1: \documentstyle[11pt,newpasp,twoside, epsf]{article}
2: \markboth{L. Page}{IAU Conference, Sept. 2000}
3: \pagestyle{myheadings}
4: \nofiles
5:
6: %
7: % \begin{quote}
8: % \verb"\documentstyle[11pt,newpasp,twoside,epsf]{article}" % dvips users ONLY
9: % \end{quote}
10: % You may have to copy {\tt epsf.tex} to {\tt epsf.sty} or create a
11: % symbolic link, since \prog{dvips} is distributed with {\tt epsf.tex}
12: % which serves both plain \TeX\ and \LaTeX. The same file works
13: % fine as a \LaTeX\ style file, it just needs the .sty extension.
14: %
15:
16: % Some definitions I use in these instructions.
17:
18: \def\emphasize#1{{\sl#1\/}}
19: \def\arg#1{{\it#1\/}}
20: \let\prog=\arg
21:
22: \def\edcomment#1{\iffalse\marginpar{\raggedright\sl#1\/}\else\relax\fi}
23: \marginparwidth 1.25in
24: \marginparsep .125in
25: \marginparpush .25in
26: \reversemarginpar
27:
28: \begin{document}
29: \title{The {\sl MAP} Satellite Mission to Map the CMB
30: Anisotropy }
31: \author{Lyman Page}
32: \affil{Princeton University, Dept. of Physics,
33: Jadwin Hall, Washington Rd}
34:
35: \begin{abstract}
36: The Microwave Anisotropy Probe ({\sl MAP}) satellite is scheduled to launch in
37: mid-2001. {\sl MAP}'s goal is to produce a map of the
38: anisotropy in the cosmic microwave background of unprecedented {\it accuracy}
39: and precision. The guiding design principle has been the minimization
40: of systematic effects. The instrument design and mapping strategy work in
41: concert to take advantage of the unique opportunities afforded
42: by deep space. We give an overview of the mission and compare
43: the projected {\sl MAP} error bars to recent measurements.
44: \end{abstract}
45:
46: \section{Introduction}
47: The cosmic microwave background (CMB) is now widely
48: recognized as one of
49: the premier probes of cosmology. Other than foreground emission, which
50: is subdominant and measurable, very little astrophysics stands between the
51: observations and that which is of cosmological import (Tegmark et al. 2000).
52: The theoretical framework is in place; the challenge before us
53: is performing accurate and unassailable measurements of the properties
54: of the CMB.
55:
56: Since the discovery of the anisotropy by COBE/DMR (Smoot et al. 1992)
57: experimental progress has been rapid.
58: By the end of the last millennium, we knew that not only was there a peak
59: in the angular spectrum, but we knew its amplitude, width, and
60: position (Hu 2000, Dodelson \& Knox 2000, Knox \& Page 2000).
61: The recent BOOMERanG (de
62: Bernardis et al. 2000) and MAXIMA (Hanany et al. 2000)
63: results have not only given us maps of the CMB but have
64: impressively shown that peak in sharp relief. Analyses of the data
65: (Bond et al. 2000, Jaffe et al. 2000) give us a glimpse of what may be learned
66: from the primary anisotropy ($l<2000$) in the context of adiabatic
67: CDM models.
68:
69: The future of CMB measurements is bright. In addition to results from
70: many ground and balloon based experiments, the {\sl MAP} satellite will
71: map the CMB over the full sky to unprecedented accuracy and precision.
72: {\sl PLANCK} will follow later in the decade.
73: The detection and characterization of the polarization will be crucial to
74: verifying the picture unveiled by the temperature anisotropy and to
75: placing further constraints on models. At $l>2000$, the CMB
76: can directly probe the formation of structures through the
77: Ostriker-Vishniac effect, the Sunyaev-Zel'dovich effect, and lensing.
78: Foreground emission at these scales will be more difficult
79: to subtract from the data (Toffolatti et al. 1998) however.
80:
81: \section{MAP in comparison with other maps}
82: The quest to make a large area map of the CMB anisotropy
83: has been in cosmologists minds for years. Early attempts
84: (reviewed by Weiss 1980 and Partridge 1995) measured the
85: CMB dipole, set limits on the anisotropy, and introduced mapping
86: and analysis techniques. The goal
87: was to produce a map with a temperature and error bar for each sky pixel.
88: All of these experiments were done from balloons
89: so that atmospheric fluctuations would not skew the map.
90:
91: The modern era of CMB studies started with
92: the $7^\circ$ resolution full-sky COBE/DMR map (Smoot et al. 1992,
93: Bennett et al. 1996) which unambiguously detected the anisotropy.
94: This is still the map with the lowest systematic error; it is also
95: the best checked map. One of the most important aspects
96: of DMR is that pixel-to-pixel correlations are small,
97: $\Sigma_{ij}=\sigma_i^2\delta_{ij}$ (where $\Sigma_{ij}$ is the
98: covariance between pixels $i$ and $j$) is an excellent
99: approximation. Lineweaver et al. (1994) showed that for $60^{\circ}$
100: lags, the DMR beam separation angle, the cross correlation was only
101: 0.45\% of the diagonal terms. In the final map, the S/N per beam
102: resolution element was roughly two.
103:
104: Next came FIRS (Meyer et al. 1991, Ganga et al. 1993) at 170 GHz
105: and $3.8^{\circ}$ resolution. For FIRS
106: $\Sigma_{ij}\ne\sigma_i^2\delta_{ij}$ but the off diagonal
107: terms were small enough for the analyses. A spot check, requested
108: by John Mather before the publication of Ganga et al. (1993), showed
109: the largest to be of order 0.1 the diagonal elements.
110: At the time, there was no way to handle the full covariance matrix for
111: FIRS's 3500 pixels. The S/N per DMR beam was roughly $\sqrt{2}$.
112:
113: QMAP (Devlin et al. 1999, Herbig et al. 1999, de Oliveira-Costa et
114: al. 1999), with a resolution of $0.8^{\circ}$ at 40 GHz,
115: came after FIRS. $\Sigma_{ij}$ was not diagonal
116: for QMAP but the full covariance matrix was taken into account in the
117: analysis. QMAP used the rotation around the NCP to achieve
118: an interlocking scan strategy.
119: The features ones sees in QMAP are hot and cold spots
120: in the CMB. The S/N is roughly 2 per beam
121: over 530 square degrees.
122:
123: Most recently, we have seen the very high
124: S/N ($>5$ per beam) BOOMERanG and MAXIMA
125: results at $\approx 0.2^{\circ}$ resolution. Here too, the covariance
126: matrix is not diagonal but has been taken into account. With their high
127: resolution, one can clearly see the angular scale of an acoustic
128: peak in the CMB. The analyzed BOOMERanG data covered 440 square degrees
129: and MAXIMA covered 124 square degrees. Both experiments
130: have more data which will be analyzed over the following few years.
131:
132: As the sophistication in data analysis grows, more and more data sets
133: are being presented as a ``map'' plus covariance matrix, so the
134: distinction made above will be blurred. Indeed, maps
135: are now commonly synthesized from
136: interferometer data and scanning beam strategies. However, the
137: ideal map is like DMR's: a set of temperatures and statistical weights,
138: with ignorable off diagonal elements in the covariance matrix.
139:
140: Producing a full-sky map with diagonal $\Sigma_{ij}$ was
141: one of {\sl MAP}'s design goals. The cornerstones of achieving this
142: are 1) low 1/f noise in the instrument and detectors, and 2)
143: a fully interlocking scan strategy\footnote{The notion of the
144: interconnectedness of the scans was stressed in the proposal in 1995
145: but had to wait to be quantified by Wright (1996) and Tegmark (1997). }.
146: Concomitant with this goal we aimed to ensure that orbit and spacecraft induced
147: systematic errors would be negligible compared to the astrophysical signal.
148:
149: \section{Instrument description}
150: {\sl MAP} was proposed in June 1995, at the height of the ``faster,
151: better, cheaper'' era; building began in June 1996. It was designed to be
152: robust, thermally and mechanically stable, built of components
153: with space heritage\footnote{Other than the HEMT amplifiers, which were
154: custom designed, all components were ``off-the-shelf.'' I think it
155: was a surprise
156: to those of us not familar with the satellite business how much testing
157: was required of, and how many problems there were with,
158: off-the-shelf hardware.}, and relatively easy to integrate and test.
159:
160: Figure 1 shows a line drawing of the {\sl MAP} satellite, though it
161: could be a photograph. The satellite is completely built and put together.
162: The long and demanding process of quality assessment is underway.
163: As of this writing, {\sl MAP} is well into its final observatory
164: level environmental test program.
165:
166: As a thumbnail sketch, the {\sl MAP} instrument is comprised of ten
167: symmetric, passively cooled, dual polarization, differential,
168: microwave receivers. There are four receivers in W-band (94 GHz),
169: two in V-band (61 GHz), two in Q-band (41 GHz), one in Ka band (33 GHz),
170: and one in K band (23 GHz). The receivers are fed by back-to-back
171: Gregorian telescopes. A more detailed description follows
172: and more information may be obtained from http://map.gsfc.nasa.gov/.
173:
174: In Figure 1, the large circular structure at the bottom
175: is comprised of solar panels and flexible aluminized mylar/kapton
176: insulation. It shields the instrument from thermal emission the Sun, Earth,
177: and Moon. At launch, this is folded up so that the S/C can fit into the rocket
178: fairing, though it deploys roughly 90 minutes after liftoff.
179:
180: A hexagonal structure, ``hex hub,'' above the solar
181: panel array holds the power supplies, instrument electronics, and
182: attitude control systems. A one meter
183: diameter thermally insulating gamma alumina cylinder (GAC) separates the
184: hex hub and the instrument. The GAC supports a 190~K thermal gradient.
185:
186: Two large (5.6 m$^2$ net) and symmetric radiators
187: passively cool the input optics and front-end microwave electronics
188: to less than 100~K. One can just make out the heat straps that connect
189: the base of the radiators to the microwave components housed below the
190: primary reflectors. There are no cryogens or mechanical refrigerators.
191:
192:
193: \begin{figure}
194: \plotone{map_line.eps}
195: \caption{Outline of the {\sl MAP} satellite. The overall height is
196: 3.6~m, the mass is 830 kg, and the diameter of the large disk on
197: the bottom is 5.1 m. Six solar arrays on the bottom of this disk
198: supply the 400 Watts to power the spacecraft and instrument.
199: Thermal blanketing between the hex hub and GAC, and
200: between the GAC and radiators, shield the instrument from thermal
201: radiation from the support electronics and attitude control systems.
202: {\sl MAP} will be launched in mid-2001 by a Delta 7425-10 from Kennedy
203: Space Center. The mission life is 27 months.}
204: \end{figure}
205:
206: \subsection{Optics} The optics comprise two back-to-back shaped
207: Gregorian telescopes. The primary mirrors are 1.4 by 1.6 m. The
208: secondaries are roughly a meter across though most of the surface
209: simply acts as a shield to prevent the feeds from directly viewing the
210: Galaxy. The telescopes illuminate ten scalar feeds on each side,
211: a few of which are visible in the figure.
212: The primary optical axes are separated by $141^{\circ}$ to allow
213: differential measurements over large angles on a fast time scale.
214: The feed centers occupy a 18 by 20 cm region in the focal plane,
215: corresponding to a $4^{\circ}$ by $4.5^{\circ}$ array on
216: the sky.
217:
218: At the base of each feed is an orthomode transducer (OMT) that sends the
219: two polarizations supported by the feed to separate receiver chains.
220: The microwave plumbing is such that a single receiver chain
221: (half of a ``differencing assembly'') differences electric fields with
222: two nearly parallel linear polarization vectors, one from each telescope.
223:
224: Precise knowledge of the beams is essential for accurately computing the
225: CMB angular spectrum and for calibration. Because of the large focal
226: plane the beams
227: are not symmetric, as shown in Table 1. Cool down distortions
228: of the optics will alter the W-band beams with respect to the
229: ground based measurements so we give only an upper bound at this time.
230: All beam profiles will be mapped in flight with Jupiter and
231: other celestial sources.
232:
233: \begin{table}
234: \begin{center}
235: \caption{Approximate Instrument Characteristics by Frequency Bands}
236: \begin{tabular}{cccccc} \hline
237: Band & $f_{center}$ (GHz) & $\Delta f_{noise}$ (GHz) & $T_{HEMT}$ (K)
238: & $N_{chan}$ & $\theta_{FWHM}$ (deg)\\
239: K & 23 & 5 & 25 &2 & $0.75^{\circ}~by~0.95^{\circ}$ \\
240: Ka & 33 & 7 & 35 &2 & $0.6^{\circ}~by~0.7^{\circ}$ \\
241: Q & 41 & 8 & 50 &4 & $0.45^{\circ}~by~0.5^{\circ}$ \\
242: V & 61 & 11 & 80 &4 & $0.3^{\circ}~by~0.35^{\circ}$ \\
243: W & 94 & 18 & 100 &8 & $<0.23^{\circ}$ \\
244: \end{tabular}
245: \end{center}
246: \end{table}
247:
248: One of the design constraints was to minimize stray radiation from the
249: Galaxy, Sun, and Earth. We use physical optics codes
250: to compute the sidelobe pattern over the full sky as well
251: as to compute the current distributions on the optics. We have also built a
252: specialized test range to make sure that, by measurement, we can
253: eliminate the Sun as a source of signal at $<1~\mu$K level in all
254: bands. This requires knowing beam profiles down to
255: roughly $-45$ dBi (gain above
256: isotropic) or $-105$ dB from the W-band peak. We find that over much of
257: the sky, the measured profiles differ from the predictions at the $-50$~dB
258: level due to scattering off the feed horns and the structure that
259: holds them.
260:
261: We use a combination of the models and measurements to place a limit on
262: the Galactic pickup in the sidelobes. At 90 GHz, less than
263: $2~\mu$K of Galactic signal should contaminate the boresite signal for
264: observations with $|b|>15^{\circ}$, before any modeling.
265: As this adds in quadrature to the CMB signal, the affect on the
266: angular spectrum will be negligible. The characterization of the
267: sidelobes has been done by Chris Barnes.
268:
269: \subsection{Receivers}
270: {\sl MAP} uses ``pseudo-correlation'' receivers (Jarosik 2000) to
271: measure the difference in
272: power coming from the outputs of the OMTs at the base of the feeds,
273: as shown in Figure 2. We use
274: the term ``pseudo'' to refer to the fact that these radiometers use
275: two hybrid tees and two square law detectors in place of the multiplier
276: in a true correlation receiver.
277: The ``correlation'' refers to the feature that the system
278: is primarily sensitive to the correlated signal in the two arms.
279:
280: The {\sl MAP} mission was made possible by the HEMT-based amplifiers developed
281: by Marian Pospieszalski (1992) at NRAO. These amplifiers achieve noise
282: temperatures of 25-100 K at 80~K physical temperature (Pospieszalski \&
283: Wollack 2000). Of equal importance is that the amplifiers can be phase matched
284: over a 20\% fractional bandwidth, as is required by the receiver design.
285:
286: \begin{figure}
287: \plotfiddle{dalp2.eps}{3.125truein}{0}{45}{45}{-110}{-10}
288: %\plotone{dalp.eps}
289: \caption{One half of a differencing assembly for detecting
290: one polarization component. Hybrids (``magic tees'') split the
291: inputs into two arms where the signal is amplified before recombining.
292: There are actually two stages of amplification. The microwave filters
293: between the lower hybrid and the detectors are not shown.}
294: \end{figure}
295:
296: In Figure 2, radiation from the two
297: feeds is combined by a hybrid into $(A+B)/\sqrt{2}$
298: and $(A-B)/\sqrt{2}$ signals,
299: where A and B refer to the amplitudes of the electric fields from one
300: linear polarization of each feed horn. In one arm of the
301: receiver, both A and B signals are amplified first by cold ($<100~$K) and
302: then by warm (290 K) HEMT amplifiers. Noise power from the amplifiers,
303: which far exceeds the input power, is added to each signal by the first
304: amplifiers. Ignoring the phase switch for
305: a moment, the two arms are then recombined in a second hybrid and both
306: outputs of the hybrid are detected after a band defining filter.
307: Thus for each differencing assembly,
308: there are four detector outputs, two for each polarization.
309: In a perfectly balanced system, one
310: detector continuously measures the power in the A signal plus the average
311: radiometer noise while the other continuously measures the power in the
312: B signal plus the average radiometer noise.
313:
314: If the amplification factors for the fields are $G_1$ and $G_2$ for the two
315: arms, the difference in detector outputs is $G_1G_2(A^2-B^2)$. Note that
316: the average power signal present in both detectors has cancelled so that
317: small gain variations, which have a $1/f$ spectrum, act on the
318: difference in powers from the two arms, which corresponds to less than
319: 1~K, rather than on the total power which corresponds to roughly 100~K in
320: W-band. The system is stabilized further by toggling one of the phase switches
321: at 2.5 kHz and coherently demodulating the detector outputs.
322: (The phase switch in the other arm is required to preserve the phase
323: match between arms and is jammed in one state.) The 2.5 kHz modulation
324: places the desired signal at a frequency above the $1/f$ knee
325: of the detectors and video amplifiers as well as rejecting any residual
326: effects due to $1/f$ fluctuations in the gain of the HEMT amplifiers.
327: The power output of
328: each detector is averaged for between 51 and 128 ms and telemetered
329: to the Earth. In total, there are forty signals (only half
330: contain independent information) plus instrument housekeeping data
331: resulting in a data rate of 110~MBy/day.
332:
333: %
334: % In addition, a 2.5~kHz phase switch in one arm modulates the
335: % difference at a frequency above the 1/f knee of the amplifiers,
336: % so that the input to one of the detectors is A-B-A-B, while the input to
337: % the other is B-A-B-A. The phase switch in the other arm is jammed in one
338: % state and preserves the phase match between the arms.
339: %
340:
341:
342: The power spectrum of the
343: noise shows that it is effectively white between 0.008 Hz,
344: the spin rate of the
345: satellite, and 2.5~kHz. The autocorrelation function shows only a spike
346: at zero lag and a 1.2\% correlation between adjacent samples (in W-band)
347: due to the antialiasing filter.
348: All tests show that the noise is stationary and Gaussian
349: for days at a time.
350:
351: Although {\sl MAP}'s differential design was driven by the $1/f$ noise in
352: the amplifiers, it is also very effective at reducing the effects of
353: $1/f$ thermal
354: fluctuations of the spacecraft itself. The thermal stability of deep space
355: combined with the insensitivity to the spacecraft's
356: slow temperature variations should result in an extremely stable
357: instrument. Outside of the antialiasing filter and finite sampling time,
358: effects which are computable, measured, and non-random, we have not been
359: able to identify other effects that will correlate one measurement to the next.
360:
361: %
362: %
363: % \plotfiddle{"\arg{file}\verb"}{"\arg{vsize}\verb"}{"\arg{rot}\verb"}
364: % {"\arg{hsf}\verb"}{"\arg{vsf}\verb"}{"\arg{htrans}\verb"}
365: % {"\arg{vtrans}\verb"}
366: %
367:
368: \subsection{Scan strategy}
369: The other key aspect of producing a map with
370: $\Sigma_{ij}=\sigma_i^2\delta_{ij}$, in addition to low system $1/f$,
371: is a highly interlocking scan
372: strategy. In any measurement, a baseline instrumental offset along with
373: its associated drift, must be subtracted. Without cross hatched scans
374: this subtraction can correlate pixels over large swaths,
375: resulting in striped maps and substantially more involved
376: analyses.
377:
378: {\sl MAP} observes from a low maintenance Lissajous orbit at L2, with the
379: Sun Earth and Moon always behind, as shown in Figure 3.
380: Corrections to the orbit are applied roughly once per
381: season through thruster jets; there is only one mode of operation.
382: {\sl MAP} will be the first satellite to stay at L2
383: for an appreciable time.
384: %
385: % corrections every three months.
386: %To our knowledge,
387: %such an orbit for CMB missions was first contemplated by the
388: %RELICT 2 team;
389:
390: {\sl MAP} spins around its axis with a period of 2 min
391: and precesses around a $22.5^{\circ}$ degree cone every hour so that the beams
392: follow a spirograph pattern. Consequently, $\approx 30$\% of the sky
393: is covered in one hour, before the instrument temperature
394: can change appreciably. This motion is accomplished with three
395: spinning momentum wheels; the net angular momentum of the satellite is
396: near zero.
397: The axis of this combined rotation/precession sweeps out approximately a
398: great circle as the Earth orbits the sun. In six months, the whole sky
399: is mapped. Systematic effects at the spin
400: period of the satellite are the most difficult to separate from true
401: sky signal. Such effects, driven by the Sun, are minimized
402: because the instrument is always in the
403: shadow of the solar array and the precession axis is fixed with
404: respect to the Earth-Sun line.
405:
406: \begin{figure}
407: \plotone{scan2.eps}
408: \caption{{\sl MAP's} scan pattern from L2. The dark circle on the lefthand
409: drawing depicts the path covered by two beams for one rotation, the
410: innermost circle is the path of the spin axis during one precession.}
411: \end{figure}
412:
413: The combination of {\sl MAP}'s four observing time scales (2.5 kHz, 2.1 min,
414: 1 hour, 6 months) and the heavily interlocked pattern result is a strong
415: spatio-temporal filter for any signal fixed in the sky. A pipeline
416: for simulating the mission, the amplifier characteristics, the beam
417: profiles, and producing maps from the time ordered
418: data is written. Using the measured amplifier
419: characteristics, we can form the two point function of a simulated noise
420: map. It shows that the pixel-pixel correlations are negligible
421: with the exception of a $<1\%$ nearest neighbor correlation from the
422: antialiasing filter and a $\sim0.1\%$ correlation at angles of
423: the beam separation. See Hinshaw (2000) for more details.
424:
425: \subsection{Science from {\sl MAP}}
426: {\it We emphasize again that the primary goal of {\sl MAP}
427: is to produce high
428: fidelity polarization-sensitive full-sky multi-frequency maps of the microwave
429: sky.} With such maps we can not only determine the anisotropy in the
430: CMB but we can test our fundamental assumptions about the cosmos.
431: For instance, perhaps the best fit model is not a simple adiabatic CDM variant
432: but has some isocurvature modes mixed in.
433:
434: If the CMB temperature fluctuations are Gaussian with random phases,
435: as the current data suggest, then the cosmological information
436: may be obtained from the angular spectrum. An example is shown in Figure
437: 4. The limit to which we can
438: know the angular spectrum is set by the number of independent patches of sky
439: corresponding to a given angular scale, or equivalently the number
440: parameters needed to describe a mode. For instance, five parameters
441: determine the quadrupole. No matter how well we measure these,
442: we cannot know the fractional variance of the parent distribution
443: better than $\sqrt{2/(2l+1)}=0.63$. This is
444: the ``cosmic variance'' limit and is one of the motivations for a
445: full-sky map. {\sl MAP} will be cosmic variance limited
446: up to $l\approx 500$ in W-band. In other words, in principle it is
447: not possible to
448: determine the angular spectrum better than this. More information
449: may be obtained from high S/N polarization measurements however.
450:
451: One of the largest uncertainties of existing measurements is
452: the calibration. Limitations come from not knowing the beam profiles
453: precisely and from the intrinsic uncertainty in the calibration sources.
454: The flux from planets is known to $\approx 5$\%
455: depending on frequency. Although the dipole has been used to calibrate
456: a number of mapping experiments (Meyer et al. 1991, de Bernardis et al. 2000,
457: Hanany et al. 2000), it is difficult to measure with limited sky coverage
458: because of its covariance with typical scan patterns. {\sl MAP} like
459: DMR, will calibrate using the change in the dipole caused by the
460: Earth's motion around the Sun. The calibration error will be $<1$\%.
461:
462: In Figure 4, we compare the projected {\sl MAP} uncertainties with the
463: state-of-the-art. We have included calibration error in both plots.
464: The right hand plot shows the
465: amplitude and position of the first peak ($50<l<420$) based on
466: Knox \& Page (2000). This is a relatively {\it cosmological
467: model independent} parametrization of the peak. It differs from the
468: work of Bond and colleagues (see these proceedings) in that a
469: CDM variant is not assumed and there
470: is no marginalization over nonrelevant parameters, consequently the
471: error bars are smaller. Though
472: one should not put too much stock in interpreting the contours
473: beyond the $\approx 2\sigma$ level, the plot shows that, despite high
474: precision, we are still not in complete agreement on the position and
475: amplitude of the first peak.
476:
477: There will be much more to do with the {\sl MAP} data than fit cosmological
478: models. A partial list includes understanding Galactic
479: foreground emission (is there spinning dust?) and extragalactic sources
480: (they are separated from the CMB through their frequency spectrum and
481: positive definiteness),
482: a search for time variable sources, a search for non-standard
483: topologies, a measurement of the polarization-temperature cross
484: correlation, a search for the large scale
485: Sunyaev-Zel'dovich effect, a measurement of the
486: cross-correlation with the Solan Digital
487: Sky Survey to assess structure formation at lower redshift ($z<1000$).
488:
489: Over the next two years, our current picture of the primary anisotropy
490: will become clearer
491: as more BOOMERanG and MAXIMA data are analyzed and as new data come in
492: from ACBAR, Arkeops, BEAST, CBI, DASI, MINT, TopHat, VSA, and others. The full
493: data set will of course be interesting on its own, and will
494: be essential for ensuring that the picture we get from {\sl MAP} extrapolates
495: as predicted in both frequency and angular scale. The next generation
496: of polarization measurements of the primary anisotropy will take
497: us beyond {\sl MAP}'s polarization sensitivity and add a new
498: dimension to CMB studies. It will not be
499: too long before CMB experiments will be calibrated on the primary
500: anisotropy, a prospect difficult to believe just a few years ago.
501:
502:
503: \begin{figure}
504: \plottwo{map_band_025_rot.eps}{like_plot_may002.eps}
505: \caption{{\it Left.} The projected {\sl MAP} error bars for a band
506: averaging of $\Delta l=50$ superimposed on a popular model
507: along with the BOOMERanG, MAXIMA, and TOCO
508: (Torbet et al. 1999, Miller et al. 1999) data. The calibration
509: uncertainty is indicated by the light lines behind the thick lines.
510: It is not usually plotted like this because the calibration is
511: common to all data in a set. {\it Right.} An analysis of the peak
512: following the Knox \& Page (2000) Gaussian/Temperature method.
513: Calibration error is included and shown in parentheses. The BOOMERanG
514: N/A data are from Mauskopf et al. (2000). The small solid dot shows the
515: projected {\sl MAP} error.}
516: \end{figure}
517:
518: \acknowledgements
519: The work on which this article is based, except for the righthand panel
520: in Figure 4,
521: was done by the {\sl MAP}
522: science team\footnote{Chuck Bennett (NASA/GSFC, PI),
523: Mark Halpern (UBC), Gary Hinshaw (NASA/GSFC), Norm Jarosik (Princeton),
524: Al Kogut (NASA/GSFC), Michele Limon (Princeton), Stephan Meyer
525: (Chicago), Lyman Page (Princeton), David Spergel (Princeton), Greg
526: Tucker (Brown), David Wilkinson (Princeton), Ed Wollack (NASA/GSFC),
527: and Ned Wright (UCLA).} and the {\sl MAP} satellite project led
528: by Liz Citrin (Project Manager) and Cliff Jackson (Systems Engineer)
529: at NASA/GSFC. Over a hundred NASA employees and contractors have
530: dedicated themselves to making {\sl MAP} work.
531:
532:
533: \begin{references}
534: %\reference Abt (1990): author, year, journal, volume, and page.,
535: %Abt, H. 1990, ApJ, 357, 1
536: \reference{Bennett C. L. et al. 1996, \apj, 464, L1}
537: \reference{Bond, J. R. et al. 2000, These proceedings and
538: astro-ph/0011381 and astro-ph/0011379}
539: \reference{de Bernardis, P. et al. 2000, Nature, 404, 955}
540: \reference{de Oliveira-Costa, A. et al. 1998, \apj, 509, L77}
541: \reference{Devlin, M. J. et al. 1998, \apj, 509, L69}
542: \reference{Dodelson, S. \& Knox, L. 2000, \prl, 84, 3523}
543: \reference{Ganga, K. et al. 1993, \apj, 410, L57}
544: \reference{Hanany, S. et al. 2000, astro-ph/0005123}
545: \reference{Herbig, T. et al. 1998, \apj, 509, L73}
546: \reference{Hinshaw, G. 2000, astro-ph/0011555}
547: \reference{Hu, W. 2000, astro-ph/0002520}
548: \reference{Jaffe, A. et al. 2000, astro-ph/0007333}
549: \reference{Jarosik, N. 2000, ``The Use of Cryogenic HEMT Amplifiers in
550: Wide Band Radiometers.'' Proceedings of the European Gallium Arsenide
551: Application Symposium, Paris, France (GAAS2000)}
552: \reference{Knox, L. \& Page, L. 2000, \prl, 85, 1366}
553: \reference{Lineweaver, C. et al. 1994, \apj, 436, 452}
554: \reference{Mauskopf, P. 2000, \apj, 536, L59}
555: \reference{Meyer, S. S., Cheng, E. S., \& Page, L. A. 1991, \apj, 371, L7}
556: \reference{Miller, A. D. et al. 1999, \apj, 524, L1}
557: \reference{Partridge, B. 1995, 3K: The Cosmic Microwave Background
558: Radiation, Cambridge University Press, New York}
559: \reference{Pospieszalski, M. W. 1992, Proc. IEEE Microwave
560: Theory Tech., MTT-3, 1369 and Pospieszalski, M. W. et al. 1994,
561: Proc. IEEE Microwave Theory Tech., MTT-3, 1345}
562: \reference{Pospieszalski, M. W. \& Wollack, E. 2000,
563: ``Ultra-Low-Noise, InP Field Effect Transistors Amplifiers
564: for Radio Astronomy Receivers,'' Proceedings
565: of the European Gallium Arsenide Application Symposium, Paris, France
566: (GAAS2000)}
567: \reference{Smoot, G. et al. 1992, \apj, 153, L1}
568: \reference{Tegmark, M. 1997, \prd, 56}
569: \reference{Tegmark, M. et al. 2000, \apj, 530, 133}
570: \reference{Toffolatti, L. et al. 1998, \mnras, 297, 117}
571: \reference{Torbet, E. et al. 1999, \apj, 521, L79}
572: \reference{Wright, E. L. 1996, astro-ph/9612006}
573: \reference{Weiss, R. 1980, \araa, 18, 489-535}
574: \end{references}
575:
576: %
577: % Biretta, J. A., Lo, K., & Young, P. J. 1982 in AIP Conf.
578: % Proc., 83, The Galactic Center, ed. G. R. Riegler & R. D.
579: % Blandford (New York: AIP), 91
580: %
581: % van der Kruit, P. C., & Shostak, G. S. 1983, in IAU Symp.
582: % 100, Internal Kinematics and Dynamics of Galaxies, ed. E.
583: % Athanassoula (Dordrecht: Reidel), 69
584: %
585: % Garcia-Lorenzo, B., Mediavilla, E., Arribas, S., & del
586: % Burgo, C. 1998, in ASP Conf. Ser. Vol. 152, Fiber Optics
587: % in Astronomy III, ed. S. Arribas, E. Mediavilla \& F.
588: % Watson (San Francisco: ASP), 185"
589: %
590:
591:
592: \end{document}
593:
594:
595:
596:
597: