1: %\documentclass[apj]{emulateapj}

2: \documentclass[12pt,preprint]{aastex}

3: \usepackage{apjfonts}

4: 

5: 

6: %% VINCENT: COMMENT THIS OUT BEFORE SUBMITTING. This allows us

7: %%          to use colorful language in our intra-paper discourse.

8: %\usepackage[usenames,dvips]{color}

9: %\newcommand\qc[1]{{\color{red} \bf #1}}  %Question comment ( \qc{question} )

10: %\newcommand\ac[1]{{\color{blue} \bf #1}} %Answer comment ( \ac{answer} )

11: %\newcommand\aeb[1]{{\color{OliveGreen} #1}} % My additions,

12:                                             % conveniently colorized for you!

13: % To remove comments, comment out above lines and uncomment out below

14: \newcommand\qc[1]{}

15: \newcommand\ac[1]{}

16: %% END

17: 

18: \def\Rs{R_{\rm S}}

19: 

20: \addtolength{\topmargin}{0.6truein}

21: %\newcommand{\nod}{\nodata}

22: %\newcommand{\lt}{\ensuremath{< 10}}

23: 

24: %% Something to make greek letters bold

25: \newcommand\bmath[1] {\mbox{\boldmath$\rm #1$}}

26: 

27: 

28: 

29: %%% Abbreviations

30: 

31: % Standard abbreviations

32: \newcommand\cf{{cf.}} %.......confer

33: \newcommand\eg{{e.g.}} %......exempli gratia

34: \newcommand\ie{{i.e.}} %......id est

35: \newcommand\etal{{et al.}} %..and others

36: \newcommand\etc{{etc}} %......et cetera

37: 

38: % Unit abbreviations

39: %%% Temperature

40: \newcommand\K{{\rm K}} %...........Kelvin

41: %%% Time

42: \newcommand\s{{\rm s}} %...........seconds

43: \newcommand\ns{{\rm n}\s} %........nanoseconds

44: \newcommand\mus{\mu\s} %...........microseconds

45: \newcommand\ms{{\rm m}\s} %........milliseconds

46: \newcommand\ks{{\rm k}\s} %........kiloseconds

47: \newcommand\hr{{\rm hr}} %.........hours

48: \newcommand\yr{{\rm yr}} %.........years

49: \newcommand\Myr{{\rm M}\yr} %......megayears

50: \newcommand\Gyr{{\rm G}\yr} %......gigayears

51: %%% Frequency

52: \newcommand\Hz{{\rm Hz}} %.........Hertz

53: \newcommand\muHz{\mu\Hz} %.........microhertz

54: \newcommand\mHz{{\rm mHz}} %.......millihertz

55: \newcommand\MHz{{\rm MHz}} %.......Megahertz

56: \newcommand\GHz{{\rm GHz}} %.......Gigahertz

57: \newcommand\THz{{\rm THz}} %.......Terahertz

58: %%% Length

59: \newcommand\m{{\rm m}} %...........meters

60: \newcommand\nm{{\rm n}\m} %........nanometers

61: \newcommand\mum{\mu\m} %...........micrometers

62: \newcommand\mm{{\rm m}\m} %........millimeters

63: \newcommand\cm{{\rm c}\m} %........centimeters

64: \newcommand\km{{\rm k}\m} %........kilometers

65: \newcommand\pc{{\rm pc}} %.........parsecs

66: \newcommand\kpc{{\rm k}\pc} %......kiloparsecs

67: \newcommand\Gpc{{\rm G}\pc} %......gigaparsecs

68: \newcommand\au{{\rm au}} %.........astronomical units

69: %%% Mass

70: \newcommand\g{{\rm g}} %...........grams

71: \newcommand\mug{\mu\g} %...........micrograms

72: \newcommand\mg{{\rm m}\g} %........milligrams

73: \newcommand\kg{{\rm k}\g} %........kilograms

74: \newcommand\Ms{M_\odot} %...........solar masses

75: %%% Energy

76: \newcommand\eV{{\rm eV}} %.........electron volts

77: \newcommand\keV{{\rm k}\eV} %......kiloelectron volts

78: \newcommand\MeV{{\rm M}\eV} %......megaelectron volts

79: \newcommand\GeV{{\rm G}\eV} %......gigaelectron volts

80: \newcommand\TeV{{\rm T}\eV} %......teraelectron volts

81: \newcommand\PeV{{\rm P}\eV} %......petaelectron volts

82: \newcommand\XeV{{\rm X}\eV} %......exaelectron volts

83: \newcommand\ZeV{{\rm Z}\eV} %......zetaelectron volts

84: \newcommand\erg{{\rm erg}} %.......ergs

85: \newcommand\J{{\rm J}} %...........joules

86: %%% Magnetic Fields

87: \newcommand\G{{\rm G}} %...........gauss

88: \newcommand\nG{{\rm n}\G} %........nanogauss

89: \newcommand\muG{\mu\G} %...........microgauss

90: \newcommand\mG{{\rm m}\G} %........milligauss

91: \newcommand\kG{{\rm k}\G} %........kilogauss

92: \newcommand\T{{\rm T}} %...........tesla

93: \newcommand\muT{\mu\T} %...........microtesla

94: \newcommand\mT{{\rm m}\T} %........millitesla

95: %%% Angular Resolution

96: \newcommand\mas{{\rm mas}} %.......milli-arcseconds

97: \newcommand\muas{\mu{\rm as}} %....micro-arcseconds

98: \newcommand\nas{{\rm nas}} %.......nano-arcseconds

99: \newcommand\e{{\rm e}}

100: 

101: 

102: 

103: \begin{document}

104: %\journalinfo{Target: Astrophysical Journal}

105: %

106: \slugcomment{Version 4.01: \today}

107: %

108: %           12345678901234567890123456789012345678901234

109: \shorttitle{Sgr A* Flare Detection with Millimeter VLBI}

110: %

111: \shortauthors{Doeleman et al.}

112: %

113: \title{Methods for Detecting Flaring Structures in Sagittarius A* with High Frequency VLBI}

114: %

115: \author{

116: %

117: Sheperd S.~Doeleman\altaffilmark{1},

118: %

119: Vincent L.~Fish\altaffilmark{1},

120: %

121: Avery E.~Broderick\altaffilmark{2},

122: %

123: Abraham Loeb\altaffilmark{3}, \&

124: %

125: Alan E.~E.~Rogers\altaffilmark{1}}

126: %

127: \altaffiltext{1}{Massachusetts Institute of Technology, Haystack

128:   Observatory, Route 40, Westford, MA 01886.}

129: %

130: \altaffiltext{2}{Canadian Institute for Theoretical Astrophysics,

131:   University of Toronto, 60 St.\ George St., Toronto, ON, M5S 3H8

132:   Canada.}

133: %

134: \altaffiltext{3}{Institute for Theory and Computation, Harvard

135:   University, Center for Astrophysics, 60 Garden St., Cambridge, MA

136:   02138.}

137: 

138: \begin{abstract}

139: The super massive black hole candidate, Sagittarius~A*, exhibits variability

140: from radio to X-ray wavelengths on time scales that correspond to < 10

141: Schwarzschild radii.  We survey the potential of millimeter-wavelength VLBI to

142: detect and constrain time variable structures that could give rise to such

143: variations, focusing on a model in which an orbiting hot spot is embedded in an

144: accretion disk.  Non-imaging algorithms are developed that use interferometric

145: closure quantities to test for periodicity, and applied to an ensemble of

146: hot-spot models that sample a range of parameter space.  We find that

147: structural periodicity in a wide range of cases can be detected on most

148: potential VLBI arrays using modern VLBI instrumentation.  Future enhancements

149: of mm/sub-mm VLBI arrays including phased array processors to aggregate VLBI

150: station collecting area, increased bandwidth recording, and addition of new

151: VLBI sites all significantly aid periodicity detection.  The methods described herein

152: can be applied to other models of Sagittarius~A*, including jet outflows and

153: Magneto-Hydrodynamic accretion simulations.

154: \end{abstract}

155: \keywords{black hole physics --- Galaxy: center --- techniques:

156:   interferometric --- submillimeter --- accretion, accretion disks}

157: 

158: \section{Introduction}

159: \label{introduction}

160: 

161: Observations of the compact radio/IR/X-ray source Sagittarius~A*

162: (Sgr~A*) make the most compelling case for the existence of super massive

163: black holes.  Both speckle imaging and adaptive optics work in the

164: near-infrared (NIR) band shows that multiple stars orbit the position

165: of Sgr~A* \citep{schodel03,ghez05} These orbits are consistent with a

166: central mass of $\sim4\times10^6 ~ M_\sun$ contained within 45~AU -

167: the closest approach of any star.  Radio interferometric proper motion

168: measurements \citep{backer99,reid04} limit the motion of Sgr~A* to

169: $<15$~km\,s$^{-1}$, implying that Sgr~A* must trace at least 10\% of

170: the mass determined from stellar orbits.  Very long baseline interferometry (VLBI) at 1.3~mm

171: wavelength \citep{doeleman08} has resolved Sgr~A*, and measures an intrinsic

172: size of $<0.3$~AU \citep[assuming a distance of 8.0~kpc, from][]{reid93}, or 4

173: times the Schwarzschild radius of the central mass ($\Rs \approx 10~\mu$as).

174: Assuming Sgr~A* marks the position of the black hole, the implied density,

175: using the proper motion lower limit on the mass and the VLBI size, is

176: $>9.3\times10^5~M_\sun\,{\mathrm{AU}}^{-3}$.  Almost any conceivable

177: aggregation of matter would, at this density, quickly collapse to a black hole

178: \citep{maoz98}.

179: 

180: The 1.3~mm VLBI result confirms the existence of structures within Sgr~A* on

181: size scales commensurate with the innermost accretion region, and matches size

182: scales inferred from light curve monitoring over a broad wavelength range.

183: Sgr~A* exhibits variability on time scales of minutes to hours in the radio,

184: millimeter, NIR, and X-ray bands \citep[e.g.,][]{baganoff01, aschenbach04,

185: genzel03, ghez04, belanger06, meyer06, yusefzadeh06, hornstein07, marrone07},

186: and flare rise times in the X-ray and NIR correspond to light-crossing times of

187: $<12~\Rs$.  Models that produce flaring X-ray flux density via Synchrotron Self

188: Compton scattering of IR photons require emission regions of diameter $<10~\Rs$

189: \citep{marrone07}.  

190: 

191: The resolution of millimeter- and submillimeter-wavelength VLBI is

192: well matched to the scale of inner disk physics.  Baselines from

193: Hawaii or Western Europe to Chile provide fringe spacings as small as

194: $30~\mu$as ($3~\Rs$) at 230~GHz and $20~\mu$as at 345~GHz.  Millimeter VLBI

195: thus has the potential to detect signatures of hot spot and jet models proposed

196: to explain the rapid variability of Sgr~A* as well as strong general

197: relativistic effects, such as the black hole silhouette or shadow

198: \citep{falcke00,broderick06b,huang07,markoff07}.  Extending the VLBI technique

199: to short (0.85~mm, 1.3~mm) wavelengths is essential for this work due to the

200: interstellar scattering towards Sgr~A*, which broadens radio images with a

201: $\lambda^2$ dependence \citep{backer78}.  VLBI at 7~mm and 3~mm wavelengths

202: \citep{rogers94,doeleman01,bower04,shen05} limits the intrinsic size of Sgr~A*

203: to be $<2$~AU and $<1$~AU respectively, but VLBI at these wavelengths is

204: strongly influenced by scattering effects.  For $\lambda < 1.3$~mm the scattering size is

205: less than the fringe spacings on the longest possible baselines.  Recent 1.3~mm

206: VLBI results \citep{doeleman08}, coupled with ongoing technical advances to

207: reach 0.85~mm, strongly suggest that it is not a question of \emph{if} but of

208: \emph{when} VLBI will directly probe Sgr~A* on event horizon scales.

209: 

210: Claims of observed periodicity in IR and X-ray light curves

211: \citep{belanger06, meyer06, eckart06} during Sgr~A* flares, can

212: potentially be explained in the context of hot spots orbiting the

213: black hole at a few times $\Rs$.  It has been proposed that the

214: fastest periodicity can be used to constrain the spin of the black

215: hole, since the period of the innermost stable circular orbit (ISCO)

216: is much shorter for a maximally rotating Kerr black hole than for a

217: nonrotating Schwarzschild black hole \citep{genzel03}.  Indeed, several

218: authors have argued that the black hole must be rotating based on

219: observed rapid X-ray and infrared periodicities

220: \citep{aschenbach04,belanger06,meyer06}.  However, it has also been

221: proposed that the flaring activity can be explained by

222: magnetohydrodynamic turbulence along with density fluctuations

223: \citep{goldston05,chan06}.  Rossby wave instabilities may naturally

224: produce periodicities on the order of tens of minutes, in which case

225: it is not necessary to appeal to a nonzero black hole spin to explain

226: the fast quasi-periodic flares seen at multiple wavelengths

227: \citep{tagger06,falanga07}.  Regardless of the source of variability

228: at millimeter wavelengths, VLBI has the potential to confirm

229: conclusively its association with the inner disk of Sgr~A*, to probe

230: the size of the region of variability (since an interferometer acts as

231: a spatial filter on the emission), and to extract periodicity.

232: 

233: Preliminary studies involving the analysis of simulated data from

234: expected models are important for a number of reasons.  Such studies

235: will highlight the abilities and limitations of millimeter VLBI in

236: regards to detecting the signatures of the physical processes in the

237: accretion disk surrounding the black hole in Sgr~A*.  Critical

238: resources (such as stable frequency standards, high-bandwidth

239: recording equipment, and phased array processors) are likely to be

240: limited initially, and telescope upgrades (such as surface accuracy

241: improvement, expanded IF bandwidth, additional receiver bands, and

242: simultaneous dual-polarization capability) must necessarily be prioritized.

243: Simulated observational data can help assess the tradeoffs that must be

244: considered for optimization of Galactic Center VLBI observations.  The ultimate

245: goal is to explore the potential of black hole parameter estimation by present

246: and future millimeter VLBI observations.

247: 

248: In this manuscript, we explore the observational signatures of an orbiting hot

249: spot embedded in a quiescent disk around Sgr~A*.  We consider a non-imaging

250: approach to analyzing millimeter VLBI data for several reasons.  First, one

251: fundamental assumption of Earth rotation aperture synthesis is that the source

252: structure is not changing.  Since orbital periods in the hot spot models are

253: much shorter than the rotation of the Earth, this assumption is clearly

254: violated.  Second, phase-referencing the data is presently not feasible at

255: millimeter wavelengths since the phase path through the atmosphere changes on a

256: time scale which is faster than the time needed to move the antennas between

257: the reference source and Sgr A*.  We note, though, that this problem could be

258: circumvented at connected-element arrays where some antennas could be dedicated

259: to simultaneously observing a reference source while others observe Sgr A*, but

260: this is also currently limited by the low SNR that can be achieved in the

261: coherence time of the atmosphere.  Potential arrays for millimeter VLBI will

262: have a small number of telescopes, initially as few as three, possibly with

263: vastly different sensitivies.  Low expected signal-to-noise ratios (SNRs) on

264: some baselines combined with the few antennas available will prevent adequate

265: self-calibration via closure relations.  Third, even if the visibility data

266: could be properly calibrated and the source structure were not changing over

267: the observation, the $(u,v)$-plane would be sparsely populated with noisy data

268: points, resulting in poor image fidelity, as shown by the simulations of

269: \citet{miyoshi04}.  At least initially, it will be more productive to analyze

270: the data by model fitting in the visibility domain rather than in the image

271: domain.

272: 

273: 

274: \section{Models of Sgr A*}

275: \label{models}

276: 

277: Models for the flaring emission of Sgr~A* at millimeter wavelengths

278: necessarily require a number of components, most succinctly decomposed

279: into models for the quiescent emission and models for the

280: short-timescale dynamical phenomena responsible for the flare.  Any

281: such model has a number of existing observational constraints that it

282: must meet, including reproducing the observed spectra \& polarization

283: properties of the quiescent \& flaring emission as well the dynamical

284: properties of the flare light curves.  Here we describe a set of flare

285: models involving orbiting hot spots embedded within a large-scale

286: accretion flow that are consistent with all existing observations,

287: based upon those described in \citet{broderick06b}.

288: 

289: \subsection{Quiescent Emission}

290: 

291: As implied by its spectrum, Sgr~A* is only starting to become optically thin at

292: millimeter wavelengths.  Due to relativistic effects this does not

293: happen isotropically \citep[e.g.,][]{broderick06a}.  As a consequence,

294: the opacity of the underlying accretion flow is important for both

295: imaging the black hole's silhouette and for the variability arising

296: from hot spots on compact orbits.

297: 

298: Despite being faint compared to the Eddington luminosity for

299: a $4\times10^6\,\Ms$ black hole, Sgr~A* is still considerably bright,

300: emitting a bolometric luminosity of approximately $10^{36}\,\erg\,

301: \s^{-1}$.  As a consequence it has been widely accepted that Sgr~A*

302: must be accretion powered, implying a minimum accretion rate of at

303: least $10^{-10}\,\Ms\, \yr^{-1}$.

304: 

305: It is presently unclear how this emission is produced.  This is

306: evidenced by the variety of models that have been proposed to explain

307: the emission characteristics of Sgr A*

308: \citep[e.g.,][]{narayan98,blandford99,falcke00,yuan02,yuan03,loeb07}.

309: Models in which the emission arises directly from the accreting

310: material have been subsumed into the general class of Radiatively

311: Inefficient Accretion Flows (RIAF), defined by the generally weak

312: coupling between the electrons, which radiate rapidly, and the ions,

313: which efficiently convert gravitational potential energy into heat

314: \citep{narayan98}. This coupling may be sufficiently weak to allow

315: accretion flows substantially in excess of the $10^{-10}\,\Ms\,

316: \yr^{-1}$ required to explain the observed luminosity with a canonical

317: radiative efficiency.

318: 

319: Nevertheless, following the detection of polarization from Sgr A*

320: above $100\,\GHz$ \citep{aitken00,bower01,bower03,marrone06}, and

321: subsequent measurements of the Faraday rotation measure

322: \citep{macquart06,marrone07b}, the accretion rate near the black hole

323: has been inferred to be significantly less than the Bondi rate,

324: implying the existence large-scale outflows \citep{agol00,quataert00}.

325: Therefore, in the absence of an unambiguous theory, we adopt a simple,

326: self-similar model for the underlying accretion flow which includes

327: substantial mass-loss.

328: 

329: Following \citet{yuan03}, this model is

330: characterized by a Keplerian velocity distribution, a population of

331: thermal electrons with density and temperature

332: \begin{equation}

333: n_{e,\rm th} = n^0_{e,\rm th} \left(\frac{r}{\Rs}\right)^{-1.1} \e^{-z^2/2 \rho^2}

334: \quad {and} \quad

335: T_{e} = T^0_{e} \left(\frac{r}{\Rs}\right)^{-0.84}\,,

336: \end{equation}

337: respectively, a population of non-thermal electrons

338: \begin{equation}

339: n_{e,\rm nth} = n^0_{e,\rm nth} \left(\frac{r}{\Rs}\right)^{-2.9} \e^{-z^2/2 \rho^2}\,,

340: \end{equation}

341: and spectral index $\alpha_{\rm disk} = 1.25$ (defined as

342: $S\propto\nu^{-\alpha_{\rm disk}}$), and a toroidal magnetic field in

343: approximate ($\beta=10$) equipartition with the ions (which produce

344: the majority of the pressure), i.e.,

345: \begin{equation}

346: \frac{B^2}{8\pi}

347: =

348: \beta^{-1} n_{e,\rm th} \frac{m_p c^2 \Rs}{12 r}\,.

349: \end{equation}

350: In all of these expressions the radial structure was taken directly

351: from \citet{yuan03} and the vertical structure was determined by

352: assuming the disk height is comparable to the polar radius, $\rho$.

353: To correct for the fact that \citet{yuan03} was a Newtonian study, we

354: determine the three coefficients ($n^0_{e,\rm th}$, $T^0_e$ and

355: $n^0_{e,\rm nth}$) by fitting the the radio, submillimeter and

356: near-infrared spectrum of Sgr~A*.  For every inclination and black

357: hole spin presented here this was possible with extraordinary accuracy

358: (reduced $\chi^2<1$ in all cases and $\lesssim0.2$ for many), implying

359: that this model is presently significantly under-constrained by the

360: quiescent spectrum alone.  It is also capable of producing the Faraday

361: rotation measures observed, and thus the polarimetric properties of

362: Sgr~A*.

363: 

364: The primary emission mechanism is synchrotron, arising from both the

365: thermal and non-thermal electrons.  We model the emission from the

366: thermal electrons using the emissivity described in \citet{yuan03},

367: appropriately altered to account for relativistic effects \citep[see,

368:   e.g.,][]{broderick04}.  Since we perform polarized radiative

369: transfer via the entire complement of Stokes parameters, we employ the

370: polarization fraction for thermal synchrotron as derived in

371: \citet{petrosian83}.  In doing so we have implicitly assumed that the

372: emission due to thermal electrons is isotropic, which while generally

373: not the case is unlikely to change our results significantly.  For the

374: non-thermal electrons we follow \citet{jones77} for a power-law

375: electron distribution, cutting the electron distribution off below a

376: Lorentz factor of $10^2$ and corresponding to a spectral index of

377: $\alpha_{\rm disk}=1.25$, both roughly in agreement with the

378: assumptions in \citet{yuan03}.  For both the thermal and non-thermal

379: electrons the absorption coefficients are determined directly via

380: Kirchoff's law.

381: 

382: 

383: \subsection{Flares}

384: 

385: We model the flare emission by a localized over-density in the

386: non-thermal electron distribution.  This naturally explains the

387: short-timescale variability and potential periodicity claimed in some

388: flares \citep{genzel03,belanger06}.  Such a feature is also a natural

389: consequence of dissipation in black hole accretion flows.

390: 

391: Sgr~A*'s quiescent radio spectrum appears to require a population of

392: non-thermal electrons.  If strong magnetic turbulence is present,

393: driven by, e.g., the magnetorotational instability, the production of

394: non-thermal electrons at strong shocks and magnetic reconnection

395: events is unavoidable.  Generally, we expect that the production of

396: non-thermal electrons will be most prominent in the innermost regions

397: of the accretion flow, where the magnetic turbulence is strongest.

398: 

399: It is important to note that the region producing the flare need not

400: be dynamically important.  Within the context of a RIAF model for the

401: accretion flow onto Sgr~A*, the pressure is overwhelmingly dominated

402: by the ions.  This is a direct consequence of the assumed weak

403: coupling between the electrons and ions, and thus the low luminosity

404: of the accretion flow.  For typical RIAF accretion rates

405: ($10^{-8}\,\Ms\, \yr^{-1}$), the luminosity would need to be increased

406: by orders of magnitude before the non-thermal electrons become

407: dynamically significant.  For the observed flares, which typically do

408: not increase the NIR luminosities by more than a single order of

409: magnitude, this means that the accelerated electrons will be frozen

410: into the accretion flow.  In this case, the size of the emitting

411: region, $2 \Delta r$, is determined by the scale of the magnetic

412: turbulence.

413: 

414: The dominant constraints upon the lifetimes of hot spots in the

415: accretion flow are Keplerian shear and synchrotron cooling.  Hot spots

416: will shear apart on roughly $r/\Delta r$ orbital periods, and thus

417: small spots will last many orbits.  Unlike shear, cooling is not

418: achromatic, with the flare cooling more rapidly at higher frequencies.

419: The cooling time is approximately

420: \begin{equation}

421: \tau_{\rm c}

422: \simeq

423: 3 \left(\frac{\lambda}{1\,\mm}\right)^{1/2} \left(\frac{B}{30\,

424:   \G}\right)^{-3/2}\,\hr\,,

425: \label{eqn-cooling}

426: \end{equation}

427: which should be compared to the period at the ISCO, ranging from

428: $30\,\min$ for a non-rotating black hole to $4\,\min$ for a maximally

429: rotating black hole.  Thus we expect that hot spots will typically

430: survive many orbits at millimeter and submillimeter wavelengths.

431: 

432: Our hot-spot model consists of a locally spherically symmetric,

433: Gaussian over-density of non-thermal electron.  Explicitly, given

434: $\Delta x^\mu = x^\mu-x_{\rm spot}^\mu$ and a scale, $R_{\rm spot}=0.75\Rs$,

435: the hot-spot density is

436: \begin{equation}

437: n_{e,\rm spot} = n^0_{e,\rm spot}

438: \exp\left[-\frac{\Delta x^\mu \Delta x_\mu + \left(u_{\rm spot}^\mu

439:     \Delta x_\mu\right)^2}{2 R_{\rm spot}^2} \right],

440: \end{equation}

441: where the hot-spot four-velocity, $u_{\rm spot}^\mu$ is assumed to be

442: the same as that of the underlying disk (which we have chosen to be

443: Keplerian).  Our description of the hot spot is completed by the

444: spectral index of the power-law distribution of electrons,

445: $\alpha_{\rm spot}=1.3$, taken from observations of NIR

446: flares \citep{eckart04}.  Like the disk, the hot-spot radiates

447: primarily via synchrotron.

448: 

449: 

450: 

451: \subsection{Generating Images}

452: 

453: The method by which images of the flaring disk are produced is

454: discussed at length in \citet{broderick03} \citep[see

455:   also,][]{broderick06a,broderick06b}.  As a result, we only briefly

456: summarize the procedure here.

457: 

458: Null geodesics are constructed by integrating a Hamiltonian

459: formulation of the geodesic equations:

460: \begin{equation}

461: \frac{dx^\mu}{d\eta} = f(x^\sigma) \, k^\mu

462: \quad{\rm and}\quad

463: \frac{dk_\mu}{d\eta} = - \left.\frac{f(x^\sigma)}{2}\frac{\partial k^\nu

464:   k_\nu}{\partial x^\mu}\right|_{k_\alpha}\,,

465: \end{equation}

466: where $f(x^\sigma)$ is an arbitrary function, corresponding to the

467: freedom inherent in the affine parametrization, $\eta$.  In order to

468: regularize the affine parametrization near the horizon, we choose

469: \begin{equation}

470: f(x^\sigma) = r^2\sqrt{1-\frac{r}{r_{\rm h}}},

471: \quad{\rm where}\quad

472: r_{\rm h } = \frac{\Rs}{2}\left( 1 + \sqrt{1-a^2} \right)

473: \end{equation}

474: is the horizon radius ($a$ is the dimensionless black hole

475: spin).  It can be explicitly shown that this does

476: reproduce the null geodesics \citep{broderick03}. 

477: 

478: The relativistic generalization of the radiative transfer problem is

479: most easily obtained by directly integrating the Boltzmann equation

480: directly \citep{lindquist66,broderick06a}.  In this case, it is the

481: photon distribution function, $N_\nu\propto I_\nu/\nu^3$ (which has

482: the virtue of being a Lorentz scalar), that is evolved.  In the case

483: of polarized transfer, it is possible to define the covariant

484: analogues of the Stokes parameters, ${\bf N}_\nu = \left( N_\nu,

485: N^Q_\nu, N^U_\nu, N^V_\nu \right)$, in terms of a parallel propagated

486: tetrad \citep{broderick04}.  In terms of these, the radiative transfer

487: equation takes on its standard form:

488: \begin{equation}

489: \frac{d{\bf N}_\nu}{d\eta}

490: =

491: {\bf \bar{j}}_\nu - \bmath{ \bar{\alpha}}_\nu {\bf N}_\nu\,,

492: \end{equation}

493: where ${\bf \bar{j}}$ and $\bmath{\bar{\alpha}_\nu}$ are the

494: appropriately generalized emission and absorption coefficients, and

495: may be trivially related to the same quantities in the local plasma

496: frame \citep{lindquist66,broderick04}.

497: 

498: Images are produced by tracing a collection of initially parallel null

499: geodesics from pixels on a distant plane backwards in time, towards

500: our model of Sgr~A*, terminating the ray when it has been captured by

501: the black hole, escaped the system, or accrued an optical depth

502: greater than $10$.  Along each ray we integrate the polarized

503: radiative transfer, obtaining ${\bf N}_\nu$ at the original plane.  We

504: construct 100 such images with resolutions of $128\times128$ pixels

505: for each hot-spot orbit.  This procedure is repeated for each

506: frequency of interest, for which we keep the underlying physical model

507: fixed.  Thus, for each spin/inclination pair, the relationship between

508: 230~GHz and 345~GHz images is dictated by the spectral properties of

509: the source.

510: 

511: 

512: \section{VLBI Analysis}

513: \label{VLBI}

514: 

515: \subsection{Techniques}

516: \label{techniques}

517: 

518: Typical VLBI analysis techniques employ an iterative

519: ``self-calibration'' loop, whereby a sky brightness model, the VLBI

520: data, and a series of complex gains for each VLBI site are brought

521: into convergence \citep{cornwell81,cornwell82}.  This process relies

522: on the assumption that all array calibration can be expressed as

523: station-based gains, which is equivalent to requiring that all

524: calibration errors 'close', or cancel when computed over suitable

525: closed loops of VLBI baselines.  In almost all cases, this assumption

526: is valid, and closure quantities can be constructed from the data,

527: which contain structural information on the observed source, but that

528: are largely immune to station-based phase and gain errors

529: \citep{jennison58,twiss60}.  However, closure quantities alone are

530: insufficient to determine baseline phases for a VLBI array.  The

531: number of independent closure phases computed over closed triangles of

532: VLBI stations, for example, grows as $\onehalf \,(N-1)\,(N-2),$ where

533: $N$ is the number of antennas, while the number baseline phases in the

534: array will be $\onehalf \, N \,(N-1).$ The fraction of visibility

535: phase information available from closure phases is thus $(N-2)/N$.

536: For millimeter VLBI $N$ will be small (initially $N = 3$ or $4$; \S

537: \ref{antennas}), and closure quantities will not be sufficiently

538: numerous to allow for full calibration of the data.  However, closure

539: quantities are robust observables and can therefore be used for model

540: fitting even when baseline-based visibilities are contaminated with

541: station-based phase and gain errors.  Indeed, closure quantities have

542: been successfully used for model fitting \citep[e.g.,][]{rogers74},

543: including recent experiments to place limits on the apparent size and

544: structure of Sgr~A* at wavelengths as short as 3~mm

545: \citep{doeleman01,bower04,shen05,markoff07}.

546: 

547: Closure phase and amplitude in the weak-detection limit are discussed

548: in detail by \citet{rogers95}.  We summarize the most relevant

549: information below.

550: 

551: The closure phase is the sum of the baseline phases along a triangle

552: of antennas.  It is independent of instrumental and atmospheric

553: complex gain terms.  The closure phase of a symmetric distribution of

554: emission is always zero or 180\degr\ \citep[e.g.,][]{monnier07}.

555: Deviations from these values are indicative of asymmetries (about the

556: origin) in the source structure on the size scales probed by the

557: baselines of the triangle.  A static asymmetric source structure will

558: show slow variations in closure phase over the course of observations

559: due to the rotation of the Earth, which changes the projected length

560: and orientation of baselines.  If the source structure is changing, as

561: would be the case for an orbiting hot spot, the closure phase on some

562: triangles will change as well.  Because the time scale of hot spot

563: orbits near the black hole is on the order of tens of minutes

564: (depending on the mass and spin of the black hole and the orbital

565: radius of the hot spot), closure phase variability will be much more

566: rapid for the case of a hot spot embedded in a disk as compared to a

567: quiescent disk alone.  Since a hot spot may survive for

568: several orbits before cooling or shearing (\S \ref{models}), closure

569: phases can exhibit approximate perodic behavior over several cycles.

570: Thus, closure phases are appropriate observables for detecting

571: periodic source structure changes on short time scales.  The SNR of

572: the closure phase is dominated by the lowest SNR of the visibility

573: data along the three segments.

574: 

575: The closure amplitude is constructed as the ratio of products of

576: visibility amplitudes ($A$) along a quadrangle of antennas.  For four

577: antennas $a$, $b$, $c$, and $d$, two independent closure amplitudes

578: can be constructed:

579: \begin{equation}

580: A_{abcd} \equiv \frac{|A_{ab}| |A_{cd}|}{|A_{ac}| |A_{bd}|}

581: \quad \mathrm{and} \quad

582: A_{adbc} \equiv \frac{|A_{ad}| |A_{bc}|}{|A_{ac}| |A_{bd}|}.

583: \end{equation}

584: For an array of $N$ antennas, there are $\onehalf\,N\,(N-3)$ independent

585: closure amplitudes.  As for closure phase, the closure amplitude is

586: unaffected by gain calibration errors on each of the antennas.

587: Deviations of the closure amplitude from unity are usually indicative

588: of resolved, asymmmetric source structure.  Large excursions in

589: closure phase and amplitude are often correlated, since small changes

590: with time in the complex visibility, due to Earth rotation or source

591: structure changes, have the largest effects on the phase when the

592: visibility amplitude is near zero.  Since a small visibility amplitude

593: results in a very small closure amplitude if it appears in the

594: numerator or a very large closure amplitdue if it appears in the denominator.

595: For this reason, closure amplitudes are generally presented on logarithmic

596: scales.

597: 

598: In the low-signal regime, the closure amplitude is a biased quantity.

599: The detected visibility amplitude on a baseline is the modulus of the

600: vector (amplitude and phase) sum of the signal and the noise and is by

601: definition nonnegative.  When the SNR is small, the visibility

602: amplitude is dominated by the noise amplitude and is therefore larger

603: than the signal amplitude.  Thus, if the source is not detected on

604: baseline $ab$ but is detected on baselines $cd$, $ac$, and $bd$, the

605: closure amplitude $A_{abcd}$ will on average be larger than the value

606: predicted by a noiseless model of the source structure.  The closure

607: phase does not suffer from a similar bias.

608: 

609: 

610: \subsection{Antennas}

611: \label{antennas}

612: 

613: \input{tab1.tex}

614: 

615: \begin{figure}                                                                                                          

616: \resizebox{\hsize}{!}{\includegraphics{f1a.eps}\includegraphics{f1b.eps}}                                               

617: \caption{Locations of candidate telescopes for millimeter-wavelength                                                    

618:   VLBI as viewed from the declination of Sgr~A*.                                                                        

619: \label{fig-locations}                                                                                                   

620: \notetoeditor{For the reader's convenience, Figure 1 should be placed                                                   

621:   near Table 1, preferably on the same page}                                                                            

622: }                                                                                                                       

623: \end{figure}             

624: 

625: Whereas centimeter-wavelength astronomy can be performed from

626: virtually any location with a clean spectrum, potential sites for

627: (sub)millimeter-wavelength astronomy are limited by the need to be

628: above most of the water vapor content in the atmosphere.

629: Consequently, possible arrays for millimeter VLBI are sparser

630: than for centimeter VLBI.  In this section, we summarize the best

631: prospects for millimeter VLBI among existing telescopes and those that

632: may come on line in the near future.

633: 

634: \emph{Hawaii}: The Caltech Submillimeter Observatory (CSO), James

635: Clerk Maxwell Telescope (JCMT), and Submillimeter Array (SMA) are all located atop Mauna

636: Kea.  Each telescope can observe in both the 230 and 345~GHz bands.

637: The JCMT has a single-polarization 230~GHz receiver and a

638: dual-polarization 345~GHz receiver.  The SMA presently does not

639: support simultaneous dual-polarization observations but is expected to

640: do so eventually at 345~GHz.  Development of instrumentation to phase

641: together the CSO, JCMT and several SMA dishes is underway, and will

642: result in an effective 23~m aperture, which we refer to as ``Hawaii,''

643: by early 2009.

644: 

645: \emph{CARMA}: The Combined Array for Research in Millimeter-wave

646: Astronomy (MWA) consists of six 10.4~m antennas and nine 6.1~m

647: antennas.  The ability to phase together up to 8 CARMA dishes

648: for VLBI is planned, and will result in an effective 27~m aperture.

649: CARMA can presently observe in the 230~GHz band.  A future upgrade to

650: include 345~GHz is planned, but it is unclear when this band will be

651: available for observations.  We also consider a single 10.4~m dish at

652: 230~GHz (``CARMA 1'') in \S \ref{results-arrays}.

653: 

654: \emph{SMTO}: The Arizona Radio Observatory Submillimeter Telescope (SMT) on Mt.

655: Graham in Arizona is a 10~m dish capable of observing at both 230 and 345~GHz.

656: 

657: \emph{LMT}: The Large Millimeter Telescope (LMT), presently under

658: construction, will be a 50~m dish capable of observing at both the 230

659: and 345~GHz bands.  When complete, it will be the most sensitive

660: millimeter telescope in the Northern hemisphere.  In anticipation that 

661: the LMT collecting area will be installed in phases, we conservatively adopt

662: an effective aperture of 32~m.

663: 

664: \emph{Chile}: Several millimeter telescopes are available in Chile.

665: The Atacama Submillimeter Telescope Experiment (ASTE) is a 10~m dish

666: with a single-polarization receiver at 345~GHz.  The Atacama

667: Pathfinder Experiment (APEX) is a 12~m dish with a double-sideband

668: (DSB) at 345~GHz.  Future two sideband (2SB) heterodyne receivers

669: capable of observing at 230~GHz and 345~GHz are under construction.

670: We refer to using one of these facilities as ``Chile 1.''  ALMA will

671: be composed of a large number of 12~m dishes with 2SB receivers at

672: both 230 and 345~GHz.  We also consider the possibility of a

673: 10-element phased ALMA station as the Chilean site (``Chile 10''),

674: since our models do not produce much expected signal on the long

675: baselines to Chile (\S \ref{methods}).

676: 

677: \emph{Pico Veleta}: The 30~m Institut de Radioastronomie

678: Millim\'{e}trique (IRAM) telescope on Pico Veleta (PV) can observe at

679: 230~GHz.

680: 

681: \emph{Plateau de Bure}: The IRAM PdB Interferometer consists of six

682: 15~m telescopes equipped with 230~GHz receivers.  An upgrade to add

683: 345~GHz capability is presently under construction and is expected to

684: be available in late 2008.  The six telescopes can currently be phased

685: up over a 256~MHz bandwidth into a single 37~m-equivalent aperture.

686: Extension to higher bandwidths will require instrumentation

687: development similar to what will be deployed to phase up the antennas

688: on Mauna Kea.

689: 

690: A summary of telescope capabilities is given in Table~\ref{tab-ants}.

691: Sensitivity is given in terms of the system equivalent flux density

692: (SEFD), which is equal to $2 k T_\mathrm{sys}/A_\mathrm{eff}$, where

693: $k$ is Boltzmann's constant and the effective area $A_\mathrm{eff}$ is

694: the product of the geometric area and the aperture efficiency.

695: Assumed $T_\mathrm{sys}$ values toward Sgr~A* include a typical

696: expected atmospheric contribution, which can be quite large for

697: Northern hemisphere telescopes due to the low elevation of the

698: Galactic Center.  Since the atmospheric contribution is highly

699: weather-dependent, actual observations may achieve significantly

700: different values of the SEFD at each telescope.  SEFD values include a

701: 10\% phasing loss factor for phased arrays.  All facilities can in

702: principle provide an IF bandwidth of at least 4~GHz, although

703: accessing the full bandwidth at some stations may be problematic.

704: Thus, 32~Gbit\,s$^{-1}$, corresponding to 2-bit Nyquist samples of a

705: 4~GHz bandwidth in two orthogonal polarizations, is the maximum

706: possible recording rate that we will consider.  At present

707: 16~Gbit\,s$^{-1}$ digital back ends are still in the planning stage,

708: so initial observations will likely employ a smaller bandwidth.

709: Except as noted, all telescopes can observe two polarizations

710: simultaneously.

711: 

712: The European telescopes (PV and PdB) and the North American telescopes

713: (Hawaii, CARMA, SMTO, and LMT) have no mutual visibility of Sgr~A*,

714: except for approximately 1~hr of overlap above 10\degr\ elevation on

715: the PV-LMT baseline.  Thus, there are only three possible types of

716: subarrays with at least three elements possible among these millimeter

717: facilities: North America only, North America plus Chile, and Europe

718: plus Chile (Fig.~\ref{fig-locations}).  For our simulations, we take our 230~GHz array to consist

719: of Hawaii, CARMA, SMTO, LMT (32~m), either APEX or a 10-element phased

720: ALMA, PV, and PdB.  Early science at 345~GHz will likely consist of

721: the single triangle of Hawaii, SMTO, and a Chilean telescope.

722: However, we consider the same set of telescopes as at 230~GHz in order

723: to illustrate what future 345~GHz upgrades might accomplish.

724: 

725: 

726: \subsection{Methods}

727: \label{methods}

728: 

729: Simulated data were obtained using task UVCON in the Astronomical

730: Image Processing System (AIPS).  Synthetic data were produced using an

731: averaging interval of 10~s.  Typical coherence times at 230~GHz are 10~s, but

732: can be as low as 2-4~s and, under good weather conditions, as long as 20~s

733: \citep{doeleman02}.  At the ALMA site in Chile, the measured coherence time of

734: the atmosphere is > 10 seconds 60\% of the time at 230GHz and 45\% of the time

735: at 345GHz \citep{holdaway97}.  Since AIPS cannot directly handle time-varying

736: source structure, for a hot spot model we simulate data from 100 static,

737: equally-spaced time slices in a single hot spot orbit and construct a data set.

738: All data assume 2-bit sampling at the Nyquist rate.  Thus, the recording rate

739: (in Gbit\,s$^{-1}$) is four times the observing bandwidth (in GHz).  For

740: continuum observations at a constant sampling rate, 2-bit quantization achieves

741: sensitivity levels very nearly approaching that of 1-bit quantization at half

742: the observing bandwidth \citep{thompson01}, which is limited by the hardware at

743: certain telescopes (\S \ref{antennas}).

744: 

745: Observations of total intensity (i.e., Stokes $I$) theoretically

746: obtain identical noise levels regardless of whether the data are taken

747: at full bandwidth in single-polarization mode or at half bandwidth in

748: dual-polarization mode, provided that circularly-polarized feeds are

749: used.  The hot spot models do not produce circular polarization

750: (Stokes $V$) but do produce large amounts of linear polarization

751: (Stokes $Q$ and $U$), especially on small spatial scales.

752: Circularly-polarized feeds measures Stokes visibilities $I \pm V$ and

753: therefore are identical for our models.  Dual-polarization

754: observations are essential if linearly-polarized feeds are used, since

755: the parallel-hand data are sensitive to the Stokes visibilities $I \pm

756: P$, where $P$ is a linear combination of Stokes $Q$ and $U$ (depending

757: on feed orientation).  In practice the observer will usually prefer to

758: observe in dual-polarization mode when available even using circular

759: feeds, since cross-hand correlation products provide polarimetric

760: information as well.  The polarimetric products are of special

761: interest, since they may show asymmetries not seen in total intensity

762: \citep[e.g.,][]{bromley01}.  However, we focus on total intensity

763: methods and results in this work and defer polarimetric considerations

764: to a future manuscript.

765: 

766: In the subsequent discussion, we consider a suite of models

767: parameterized by black hole spin ($a = 0$ nonrotating or $a = 0.9$

768: highly rotating), disk orientation, and hot spot orbital radius ($r$).

769: For each spin, the radius of the ISCO ($r_{ISCO} = 6 \, R_G$ for $a =

770: 0$ and $2.32 \, R_G$ for $a = 0.9$)

771: and one larger radius are chosen; the larger radius for $a = 0.9$ is

772: chosen to have the same period as the $a = 0$, ISCO model.  Disk

773: models assume that the spin axes of the black hole and accretion disk

774: are aligned, with the major axis of the projected disk aligned

775: east-west except in Model C, in which the disk and hot spot are

776: aligned north-south instead.  Flux densities have been scaled to match

777: connected-element interferometric observations.

778: Contributions from the quiescent disk are assumed to be approximately

779: 3~Jy, depending slightly on the specific model parameters chosen, as

780: listed in Table~\ref{tab-models}.  A single prograde orbiting hot spot

781: contributes a variable flux density component, depending on orbital

782: phase and the specific model, consistent with submillimeter

783: observations \citep{marrone07b}.  The hot

784: spot contributes some flux density even at the minimum in the light

785: curve, an effect that is most pronounced in the $i = 30$\degr\ models.

786: All models are convolved with the expected interstellar scattering given by

787: \citep{bower06}.  While these model parameters do not span the entire range of

788: possibilities of the Sgr A* disk system, they do show how changes in model

789: parameters affect the observable quantities.  Non-imaging VLBI methods are

790: applicable to all reasonable models for the Sgr A* system.  Figure~\ref{fig-modelb} shows 

791: a series of images from Model B at both 230~GHz and 345~GHz.

792: 

793: \input{tab2.tex}

794: 

795: \begin{figure*}                                                                                                         

796: \resizebox{0.19\hsize}{!}{\includegraphics{f2a.eps}}                                                                    

797: \resizebox{0.19\hsize}{!}{\includegraphics{f2b.eps}}                                                                    

798: \resizebox{0.19\hsize}{!}{\includegraphics{f2c.eps}}                                                                    

799: \resizebox{0.19\hsize}{!}{\includegraphics{f2d.eps}}                                                                    

800: \resizebox{0.19\hsize}{!}{\includegraphics{f2e.eps}}\\                                                                  

801: \resizebox{0.19\hsize}{!}{\includegraphics{f2f.eps}}                                                                    

802: \resizebox{0.19\hsize}{!}{\includegraphics{f2g.eps}}                                                                    

803: \resizebox{0.19\hsize}{!}{\includegraphics{f2h.eps}}                                                                    

804: \resizebox{0.19\hsize}{!}{\includegraphics{f2i.eps}}                                                                    

805: \resizebox{0.19\hsize}{!}{\includegraphics{f2j.eps}}\\                                                                  

806: \resizebox{0.19\hsize}{!}{\includegraphics{f2k.eps}}                                                                    

807: \resizebox{0.19\hsize}{!}{\includegraphics{f2l.eps}}                                                                    

808: \resizebox{0.19\hsize}{!}{\includegraphics{f2m.eps}}                                                                    

809: \resizebox{0.19\hsize}{!}{\includegraphics{f2n.eps}}                                                                    

810: \resizebox{0.19\hsize}{!}{\includegraphics{f2o.eps}}\\                                                                  

811: \resizebox{0.19\hsize}{!}{\includegraphics{f2p.eps}}                                                                    

812: \resizebox{0.19\hsize}{!}{\includegraphics{f2q.eps}}                                                                    

813: \resizebox{0.19\hsize}{!}{\includegraphics{f2r.eps}}                                                                    

814: \resizebox{0.19\hsize}{!}{\includegraphics{f2s.eps}}                                                                    

815: \resizebox{0.19\hsize}{!}{\includegraphics{f2t.eps}}                                                                    

816: \caption{Images of Model B.  Columns show orbital phases 0, 0.2, 0.4,                                                   

817:   0.6, and 0.8 from left to right.  From top to bottom, rows show the                                                   

818:   input model at 230~GHz, the same model after convolution with                                                         

819:   expected interstellar scattering at 230~GHz, the model at 345~GHz,                                                    

820:   and the same model after scattering at 345~GHz.  

821: \label{fig-modelb}

822: }                                                                                                                       

823: \end{figure*}               

824: 

825: The angular resolution available to potential millimeter VLBI arrays

826: is well matched to potentially interesting scales for observing Sgr~A*

827: (Fig.~\ref{fig-uvplt}).  The longest baselines, Hawaii-Chile and

828: PdB-Chile, provide fringe spacings of 30--35~$\mu$as at 230~GHz and

829: 19--23~$\mu$as at 345~GHz, slightly larger than the expected

830: interstellar scattering and only several times $\Rs$.  Our

831: models do not produce much detectable signal on these small angular

832: scales (Fig.~\ref{fig-uv}), but it is possible that smaller hot spots

833: or disk instabilities (not modelled) will produce greater amplitudes

834: on small angular scales.  The shortest baselines, SMTO-CARMA and

835: PV-PdB, provide fringe spacings of 230--1000~$\mu$as at 230~GHz and

836: 160--700~$\mu$as at 345~GHz.

837: 

838: \begin{figure}

839: \resizebox{\hsize}{!}{\includegraphics{f3.eps}}

840: \caption{Possible $(u,v)$ tracks for millimeter VLBI.  Tracks are

841:   labelled by baseline (\emph{A}: ALMA/APEX/ASTE, \emph{B}: Plateau de

842:   Bure, \emph{C}: CARMA, \emph{H}: Hawaii, \emph{L}: LMT, \emph{S}:

843:   SMTO, \emph{V}: Pico Veleta).  Unlabelled tracks correspond to the

844:   baseline indicated by $(-u,-v)$.  Axes are in units of

845:   gigawavelengths at $\nu = 230$~GHz.

846: \label{fig-uvplt}}

847: \end{figure}

848: 

849: \begin{figure}

850: \resizebox{\hsize}{!}{\includegraphics{f4.eps}}

851: \caption{Plot of expected visibility amplitude as a function of

852:   baseline length for two frames of Model B230.  Symbols show

853:   the noiseless visibility amplitudes that would be obtained if the

854:   source structure were frozen near minimum (black) and maximum

855:   (color) flux in the hot spot orbit.  At maximum flux, points are

856:   color-coded and labelled by baseline as in Figure~\ref{fig-uvplt}.

857:   Visibility amplitude falls off rapidly with baseline length, and the

858:   fractional variability on long baselines can be significantly

859:   different than that seen at zero spacing. 

860: \label{fig-uv}}

861: \end{figure}

862: 

863: \section{Results}

864: \label{results}

865: 

866: \subsection{Closure Phases and Amplitudes}

867: 

868: \begin{figure*}

869: \resizebox{\hsize}{!}{\rotatebox{-90}{\includegraphics[]{f5_gimp.eps}}}

870: \caption{Closure phases on selected triangles at 230~GHz.  The solid

871:   line (red in the online edition) shows the predicted closure phase

872:   in the absence of noise.  Each point indicates 10~s of

873:   coherently-integrated data.  The fourth column shows the same

874:   triangle as the third column but with a higher data rate and

875:   substitution of Chile~10 for Chile~1.  The same 2-hour period,

876:   corresponding to 4.5 periods (14.8 periods for Model E) is

877:   shown in all panels excepting PV-PdB-Chile~10.

878: \label{fig-phase230}}

879: \end{figure*}

880: 

881: Figure~\ref{fig-phase230} shows predicted closure phases at 230~GHz on

882: selected triangles.  The closure phase responses are highly dependent

883: on the physical parameters of the Sgr~A* system, but several patterns

884: emerge.  Data from the smallest triangles (i.e., those composed of

885: only North American telescopes) achieve a high SNR even at the present

886: maximum recording rate of 4~Gbit\,s$^{-1}$.  Models with a greater

887: north-south extent (e.g., Models A and C), produce a larger closure

888: phase signature on the small triangles.  The SNR on all triangles to

889: Chile~1 is low owing to the small flux on small angular scales.

890: Higher bit rates and the substitution of phased ALMA for Chile~1

891: greatly increase detectability, as shown by the third and fourth

892: columns, which correspond to the expected array and recording

893: capabilities in the next few years.  Net offsets from zero closure

894: phase on some triangles result from apparent asymmetric structure in

895: the disk caused by lensing and opacity effects.

896: 

897: Figure~\ref{fig-phase345} shows modelled closure phases at 345~GHz.

898: The only available triangle in the near future at 345~GHz will be

899: Hawaii-SMTO-Chile~1, which produces rather low SNR closure phases at

900: 4~Gbit\,s$^{-1}$.  Of the remaining telescopes, both CARMA and Plateau

901: de Bure plan for 345~GHz receivers in the future.  In addition to

902: these, we show triangles including the LMT and Pico Veleta, neither of

903: which has planned 345~GHz capability, to demonstrate what might be

904: seen in the event of future upgrades to those facilities.  At both

905: 230~GHz and 345~GHz, closure phases on several baseline triangles show

906: clear evidence for periodicity associated with hot-spot orbits.

907: 

908: \begin{figure*}

909: \resizebox{\hsize}{!}{\rotatebox{-90}{\includegraphics[]{f6_gimp.eps}}}

910: \caption{Closure phases on selected triangles at 345~GHz.  See

911:   Figure~\ref{fig-phase230} for details.

912: \label{fig-phase345}}

913: \end{figure*}

914: 

915: Four telescopes are required to obtain a closure amplitude, which

916: effectively means that closure amplitudes can only be measured on

917: Western hemisphere arrays.  Figures~\ref{fig-amp230} and

918: \ref{fig-amp345} show closure amplitudes on selected quadrangles at

919: 230 and 345~GHz, respectively.  Periodicity associated with the

920: orbiting hot-spot is evident on most quadrangles.  In some models,

921: closure amplitudes including the Hawaii-Chile~1 baseline show bias (as

922: described in \S \ref{techniques}) due to the low visibility amplitude

923: on this baseline.  The substitution of phased ALMA for Chile~1

924: combined with higher recording rates suffice to clearly detect Sgr~A*

925: on this baseline, resulting in unbiased closure amplitudes.  It is

926: possible (and necessary for model fitting) to de-bias closure

927: amplitudes by correcting the visibility amplitudes for the expected

928: noise levels, but the procedure can be difficult when the SNR on a

929: baseline is low \citep[for more details, see][]{trotter98}.  In any

930: case, the presence of bias does not hinder detection of periodicity.

931: 

932: 

933: \begin{figure*}

934: \resizebox{\hsize}{!}{\rotatebox{-90}{\includegraphics[]{f7_gimp.eps}}}

935: \caption{Closure amplitudes on selected quadrangles at 230~GHz.  See

936:   Figure~\ref{fig-phase230} for details.

937: \label{fig-amp230}}

938: \end{figure*}

939: 

940: \begin{figure*}

941: \resizebox{\hsize}{!}{\rotatebox{-90}{\includegraphics[]{f8_gimp.eps}}}

942: \caption{Closure amplitudes on selected quadrangles at 345~GHz.  See

943:   Figure~\ref{fig-phase230} for details.

944: \label{fig-amp345}}

945: \end{figure*}

946: 

947: 

948: \subsection{Autocorrelation Functions}

949: \label{ACF}

950: 

951: Signatures of time periodic structure associated with orbiting

952: hot-spots can be derived from autocorrelations of closure quantity

953: time series.  The autocorrelation function of a time series of $n$

954: closure amplitudes $A$ on a quadrangle of telescopes is given by

955: \begin{equation}

956: \mathrm{ACF}_A(k) \equiv \frac{1}{(n-k)\,\sigma^2}

957: \sum\limits_{i=1}^{n-k}\left[\left(\log A_i-\mu\right)\left(\log

958: A_{i+k}-\mu\right)\right],

959: \end{equation}

960: where $\mu$ and $\sigma^2$ are the mean and variance of the

961: distribution of the logarithm of the closure amplitudes, respectively.

962: The logarithm is used in preference to the closure amplitude itself

963: due to the tendency for the closure amplitude to obtain both very

964: small (near 0) and very large values.  We also define a variant of the

965: autocorrelation function for closure phases, $\phi$:

966: \begin{equation}

967: \mathrm{ACF}_\phi(k) \equiv \frac{1}{n-k}

968: \sum\limits_{i=1}^{n-k}\cos\left(\phi_i-\phi_{i+k}\right).

969: \end{equation}

970: This definition has the advantage that it handles phase-wrap ambiguity

971: gracefully, since $\cos$ is a periodic function.  The normalizations

972: are such that $\mathrm{ACF}(k) = 1$ when $k$ is an integer multiple of

973: the period for a noiseless periodic function.  In practice, the peaks

974: of an ACF (other than the trivial peak at $k = 0$, where

975: $\mathrm{ACF}(0) = 1$ by definition) of closure phases or amplitudes

976: fall off with lag rather than returning precisely to unity, because

977: the projected baseline geometries change with Earth rotation.  In the

978: case of long periodicities ($\gtrsim 2$~hr), this effect can be large

979: enough to obscure any periodicity at all on some

980: triangles/quadrangles.  While the duration of some SgrA* flare events

981: exceed this time interval, there are claims of modulation within some

982: NIR flares with characteristic time scales of $\sim17$~min

983: \citep{genzel03}.  From an observational perspective, periodicity is

984: not convincingly detected until at least two full periods (preferably

985: more) have been observed.  The longest mutual visibility appears on

986: the SMTO-LMT-Chile triangle, which can see Sgr~A* for less than 7 full

987: hours.  Triangles including Hawaii or Europe have significantly

988: smaller windows of mutual visibility.  The distribution of millimeter

989: telescopes is not optimal for detecting slow periodic variability in

990: Sgr~A*.

991: 

992: The largest nontrivial peak in the ACF indicates the period of the hot

993: spot orbit, as indicated in Figure~\ref{fig-acfs}, excluding the slow

994: periodic case.  The ACF correctly identifies the period in all models

995: at all recording rates provided that the average SNR $\gtrsim 1$ and

996: that the triangle contains at least one baseline long enough to be

997: sensitive to the changing source structure.  On the small triangles,

998: $\mathrm{ACF}_\phi(k)$ shows little variation as a function of lag

999: $k$, since the phase autocorrelation function of a nearly-constant

1000: function is itself nearly constant.  However, small peaks are visible

1001: in the ACF on the Hawaii-SMTO-CARMA and Hawaii-CARMA-LMT triangles in

1002: Figure~\ref{fig-acfs}.  The large SNR on the small triangles, due to

1003: the fact that the shorter baselines do not resolve out much of the

1004: source flux density, permits detectability for reasonable bandwidths.

1005: 

1006: In addition to the peak at the period, the ACF has peaks at integer

1007: multiples of the period.  In the weak-detection limit, this may be an

1008: important discriminator indicating that the detected period is real.

1009: While purely random noise may produce peaks in the ACF, there is no

1010: reason why it should produce periodic peaks.  At the other end of time

1011: scales, intraperiod sub-peaks in the ACF may be produced depending on

1012: the details of the source structure and array geometry.  Excluding

1013: long-period orbits, it is usually clear which peak indicates the

1014: orbital period of the hot spot, since the sub-peaks are of smaller

1015: amplitude.  Nevertheless, it is possible that pathological cases exist

1016: in which the observer may misidentify the orbital period.  For

1017: instance, two identical hot spots at the same radius but separated by

1018: 180\degr\ in azimuth would produce strong peaks in the ACF at both

1019: integer and half-integer multiples of the period, which would lead to

1020: a conclusion that the orbital period is half of its true value.  A

1021: similar ambiguity could arise if the observed variability is produced,

1022: for instance, by the rotation of a pattern produced by a two-armed

1023: Rossby wave instability \citep[e.g.,][]{falanga07}.

1024: 

1025: The key point about the ACF plots is that for the same model,

1026: \emph{all triangles and quadrangles indicate the same period}.  It

1027: will be important to observe with as many telescopes as possible

1028: simultaneously, since the extra information provided by the additional

1029: telescopes may be important for detecting or confirming

1030: marginally-detected variability.  Three antennas provide a single

1031: closure phase, for a total of 1 ACF; four antennas provide three

1032: independent closure phases and two independent closure amplitudes, for

1033: a total of 5 independent ACFs; and five antennas provide six closure

1034: phases and five closure amplitudes, for a total of 11 independent

1035: ACFs.  Even if the source is not detected on one long baseline,

1036: resulting in a biased closure amplitude, the ACF may still clearly

1037: indicate periodicity (e.g., the Hawaii-Chile~1-CARMA-SMTO panel for

1038: Model A230 in Fig.~\ref{fig-acfs}).

1039: 

1040: \begin{figure*}

1041: \resizebox{\hsize}{!}{\rotatebox{-90}{\includegraphics{f9.eps}}}

1042: \caption{Autocorrelation function plots of selected triangles and

1043:   quadrangles.  Representative models are shown.  Black, red, and blue

1044:   lines indicate ACFs for noiseless data, 16~Gbit\,s$^{-1}$, and

1045:   4~Gbit\,s$^{-1}$, respectively.  Panels marked with a red star show

1046:   $\mathrm{ACF}_\phi$ on a different ordinate scale, as indicated to

1047:   the left in that row; other panels in the same row are on the scale

1048:   show in the upper left panel.  The dotted line shows the period of

1049:   the hot spot orbit.

1050: \label{fig-acfs}}

1051: \end{figure*}

1052: 

1053: 

1054: \subsection{Detecting Periodicity with Likely Arrays}

1055: \label{results-arrays}

1056: 

1057: We consider potential observing arrays that may be employed in

1058: observations of Sgr~A*.  The minimum number of telescopes required to

1059: produce a closure quantity is 3, and the maximum number of telescopes

1060: located at substantially different sites with mutual visibility of

1061: Sgr~A* is 5.  At 230~GHz, we consider western hemisphere arrays

1062: consisting of the US triangle; US and Chile-1; US, Chile-1, and the

1063: LMT; and the same array with Chile-10 instead.  We also consider

1064: European-Chile arrays.  Since PV is not expected to have 345~GHz

1065: capability in the near future, we expect that Chile-10 will exist

1066: before 345~GHz observations are possible on the Europe-Chile triangle.

1067: Likewise, we expect that western hemisphere arrays will include

1068: Chile-10 before the LMT is available at 345~GHz, since 345~GHz

1069: capability is not currently planned at the LMT.

1070: 

1071: For each array, we generate $\mathrm{ACF}_\phi$ for each triangle of

1072: three antennas and $\mathrm{ACF}_A$ for each quadrangle of four

1073: antennas.  The ACFs can be combined to produce a single combined ACF

1074: of the entire array.  For optimum detectability, it is necessary to

1075: weight the individual ACFs by the square of their effective

1076: $\mathrm{SNR}_\mathrm{ACF} \equiv X/\xi$, where $X$ and $\xi$ quantify

1077: the effective signal and noise in the ACF, respectively.  For

1078: definiteness, we take $\xi$ as the rms deviation in the ACF after a

1079: 5-channel boxcar-smoothed SNR has been subtracted.  We define $X

1080: \equiv \max(\mathrm{ACF}(k)) - \max(\min(\mathrm{ACF}(k)),0), k \neq

1081: 0$ as a measure of the range of the ACF.  The quantity $X$ reflects

1082: the contrast of the ACF and is especially critical for proper

1083: weighting of the phase ACFs, since on many triangles

1084: $\mathrm{ACF}_\phi$ is approximately constant though with small

1085: periodic peaks (Fig.~\ref{fig-acfs}).  Effectively, $X/\xi$ quantifies

1086: the significance of a peak in the ACF.  In practice, $X$ is frequently

1087: small for $\mathrm{ACF}_\phi$ on certain triangles, although the very

1088: high SNR of the closure phases on the small SMTO-CARMA-LMT and

1089: Hawaii-SMTO-CARMA triangles (which do not resolve out much of the

1090: total flux) often compensates by producing a very clean ACF (i.e.,

1091: $\xi$ is small as well).  In order to identify the period in an

1092: algorithmic way, we created a ``folded'' version of the composite

1093: array ACF, defined as

1094: \begin{equation}

1095: \mathrm{Fold}(k) \equiv \frac{1}{\lfloor

1096:   m/k\rfloor}\sum\limits_{i=1}^{\lfloor m/k\rfloor}

1097: \mathrm{ACF}(i\cdot k),

1098: \end{equation}

1099: where $m$ is the number of points in the ACF.  Folding the ACF

1100: effectively suppresses the trivial peak at zero lag.  In the absence

1101: of noise, the peak of $\mathrm{Fold}(k)$ is the period.

1102: 

1103: In order to test for the significance of periodicity detection, we ran

1104: Monte Carlo simulations for each model and array at bit rates from 1

1105: to 32~Gbit\,s$^{-1}$ by powers of two.  We ran 10000 simulations of

1106: the data with different noise instantiations and obtained the value of

1107: $k$ maximizing $\mathrm{Fold}(k)$.  The distribution of

1108: $\mathrm{Fold}(k)$ provides information on the probability of false

1109: detection of periodicity.  The results are summarized in

1110: Table~\ref{tab-results}, which lists the number of trials for which

1111: $\max(\mathrm{Fold}(k))$ misidentifies the period, rounded to the

1112: nearest 10~s, by more than one minute.  Four and a half orbital

1113: periods of simulated data were used, corresponding to 2~hr for Models

1114: A-D and 37~min for Model E.

1115: 

1116: \clearpage \input{tab3.tex} \clearpage

1117: 

1118: It is clear that the inclusion of a fourth or fifth telescope in the western

1119: hemisphere array produces a large improvement in periodicity detectability.

1120: This is due both to the much larger number of closure quantities that can be

1121: averaged together to produce a detection and also to the fact that a 4- or

1122: 5-element array will necessarily probe a larger range of spatial scales than

1123: possible 3-element arrays, which is important because it is not clear a priori

1124: which triangle will be best matched to the angular resolution of the variable

1125: emission in Sgr~A*.  Additional bandwidth is important as well, but less so

1126: than additional telescopes in the observing array.  The Hawaii-SMTO-CARMA

1127: array, for example, is insensitive to periodic structure in several models,

1128: even at high recording rates.  But inclusion of a Chilean telescope, and/or the

1129: LMT, produces a robust detection of periodicity at modest recording rates.  It

1130: is especially worth noting that a bit rate of 8~Gbit\,s$^{-1}$, corresponding

1131: to 2~GHz total bandwidth (e.g., 1~GHz in each of two orthogonal polarizations)

1132: is sufficient to produce clear periodicity detections (with error rate $< 1

1133: \times 10^{-4}$) in nearly all models provided that at least four telescopes

1134: are used in the array.  For the Europe-Chile triangle, where no clear fourth

1135: telescope is presently available, the use of high bit rates and phased ALMA

1136: will be critical for periodicity detection.

1137: 

1138: Array options at 345~GHz are relatively limited, since neither the LMT

1139: nor PV have planned 345~GHz capability.  We include the Europe-Chile

1140: triangle to show what might be expected if PV is ever upgraded to

1141: include a 345~GHz receiver.  The largest simultaneous array that can

1142: be deployed at 345~GHz in the near future consists of the three US

1143: telescopes and a Chilean station.  Should phased ALMA not become

1144: available, an array consisting of Hawaii, SMTO, CARMA, Chile~1, and

1145: the LMT with a 345~GHz receiver (but not otherwise optimized for

1146: 345~GHz performance) would suffice to detect periodicity in Models A-E

1147: at bit rates of 2~Gbit\,s$^{-1}$ or higher.

1148: 

1149: Observations have recently been taken with an array consisting of the

1150: JCMT, SMTO, and CARMA-1 at 230~GHz \citep{doeleman08}.  We run

1151: simulations of this array, which is usable already, as well as the

1152: same three telescopes with the addition of Chile-1, since APEX is

1153: likely to have 230~GHz capability in the near future.  We find that

1154: periodicity may be just marginally detectable at the current maximum

1155: bit rate of 4~Gbit\,s$^{-1}$ if the source geometry in Sgr~A* is

1156: favorable.  Larger bandwidths are critical for detecting periodicity

1157: on these arrays (and may not suffice if only the US telescopes are

1158: used, depending on source geometry).  It is likely that phased-array

1159: capability at Hawaii and CARMA will become available on the same time

1160: scale as higher bit rate capability.  At 345~GHz, an array consisting

1161: of the JCMT, SMTO, and Chile-1 (APEX or ASTE) would not suffice to

1162: detect periodicity except at very high bit rates ($\geq

1163: 32$~Gbit\,s$^{-1}$, depending on the model).

1164: 

1165: 

1166: \subsection{Long-Period Models}

1167: 

1168: \begin{figure}

1169: \resizebox{\hsize}{!}{\includegraphics{f10.eps}}

1170: \caption{Closure phases and amplitudes for Model F at 230~GHz (period

1171:   2.8~hr).  The colored lines indicate the closure quantities that

1172:   would be obtained in the absence of noise.  The abscissa shows the

1173:   time range over which Sgr~A* is above $5\degr$ elevation at at least

1174:   three western hemisphere telescopes.  Due to changing baseline

1175:   orientations, closure phases and amplitudes do not approximately

1176:   repeat, although significant deviations are seen simultaneously on

1177:   most subarrays.  Very small spikes seen in some of the lines are

1178:   artifacts of the modelling.

1179:   \label{fig-modelF}

1180: }

1181: \end{figure}

1182: 

1183: Detecting periodicity in the closure quantities will be more difficult

1184: if the hot spot orbital period is long (e.g., several hours).  The

1185: change in baseline orientation is significant enough within the

1186: 167-minute orbital period in Model F that the closure quantities do

1187: not approximately repeat, as shown in Figure~\ref{fig-modelF}.

1188: Consequently, the corresponding ACFs on most triangles and quadrangles

1189: fall off with increasing lag, so the period cannot be determined by

1190: finding the largest peak in the ACF.  Closure phases and amplitudes

1191: show periods of relative quiescence punctuated by periods of large

1192: simultaneous variability on most subarrays.

1193: 

1194: The mutual visibility on most triangles and quadrangles is comparable

1195: to twice the orbital period of the long-period models, so periodic

1196: behavior in closure phases and amplitudes may not be convincing in

1197: demonstrating source structure periodicity.  For instance, it may not

1198: be clear whether closure phases and amplitudes such as those shown in

1199: Figure~\ref{fig-modelF} are due to a hot spot in a large orbit or due

1200: to two unrelated flaring episodes of an aperiodic nature.  The LMT may

1201: be critical for establishing long-period periodicity, since its

1202: inclusion lengthens the total time range for observing, increasing the

1203: chance that a third period will be detected.  While present

1204: millimeter-VLBI arrays are thus not optimal for detecting long-period

1205: orbiting hot spots, observational evidence suggests that hot spot

1206: periods may be significantly shorter than that in Model~F.  VLBI

1207: measurements of the position of the centroid of emission from Sgr~A*

1208: presently constrain the orbital period of hot spots to be $\lesssim

1209: 120$~min if the hot spot flux dominates the disk flux, with

1210: progressively longer periods allowed as the hot-spot-to-disk flux

1211: ratio decreases \citep{reid08}.

1212: 

1213: However, there are fundamental constraints upon the existence of such

1214: long-period hot spots.  The synchrotron cooling time at millimeter

1215: wavelengths is roughly $3\,\hr$ (see eq.\,[\ref{eqn-cooling}]), and

1216: thus in the absence of a continuous mechanism for injecting energetic

1217: electrons in the hot spot, we would not expect to see single hot spots

1218: persist longer than this.  Additionally, the available energy

1219: (magnetic \& hydrodynamic) that may be tapped to generate substantial

1220: emission decreases rapidly with radius, implying that the brightest

1221: (and therefore dominant) events will preferentially occur at small

1222: radii, with correspondingly short dynamical timescales.

1223: 

1224: 

1225: 

1226: 

1227: \section{Discussion}

1228: \label{discussion}

1229: 

1230: \subsection{Observing Strategies and Telescope Prioritization}

1231: 

1232: Ultimately, observations are limited by the telescopes that are

1233: available.  Individual observers often have little influence over

1234: telescope construction priorities.  Nevertheless, it is important to

1235: consider the relative advantages of planned and potential instruments.

1236: Future observers may be limited by scarce resources, such as DBEs and

1237: recording equipment, instead of by available telescopes.

1238: 

1239: Array phasing will be crucial for maximizing the potential of

1240: millimeter VLBI arrays.  Phased versions of CARMA and the millimeter

1241: telescopes on Mauna Kea are necessary in order to increase the

1242: detectability of very weak signals on the long baselines.  However,

1243: further observations on the Hawaii-CARMA-SMTO triangle should not be

1244: deferred due to the lack of a phased-array processor at Hawaii and/or

1245: CARMA.  In some models, periodicity could be detected on the triangle

1246: consisting of the JCMT, a single CARMA dish, and the SMTO at high

1247: bandwidth.  Most of our models also indicate that periodicity may be

1248: detectable on the Hawaii-CARMA-Chile~1 triangle, although the use of

1249: phased ALMA rather than a single dish as the Chilean telescope may

1250: also be important for detecting periodicity if the black hole is

1251: highly rotating or if hot spot flux density signatures are typically

1252: smaller than assumed in this work.

1253: 

1254: Given the schedule of proposed telescope upgrades, observations

1255: utilizing closure techniques in the near future are most likely to use

1256: the Hawaii-CARMA-SMTO triangle at 230~GHz and the Hawaii-SMTO-Chile

1257: triangle at 345~GHz, with the possibility that the JCMT and/or CARMA-1

1258: may necessarily be used in place of Hawaii and CARMA for the earliest

1259: observations.  As illustrated in Figure~\ref{fig-phase230} (and

1260: Table~\ref{tab-results}), the Hawaii-CARMA-SMTO triangle at 230~GHz

1261: may not provide adequate spatial resolution to detect an orbiting hot

1262: spot, depending on the parameters of the Sgr~A* system.  At the other

1263: extreme, the currently available array of Hawaii, SMTO, and Chile-1 at

1264: 345~GHz may resolve out most of the emission (Fig.~\ref{fig-uv}) and

1265: is unlikely to be useful for periodicity detection by itself at data

1266: rates less than about 16~Gbit\,s$^{-1}$.  We therefore conclude that

1267: observations of Sgr~A* in the near future to detect periodicity should

1268: be taken at 230~GHz.

1269: 

1270: It will be important to include a fourth antenna in the array at each

1271: frequency.  As demonstrated in \S \ref{ACF}, the additional closure

1272: phases and amplitudes obtained by a 4-element array may be important

1273: for detecting periodicity.  Even if periodicity can be marginally

1274: detected by a 3-element array, the extra 4 independent closure

1275: quantities provided by inclusion of a fourth telescope will vastly

1276: increase the significance of periodicity detection.  Early

1277: observations at 230~GHz should include APEX or a single ALMA dish, if

1278: at all possible, in order to provide a fourth station in the array

1279: (otherwise assumed to include Hawaii, SMTO, and CARMA).  A key benefit

1280: of a 4-station US-Chile array, beyond the much larger number of

1281: closure quantities provided, is that the constituent baselines will be

1282: sensitive to a large range of spatial scales, making it less likely

1283: that periodicity will be missed due to a mismatch between the angular

1284: scale of source structure variability and the angular resolution of

1285: the array.

1286: 

1287: The LMT will be a critical telescope for two reasons.  First, it fills

1288: in an important hole in the $(u,v)$ plane, allowing for a large range

1289: of angular scales to be probed by the full array.  Indeed, the LMT

1290: produces excellent closure phase data on some triangles for every

1291: model considered in this work.  Second, inclusion of the LMT allows

1292: for simultaneous visibility of Sgr~A* from up to 5 telescopes.  The

1293: utility of the LMT derives especially from its location and its size.

1294: Even an incomplete LMT (e.g., a 32 m dish, as assumed in this work)

1295: whose surface has not yet been fully tuned to maximize aperture

1296: efficiency would be highly useful for observations of the Galactic

1297: center at 230~GHz.  Thus, we conclude that millimeter-wavelength VLBI

1298: arrays observing Sgr~A* should include the LMT as soon as is

1299: practical.

1300: 

1301: There is also a need for 345~GHz capability at a greater number of

1302: telescopes.  CARMA would be especially useful due both to its large

1303: effective collecting area as well as its short baselines to the SMTO

1304: and Hawaii.  The same comments that apply to the LMT in its planned

1305: 230~GHz band also apply to a future 345~GHz system, if ever planned.

1306: Expanding PV to include 345~GHz would allow for observations on the

1307: Europe-Chile triangle, which would provide only one additional closure

1308: phase and is therefore not a high priority for periodicity detection.

1309: However, the long Europe-Chile baselines cover an otherwise unsampled

1310: region of the $(u,v)$ plane and may therefore prove to be important

1311: for eventual modelling of the Sgr~A* quiescent disk.

1312: 

1313: An alternative observing strategy for the Chilean site, if phased ALMA

1314: is unavailable, would be to use multiple individual telescopes.

1315: Baselines from the Chilean telescopes (APEX, ASTE, or a single ALMA

1316: dish) to North American or European telescopes would effectively be

1317: redundant with each other in the $(u,v)$ plane but would offer

1318: independent data, thus increasing the number of closure quantities

1319: available.  Baselines between Chilean telescopes would be too short to

1320: resolve the emission from Sgr~A* but would provide a valuable

1321: simultaneous (near-)zero-spacing flux measurement, allowing estimation

1322: of the fraction of flux resolved out by the short PV-PdB or SMTO-CARMA

1323: baselines.

1324: 

1325: 

1326: \subsection{Bandwidth Considerations}

1327: 

1328: Our simulations indicate that the periodicity of reasonable hot spot

1329: models is detectable on most millimeter VLBI triangles/quadrangles at

1330: data rates from 1 to 32~Gbit\,s$^{-1}$.  However, we note several

1331: caveats whose applicability depends on the exact physical parameters

1332: of the Sgr~A* system.  It may be difficult to detect periodicity on

1333: the smallest triangles due to insufficient spatial resolution.

1334: Conversely, the largest triangles may resolve out much of the

1335: emission.  Very high data rates may be required to detect periodicity

1336: on long triangles to Chile~1.  During any particular track of

1337: observations, it is possible that no hot spot will be present, or that

1338: the flux density of the hot spot may be substantially smaller than

1339: assumed here.

1340: 

1341: We have assumed throughout that the atmospheric coherence time is 10~s.  In

1342: poorer weather conditions, the coherence time may be significantly shorter.

1343: Since the SNR of a coherently-integrated visibility grows as $(Bt)^{1/2}$,

1344: where $B$ is the bandwidth, the SNR for a single visibility integrated over a

1345: 2.5~s coherent time interval will be half that of a visibility integrated over

1346: 10~s.  So long as the signal is strong enough compared to the noise (i.e., the

1347: SNR $\gg 1$ in a single coherent integration), there is no net loss of signal

1348: because the decrease in SNR is exactly compensated by the increase in the

1349: number of data points obtained.  The SNR of the closure phase can then be built

1350: up over a longer period of time, if desired, by averaging consecutive closure

1351: phases (provided that the noiseless closure phase is not changing rapidly over

1352: the time scale of integration).  Periodicity may still be weakly detected if

1353: the average SNR exceeds unity over only a portion of the hot spot orbit, as

1354: could be the case if the visibility amplitude on the weakest baseline changes

1355: by a factor of several over an orbit.

1356: 

1357: Due to this effective SNR cutoff, it will be important to obtain data

1358: at the maximum bandwidth (or recording rate) possible at the time of

1359: observations.  Initial observations may be bandwidth-limited to 1~GHz

1360: (4~Gbit\,s$^{-1}$).  If subsequent observations are limited by average

1361: recording rate rather than IF bandwidth, it will be advantageous to

1362: use burst-mode recording, if available.  In the limit of marginal

1363: detections, it is far preferable to have a reduced number of good data

1364: points rather than a full set of poor data.

1365: 

1366: 

1367: \subsection{Black Hole Parameter Estimation}

1368: 

1369: The observation of periodicity in the closure quantities would provide

1370: important evidence, independent of observations of periodicity in

1371: flare light curves, that at least some subset of Sgr~A*'s flares

1372: are due to bright orbiting structures.  This is critical to

1373: efforts to use such structures ( e.g., hot spots) to infer the

1374: properties of the black hole spacetime.

1375: 

1376: In principle, combined with the flare amplitude (as measured via the unresolved

1377: light curves), the degree of variability in the closure quantities is

1378: indicative of the size of the hot-spot orbit.  Combined with the period, this

1379: provides a straightforward way in which to estimate the spin.  However, the

1380: precision with which submillimeter closure quantities can constrain the spin

1381: has yet to be determined, and will likely have to await a detailed

1382: parameter-space study.  On the topic of black hole spin, it should be pointed

1383: out that arguments for a non-zero spin of the SgrA* black hole can be made

1384: based on the observed intrinsic size of Sgr~A* at 1.3~mm \citep{doeleman08},

1385: and from NIR variability results \citep{genzel03}.

1386: 

1387: Our methods are generalizable to any mechanism of flare production.

1388: There will almost surely be asymmetric structure on scales of a few to

1389: a few tens of $\Rs$ due to general relativistic effects from the

1390: accretion disk.  Regardless of whether flares are produced by orbiting

1391: or spiralling hot spots, jets, magnetohydrodynamic instabilities, or

1392: some other mechanism, there will also be asymmetric structures in the

1393: inner disk region on scales of a few $\Rs$.  Our models

1394: demonstrate that millimeter-wavelength VLBI arrays will be sensitive

1395: to changes in the source structure no matter how these asymmetries are

1396: oriented relative to each other on the sky.

1397: 

1398: \subsection{Summary of Simulation Findings}

1399: 

1400: \begin{itemize}

1401: 

1402: \item The currently useable 230~GHz array consisting of SMTO, CARMA-1, and JCMT

1403: at 4~Gbit\,s$^{-1}$ cannot detect periodicity in any of the models tested.

1404: \item Phasing connected element arrays (CARMA, Hawaiian telescopes, ALMA) to

1405: form large effective apertures, significantly improves the probability of

1406: detecting the hot spot period.  When the CARMA array is phased, and the

1407: Hawaiian telescopes are coherently summed, periodicity in two out of five of

1408: the 230~GHz models can be detected.  When 10 elements of ALMA are phased

1409: together, periodicity can be reliably extracted in all of the 230~GHz models

1410: tested.  

1411: \item Higher recording bandwidth increases signal-to-noise on all baselines,

1412: thereby improving periodicity detection, and making the array more robust

1413: against loss of VLBI signal coherence due to atmospheric turbulence.  In three

1414: out of five 230~GHz models tested, a recording bandwidth of 16~Gbit\,s$^{-1}$

1415: allows detection of periodicity using just a three-station VLBI array.  

1416: \item Adding a fourth or fifth telescope to the array enables detection of the

1417: hot-spot period \emph{in every} 230~GHz model tested.  The added baseline

1418: coverage in larger arrays allows more complete sampling of spatial scales in

1419: SgrA*.  

1420: \item Inclusion of the LMT in 230~GHz arrays will be very important as

1421: baselines to the LMT fill critical voids in baseline coverage.  When long

1422: baselines between the US and Chile heavily resolve SgrA*, baselines between

1423: each of these regions and the LMT can still provide high signal-to-noise

1424: detections that connect all telescopes in the array.  

1425: \item A 345~GHz capability at more sites is needed.  The currently available

1426: 345~GHz array, consisting of SMTO, Hawaii, and Chile has long baselines that

1427: will resolve much of the flux density of SgrA*.  Enhancing either CARMA or the

1428: LMT with low-noise 345~GHz receivers, puts periodicity detection of all the

1429: tested 345~GHz models within reach when ALMA comes on line.

1430: 

1431: \end{itemize}

1432: 

1433: \subsection{Conclusions}

1434: 

1435: Motivated by recent detection of intrinsic structure within SgrA* on $<4\Rs$

1436: scales, this work explores the feasibility of detecting time variable structure

1437: and periodicity in the context of flare models in which a hot-spot orbits the

1438: central black hole.  Algorithms are described that use interferometric closure

1439: quantities, direct VLBI observables which reflect asymmetries in source

1440: structure that can be tracked on time scales that are short compared to

1441: presumed hot-spot orbital periods.  We find that periodicity from these models

1442: over a representative range of parameters can be reliably extracted using mm

1443: and submm VLBI arrays that are planned over the next 5 years.  Thus, mm and sub-mm

1444: VLBI has matured to the level where one can envisage studying fundamental black

1445: hole parameters, accretion physics, and General Relativity in the strong field

1446: regime at meaningful angular resolutions.

1447: 

1448: The techniques and concepts described here are applicable to a broad range of

1449: Sgr~A* models, including emission due to jets and outflows, MHD simulations,

1450: and adiabadically expanding flaring structures.  Closure amplitude analysis,

1451: for example, is capable of tracking the size of SgrA* over time and sensitively

1452: testing for expansion during a flare event.  Short wavelength VLBI, when

1453: coupled with wide spectrum simultaneous monitoring, can thus make detailed

1454: tests of emission models for Sgr~A* in which low frequency synchrotron photons

1455: are up-scattered to produce X-rays.

1456: 

1457: The three fundamental technical efforts to improve mm and submm arrays include

1458: increasing the number of VLBI sites, achieving the capability to phase

1459: connected element arrays into a large effective aperture, and increasing the

1460: VLBI recording data rate.  Projects in each of these areas, supported by an

1461: international collaboration, are underway, and it is expected that observations

1462: on VLBI arrays at 1.3~mm and 0.85~mm wavelength with dramatically improved

1463: sensitivities will begin by 2009.  The increase of bandwidth, while of

1464: general importance to lowering VLBI detection thresholds, may be most useful by

1465: enabling full polarization observations.  The techniques developed in this work

1466: can be extended to non-imaging VLBI polarimetry analysis, allowing tests for

1467: small scale polarization structure to be carried out.  The prospect of using

1468: VLBI to probe such polarized sub-structure emphasizes the power of the

1469: technique for studying Sgr~A*.

1470: 

1471: \acknowledgments

1472: 

1473: The high-frequency VLBI program at Haystack Observatory is funded

1474: through a grant from the National Science Foundation. 

1475: 

1476: \begin{thebibliography}{}

1477: 

1478: \bibitem[Agol(2000)]{agol00} Agol, E.\ 2000, \apjl, 538, L121

1479: 

1480: \bibitem[Aitken et al.(2000)]{aitken00} Aitken, D.~K., et al.\ 2000,

1481:   \apj, 534, L173

1482: 

1483: \bibitem[Aschenbach et al.(2004)]{aschenbach04} Aschenbach, B.,

1484:   Grosso, N., Porquet, D., \& Predehl, P.\ 2004, \aap, 417, 71

1485: 

1486: \bibitem[Backer(1978)]{backer78} Backer, D.~C.\ 1978, \apjl, 222, L9

1487:  

1488: \bibitem[Backer \& Sramek(1999)]{backer99} Backer, D.~C., \& Sramek,

1489:   R.~A.\ 1999, \apj, 524, 805

1490: 

1491: \bibitem[Baganoff et al.(2001)]{baganoff01} Baganoff, F.~K., et al.\

1492:   2001, \nat, 413, 45 

1493: 

1494: \bibitem[B\'{e}langer et al.(2006)]{belanger06} B\'{e}langer, G.,

1495:   Terrier, R., de Jager, O.~C., Goldwurm, A., \& Melia, F.\ 2006,

1496:   Journal of Physics Conference Series, 54, 420

1497: 

1498: \bibitem[Blandford \& Begelman(1999)]{blandford99} Blandford, R.~D. \&

1499:   Begelman, M.~C.\ 1999, \mnras, 303, L1

1500: 

1501: \bibitem[Bower et al.(2004)]{bower04} Bower, G.~C., Falcke, H.,

1502:   Herrnstein, R.~M., Zhao, J.-H., Goss, W.~M., \& Backer, D.~C.\ 2004,

1503:   Science, 304, 704

1504: 

1505: \bibitem[Bower et al.(2001)]{bower01} Bower, G.~C., Wright, M.~C.~H.,

1506:   Falcke, H. \& Backer, D.~C.\ 2001, \apj, 555, 103

1507: 

1508: \bibitem[Bower et al.(2003)]{bower03} Bower, G.~C., Wright, M.~C.~H.,

1509:   Falcke, H. \& Backer, D.~C.\ 2003, \apj, 588, 331

1510: 

1511: \bibitem[Bower et al.(2006)]{bower06} Bower, G.~C., Goss, W.~M., 

1512:   Falcke, H. \& Backer, D.~C., \& Lithwick, Y.\ 2006, \apjl, 648, L127

1513: 

1514: \bibitem[Broderick (2006)]{broderick06} Broderick, A.~E.\ 2006,

1515:   \mnras, 366, L10

1516: 

1517: \bibitem[Broderick \& Blandford(2003)]{broderick03} Broderick, A. \&

1518:   Blandford, R.\ 2003, \mnras, 342, 1280

1519: 

1520: \bibitem[Broderick \& Blandford(2004)]{broderick04} Broderick,

1521:   A.~E. \& Blandford, R.~D.\ 2004, \mnras, 349, 994

1522: 

1523: \bibitem[Broderick \& Loeb(2005)]{broderick05} Broderick, A.~E. \&

1524:   Loeb, A.\ 2005, \mnras, 363, 353

1525: 

1526: \bibitem[Broderick \& Loeb(2006a)]{broderick06a} Broderick, A.~E. \&

1527:   Loeb, A.\ 2006a, \apjl, 636, L109

1528: 

1529: \bibitem[Broderick \& Loeb(2006b)]{broderick06b} Broderick, A.~E. \&

1530:   Loeb, A.\ 2006b, \mnras, 367, 905

1531: 

1532: \bibitem[Bromley et al.(2001)]{bromley01} Bromley, B.~C., Melia, F.,

1533:   \& Liu, S.\ 2001, \apjl, 555, L83

1534: 

1535: \bibitem[Chan et al.(2006)]{chan06} Chan, C.-k., Liu, S., Fryer,

1536:   C.~L., Psaltis, D., Ozel, F., Rockefeller, G., \& Melia, F.\ 2006,

1537:   ArXiv Astrophysics e-prints, arXiv:astro-ph/0611269

1538: 

1539: \bibitem[Cornwell(1982)]{cornwell82} Cornwell, T.\ 1982, Synthesis

1540:   Mapping, 13

1541: 

1542: \bibitem[Cornwell \& Wilkinson(1981)]{cornwell81} Cornwell, T.~J., \&

1543:   Wilkinson, P.~N.\ 1981, \mnras, 196, 1067

1544: 

1545: \bibitem[Doeleman et al.(2002)]{doeleman02} Doeleman, S., et

1546:   al.\ 2002, Proceedings of the 6th EVN Symposium, 223

1547: 

1548: \bibitem[Doeleman et al.(2001)]{doeleman01} Doeleman, S.~S., et

1549:   al.\ 2001, \aj, 121, 2610

1550: 

1551: \bibitem[Doeleman et al.(2008)]{doeleman08} Doeleman, S.~S., et

1552:   al.\ 2008, \nat, submitted

1553: 

1554: \bibitem[Eckart et al.(2006)]{eckart06} Eckart, A., Sch{\"o}del, R.,

1555:   Meyer, L., Trippe, S., Ott, T. \& Genzel, R.\ 2006, \aap, 455, 1.

1556: 

1557: \bibitem[Eckart et al.(2004)]{eckart04} Eckart, A., et al.\ 2004, \aap,

1558:   427, 1

1559: 

1560: \bibitem[Falanga et al.(2007)]{falanga07} Falanga, M., Melia, F.,

1561:   Tagger, M., Goldwurm, A., \& B{\'e}langer, G.\ 2007, \apjl, 662, L15

1562: 

1563: \bibitem[Falcke \& Markoff(2000)]{facke00a} Falcke, H. \& Markoff,

1564:   S.\ 2000, \aap, 362, 113

1565: 

1566: \bibitem[Falcke et al.(2000)]{falcke00} Falcke, H., Melia, F., \&

1567:   Agol, E.\ 2000, \apjl, 528, L13

1568: 

1569: \bibitem[Genzel et al.(2003)]{genzel03} Genzel, R., Sch{\"o}del, R.,

1570:   Ott, T., Eckart, A., Alexander, T., Lacombe, F., Rouan, D., \&

1571:   Aschenbach, B.\ 2003, \nat, 425, 934

1572: 

1573: \bibitem[Ghez et al.(2005)]{ghez05} Ghez, A.~M., Salim, S., Hornstein,

1574:   S.~D., Tanner, A., Lu, J.~R., Morris, M., Becklin, E.~E., \&

1575:   Duch{\^e}ne, G.\ 2005, \apj, 620, 744

1576: 

1577: \bibitem[Ghez et al.(2003)]{ghez03} Ghez, A.~M.\, et al.\ 2003, \apjl,

1578:   586, L127

1579: 

1580: \bibitem[Ghez et al.(2004)]{ghez04} Ghez, A.~M., et al.\ 2004, \apjl,

1581:   601, L159

1582: 

1583: \bibitem[Goldston et al.(2005)]{goldston05} Goldston, J.~E., Quataert,

1584:   E., \& Igumenschev, I.~V.\ 2005, \apj, 621, 785

1585: 

1586: \bibitem[Holdaway(1997)]{holdaway97} Holdaway, M.~A.\ 1997, MMA Memo Series \#169.

1587: 

1588: \bibitem[Hornstein et al.(2007)]{hornstein07} Hornstein, S.~D.,

1589:   Matthews, K., Ghez, A.~M., Lu, J.~R., Morris, M., Becklin, E.~E.,

1590:   Rafelski, M., \& Baganoff, F.~K.\ 2007, \apj, 667, 900

1591: 

1592: \bibitem[Huang et al.(2007)]{huang07} Huang, L., Cai, M., Shen, Z.-Q.,

1593:   \& Yuan, F.\ 2007, \mnras, 379, 833

1594: 

1595: \bibitem[Jennison(1958)]{jennison58} Jennison, R.~C.\ 1958, \mnras,

1596:   118, 276

1597: 

1598: \bibitem[Jones \& O'Dell(1977)]{jones77} Jones, T.~W. \& O'Dell,

1599:   S.~L.\ 1977, \apj, 214, 522

1600: 

1601: \bibitem[Lindquist(1966)]{lindquist66} Lindquist, R.~W.\ 1966, Annals

1602:   of Physics, 37, 487

1603: 

1604: \bibitem[Lo et al.(1998)]{lo98} Lo, K.~Y., Shen, Z.-Q., Zhao, J.-H.,

1605:   \& Ho, P.~T.~P.\ 1998, \apjl, 508, L61

1606: 

1607: \bibitem[Loeb \& Waxman(2007)]{loeb07} Loeb, A. \& Waxman, E.\ 2007,

1608:   JCAP, 3, 11

1609: 

1610: \bibitem[Macquart et al.(2006)]{macquart06} Macquart, J.-P. et

1611:   al.\ 2006, \apj, 646, 111

1612: 

1613: \bibitem[Maoz(1998)]{maoz98} Maoz, E.\ 1998, \apjl, 494, L181

1614: 

1615: \bibitem[Markoff et al.(2007)]{markoff07} Markoff, S., Bower, G.~C.,

1616:   \& Falcke, H.\ 2007, \mnras, 379, 1519

1617: 

1618: \bibitem[Marrone et al.(2007)]{marrone07} Marrone, D.~P.\ 2007,

1619:   astro-ph/0712.2877

1620: 

1621: \bibitem[Marrone et al.(2006)]{marrone06} Marrone, D.~P., Moran,

1622:   J.~M., Zhao, J.-H. \& Rao, R.\ 2006, Journal of Physics Conference

1623:   Series, 54, 354

1624: 

1625: \bibitem[Marrone et al.(2007)]{marrone07b} Marrone, D.~P., Moran,

1626:   J.~M., Zhao, J.-H. \& Rao, R.\ 2007, \apj, 654, 57

1627: 

1628: \bibitem[Melia et al.(2001)]{melia01} Melia, F., Bromley, B.~C., Liu,

1629:   S., \& Walker, C.~K.\ 2001, \apjl, 554, L37

1630: 

1631: \bibitem[Meyer et al.(2006)]{meyer06} Meyer, L., Sch{\"o}del, R.,

1632:   Eckart, A., Karas, V., Dov{\v c}iak, M., \& Duschl, W.~J.\ 2006,

1633:   \aap, 458, L25

1634: 

1635: \bibitem[Miyoshi et al.(2004)]{miyoshi04} Miyoshi, M., Ishitsuka,

1636:   J.~K., Kameno, S., Shen, Z., \& Horiuchi, S.\ 2004, Progress of

1637:   Theoretical Physics Supplement, 155, 186

1638: 

1639: \bibitem[Monnier(2007)]{monnier07} Monnier, J.~D.\ 2007, New Astronomy

1640:   Review, 51, 604

1641: 

1642: \bibitem[Narayan et al.(1998)]{narayan98} Narayan, R., Mahadevan, R.,

1643:   Grindlay, J.~E., Popham, R.~G. \& Gammie, C.\ 1998, \apj, 492, 554

1644: 

1645: \bibitem[Petrosian \& McTiernan(1983)]{petrosian83} Petrosian, V. \&

1646:   McTiernan, J.~M.\ 1983, Phys. Fluids, 26, 3023

1647: 

1648: \bibitem[Quataert \& Gruzinov(2000)]{quataert00} Quataert, E. \&

1649:   Gruzinov, A.\ 2000, \apj, 545, 842

1650: 

1651: \bibitem[Reid(1993)]{reid93} Reid, M.~J.\ 1993, \araa, 31, 345

1652: 

1653: \bibitem[Reid et al.(2008)]{reid08} Reid, M.~J., Broderick, A.~E.,

1654:   Loeb, A., Honma, M., \& Brunthaler, A.\ 2008, \apj, submitted,

1655:   arXiv:0801.4505

1656: 

1657: \bibitem[Reid \& Brunthaler(2004)]{reid04} Reid, M.~J., \& Brunthaler,

1658:   A.\ 2004, \apj, 616, 872

1659: 

1660: \bibitem[Rogers et al.(1995)]{rogers95} Rogers, A.~E.~E.,

1661:   Doeleman, S.~S., \& Moran, J.~M.\ 1995, \aj, 109, 1391

1662: 

1663: \bibitem[Rogers et al.(1974)]{rogers74} Rogers, A.~E.~E., et

1664:   al.\ 1974, \apj, 193, 293

1665: 

1666: \bibitem[Rogers et al.(1994)]{rogers94} Rogers, A.~E.~E., et 

1667:   al.\ 1994, \apjl, 434, L59

1668: 

1669: \bibitem[Sch{\"o}del et al.(2003)]{schodel03} Sch{\"o}del, R., Ott,

1670:   T., Genzel, R., Eckart, A., Mouawad, N., \& Alexander, T.\ 2003,

1671:   \apj, 596, 1015

1672: 

1673: \bibitem[Sch\"{o}del et al.(2002)]{schodel02} Sch\"{o}del, R., et

1674:   al.\ 2002, \nat, 419, 694

1675: 

1676: \bibitem[Shen et al.(2005)]{shen05} Shen, Z.-Q., Lo, K.~Y., Liang,

1677:   M.-C., Ho, P.~T.~P., \& Zhao, J.-H.\ 2005, \nat, 438, 62

1678: 

1679: \bibitem[Tagger \& Melia(2006)]{tagger06} Tagger, M., \& Melia,

1680:   F.\ 2006, \apjl, 636, L33

1681: 

1682: \bibitem[Thompson et al.(2001)]{thompson01} Thompson, A.~R., Moran,

1683:   J.~M., \& Swenson, G.~W., Jr.\ 2001, Interferometry and Synthesis in

1684:   Radio Astronomy (New York: Wiley)

1685: 

1686: \bibitem[Trotter et al.(1998)]{trotter98} Trotter, A.~S., Moran,

1687:   J.~M., \& Rodr\'{\i}guez, L.~F.\ 1998, \apj, 493, 666

1688: 

1689: \bibitem[Twiss et al.(1960)]{twiss60} Twiss, R.~Q., Carter, A.~W.~L.,

1690:   \& Little, A.~G.\ 1960, The Observatory, 80, 153

1691: 

1692: \bibitem[Yuan et al.(2002)]{yuan02} Yuan, F., Markoff, S. \& Falcke,

1693:   H.\ 2002, \aap, 383, 854

1694: 

1695: \bibitem[Yuan et al.(2003)]{yuan03} Yuan, F., Quataert, E. \& Narayan,

1696:   R.\ 2003, \apj, 598, 301

1697: 

1698: \bibitem[Yusef-Zadeh et al.(2006)]{yusefzadeh06} Yusef-Zadeh, F.,

1699:   Roberts, D., Wardle, M., Heinke, C.~O., \& Bower, G.~C.\ 2006, \apj,

1700:   650, 189

1701: 

1702: \end{thebibliography}

1703: 

1704: \end{document}

1705: