{hainline04}}
856: \tablenotetext{c}{\citet{guilloteau99}}
857: \tablenotetext{d}{\citet{weiss03}}
858: \label{tab:results}
859: \end{deluxetable}
860:
861:
862: \clearpage
863:
864: \begin{figure}
865: \epsscale{1.}
866: \plotone{f1.eps}
867: \caption{Spitzer-IRS spectra of the sample of (sub)mm-bright high-z QSO1s.
868: Vertical lines indicate the location of the 6.2 and 7.7$\mu$m PAH features at
869: the redshift of the sources with PAH detections.}
870: \label{fig:allhighz}
871: \end{figure}
872:
873: \clearpage
874:
875: \begin{figure}
876: \epsscale{1.}
877: \plotone{f2.eps}
878: \caption{Decomposition of IRS spectra. The individual panels are
879: shown in the same ordering and scaling as in Fig.~\ref{fig:allhighz}, with the
880: thin black lines retracing the spectra as observed. Thick blue lines show
881: the AGN continuum and silicate emission remaining after subtraction of the PAH
882: component obtained in the decomposition. The green line indicates the 1$\sigma$
883: noise measured in IRS data obtained with same observing setup and integration
884: time but lacking a source (one of the nondetections among the SMGs of
885: \citet{valiante07}). While noise from this curve in good agreement with
886: expected IRS
887: sensitivity and measured noise for our continuum-weaker QSOs and gives an
888: indication of
889: trends with wavelength, it misses possible sources of noise in spectra
890: with bright continuum that can arise e.g. due to more
891: difficult identification of outlying values. Instead, we conservatively
892: present in Table~\ref{tab:results} noise measured in the PAH region of the
893: spectra after subtraction of the PAH template. This noise is the RMS after
894: subtracting a linear fit to the rest wavelength 6.5--8$\mu$m region (before
895: the main upturn of silicate emission).}
896: \label{fig:allhighzdecomp}
897: \end{figure}
898:
899: \clearpage
900:
901: \begin{figure}
902: \epsscale{.70}
903: \plotone{f3.eps}
904: \caption{Average rest frame mid-IR spectrum of the sample high-z QSO1s. The
905: individual spectra have been scaled before averaging to the same flux of
906: 15mJy at rest frame wavelength 222$\mu$m. This allows a direct comparison
907: of spectrum and the relation of PAH and far-infrared emission to the stacked
908: submm galaxy spectrum
909: presented by \citet{valiante07}. A direct average of the QSO spectra has
910: a very similar shape. A
911: scaled average spectrum of the local PG QSOs studied by \citet{netzer07} is
912: shown for comparison.}
913: \label{fig:averagehighz}
914: \end{figure}
915:
916: \clearpage
917:
918: \begin{figure}
919: \epsscale{.70}
920: \plotone{f4.eps}
921: \caption{Ratio of PAH peak flux density and rest frame 222$\mu$m far-infrared
922: continuum for the sample QSOs (top panel). z$\sim$2 SMGs and local ULIRGs
923: are added for comparison in the lower panels
924: \citep[SMGs and ULIRGs adopted from][]{valiante07}.}
925: \label{fig:pah222}
926: \end{figure}
927:
928: \clearpage
929:
930: \begin{figure}
931: \epsscale{.80}
932: \plotone{f5.eps}
933: \caption{Relation of the PAH 7.7$\mu$m luminosity and far-infrared continuum
934: luminosities for our sample and for local comparison objects. PG QSOs as
935: well as the comparison sample of starbursting ULIRGs are from
936: the studies of \citet{schweitzer06} and \citet{netzer07}. For lensed
937: high-z objects,
938: luminosities have been corrected for the magnifications listed in
939: Table~\ref{tab:sample}.}
940: \label{fig:pahtofir}
941: \end{figure}
942:
943:
944: \clearpage
945:
946: \begin{figure}
947: \epsscale{.80}
948: \plotone{f6.eps}
949: \caption{Top panel: Relation of PAH 7.7$\mu$m luminosity and AGN 5100\AA\
950: continuum
951: luminosity. The large `Average high-z QSO' upper limit symbol refers to
952: the limit derived by
953: \citet{maiolino07} for the average spectrum of a high redshift QSO sample
954: which is not selected for bright (sub)mm emission. Bottom panel: To help
955: identifying trends, the same data are shown plotting the ratio of
956: PAH 7.7$\mu$m luminosity and AGN 5100\AA\ continuum luminosity as a function
957: of AGN 5100\AA\ continuum luminosity.}
958: \label{fig:pahto5100}
959: \end{figure}
960:
961: \clearpage
962:
963: \begin{figure}
964: \epsscale{.75}
965: \plotone{f7.eps}
966: \caption{Top panel: Relation of 60$\mu$m continuum luminosity and AGN 5100\AA\
967: continuum luminosity. The 60$\mu$m luminosities of high redshift QSOs are
968: extrapolated from their (sub)mm photometry using a T=40K $\beta$=1.5 greybody
969: approximation. The green hexagon indicates the locus for the median of the
970: full z$\sim$2 L$_{5100}>10^{46}$erg\,s$^{-1}$ population. Bottom panel: To
971: help identifying trends, the same data are shown plotting the ratio of
972: FIR luminosity and AGN 5100\AA\ continuum luminosity as a function
973: of AGN 5100\AA\ continuum luminosity. `Other' high-z QSOs refers to objects
974: in the studies of \citet{omont03}, \citet{priddey03}, and \citet{barvainis02}
975: that are not part of our mm-bright sample.}
976: \label{fig:firto5100}
977: \end{figure}
978:
979: \end{document}
980: