1:
2: %% --------------------------------------------------------------------
3: %% Thu Feb 12 21:42:15 2004
4: %% This file was generated automagically from the files
5: %% quadmicrolens.bbl and quadmicrolens.tex using
6: %% /scr0/gtr/astronat/nat2jour.pl
7: %% All citations have been inlined and dependencies on the natbib
8: %% package have been removed so that this file (together with
9: %% quadmicrolens-aas.bbl) should be suitable for submission to journals with
10: %% the citation styles of ApJ or MNRAS.
11: %% --------------------------------------------------------------------
12:
13: \documentclass[12pt,preprint]{aastex}
14: %\usepackage{natbib}
15: %\citestyle{aa}
16:
17: \newcommand\ReinL{R_{\rm Ein,L}}
18: \newcommand\ReinS{R_{\rm Ein,S}}
19:
20: \received{}
21: \accepted{}
22: %\journalid{}{}
23: %\articleid{}{}
24:
25: \slugcomment{12 February 2004; submitted to ApJ}
26:
27: \author{
28: Gordon T. Richards,\altaffilmark{1}
29: Charles R. Keeton,\altaffilmark{2,3}
30: Bartosz Pindor,\altaffilmark{1,4}
31: Joseph F. Hennawi,\altaffilmark{1}
32: Patrick B. Hall,\altaffilmark{1}
33: Edwin L. Turner,\altaffilmark{1}
34: Naohisa Inada,\altaffilmark{5}
35: Masamune Oguri,\altaffilmark{5}
36: Shin-Ichi Ichikawa,\altaffilmark{6}
37: Robert H. Becker,\altaffilmark{7,8}
38: Michael D. Gregg,\altaffilmark{7,8}
39: Richard L. White,\altaffilmark{9}
40: J. Stuart B. Wyithe,\altaffilmark{1,10}
41: Donald P. Schneider,\altaffilmark{11}
42: David E. Johnston,\altaffilmark{1,2,12}
43: Joshua A. Frieman,\altaffilmark{2,12,13}
44: and J. Brinkmann\altaffilmark{14}
45: }
46:
47: \altaffiltext{1}{Princeton University Observatory, Peyton Hall, Princeton, NJ 08544.}
48: \altaffiltext{2}{Department of Astronomy and Astrophysics, The University of Chicago, 5640 South Ellis Avenue, Chicago, IL 60637.}
49: \altaffiltext{3}{Hubble Fellow.}
50: \altaffiltext{4}{Canadian Institute for Theoretical Astrophysics, University of Toronto, 60 St.\ George Street, Toronto, Ontario, M5S 3H8, Canada.}
51: \altaffiltext{5}{Department of Physics, University of Tokyo, 7-3-1 Hongo, Bunkyo, Tokyo 113-0033, Japan.}
52: \altaffiltext{6}{National Astronomical Observatory, 2-21-1 Osawa, Mitaka, Tokyo 181-8588, Japan.}
53: \altaffiltext{7}{Physics Department, University of California, Davis, CA 95616.}
54: \altaffiltext{8}{IGPP-LLNL, L-413, 7000 East Avenue, Livermore, CA 94550.}
55: \altaffiltext{9}{Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218.}
56: \altaffiltext{10}{University of Melbourne, Parkville, VIC 3010, Australia.}
57: \altaffiltext{11}{Department of Astronomy and Astrophysics, The Pennsylvania State University, 525 Davey Laboratory, University Park, PA 16802.}
58: \altaffiltext{12}{Center for Cosmological Physics, The University of Chicago, 5640 South Ellis Avenue Chicago, IL 60637.}
59: \altaffiltext{13}{Fermi National Accelerator Laboratory, P.O. Box 500, Batavia, IL 60510.}
60: \altaffiltext{14}{Apache Point Observatory, P.O. Box 59, Sunspot, NM 88349.}
61:
62: \begin{document}
63:
64: \title{Microlensing of the Broad Emission Line Region in the
65: Quadruple Lens SDSS~J1004+4112}
66:
67: \begin{abstract}
68:
69: We present seven epochs of spectroscopy on the quadruply imaged quasar
70: SDSS~J1004+4112, spanning observed-frame time delays from 1 to 322
71: days. The spectra reveal differences in the emission lines between
72: the lensed images. Specifically, component A showed a strong
73: enhancement in the blue wings of several high-ionization lines
74: relative to component B, which lasted at least 28 days (observed
75: frame) then faded. Since the predicted time delay between A and B is
76: $\lesssim$30 days, our time coverage suggests that the event was not
77: intrinsic to the quasar. We attribute these variations to
78: microlensing of part of the broad emission line region of the quasar,
79: apparently resolving structure in the source plane on a scale of
80: $\sim\!10^{16}\,{\rm cm}$ at $z=1.734$. In addition, we observed
81: smaller differences in the emission line profiles between components A
82: and B that persisted throughout the time span, which may also be due
83: to microlensing or millilensing. Further spectroscopic monitoring of
84: this system holds considerable promise for resolving the structure of
85: the broad emission line region in quasars.
86:
87: \end{abstract}
88:
89: \keywords{gravitational lensing --- quasars: general ---
90: quasars: emission lines --- quasars: individual (SDSS~J100434.91+411242.8)}
91:
92: \section{Introduction}
93:
94: Microlensing in the images of a multiply-imaged quasar was first
95: reported by \markcite{iwh+89}{Irwin} {et~al.} (1989) for the quadruple lens Q2237+0305
96: \markcite{huchra85}({Huchra} {et~al.} 1985). Most quasar microlensing studies have been based on
97: broad-band photometric monitoring \markcite{wus+00,sus+03,cs03}(e.g., {Wo{\' z}niak} {et~al.} 2000; {Schechter} {et~al.} 2003; {Colley} \& {Schild} 2003),
98: which is sensitive primarily to variations in the continuum.
99: Microlensing of the continuum is expected since the optical/UV
100: continuum emission is thought to originate in a region that is
101: comparable in size to the Einstein radius of a typical star in a
102: typical lens galaxy.
103:
104: Microlensing of the broad emission line region (BELR) is also
105: possible, if the BELR has structure on scales comparable to the
106: Einstein radius of a star \markcite{nem88,sw90}({Nemiroff} 1988; {Schneider} \& {Wambsganss} 1990). The possibility of BELR
107: microlensing seemed rather remote until recent reverberation mapping
108: work revised the estimate of the BELR size downward from
109: $\sim\!10^{18}$ cm to $\sim\!10^{16}$ cm \markcite{wpm99,ksn+00}({Wandel}, {Peterson}, \& {Malkan} 1999; {Kaspi} {et~al.} 2000).
110: Inspired by these numbers, \markcite{amm+02}{Abajas} {et~al.} (2002) and \markcite{la03}{Lewis} \& {Ibata} (2003) revived the
111: idea of looking for microlensing of the BELR and computed possible
112: line profile variations for various BELR models.
113:
114: Possible examples of microlensing of a quasar emission line have been
115: presented by \markcite{fil89}{Filippenko} (1989) for Q2237+0305, and by \markcite{cea+04}{Chartas et al.} (2003) for
116: H1413+117. In particular, \markcite{cea+04}{Chartas et al.} (2003) detected a strong, redshifted
117: Fe K$\alpha$ emission line in the X-ray spectrum of only one of the
118: components in the quadruple lens H1413+117. Although they did not have
119: multiple epochs to look for variability, \markcite{cea+04}{Chartas et al.} (2003) invoked a short
120: predicted time delay between the components to argue that microlensing
121: is the preferred explanation for seeing the Fe K$\alpha$ line in only
122: one component.
123:
124: In this paper we present results from spectroscopic monitoring of the
125: recently discovered quadruple lens SDSS~J1004+4112
126: \markcite{inada03,oguri03}({Inada et al.} 2003; {Oguri et al.} 2003). We observe variability in the broad emission
127: line profiles of one of the lensed images that provides strong
128: evidence for microlensing of the BELR, suggesting that the theoretical
129: predictions for microlensing were correct and confirming that the BELR
130: has structure on the scale of the Einstein radius of a star.
131:
132: \section{The Data and Spectral Analysis}
133:
134: Component B of SDSS~J1004+4112 was first identified as a quasar by
135: \markcite{cwh99}{Cao}, {Wei}, \& {Hu} (1999) and was also targeted as a quasar candidate
136: \markcite{rfn+02}({Richards} {et~al.} 2002) as part of the Sloan Digital Sky Survey (SDSS;
137: \markcite{yaa+00}{York} {et~al.} 2000); the SDSS spectrum was taken on 2003 February 3 and
138: the object was identified as a $z=1.734$ quasar. Component A was also
139: identified as a possible quasar candidate based on its colors in the
140: SDSS imaging data \markcite{fig+96,gcr+98}({Fukugita} {et~al.} 1996; {Gunn} {et~al.} 1998); it was first confirmed as a
141: quasar (also at $z=1.734$) from observations taken on 2003 May 3 using
142: the ARC 3.5m telescope at Apache Point Observatory. Higher quality
143: spectra of each of the four components and the lens galaxy were taken
144: on 2003 May 31 with the LRIS \markcite{occ+95}({Oke} {et~al.} 1995) spectrograph on the Keck I
145: telescope at the W.~M.~Keck Observatory. These spectra are further
146: described in \markcite{inada03}{Inada et al.} (2003), \markcite{oguri03}{Oguri et al.} (2003), and
147: Table~\ref{tab:tab1}. Table~\ref{tab:tab1} gives the UT date of the
148: observations, the telescope used, the components observed, the
149: exposure time, the spectral dispersion, and the wavelength range
150: covered.
151:
152: \markcite{inada03}{Inada et al.} (2003) and \markcite{oguri03}{Oguri et al.} (2003) noted that the Keck/LRIS spectra
153: showed that all four components had the same redshift and similar
154: spectra, but there are some obvious differences --- with component A
155: showing the largest differences relative to the other components (see
156: Fig.~\ref{fig:fig1}). Specifically, component A has a much stronger
157: blue emission line wing in the high-ionization lines
158: (\ion{Si}{4}/\ion{O}{4}], \ion{C}{4}, and \ion{He}{2}). Component B
159: appears to have a slightly enhanced red wing as compared with the
160: other three components. In Figure~\ref{fig:fig1}, we have sought to
161: emphasize the differences in the \ion{C}{4} emission line profile by
162: subtracting a continuum fit between $\lambda\,1450$~\AA\ and
163: $\lambda\,1690$~\AA, and then normalizing to the peak of \ion{C}{4}
164: emission. Other choices in presentation could enhance or reduce the
165: appearance of similarities and differences between the components.
166:
167: Based on the puzzling differences in the broad emission line profiles
168: of the four components of the Keck/LRIS spectra, in 2003 November and
169: 2003 December we obtained additional spectra of images A and B to
170: monitor the differences; see Table~\ref{tab:tab1}. We used the
171: DIS\,III spectrograph on the ARC 3.5m telescope at Apache Point
172: Observatory, using the same instrument setup and data reduction for
173: each epoch. The dispersion was $2.4\,{\rm \AA\,pixel}^{-1}$ and the
174: spectra cover the range $\lambda\lambda 3890$--$9350$~\AA. Flux
175: calibration was performed with respect to a hot standard star: either
176: Feige~34 or G~191-B2B. Wavelength calibration was performed with
177: respect to a helium-neon-argon comparison lamp. Both components
178: (which are separated by $3\farcs73$) were observed at the same time in
179: order to minimize any differences between them, at the expense of loss
180: of spectrophotometric accuracy (slit not positioned at the parallactic
181: angle). At each epoch we took two spectra of either 40 or 45 minutes
182: each. These spectra were extracted separately and then combined to
183: yield 2003 November/December epoch spectra shown in
184: Figure~\ref{fig:fig2}. The 2D spectra are only moderately resolved,
185: so we deblended the 2D spectra by fitting a double Gaussian profile
186: before extracting the spectra. We estimate that contamination of the
187: faint component by the brighter component is less than 3\%.
188:
189: Figure~\ref{fig:fig2} shows all seven epochs of spectroscopic data for
190: components A and B in the \ion{C}{4} emission line region. The 2003
191: May 3 epoch observation of component A at APO confirms the reality of
192: the excess emission in the blue wing of component A in the Keck
193: spectrum. The excess emission in \ion{He}{2} spans a velocity range
194: $\sim$2500--8500~km~s$^{-1}$ (blueshifted). Both \ion{Si}{4} and
195: \ion{C}{4} have excesses over a slightly higher velocity range. The
196: excess in \ion{He}{2} is much stronger in terms of equivalent width
197: than that in \ion{C}{4}.
198:
199: \section{Discussion}
200:
201: \subsection{Time-dependent Blue Wing Differences: Microlensing of the BELR}
202:
203: \subsubsection{Discounting Alternatives and Objections}
204:
205: To argue that the excess in the blue wings of the high-ionization
206: emission lines of component A is caused by microlensing of the BELR,
207: we must rule out alternative explanations. The first question is
208: whether the flux enhancements are an artifact of the data acquisition
209: or reduction procedures. We view this explanation as highly unlikely
210: because the enhancement appeared at two different epochs, in spectra
211: taken by different observers with different telescopes, and reduced
212: independently.
213:
214: The second question is whether the spectral variations could be
215: intrinsic to the quasar, rather than induced by lensing. If so, then
216: the same variations should be seen in the other lensed images, offset
217: in time by the lens time delays. While the time delays are still
218: uncertain, in nearly all of the lens models presented by
219: \markcite{oguri03}{Oguri et al.} (2003) the delay between components A and B is predicted to
220: be $\lesssim$30 days. (The Oguri et al.\ models do not form an
221: exhaustive set, but the prediction of a short time delay between A and
222: B is generic.) In $\sim$10\% of the models component A leads component
223: B. In this case seeing the flux enhancement in the 2003 May 3 and
224: 2003 May 31 spectra of component A but not in the 2003 May 31 spectrum
225: of component B means that intrinsic variability cannot explain the
226: data. In the other $\sim$90\% of the models component B leads
227: component A. In this case, the intrinsic variability hypothesis would
228: imply that the variations must have appeared in component B sometime
229: after 2003 February 3, been present during 2003 April (in advance of
230: the variations observed in component A in 2003 May), and then
231: disappeared before 2003 May 31. In other words, the event must have
232: lasted $<$117 days in the observed frame, or $<$43 days in the rest
233: frame --- and we were fortunate to catch the event in component A just
234: before it disappeared.
235:
236: Since the time separation of our spectral coverage does not completely
237: rule out intrinsic variability, this issue deserves further
238: discussion. Specifically, we must address the likelihood of intrinsic
239: variability of the BELR on timescales of less than 43 days in the rest
240: frame. Reverberation timescales can be less than 43 days since the
241: radius of the BELR is likely to be on the order of or less than this
242: size. However, in that case the entire emission line profile will
243: vary with respect to any significant change in the continuum. Since
244: we do not observe a significant change in the continuum level (see
245: below) and since the enhancement of the emission lines is only in the
246: blue wing and not over the entire profile, a reverberation effect is
247: unlikely. Thus we are left with the possibility of a dynamical change
248: --- such as is seen in the so-called double-peaked emission line
249: quasars \markcite{eh03,ssh+03}(e.g., {Eracleous} \& {Halpern} 2003; {Strateva} {et~al.} 2003) --- which could better explain
250: the blue-wing-only nature of the enhancement in the high-ionization
251: emission lines. Indeed the dynamical timescale ($\sim\!6$\,months;
252: \markcite{era04}{Eracleous} 2003) can be in the range needed to explain our
253: observations. However, double-peaked emission is typically only seen
254: in the Balmer lines (and sometimes \ion{Mg}{2}), and is absent or weak
255: at best in high-ionization lines like \ion{C}{4} and \ion{He}{2}.
256: Furthermore, it would be very unusual for a double-peaked emission
257: line object to show a blue-wing enhancement in both our 2003 May 3 and
258: 2003 May 31 epochs, but no strong blue- or red-wing enhancement of our
259: four 2003 November/December epochs. Thus, although an explanation of
260: the data in terms of intrinsic variability is not rigorously excluded,
261: such a model would have to be rather improbably contrived.
262:
263: One possible objection to the microlensing hypothesis is if the BELR
264: was microlensed, why not the continuum as well? It is difficult to
265: put constraints on any enhancement of the continuum of component A
266: relative to B during the time spanned by our observations, because not
267: all of the spectra were taken at the parallactic angle. Still, we can
268: estimate that component A was no more than $\sim$20\% brighter than B
269: in the continuum in 2003 May (as compared to 2003 November/December).
270: This is not obviously inconsistent with the microlensing hypothesis,
271: however. It is easy to imagine configurations in which part of the
272: BELR is close enough to the caustic in the source plane of a star in
273: the lens to be microlensed, while the continuum source was far enough
274: to feel little effect. In fact, this picture is consistent with our
275: hypothesis that only {\em part} of the BELR was being microlensed in
276: 2003 May (see \S~\ref{sec:belr}).
277:
278: Another possible complication with the microlensing hypothesis is that
279: cluster galaxies tend to be stripped of their halos through mergers
280: and interactions. Microlensing would therefore require either a
281: cluster member very close to the line of sight to component A or a
282: population of intracluster stars or massive compact halo objects
283: (MACHOs; \markcite{tot03,blz+04}e.g., {Totani} 2003; {Baltz et al.} 2003). The presence of
284: intracluster MACHOs might not be surprising, because tidal forces
285: would naturally strip MACHOs from cluster galaxies along with the
286: galaxy halos. On the other hand, there is evidence for a galaxy
287: superimposed on component A that could host the microlensing object,
288: as shown in Figure~\ref{fig:fig3}. Proper PSF subtraction to confirm
289: this hypothesis is not possible since component A is saturated in this
290: image, but our best efforts do reveal residual flux at the center of
291: the circle in Figure~\ref{fig:fig3} with an estimated magnitude of
292: $i\sim24.5$.
293:
294: \subsubsection{Examining the Microlensing Hypothesis}
295:
296: We can also reverse the argument and ask if our microlensing
297: hypothesis makes sense given our current understanding of the
298: structure of quasars and the details of microlensing. Without
299: invoking detailed microlensing and BELR models (which would be
300: premature since we have observed only one microlensing event), we
301: still find that qualitative and quantitative arguments reveal good
302: consistency between the microlensing hypothesis and the data.
303:
304: The scale for microlensing is given by the Einstein radius of a star
305: \markcite{sef92}({Schneider}, {Ehlers}, \& {Falco} 1992),
306: \begin{eqnarray}
307: \ReinS &=& \left(4\frac{GM}{c^2}\frac{D_{s} D_{ls}}{D_{l}}\right)^{1/2} \\
308: \ReinL &=& \frac{D_{l}}{D_{s}}\,\ReinS
309: \end{eqnarray}
310: where $\ReinL$ and $\ReinS$ are the Einstein radius projected into the
311: lens and source planes, respectively, and $D_{l}$, $D_{s}$, and
312: $D_{ls}$ are the angular diameter distances to the lens, to the
313: source, and from the lens to the source, respectively. Microlensing
314: is said to occur at low optical depth if the mean separation between
315: stars is $d \gg \ReinL$ and stars contribute a small fraction of the
316: surface mass density, or at high optical depth if $d \sim \ReinL$ and
317: stars contribute a substantial fraction of the surface mass density.
318: The microlensing probability can be high even if the optical depth is
319: low \markcite{sw02}({Schechter} \& {Wambsganss} 2002). Regardless of whether the optical depth is high or
320: low, the Einstein radius sets the characteristic scale, and the BELR
321: can be microlensed if it has structure on scales $\lesssim\!\ReinS$
322: \markcite{amm+02,la03}({Abajas} {et~al.} 2002; {Lewis} \& {Ibata} 2003). For SDSS~J1004+4112, the lens redshift is
323: $z_l=0.68$ and the source redshift is $z_s=1.734$, so the Einstein
324: radius of a $0.1\,M_{\sun}$ star associated with the lensing cluster
325: is
326: \begin{equation}
327: \ReinS
328: \sim 1.4\times10^{16} \left(\frac{M}{0.1M_{\sun}}\right)^{1/2}
329: h_{70}^{-1/2} \mbox{ cm}
330: \sim 5.3 \left(\frac{M}{0.1M_{\sun}}\right)^{1/2}
331: h_{70}^{-1/2} \mbox{ lt-days}
332: \end{equation}
333: in a cosmology with $\Omega_M=0.3$, $\Omega_\Lambda=0.7$, and
334: $H_0=70\,h_{70}\,{\rm km\,s^{-1}\,Mpc^{-1}}$. Thus, the typical
335: stellar Einstein radius is indeed comparable to the currently favored
336: size of the BELR, and microlensing is not unexpected.
337:
338: The variability is caused by relative motion between the caustic
339: network and the source. If the motion is dominated by the proper
340: motion of the lens galaxy (with transverse velocity $v_{\perp}$), then
341: the effective transverse velocity projected into the source plane and
342: expressed in distance per unit observed-frame time is
343: \begin{equation}
344: v_{\rm eff} = \frac{v_{\perp}}{1+z_{l}}\,\frac{D_{s}}{D_{l}}\ .
345: \end{equation}
346: If the microlensing is dominated by caustic crossings, then the
347: characteristic event duration is the time for the caustic to sweep
348: across the source, $t_{\rm src} \sim 2 R_{\rm src} / v_{\rm eff}$. If
349: we estimate $v_{\perp} \sim \sigma$ where $\sigma$ is the velocity
350: dispersion of the lens, then $\sigma \sim 700$~km~s$^{-1}$ for the
351: lensing cluster in SDSS~J1004+4112 \markcite{oguri03}({Oguri et al.} 2003) yields
352: \begin{equation}
353: t_{\rm src} \sim 12.7 \left(\frac{R_{\rm src}}{10^{16}\mbox{ cm}}\right)
354: \left(\frac{v_{\perp}}{700\mbox{ km s}^{-1}}\right)^{-1}
355: \mbox{ yr}
356: \end{equation}
357: The fact that we see variability in the broad emission lines on a time
358: scale of $\sim$6 months suggests either that we happened to catch the
359: end of a long event, or that microlensing is affecting only part of
360: the BELR with a characteristic size smaller than $10^{16}$~cm (or
361: both).
362:
363: A second interesting time scale is the typical time between
364: microlensing events. Here the key unit is the time to cross an
365: Einstein radius,
366: \begin{eqnarray}
367: t_{\rm Ein} &\sim& \ReinS / v_{\rm eff} \;\; \sim \; 8.6 \left(\frac{M}{0.1M_{\sun}}\right)^{1/2}
368: \left(\frac{v_{\perp}}{700\mbox{ km s}^{-1}}\right)^{-1}
369: h_{70}^{-1/2}\mbox{ yr}
370: \end{eqnarray}
371: If even a few percent of the mass is in stars, microlensing is
372: associated with a caustic network rather than a single star, and so
373: naive estimates of the time between events are difficult. To obtain a
374: better estimate, we have used the standard ray shooting technique to
375: compute microlensing magnification maps and generate sample light
376: curves \markcite{krs86,wpk90}({Kayser}, {Refsdal}, \& {Stabell} 1986; {Wambsganss}, {Paczynski}, \& {Katz} 1990). We find that starting from a
377: non-microlensed position, the average wait time until the
378: magnification changes by $\ge$30\% is $\sim\!(0.2\mbox{--}0.8) \times
379: t_{\rm Ein}$. The range represents uncertainties in the source size
380: and in the number density of microlenses. Although there are many
381: uncertainties, it seems reasonable to expect that microlensing events
382: could be observed in SDSS~J1004+4112 on an approximately yearly basis.
383: Incidentally, if $t_{\rm src} \gtrsim t_{\rm Ein}$ then microlensing
384: events will blur together and the light curves will show continuous
385: smooth variations \markcite{koc04}(see, e.g., {Kochanek} 2003). The apparent lack of
386: microlensing in the 2003 November/December epochs therefore adds
387: further support to the hypothesis that the region being microlensed is
388: smaller than $\sim\!10^{16}$~cm.
389:
390: The microlensing hypothesis nicely explains one of the more
391: interesting observational results, namely that different emission
392: lines show different amounts of variability. From reverberation
393: mapping results, we know that the BELR is stratified by ionization and
394: that higher ionization lines are found closer to the center
395: \markcite{pw99}({Peterson} \& {Wandel} 1999). As a result, the highest ionization regions have the
396: smallest effective sizes and should be the most sensitive to
397: microlensing. Our finding that the excess in the \ion{He}{2} line is
398: stronger than the excess in the \ion{C}{4} line (in terms of
399: equivalent width) is consistent with their relative reverberation
400: mapping sizes \markcite{pw99}({Peterson} \& {Wandel} 1999). Furthermore, there is a suggestion of a
401: weaker excess in the blue wings of the
402: \ion{C}{3}]/\ion{Si}{3}]/\ion{Al}{3}/FeIII\,UV34 complex, \ion{Mg}{2},
403: and \ion{Fe}{3} UV48 (a triplet at $\lambda\lambda\lambda$ 2062.211,
404: 2068.904, 2079.652); see Figure~\ref{fig:fig4}, and also the ratio
405: spectra in Figure~3 of \markcite{oguri03}{Oguri et al.} (2003). All the above lines are lower
406: ionization lines than \ion{C}{4} and \ion{He}{2}, so their weaker
407: excesses are also consistent with being more weakly microlensed as a
408: result of their emitting regions being larger with respect to the
409: projected Eistein radius of the lens.
410:
411: An obvious question is why should microlensing of the BELR be seen in
412: this lens system, but not in others? One possibility has to do with
413: the quasar's intrinsic luminosity. The observed absolute $i$
414: magnitude of component A is $M_i=-26.9$, but this component is
415: amplified by a factor of $\sim$20 or more,\footnote{An isothermal
416: ellipsoid plus shear lens model, which is simple but fits the data
417: well, implies an amplification factor of 23 for component A. More
418: complicated models yield a broad range of amplifications where the
419: median is 18 but there is a long tail to amplifications of 100 or
420: more. See \markcite{oguri03}{Oguri et al.} (2003) for modeling details.} so the intrinsic
421: absolute $i$ magnitude is $M_i \gtrsim -23.8$. Many other lensed
422: quasars have similar observed magnitudes but smaller amplifications,
423: and hence higher luminosities. Since the size of the BELR scales as
424: the 0.5--0.7 power of the luminosity \markcite{ksn+00}(e.g., {Kaspi} {et~al.} 2000),
425: SDSS~J1004+4112 being relatively under-luminous would make it more
426: sensitive to microlensing than many other lensed quasars.
427:
428: The large amplification helps make microlensing unsurprising in
429: another way. Large amplifications are associated with significant
430: distortions that increase the size of the caustics (see Fig.~2 of
431: \markcite{sw02}{Schechter} \& {Wambsganss} 2002). Thus, with a given star field the microlensing
432: probability increases as the amplification increases. This may
433: explain not only why microlensing has been detected in
434: SDSS~J1004+4112, but also why it was seen first in component A (the
435: highest-amplification component).
436:
437: \subsubsection{Implications for BELR Structure\label{sec:belr}}
438:
439: Obtaining concrete constraints on the structure of the BELR will
440: require detailed modeling, and would greatly benefit from observations
441: of additional microlensing events. Nevertheless, combining the data
442: in hand with general arguments already permits some strong and
443: valuable conclusions. First, unless the number of clouds is orders of
444: magnitude smaller than required by other high-resolution Keck
445: observations \markcite{abl98}({Arav} {et~al.} 1998), the asymmetric nature of the microlensing
446: (lensing of the blue wing only) rules out a spherically symmetric
447: distribution of dynamically virialized, thermal line-width clouds
448: \markcite{nem88,amm+02,la03}({Nemiroff} 1988; {Abajas} {et~al.} 2002; {Lewis} \& {Ibata} 2003).
449:
450: Pure radial outflow models can be ruled out because they would produce
451: symmetric line profile changes with most of the variation in the line
452: core \markcite{nem88}({Nemiroff} 1988). Pure radial inflow models cannot be fully
453: excluded, but asymmetric microlensing would require the accretion disk
454: to have a radial extent comparable to the radial extent of the BELR
455: and to be optically thick (such that we do not see clouds at all
456: velocities).
457:
458: The asymmetry therefore seems to imply that a strong rotational
459: component is needed in the high-ionization region of the BELR. Such a
460: component could come in the form of a pure Keplerian disk or a
461: rotating disk-wind \markcite{mc98,elv00}(e.g., {Murray} \& {Chiang} 1998; {Elvis} 2000). Microlensing of the
462: part of the BELR that is rotating toward us would then easily explain
463: the features that we observe.
464:
465: The most robust statement we can make is that the observations confirm
466: that the BELR has structure on the scale of the Einstein radius.
467: Because we are sampling a region on the order of
468: $\sim\!1.4\times10^{16}$~cm, we effectively have a ``telescope'' with
469: a resolution of $\sim\!5\times10^{-7}$~arcseconds (given an angular
470: diameter distance to the quasar of $\sim5.4\times10^{27}$~cm). In
471: other words, in SDSS~J1004+4112, nature has provided us with an
472: extremely powerful tool for the study of BELR structure.
473:
474: \subsubsection{Some Predictions}
475:
476: The microlensing hypothesis leads to several predictions that may
477: guide further observations. First, the nature of the object(s)
478: responsible for the microlensing is unknown. Two obvious
479: possibilities are stars or MACHOs associated with a galaxy (presumably
480: a cluster member) with a small projected impact parameters to
481: component A, or stars or MACHOs in the intracluster medium. While the
482: Subaru $i$-band image of the field (Fig.~\ref{fig:fig3}) suggests that
483: there is indeed a galaxy near component A, the saturation of the
484: quasar image makes PSF subtraction uncertain. Deep, high-resolution
485: images, preferably in the near-IR, would better reveal whether this
486: galaxy is real. If so, then the microlensing optical depth would be
487: relatively high for component A and (perhaps quite) low for the other
488: components, so further monitoring would probably reveal additional
489: microlensing events in component A but not in the other components.
490:
491: Alternatively, if the microlensing is caused by intracluster stars
492: or MACHOs then the microlensing optical depth for component A is
493: probably fairly low --- but it is likely to be similar for the other
494: components. In this case further monitoring could well reveal
495: microlensing events in the other components, and the frequency of
496: events would reveal the number density of intracluster objects.
497:
498: If additional microlensing events are detected, it will be extremely
499: interesting to track which lines are microlensed as a function of
500: time. That would provide a unique and powerful new probe of
501: ionization stratification in the BELR. In addition, the velocity
502: dependence of any enhancements of the BELR in future microlensing
503: events will help to reveal the kinematic structure of the BELR.
504:
505: \subsection{Time-independent Emission Line Differences}
506:
507: In addition to the time-dependent differences in the blue wings of the
508: high-ionization emission lines, there are also subtle differences in
509: the \ion{C}{4} emission line profiles between the components that
510: persist over at least 322 days in the observed frame. In
511: Figure~\ref{fig:fig2} we have overplotted a Gaussian at the position
512: of \ion{C}{4} emission to guide the eye and help illustrate these
513: differences.
514:
515: In particular, we note that there is a slight excess of
516: high-velocity redshifted \ion{C}{4} emission in component B as
517: compared with component A. This excess is best seen in the 2003
518: May 31 Keck spectra, where there is a kink in the profile of B
519: near $\lambda\,1560$~\AA\ that is not present in component A. In
520: addition, in all observations of component B the fall-off in the
521: red wing toward $\lambda\,1600$~\AA\ is more gradual than in
522: component A. We also see similar (but opposite) differences in the
523: blue wing with the most extreme blue wing flux falling off more
524: gradually in component A than in component B. In other words, at
525: all epochs the \ion{C}{4} emission line profile is somewhat
526: blueward asymmetric in component A, and somewhat redward asymmetric
527: in component B.
528:
529:
530: The cause of these differences is unclear. Because the predicted A--B
531: time delay is $\lesssim$30 days, they are unlikely to be due to
532: intrinsic variability (unless the time delays are grossly
533: underestimated). One interesting possibility is that they are also
534: due to microlensing, but with a mass scale that is much larger than
535: for the blue-wing BELR microlensing discussed above (in order to make
536: the variability time scale longer than 322 days in the observed
537: frame). The responsible objects could perhaps be globular clusters,
538: or clumps of dark matter of mass $\sim\!10^4$--$10^8\,M_{\sun}$, in
539: which case the phenomenon would be termed millilensing (rather than
540: microlensing) and could provide a unique probe of dark matter
541: substructure of the type predicted in the Cold Dark Matter paradigm
542: \markcite{mm01,dk02,wbc+03,mmb+03}({Metcalf} \& {Madau} 2001; {Dalal} \& {Kochanek} 2002; {Wisotzki} {et~al.} 2003; {Metcalf et al.} 2003).
543:
544: One way to test the millilensing hypothesis would be to normalize the
545: spectra using narrow lines since they should be insensitive to
546: small-scale structure \markcite{mm03}({Moustakas} \& {Metcalf} 2003). Any differences in the broad
547: lines would then indicate millilensing. This might be possible with
548: either J-band IR spectra of [\ion{O}{3}] $\lambda\lambda4959,5007$ or
549: with the unusually strong (but not apparently microlensed) nitrogen
550: lines seen in our optical spectra, since the nitrogen lines appear
551: relatively narrow and may lack a broad component.
552:
553: \section{Conclusions}
554:
555: We have presented seven epochs of spectroscopic data on the two
556: brightest components of the wide-separation, quadruply imaged quasar
557: SDSS J1004+4112. Although the simplest lensing scenarios predict that
558: the four components should have identical spectra, the data reveal
559: significant differences in the emission line profiles of the
560: components. In particular, the \ion{C}{4} emission line profile in
561: components A and B show both variable differences and differences that
562: are constant over 322 observed-frame days. The \ion{He}{2} and
563: \ion{Si}{4}/\ion{O}{4}] lines in component A also show variability
564: similar to that seen in the \ion{C}{4} line.
565:
566: Because the predicted time delay between A and B is $\lesssim$30 days,
567: we argue that the differences are not due to intrinsic variability in
568: the quasar coupled with a lensing time delay. Instead, we suggest
569: that the variability in the blue wing of component A is best explained
570: by microlensing of part of the broad emission line region, resolving
571: BELR structure on the order of a few light days. This represents the
572: first robust detection of BELR microlensing, with evidence based on
573: multiple emission lines and involving observed variability. The
574: nature of the time-independent differences is less clear, but they may
575: also be the result of a lensing event. In any case, it is clear that
576: continued spectroscopic monitoring of SDSS~J1004+4112 should be
577: carried out in an attempt to map the structure of its broad emission
578: line region through additional microlensing events.
579:
580: \acknowledgements
581:
582: Funding for the creation and distribution of the SDSS Archive has been
583: provided by the Alfred P. Sloan Foundation, the Participating
584: Institutions, the National Aeronautics and Space Administration, the
585: National Science Foundation, the U.S.\ Department of Energy, the
586: Japanese Monbukagakusho, and the Max Planck Society. The SDSS Web
587: site is http://www.sdss.org/. The SDSS is managed by the
588: Astrophysical Research Consortium (ARC) for the Participating
589: Institutions. The Participating Institutions are The University of
590: Chicago, Fermilab, the Institute for Advanced Study, the Japan
591: Participation Group, The Johns Hopkins University, Los Alamos National
592: Laboratory, the Max-Planck-Institute for Astronomy (MPIA), the
593: Max-Planck-Institute for Astrophysics (MPA), New Mexico State
594: University, University of Pittsburgh, Princeton University, the United
595: States Naval Observatory, and the University of Washington.
596:
597: Based on observations obtained with the Apache Point Observatory
598: 3.5-meter telescope, which is owned and operated by the Astrophysical
599: Research Consortium. Based on observations obtained at the W.~M. Keck
600: Observatory, which is operated as a scientific partnership among the
601: California Institute of Technology, the University of California and
602: the National Aeronautics and Space Administration, made possible by
603: the generous financial support of the W. M. Keck Foundation. Based in
604: part on data collected at Subaru Telescope, which is operated by the
605: National Astronomical Observatory of Japan. G.~T.~R. was supported in
606: part by HST grant HST-GO-09472.01-A. C.~R.~K. is supported by NASA
607: through Hubble Fellowship grant HST-HF-01141.01-A from the Space
608: Telescope Science Institute, which is operated by the Association of
609: Universities for Research in Astronomy, Inc., under NASA contract
610: NAS5-26555.
611:
612: \clearpage
613:
614: %% \bibliography
615: \begin{thebibliography}{}
616:
617: \bibitem[{Abajas}, {Mediavilla}, {Mu{\~ n}oz}, {Popovi{\' c}}, \& {Oscoz} 2002]{amm+02}
618: {Abajas}, C., {Mediavilla}, E., {Mu{\~ n}oz}, J.~A., {Popovi{\' c}}, L.~{\v C}., \& {Oscoz}, A. 2002, \apj, 576, 640
619:
620: \bibitem[{Arav}, {Barlow}, {Laor}, {Sargent}, \& {Blandford} 1998]{abl98}
621: {Arav}, N., {Barlow}, T.~A., {Laor}, A., {Sargent}, W.~L.~W., \& {Blandford}, R.~D. 1998, \mnras, 297, 990
622:
623: \bibitem[{Baltz et al.} 2003]{blz+04}
624: {Baltz et al.} 2003, astroph/0310845
625:
626: \bibitem[{Cao}, {Wei}, \& {Hu} 1999]{cwh99}
627: {Cao}, L., {Wei}, J.-Y., \& {Hu}, J.-Y. 1999, \aaps, 135, 243
628:
629: \bibitem[{Chartas et al.} 2003]{cea+04}
630: {Chartas et al.} 2003, astroph/0401240
631:
632: \bibitem[{Colley} \& {Schild} 2003]{cs03}
633: {Colley}, W.~N. \& {Schild}, R.~E. 2003, \apj, 594, 97
634:
635: \bibitem[{Dalal} \& {Kochanek} 2002]{dk02}
636: {Dalal}, N. \& {Kochanek}, C.~S. 2002, \apj, 572, 25
637:
638: \bibitem[{Elvis} 2000]{elv00}
639: {Elvis}, M. 2000, \apj, 545, 63
640:
641: \bibitem[{Eracleous} 2003]{era04}
642: {Eracleous}, M. 2003, astroph/0310649
643:
644: \bibitem[{Eracleous} \& {Halpern} 2003]{eh03}
645: {Eracleous}, M. \& {Halpern}, J.~P. 2003, \apj, 599, 886
646:
647: \bibitem[{Filippenko} 1989]{fil89}
648: {Filippenko}, A.~V. 1989, \apjl, 338, L49
649:
650: \bibitem[{Fukugita}, {Ichikawa}, {Gunn}, {Doi}, {Shimasaku}, \& {Schneider} 1996]{fig+96}
651: {Fukugita}, M., {Ichikawa}, T., {Gunn}, J.~E., {Doi}, M., {Shimasaku}, K., \& {Schneider}, D.~P. 1996, \aj, 111, 1748
652:
653: \bibitem[{Gunn}, {Carr}, {Rockosi}, {Sekiguchi}, {Berry}, {Elms}, {de Haas}, {Ivezi{\' c} }, {Knapp}, {Lupton}, {Pauls}, {Simcoe}, {Hirsch}, {Sanford}, {Wang}, {York}, {Harris}, {Annis}, {Bartozek}, {Boroski}, {Bakken}, {Haldeman}, {Kent}, {Holm}, {Holmgren}, {Petravick}, {Prosapio}, {Rechenmacher}, {Doi}, {Fukugita}, {Shimasaku}, {Okada}, {Hull}, {Siegmund}, {Mannery}, {Blouke}, {Heidtman}, {Schneider}, {Lucinio}, \& {Brinkman} 1998]{gcr+98}
654: {Gunn}, J.~E., {Carr}, M., {Rockosi}, C., {Sekiguchi}, M., {Berry}, K., {Elms}, B., {de Haas}, E., {Ivezi{\' c} }, {\v Z}., {et al.} 1998, \aj, 116, 3040
655:
656: \bibitem[{Huchra}, {Gorenstein}, {Kent}, {Shapiro}, {Smith}, {Horine}, \& {Perley} 1985]{huchra85}
657: {Huchra}, J., {Gorenstein}, M., {Kent}, S., {Shapiro}, I., {Smith}, G., {Horine}, E., \& {Perley}, R. 1985, \aj, 90, 691
658:
659: \bibitem[{Inada et al.} 2003]{inada03}
660: {Inada et al.} 2003, Nature, 426, 810
661:
662: \bibitem[{Irwin}, {Webster}, {Hewett}, {Corrigan}, \& {Jedrzejewski} 1989]{iwh+89}
663: {Irwin}, M.~J., {Webster}, R.~L., {Hewett}, P.~C., {Corrigan}, R.~T., \& {Jedrzejewski}, R.~I. 1989, \aj, 98, 1989
664:
665: \bibitem[{Kaspi}, {Smith}, {Netzer}, {Maoz}, {Jannuzi}, \& {Giveon} 2000]{ksn+00}
666: {Kaspi}, S., {Smith}, P.~S., {Netzer}, H., {Maoz}, D., {Jannuzi}, B.~T., \& {Giveon}, U. 2000, \apj, 533, 631
667:
668: \bibitem[{Kayser}, {Refsdal}, \& {Stabell} 1986]{krs86}
669: {Kayser}, R., {Refsdal}, S., \& {Stabell}, R. 1986, \aap, 166, 36
670:
671: \bibitem[{Kochanek} 2003]{koc04}
672: {Kochanek}, C.~S. 2003, astroph/0307422
673:
674: \bibitem[{Lewis} \& {Ibata} 2003]{la03}
675: {Lewis}, G.~F. \& {Ibata}, R.~A. 2003, MNRAS, astroph/0310818
676:
677: \bibitem[{Metcalf} \& {Madau} 2001]{mm01}
678: {Metcalf}, R.~B. \& {Madau}, P. 2001, \apj, 563, 9
679:
680: \bibitem[{Metcalf et al.} 2003]{mmb+03}
681: {Metcalf et al.} 2003, astroph/0309738
682:
683: \bibitem[{Moustakas} \& {Metcalf} 2003]{mm03}
684: {Moustakas}, L.~A. \& {Metcalf}, R.~B. 2003, \mnras, 339, 607
685:
686: \bibitem[{Murray} \& {Chiang} 1998]{mc98}
687: {Murray}, N. \& {Chiang}, J. 1998, \apj, 494, 125
688:
689: \bibitem[{Nemiroff} 1988]{nem88}
690: {Nemiroff}, R.~J. 1988, \apj, 335, 593
691:
692: \bibitem[{Oguri et al.} 2003]{oguri03}
693: {Oguri et al.} 2003, ApJ, astroph/0312429
694:
695: \bibitem[{Oke}, {Cohen}, {Carr}, {Cromer}, {Dingizian}, {Harris}, {Labrecque}, {Lucinio}, {Schaal}, {Epps}, \& {Miller} 1995]{occ+95}
696: {Oke}, J.~B., {Cohen}, J.~G., {Carr}, M., {Cromer}, J., {Dingizian}, A., {Harris}, F.~H., {Labrecque}, S., {Lucinio}, R., {et al.} 1995, \pasp, 107, 375
697:
698: \bibitem[{Peterson} \& {Wandel} 1999]{pw99}
699: {Peterson}, B.~M. \& {Wandel}, A. 1999, \apjl, 521, L95
700:
701: \bibitem[{Richards}, {Fan}, {Newberg}, {Strauss}, {Vanden Berk}, {Schneider}, {Yanny}, {Boucher}, {Burles}, {Frieman}, {Gunn}, {Hall}, {Ivezi{\' c}}, {Kent}, {Loveday}, {Lupton}, {Rockosi}, {Schlegel}, {Stoughton}, {SubbaRao}, \& {York} 2002]{rfn+02}
702: {Richards}, G.~T., {Fan}, X., {Newberg}, H.~J., {Strauss}, M.~A., {Vanden Berk}, D.~E., {Schneider}, D.~P., {Yanny}, B., {Boucher}, A., {et al.} 2002, \aj, 123, 2945
703:
704: \bibitem[{Schechter}, {Udalski}, {Szyma{\' n}ski}, {Kubiak}, {Pietrzy{\' n}ski}, {Soszy{\' n}ski}, {Wo{\' z}niak}, {{\. Z}ebru{\' n}}, {Szewczyk}, \& {Wyrzykowski} 2003]{sus+03}
705: {Schechter}, P.~L., {Udalski}, A., {Szyma{\' n}ski}, M., {Kubiak}, M., {Pietrzy{\' n}ski}, G., {Soszy{\' n}ski}, I., {Wo{\' z}niak}, P., {{\. Z}ebru{\' n}}, K., {et al.} 2003, \apj, 584, 657
706:
707: \bibitem[{Schechter} \& {Wambsganss} 2002]{sw02}
708: {Schechter}, P.~L. \& {Wambsganss}, J. 2002, \apj, 580, 685
709:
710: \bibitem[{Schneider}, {Ehlers}, \& {Falco} 1992]{sef92}
711: {Schneider}, P., {Ehlers}, J., \& {Falco}, E.~E. 1992, {Gravitational Lenses} (Springer-Verlag: Berlin)
712:
713: \bibitem[{Schneider} \& {Wambsganss} 1990]{sw90}
714: {Schneider}, P. \& {Wambsganss}, J. 1990, \aap, 237, 42
715:
716: \bibitem[{Strateva}, {Strauss}, {Hao}, {Schlegel}, {Hall}, {Gunn}, {Li}, {Ivezi{\' c}}, {Richards}, {Zakamska}, {Voges}, {Anderson}, {Lupton}, {Schneider}, {Brinkmann}, \& {Nichol} 2003]{ssh+03}
717: {Strateva}, I.~V., {Strauss}, M.~A., {Hao}, L., {Schlegel}, D.~J., {Hall}, P.~B., {Gunn}, J.~E., {Li}, L., {Ivezi{\' c}}, {\v Z}., {et al.} 2003, \aj, 126, 1720
718:
719: \bibitem[{Totani} 2003]{tot03}
720: {Totani}, T. 2003, \apj, 586, 735
721:
722: \bibitem[{Wambsganss}, {Paczynski}, \& {Katz} 1990]{wpk90}
723: {Wambsganss}, J., {Paczynski}, B., \& {Katz}, N. 1990, \apj, 352, 407
724:
725: \bibitem[{Wandel}, {Peterson}, \& {Malkan} 1999]{wpm99}
726: {Wandel}, A., {Peterson}, B.~M., \& {Malkan}, M.~A. 1999, \apj, 526, 579
727:
728: \bibitem[{Wisotzki}, {Becker}, {Christensen}, {Helms}, {Jahnke}, {Kelz}, {Roth}, \& {Sanchez} 2003]{wbc+03}
729: {Wisotzki}, L., {Becker}, T., {Christensen}, L., {Helms}, A., {Jahnke}, K., {Kelz}, A., {Roth}, M.~M., \& {Sanchez}, S.~F. 2003, \aap, 408, 455
730:
731: \bibitem[{Wo{\' z}niak}, {Udalski}, {Szyma{\' n}ski}, {Kubiak}, {Pietrzy{\' n}ski}, {Soszy{\' n}ski}, \& {{\. Z}ebru{\' n}} 2000]{wus+00}
732: {Wo{\' z}niak}, P.~R., {Udalski}, A., {Szyma{\' n}ski}, M., {Kubiak}, M., {Pietrzy{\' n}ski}, G., {Soszy{\' n}ski}, I., \& {{\. Z}ebru{\' n}}, K. 2000, \apjl, 540, L65
733:
734: \bibitem[{York}, {Adelman}, {Anderson}, {Anderson}, {Annis}, {Bahcall}, {Bakken}, {Barkhouser}, {Bastian}, {Berman}, {Boroski}, {Bracker}, {Briegel}, {Briggs}, {Brinkmann}, {Brunner}, {Burles}, {Carey}, {Carr}, {Castander}, {Chen}, {Colestock}, {Connolly}, {Crocker}, {Csabai}, {Czarapata}, {Davis}, {Doi}, {Dombeck}, {Eisenstein}, {Ellman}, {Elms}, {Evans}, {Fan}, {Federwitz}, {Fiscelli}, {Friedman}, {Frieman}, {Fukugita}, {Gillespie}, {Gunn}, {Gurbani}, {de Haas}, {Haldeman}, {Harris}, {Hayes}, {Heckman}, {Hennessy}, {Hindsley}, {Holm}, {Holmgren}, {Huang}, {Hull}, {Husby}, {Ichikawa}, {Ichikawa}, {Ivezi{\' c}}, {Kent}, {Kim}, {Kinney}, {Klaene}, {Kleinman}, {Kleinman}, {Knapp}, {Korienek}, {Kron}, {Kunszt}, {Lamb}, {Lee}, {Leger}, {Limmongkol}, {Lindenmeyer}, {Long}, {Loomis}, {Loveday}, {Lucinio}, {Lupton}, {MacKinnon}, {Mannery}, {Mantsch}, {Margon}, {McGehee}, {McKay}, {Meiksin}, {Merelli}, {Monet}, {Munn}, {Narayanan}, {Nash}, {Neilsen}, {Neswold}, {Newberg}, {Nichol}, {Nicinski}, {Nonino}, {Okada}, {Okamura}, {Ostriker}, {Owen}, {Pauls}, {Peoples}, {Peterson}, {Petravick}, {Pier}, {Pope}, {Pordes}, {Prosapio}, {Rechenmacher}, {Quinn}, {Richards}, {Richmond}, {Rivetta}, {Rockosi}, {Ruthmansdorfer}, {Sandford}, {Schlegel}, {Schneider}, {Sekiguchi}, {Sergey}, {Shimasaku}, {Siegmund}, {Smee}, {Smith}, {Snedden}, {Stone}, {Stoughton}, {Strauss}, {Stubbs}, {SubbaRao}, {Szalay}, {Szapudi}, {Szokoly}, {Thakar}, {Tremonti}, {Tucker}, {Uomoto}, {Vanden Berk}, {Vogeley}, {Waddell}, {Wang}, {Watanabe}, {Weinberg}, {Yanny}, \& {Yasuda} 2000]{yaa+00}
735: {York}, D.~G., {Adelman}, J., {Anderson}, J.~E., {Anderson}, S.~F., {Annis}, J., {Bahcall}, N.~A., {Bakken}, J.~A., {Barkhouser}, R., {et al.} 2000, \aj, 120, 1579
736:
737: \end{thebibliography}
738: %\bibliographystyle{apj}
739:
740: \clearpage
741:
742: \begin{figure}[p]
743: \epsscale{1.0}
744: \plotone{fig1.eps}
745: \caption{Keck/LRIS spectra of components A, B, C, and D of
746: SDSS~J1004+4112. Dashed vertical lines indicate the expected peaks of
747: \ion{Si}{4}/\ion{O}{4}], \ion{C}{4}, and \ion{He}{2} for $z=1.734$.
748: A power law continuum, fit between $\lambda\,1450$~\AA\ and
749: $\lambda\,1690$~\AA, has been subtracted from each spectrum, and the
750: spectra are all normalized to the peak of \ion{C}{4}. The spectra
751: are smoothed by a seven pixel boxcar filter.
752: \label{fig:fig1}}
753: \end{figure}
754:
755: \begin{figure}[p]
756: \epsscale{1.0}
757: \plotone{fig2.eps}
758: \caption{Seven epochs of data in the \ion{C}{4} emission line region
759: of SDSS~J1004+4112. A power law continuum, fit between
760: $\lambda\,1450$~\AA\ and $\lambda\,1690$~\AA, has been subtracted from
761: each spectrum, and the spectra are all normalized to the peak of
762: \ion{C}{4}. We also overplot a scaled Gaussian at the center of
763: \ion{C}{4} to guide the eye toward emission line differences that are
764: persistent with time. The dashed vertical line indicates the expected
765: \ion{C}{4} peak at $\lambda\,1549.06$~\AA. All spectra have been
766: smoothed to similar resolutions. Note that this smoothing hides the
767: associated absorption system that is observed just blueward of the
768: \ion{C}{4} emission line peak.
769: \label{fig:fig2}}
770: \end{figure}
771:
772: \begin{figure}[p]
773: \epsscale{0.5}
774: \plotone{fig3.eps}
775: \caption{Central ($22\farcs6$) region of the 1340\,s $i$-band Subaru
776: Prime Focus Camera image from \markcite{oguri03}{Oguri et al.} (2003, Fig.~8). North is up,
777: East is left. Although the image is saturated and proper PSF
778: subtraction of component A is not possible, there appears to be a
779: superimposed galaxy (with an estimated $i\sim24.5$) just to the
780: North-East of component A that could host the microlensing object.
781: The white circle is centered on the possible microlensing galaxy and
782: represents a $20\,h_{70}^{-1}\,{\rm kpc}$ radius at $z=0.68$ assuming
783: $\Omega_M=0.3$, $\Omega_\Lambda=0.7$, and $H_0=70\,h_{70}\,{\rm
784: km\,s^{-1}\,Mpc^{-1}}$.
785: \label{fig:fig3}}
786: \end{figure}
787:
788: \begin{figure}[p]
789: \epsscale{1.0}
790: \plotone{fig4.eps}
791: \caption{Comparison of Keck spectra of components A and B in the
792: \ion{C}{3}] to \ion{Mg}{2} region of the spectra, showing the weaker
793: enhancement (relative to the higher ionization lines) of the blue
794: wings of these emission lines. A power law continuum, fit between
795: $\lambda\,1690$~\AA\ and $\lambda\,2200$~\AA, has been subtracted
796: from each spectrum, and the spectra are all normalized to the peak of
797: \ion{C}{3}]. No constraint is placed on \ion{Mg}{2}. The dashed
798: vertical lines indicate the expected \ion{Al}{3}, \ion{C}{3}] and
799: \ion{Mg}{2} peaks in addition to the \ion{Fe}{3} UV34 and UV48
800: complexes. The spectra are smoothed by a seven pixel boxcar filter.
801: \label{fig:fig4}}
802: \end{figure}
803:
804: \clearpage
805:
806: \begin{deluxetable}{llllll}
807: \tabletypesize{\small}
808: \tablewidth{0pt}
809: \tablecaption{Summary of Observations\label{tab:tab1}}
810: \tablehead{
811: \colhead{Date (UT)} &
812: \colhead{Telescope} &
813: \colhead{Components} &
814: \colhead{Exp. Time} &
815: \colhead{Dispersion} &
816: \colhead{$\lambda$ Range} \\
817: \colhead{} &
818: \colhead{} &
819: \colhead{} &
820: \colhead{(sec)} &
821: \colhead{(${\rm \AA}\,{\rm pix}^{-1}$)} &
822: \colhead{(\AA)}
823: }
824: \startdata
825: 2003 February 3 & SDSS 2.5m & B & 2700 & $\sim1$ & 3800-9200 \\
826: 2003 May 3 & ARC 3.5m (DISIII) & A & 1800 & $\sim2.4$ & 3820-5630 \\
827: 2003 May 31 & Keck I (LRIS) & ABCD & 900 & $\sim1$ & 3028-9700 \\
828: 2003 November 21 & ARC 3.5m (DISIII) & AB & 2700+2700 & $2.4$ & 3890-9350 \\
829: 2003 November 30 & ARC 3.5m (DISIII) & AB & 2700+2700 & $2.4$ & 3890-9350 \\
830: 2003 December 1 & ARC 3.5m (DISIII) & AB & 2400+2400 & $2.4$ & 3890-9350 \\
831: 2003 December 22 & ARC 3.5m (DISIII) & AB & 2400+2400 & $2.4$ & 3890-9350 \\
832: \enddata
833: \end{deluxetable}
834:
835: \end{document}
836: