Filename
stringlengths 22
64
| Paragraph
stringlengths 8
5.57k
|
|---|---|
Processed_In_search_of_cool_flow_accretion_onto_galaxies_$-$.txt | In order to exclude faint starlight, the resultant images must be compared with broadband images using a red filter presum- ably from a companion Dragonfly system. The combined sig- nal could reasonably get down to a few mR smoothed over a 10(cid:48)(cid:48) patch of sky for one or more fields in several dark nights. This is the sensitivity needed if we are to detect the ionized layer in the proto-disk region. Our estimate comes from the 3σ broadband (B, I) limit (30 mag arcsec−2) of the Dragon- fly system. Here we assume this performance holds for the R band. For background limited observations (sky+galaxy con- tinuum), an emission-line region with a surface brightness of 500 mR detected within the band (∆λR ≈ 140 nm) is detected If we replace the R filter with a with the same sensitivity. 100× narrower bandpass filter, as in the UKST experiment, our detection limit falls to 50 mR. We get another factor of 10 gain (5 mR at 3σ) by smoothing the diffuse data over a 10(cid:48)(cid:48) patch. Our wild extrapolations assume the instrument re- sponse (e.g. flatfielding) can be removed to better than 0.1% or so (Bland-Hawthorn et al 1994, 1995). Faint H I emission. If the goal is to ‘image’ the proto-disk re- gion, we argue here that the best prospect is to go after the clumpy cool gas emission. Support for this idea comes from the simulations of Nuza et al (2014) where clumpy H I gas is "observed" to occupy a complex, roughly planar network of filaments - we refer to this scree of material as the proto-disk region. Their simulations are broadly tuned to the character- istics of the Local Group. The cool proto-disks for M31 and the Galaxy are seen to overlap and to extend out to at least 100 kpc from both galaxies.10 Interestingly, Wolfe et al. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | Publications of the Astronomical Society of Australia (PASA) c(cid:13) Astronomical Society of Australia 2021; published by Cambridge University Press. doi: 10.1017/pas.2021.xxx. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | Abstract Noise considerations for experiments that aim to statistically estimate the 21 cm signal from high redshift neutral hydrogen during the Epoch of Reionisation (EoR) using interferometric data are typically computed assuming a tracked observation, where the telescope pointing centre and instrument phase centre are the same over the observation. Current low frequency interferometers use aperture arrays of fixed dipoles, which are steered electronically on the sky, and have different properties to mechanically-steered single apertures, such as reduced sensitivity away from zenith, and discrete pointing positions on the sky. These properties encourage the use of two additional observing modes: (1) zenith drift, where the pointing centre remains fixed at the zenith, and the phase centre tracks the sky, and (2) drift+shift, a hybrid mode where the telescope uses discrete pointing centres, and the sky drifts during each fixed pointing. These three observing modes view the sky differently, and therefore yield different uncertainties in the power spectrum according to the balance of radiometric noise and cosmic variance. The coherence of measurements made by the instrument in these modes dictates the optimal reduction in thermal noise by combination of coherent modes, and the reduction in cosmic variance by combination of incoherent modes (views of different patches of the sky). Along with calibration and instrument stability considerations, the balance between these noise components provides one measure for the utility of these three modes for measuring a statistical signature of the EoR signal. We provide a general framework for estimating the uncertainty in the power spectrum for a given observing mode, telescope beam shape, and interferometer antenna distribution. We then apply this framework to the Murchison Widefield Array (MWA) using an analysis of the two-dimensional (2D) and one-dimensional (1D) power spectra for 900 hours of observing. We demonstrate that zenith drift scans can yield marginally lower uncertainty in the signal power compared with tracked scans for the MWA EoR experiment, and that moderately higher signal-to-noise ratio (S/N) estimates of the amplitude (3%) and slope (1%) of the 1D power spectrum are accessible, translating directly into a reduction in the required observing time to reach the same estimation precision. We find that the additional sensitivity of pointing at zenith, and the reduction in cosmic variance available with a zenith drift scan, makes this an attractive observing mode for current and future arrays. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | galaxies) can exceed 100s K (Pritchard & Loeb 2008; Jeli´c et al. 2010; Mesinger et al. 2011). Careful discrim- ination of spectral line signal from the early Universe from contaminating sources relies on the distinct spatial and spectral structure for these sources; radio galax- ies and other point source-like foregrounds contribute power on small angular scales, while the EoR sig- nal is expected to peak on larger scales (a few de- grees). Diffuse emission from the Galaxy, in particular in the Galactic plane, are spatially similar to this signal (Jeli´c et al. 2010). Conversely, the spectral structure of foreground continuum sources versus spatially-variable spectral line emission from the EoR provides a strong discriminator for the signals. Coupled with the weak- ness of the expected EoR signal, and the need to com- bine information from across the sky (stack the signal to increase detectability), statistical estimates are a use- ful tool for EoR detection and estimation. In particular, the two-dimensional power spectrum (signal variance as a function of angular and line-of-sight mode) is a useful metric for discriminating EoR signal from foregrounds on the basis of spatial and spectral differences, and for increasing detectability by combining signal statistically (Morales & Hewitt 2004; McQuinn et al. 2006). Mea- suring the statistical EoR signal relies on (1) a stable and calibratable instrument, and (2) sufficient data to reduce thermal noise, measure and remove foregrounds, and reduce cosmic variance (Mellema et al. 2013). |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | Statistical measurement of the EoR signal has many challenges, including reduction of noise to yield a de- tection. The two primary sources of noise for a stable and well-calibrated instrument (i.e., one where radio- metric noise exceeds calibration uncertainties and cali- bration errors are minimal) are radiometric (stochastic measurement noise because we are sampling the signal) and cosmic variance (uncertainty in measurement of a stochastic global signal because it is estimated by finite data). The most effective methods for reducing these noise contributions are not the same. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | The expected EoR signal power is low at very large angular scales, with the peak power occurring on scales of 2 degrees (see Mellema et al. 2013, and references therein). Large scales are sampled by the shortest base- lines of an interferometer, and most arrays have few. Therefore, at low k, we expect thermal noise to domi- nate. At larger k, where the EoR signal is expected to be strongest, and baselines are numerous (by design for an EoR experiment), cosmic variance becomes relatively more important. The balance between these coherent and incoherent modes is the subject of this note. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | estimate. In this work we build on the existing literature by: (1) deriving a framework that correctly accounts for the telescope’s beam pattern (and therefore the level of coherence/incoherence in a measurement set. See Thyagarajan et al. 2013, who also considered this aspect); (2) including covariances between angular modes in the calculation, and (3) applying it to the specific case of comparing different observing modes. In their work, Thyagarajan et al. (2013) considered both tracked scans and zenith scans with independent pointing centres. Parsons et al. (2012) also considered drift scans for exploring observing options for PAPER. We consider a scan with a constant evolution of the pointing centre, as the Earth rotates, requiring a robust framework to capture the time-dependent coherence of two baselines relative to each other. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | Traditional tracked observing modes, where the pointing and phase centres align and track a single sky position, are undesirable for aperture arrays with ana- logue beamformers, where such modes produce a time- variable sky beam because the individual antenna el- ements have discrete timing offset settings (‘delays’). Instead, two modes can be employed, which have differ- ent calibration and EoR measurement properties: (1) a “zenith drift”, where the pointing centre remains fixed at the zenith, and the phase centre tracks a single po- sition on the sky, and (2) “drift+shift”, a hybrid mode where the beam has fixed, discrete pointing centres, with the sky drifting during each fixed pointing. While providing a stable, constant beam, the former sees a time-variable sky, thereby reducing the coherence of measurements across a long observation. I.e., fewer in- dividual visibility measurements sample the same sky, at the same baseline vector. For EoR statistical esti- mation, sample coherence is used to combine visibilities directly. Coherent addition is desirable because ther- mal noise is reduced the most effectively. On the other hand, cosmic variance does not benefit from redundant samples of the same sky; coherent sampling maintains cosmic variance. Measurements that sample the same modes on the sky, but at different orientations or point- ing in different directions, can be added incoherently (powers added). This has less benefit for reducing ther- mal noise, but does reduce cosmic variance (indepen- dent samples of the sky). |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | and in relatively cold patches of the sky, over two frequency ranges (139–169 MHz, 167–197 MHz) with 40 kHz spectral resolution. The initial experiment will be based on two fields with a combined observing time of 1000 hours. Beardsley et al. (2013) estimate that a 14σ detection is possible in the spherically-averaged power spectrum for this experiment. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | In section 2 we describe the power spectrum, and its derivation from an interferometric dataset. We then connect the measured data to the information it pro- vides about the temperature fluctuations in a given sky mode, and use this model to propagate the noise terms to the final power spectrum. In section 3 we describe the differences in the noise estimates for each observing mode, based on the way in which the telescope’s beam pattern samples the sky. We then apply the model to a simulation of the EoR experiment for MWA in sec- tion 4. Throughout we use comoving co-ordinates in a ΛCDM cosmology, with Ωm = 0.27, ΩΛ = 0.73, Ωk=0.0, and H0 = 100h (Bennett et al. 2012). Matrices are de- noted in bold font, and vectors are denoted by an over- line. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | The aim of this paper is to quantify the information our data contain about the power in mode u, P (u). Typically, one would write a description of the forward model connecting the sky with the measurements, and compare the noise properties of the measurement set. However, because the number of measurments (visibil- ities) is very large compared with the number of sky modes in which we are interested (millions per day com- pared with hundreds for the whole observation), it is prohibitive to describe the full covariance in the mea- surement basis. Instead, we explore the inverse problem, where we describe the information available in a partic- ular sky mode due to the underlying measurement set. Working in this basis provides a simpler framework with which to tackle the problem. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | where the square matrix inversion involves inverting an (M M ) Hermitian complex-valued matrix, which will be (almost) diagonal for independent modes, u. Note that this procedure does not recover full sky informa- tion; the sampling function of the interferometer does not provide complete uv-coverage, and the problem is under-constrained. Instead, this procedure distributes the measured visibility data across contributing sky modes, to represent the smearing of Fourier space by the primary beam (grids and weights). It also represents a continuous transform with a discrete approximation, which destroys some information. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | where the first two terms correspond to the cosmic vari- ance of the 21 cm signal and foregrounds, and the final term is the thermal noise contribution. In the analysis that follows we will retain the contribution from fore- ground signal as a separate term, and focus on the ther- mal noise and cosmological signal terms. As a signal, foreground contributions exhibit the same behaviour as cosmological signal (cosmic variance noise consid- erations). We will assume herein that the foregrounds have been removed to a certain level, but their residu- als (which may be significant compared with the signal of interest) remain in the data. In the final analysis of the estimation performance of each method, we will consider uncontaminated regions of parameter space. For this work, where we focus on small k modes (short baselines), this will correspond to omitting the kk=0 mode in the power spectrum. We revisit the impact of foregrounds in sections 4.1 and 6. An excellent recent discussion of the impact of foregrounds can be found in Thyagarajan et al. (2013). |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | of the matrix removes (normalises) the im- provement one would otherwise compute by sampling the sky coherently (i.e., sampling the same sky signal does not improve cosmic variance). |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | For an ideal instrument, a visibility measures a single sky mode, with a delta-function footprint in the Fourier plane. For a real instrument with a finite primary beam and bandpass, modes are mixed into an individual visi- bility measurement due to leakage. For a small FOV, the corresponding beam in Fourier space is broad, because the beam convolution distributes the power from adja- cent modes. This beam footprint determines the coher- ence length of Fourier modes, as demonstrated schemat- ically in Figure 1. Here, the beam footprint from indi- vidual visibility measurements determines the joint sky information available on a given angular scale. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | ui∈k X where N (ui) is the number of measurements contribut- ing to mode ui. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | Figure 1. Schematic representation of the uv-plane (Fourier plane), and the footprint of the primary beam (station beam) from individual visibility measurements (red circles). Also shown in the blue annular region is an example area over which the cos- mological signal is not expected to evolve substantially, and the powers are averaged incoherently. The finite size of the beam in Fourier space mixes information from different sky modes into an individual visibility measurement. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | integrate visibilities that contribute coher- of ently, while the off-diagonal components encode the cor- relations between sky modes introduced by the finite Fourier beam. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | Together, equations 22 and 24 provide the quantity in which we are interested; the uncertainty in the mea- sured power in mode k. These expressions describe the normal ML incoherent power (equation 15). For ad- dition of independent modes, this expression reduces to the approximate form from McQuinn et al. (2006) (their equation 24). |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | in steps across the sky, phase centre = EoR field centre (approximates a tracked scan). |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | The key differences here lie in the influence of the an- tenna primary beam pattern on the sampling of the sky, and the effect this has on the coherence of measured data. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | The beam is the Fourier transform of the primary beam (PB) pattern on the sky, relative to the phase centre of the observation (for antennas with different beams, in Fourier-space it is the convolution of the an- tenna illumination pattern for each — for simplicity here, we will consider identical beams). |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | offset, θ. For the purposes of this calculation (to compute Ωu), this approximation is appropriate. 3. Drift+shift: the drift+shift observation is an in- termediate protocol between the tracked and pure- drift scans. It is characterised by a constant sky phase centre throughout the observation, but pe- riods of drift at fixed pointing centres. For the MWA, the field drifts for 30 minutes ( 8 degrees) before re-pointing. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | Figure 2. Observation volume, Ωu, for a zenith drift scan, rela- tive to a tracked scan for the MWA, as a function of the maximum offset between the pointing centre and phase centre. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | This value is used for each visibility measured by the 1010 visibilities com- experiment, with an integrated bined over the course of the experiment (10 second cadence, with 128 tiles and 900 hours total observing time). |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | We explicitly included the contribution from unsub- tracted foregrounds in the signal covariance matrix, but have, as yet, not distinguished between cosmological sig- nal and foregrounds in the noise analysis. Most current and future experiments aim to remove the strongest of these sources to high precision (others, such as PA- PER, Parsons et al. (2010), try to work in a region of parameter space where foreground contamination is minimal), but emission will remain due to unremoved sources (often below the confusion limit) and imprecise source subtraction (e.g., Trott et al. 2012). The infor- mation available in the data limits the level and pre- cision to which foregrounds can be removed. Although foregrounds are contaminants to the signal of interest, they have a different signature in the 2D power spec- trum parameter space. Previous theoretical work has demonstrated the expected existence of a wedge feature, which is the signature of imprecisely- and un-subtracted smooth-spectrum point sources (Bowman et al. 2009; Datta et al. 2010; Trott et al. 2012; Vedantham et al. 2012; Morales et al. 2012; Thyagarajan et al. 2013). |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | For an achromatic instrument, spectrally-smooth fore- grounds occupy the lowest kk modes (flat-spectrum sources strictly occupy only the kk = 0 mode). However, interferometers are inherently chromatic, with each physical baseline sampling signal at different wavenum- bers for each spectral channel. This yields the wedge- like structure in the 2D power spectrum, where larger k⊥ modes have greater leakage into higher kk modes. For the small k⊥ modes studied in this work, and the coarse spectral sampling, the wedge occupies only the kk = 0 mode. To avoid the contribution of foregrounds, in the foregoing analysis, we omit signal in this mode. |
Processed_Comparison_of_observing_modes_for_statistical_esti.txt | evident at low kk, but the overall improvement in these modes is less than a factor of two in noise uncertainty. At large k⊥, the drift+shift and tracked modes per- form better because the experiment is thermal noise- dominated and there is less coherence for the zenith drift scan mode. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | We study the liquidity commonality impact of local and foreign institutional investment in the Australian equity market in the cross-section and over time. We find that commonality in liquidity is higher for large stocks compared to small stocks in the cross-section of stocks, and the spread between the two has increased over the past two decades. We show that this divergence can be explained by foreign institutional ownership. This finding suggests that foreign institutional investment contributes to an increase in the exposure of large stocks to unexpected liquidity events in the local market. We find a positive association between foreign institutional ownership and commonality in liquidity across all stocks, particularly in large- and mid-cap stocks. Correlated trading by foreign institutions explains this association. However, local institutional ownership is positively related to the commonality in liquidity for large-cap stocks only. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | Corresponding author. Tel: +61 8 8302 0523. Fax: +61 8 8302 0992. Email: [email protected]. We thank two anonymous referees for the valuable comments and suggestions that have helped us to make significant improvements to this paper. We also thank the Securities Industry Research Centre of Asia-Pacific (SIRCA) and the Australian Securities Exchange (ASX) for providing the data used in this study. This research was supported by the Australian Research Council (ARC DP130103517). |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | register system that the ASX operates to register and settle share trading and to transfer share titles. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | investors must first have their shares transferred to CHESS before they can trade them on the ASX. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | value of a stock’s daily return divided by the dollar volume and captures the price impact of trades. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | that 𝐻0: 𝛽𝐿,𝑙𝑎𝑟𝑔𝑒 − 𝛽𝐿,𝑠𝑚𝑎𝑙𝑙 = 0 is rejected for almost every year (exceptions are 2002 and 2013). |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | and the time series of the spread in market liquidity beta between large and small firms in Fig. 2(b). |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | As Fig 2(a) shows, the evolution of liquidity beta over time is different for large and small firms. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | lowest value, −0.08, in 2000. It then gradually increases to 0.903 in the second quarter of 2016. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | However, our analysis is limited to the period 2007–2014, for which we have the ownership data. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | statistic=2.32) for large stocks, but it is insignificant for all stocks as well as for other size groups. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | The results in Table 7, Panel A, show that βHI increases monotonically with foreign ownership. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | as in Eq. 2. It is important to show that our results are robust if we use equal-weighted portfolios. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | Table A1 Market liquidity beta estimated with an equal-weighted market portfolio and stock ownership in the cross- section. We estimate βL in a regression similar to Eq. 2, in which we use an equal-weighted market portfolio to compute changes in the daily illiquidity measure, and repeat Table 5. Numbers in brackets are t-statistics, adjusted for heteroskedasticity and autocorrelation using Newey and West (1987) standard errors with two lags. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | Table A2 Market liquidity beta and stock ownership in the cross-section with alternative sample filters. We repeat Table 5 with the following filters applied to the sample. Panel A excludes stocks with a price lower than 2 cents from the sample. Panel B excludes stocks with a price lower than 5 cents from the sample. Panel C retains firm-quarters in the sample with a minimum of 40 observations. Numbers in brackets are t-statistics, adjusted for heteroskedasticity and autocorrelation using Newey and West (1987) standard errors with two lags. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | Table A3 Market liquidity beta and stock ownership in the cross-section with alternative illiquidity controls. We repeat Table 5 with alternative illiquidity measures as a control. Panel A replaces the daily Amihud (2002) illiquidity measure with the Pastor and Stambaugh (2003) illiquidity measure as a control. Panel B replaces the daily Amihud (2002) illiquidity measure with the quoted percentage spread as a control. Numbers in brackets are t-statistics, adjusted for heteroskedasticity and autocorrelation using Newey and West (1987) standard errors with two lags. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | Table A4 Market liquidity beta and stock ownership in the cross-section with additional controls. We repeat Table 5 and control for additional firm-specific variables. BM is the natural log of book to market ratio, computed as one plus the ratio of the book equity at the end of last June to the market capitalization at the end of last December. DY is the dividend yield, computed as the ratio of total cash dividend payments during the last calendar year to the stock price at the end of last December. STDRET is the standard deviation of daily returns over the previous quarter. RE is the log difference of stock prices between the beginning and the end of the previous quarter. Numbers in brackets are t-statistics, adjusted for heteroskedasticity and autocorrelation using Newey and West (1987) standard errors with two lags. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | Le, N. N, A., Yin, X., & Zhao, J. (2019). Effects of investor tax heterogeneity on stock prices and trading behaviour around the ex-dividend day: the case of Australia. Accounting and Finance, forthcoming. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | Fig. 1. Quarterly standard deviation of changes in market illiquidity. We compute Δilliqi,d as the logarithmic change in daily Amihud (2002) illiquidity measure for each firm i. Δilliqmkt,d is the daily value-weighted average of Δilliqi,d for all stocks in our sample from January 1990 to December 2016. We present the quarterly standard deviation of Δilliqmkt,d. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | Fig. 2. Time series of market liquidity beta. In each quarter and for each firm, we regress the firm’s daily changes in the Amihud (2002) illiquidity measure on the daily changes in the value-weighted illiquidity measure of the market portfolio and a set of controls. For each regression, we remove firm i from the market portfolio in computing the daily changes in value-weighted illiquidity. For each quarter, we retain firms with at least 25 valid observations. We sort firms into terciles each quarter based on the market capitalization at the end of the previous quarter: small (bottom tercile), mid-cap (middle tercile) or large (top tercile). Panel (a) plots the quarterly cross-sectional average of βL for large and small firms, and Panel (b) presents the difference in average βL between large and small firms. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | Fig. 3. Time series of high ownership liquidity beta for foreign institutional investors. The high foreign institutional ownership liquidity beta (𝛽𝐻𝐼,𝑖,𝑡 ) for stock i is estimated as in Eq. 7. We plot the time series of cross-sectional means of βHI and the percentage of positive βHI among all sample stocks. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | Table 1 Summary Statistics. This table reports summary statistics for the sample stocks trading on the ASX at any point between 1 January1990 and 31 December 2016. Size is the market capitalization of the stock’s equity at the end of the previous quarter. Price is the average close price of the stock over the quarter. Dollar Volume is the average daily dollar volume over the quarter. We sort firms into terciles each quarter based on the market capitalization at the end of the previous quarter: small (bottom tercile), mid-cap (middle tercile) or large (top tercile). Panel A reports means, medians, standard deviations, minimums and maximums for the full sample of stock quarters. Panel B reports the summary statistics on ownership percentage by each investor category. For each quarter, we compute ownership percentages as the number of shares held by each investor category divided by the total number of shares outstanding at the end of the previous quarter between 2007 and 2014. Panel C reports the summary statistics on changes in stock ownership. We compute the dollar value of shareholdings by multiplying the number of shares held by each investor type with the close price. We then aggregate the holdings across the sample stocks to compute total dollar shareholdings by each investor category. We present the daily average dollar shareholdings and ownership percentage for the year 2007 and the year 2014, as well as changes in dollar and percentage shareholdings. We delete daily observations from the sample if the stock price crossed a minimum tick level during the day in either direction. Stock price must be equal or greater than $0.01 to be included in the sample. We retain stock quarters with at least 25 valid observations from the sample remaining after applying previous filters. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | For each regression, we remove firm i from the market portfolio in computing Δilliqmkt,d. We control for the variables (𝑍𝑗𝑖,𝑑) which may correlate with the firm’s liquidity: lead and lag changes in the market portfolio; lead, lag, and contemporaneous market returns; and contemporaneous squared stock return. For each quarter, we retain firms with at least 25 valid observations. We sort firms into terciles each quarter based on the market capitalization at the end of the previous quarter: small (bottom tercile), mid-cap (middle tercile) or large (top tercile). For each firm year, we compute the mean of quarterly betas. We report annual cross-sectional average of these means as well as t-statistics of βL and the percentage that are positive. We also report the difference in means and t-statistics of βL between large and small firms. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | Table 4 Implied trading activity by institutional investor groups. Each quarter, we sort stocks into quintiles based on the fraction of ownership by foreign institutions and local institutions at the end of the previous quarter, and compute average daily turnover as volume divided by shares outstanding for stocks in each ownership quintile. Panel A reports time-series means of average daily turnover (in percentage) for each ownership quintile. Panel B reports ownership-weighted turnover and holdings periods computed as follows. First, we compute average daily turnover (in percentage) as volume divided by shares outstanding for stocks over a quarter. Then, for each quarter, we compute the cross-sectional average of turnover, weighted by ownership percentage, at the end of the previous quarter for both investor groups, and compute the cross-sectional holding period (in days) as the inverse of cross-sectional average of turnover. We report time series means of cross-sectional averages of turnover and holding period for each investor group. Panel C presents average of daily turnover and holding period of ASX200 stocks and non-ASX200 stocks. t-statistics are in brackets. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | Table 7 High ownership liquidity beta of quintile portfolios sorted by foreign institutional ownership. Panel A reports the average βHI sorted by foreign institutional ownership. Each quarter we sort stocks into quintiles based on the number of shares held by the foreign institutions divided by the total number of shares outstanding at the end of the previous quarter. For each quintile we report average βHI and the difference in mean βHI between the top and the bottom quintiles. βHI is estimated for foreign institutional investors as in Eq. 7. Panel B presents the dependent sort results. We first sort stocks into terciles based on the market capitalization at the end of the previous quarter: small (bottom tercile), mid-cap (middle tercile) or large (top tercile). Within each size group, we sort stocks into quintiles based on the foreign institutional ownership. Numbers in brackets are t-statistics, adjusted for heteroskedasticity and autocorrelation using Newey and West (1987) standard errors with two lags. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | where βHI,i,t is estimated for foreign institutional investors as in Eq. 7, Fowni,t-1 is the foreign institutional ownership and computed as the number of shares held by foreign institutional investors divided by the total shares outstanding, Sizei,t-1 is the natural logarithm of the firm’s market capitalization (in millions) and Illiqi,t-1 is the daily average of the firm’s Amihud (2002) illiquidity measure. Fowni,t-1, Sizei,t-1, Illiqi,t-1 are computed at the end of the previous quarter. We sort firms into terciles each quarter based on the market capitalization at the end of the previous quarter: small (bottom tercile), mid-cap (middle tercile) or large (top tercile). We present the time-series averages of the coefficients for all stocks as well as for each size group. Numbers in brackets are t-statistics, adjusted for heteroskedasticity and autocorrelation using Newey and West (1987) standard errors with two lags. |
Processed_Institutional_ownership_and_liquidity_commonality:.txt | Table 9 Correlated trading: ASX200 vs. non-ASX200 stocks. This table reports summary statistics on a number of proxies for correlated trading for ASX200 stocks and non-ASX200 stocks. We split the non-ASX200 stocks into medium and small groups based on their market capitalizations at the end of the previous quarter. Medium group represents top 50% of the non-ASX200 stocks and small group represents bottom 50% of the non-ASX200 stocks. βL is the market liquidity beta estimated as in Eq. 2. βTO is the trading volume comovements measure or volume beta and computed as follows. Each quarter, we use a market model for volume and regress the percentage change in the daily turnover ratio for the individual stock on the concurrent percentage change in the market wide, value-weighted turnover. Dollar Volume and Turnover are the average daily dollar volume and turnover, respectively, and Autocorrelation is the daily return autocorrelations. The columns present pooled mean, median, minimum, 25th percentile, 75th percentile, maximum and standard deviation of the variables for all firms. |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | ABSTRACT We report a λ > ∼ 1 cm search for rotational molecular absorption towards quasars, now possible with the upgraded Australia Telescope Compact Array (ATCA). The targets were PKS 0201+113, PKS 0336–017 and Q 0537–286, where known damped Lyman-alpha absorption systems (DLAs) could cause redshifted molecular absorption in the 12 mm band of the telescope. We place 3σ upper limits on any HCO+ 0 → 1 −1 channel. The non-detections could absorption features of < 30 mJy per 3.4 km s be attributed to the inherent low metallicities in DLAs leading to generally low H2, and thus HCO+, column densities. In general, the detection of molecular rotational transitions in DLAs could be further hindered by a lower than expected CO-to-H2 conversion ratio, whether due either to photoionization of carbon or its relative under- abundance at high redshift. |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | tive detection, only lead to upper limits for 18 transitions (Curran et al. 2002a). With the upgraded ATCA it is now possible to improve on these previous attempts. In this paper we present the results of our first search with this telescope – observations towards the known southern centimetre-loud quasars occulted by DLAs in which a commonly detected transition falls into the 12 mm band. |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | Table 1. The 22 GHz illuminated DLAs where the redshifted HCO+ 0 → 1 falls into either the 16.089–18.888 or 20.089– 22.488 GHz ATCA bands. zabs is the DLA redshift and νobs is the expected frequency of the redshifted HCO+ 0 → 1 line, given to the number of significant figures available from the optical data. The red, visible and blue magnitudes of the quasar are given as well as the measured 22 GHz flux densities from both the literature (Lit.) and our observations (Obs.). Any difference between these two values may be due to quasar variability. |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | a The flux density for 0201+113 is from Ter¨asranta et al. (2001) and for 0202+149, 0336–019 and 0537–286 from the ATCA calibrators site (http://www.narrabri.atnf.csiro.au/calibrators/). The uncertainty of ≈ 15% in these calibrators is the major contributor to the errors in the measured flux densities. |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | DLAs at redshifts of ∼ 3 (Table 1)3. In order to minimize the bias introduced by using such optically selected sources, ideally we would have selected the most visually faint (and hence dusty) objects. However, the ATCA still has a rela- tively narrow tuning range in the 12 mm band. We therefore selected all of the “high flux” DLAs which can be observed with this restriction (see Table 1). Fortunately, all three are relatively faint (by DLA standards) with 0201+113 and 0537–286 also being red (B − R = 2.9 and 0.9, respectively). Because of the lack of precision in the optical redshifts (most significant for 0537–286, Table 1), we sacrificed one polarisation in order to overlap two of the widest available (64 MHz) bands. This enabled us to cover an uncertainty of ≈ ±0.01 in redshift while retaining a relatively high spectral −1. Finally, post observation, baseline resolution of 3.4 km s 2–4 was discarded because of phase referencing problems. No other flagging of bad data was required. For 0201+113 and 0336–017 the bandpass of the calibrator was removed from the spectra and in the case of 0537–286, which was self calibrated, we removed a low order polynomial to “flatten” the bandpass. |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | Figure 1. HCO+ 0 → 1 at z = 3.386 +0.013 −0.009 towards 0201+113. The 1σ r.m.s. noise is 10 mJy. The overlap be- tween the two spliced bands is 14 MHz. In this and Figs. 2 and 3 the arrow shows the expected absorption frequency according to the published DLA redshift (Table 1). |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | In Figs. 1 to 3 we show the time averaged spectra over both good baselines and note that there are no HCO+ 0 → 1 −1 channel in these absorption features of > 3σ per 3.4 km s DLAs. From the r.m.s. noise levels we derive optical depth −1 and column density limits limits for a resolution of 1 km s according to Nmm ∝ R τ dv for an excitation temperature of ≈ 10 K (see Curran et al. 2002a). These are listed in Table 2 together with all previously published results. |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | density, NH2 . In order to bypass this uncertainty, we use the best NCO limit (from Table 2), to estimate an upper limit, though we are still forced to assign an expected line-width to a non-detected feature. For three of the four known high −1 (the FWHM being redshift absorbers this is ≈ 20 km s −1 towards PKS 1413+1356, Wiklind & Combes ≈ 7 km s 1994a, 1995, 1996b, 1998). Also, as discussed in Curran et al. |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | Figure 2. HCO+ 0 → 1 at z = 3.062 +0.011 −0.007 towards 0336– 017. The 1σ r.m.s. noise is 6 mJy. The overlap between the two spliced bands is 12 MHz. |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | Figure 3. HCO+ 0 → 1 at z = 2.976 +0.010 −0.004 towards 0537– 286. The 1σ r.m.s. noise is 8 mJy. The gap in the spectrum occurs since there was no bandpass calibration and the fitted polynomial is unreliable at these frequencies. |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | ing that molecular absorption lines have a FWHM of < km s deeply searched for rotational molecular absorption7. |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | For high redshift (z > 1.8) sources, the ultra-violet H2 lines are redshifted to optical wavelengths and indeed H2 has been found in absorption in 7 DLAs at NH2 /NHI ratios of 6 1% (Ge & Bechtold 1997; Srianand & Petitjean 1998; Levshakov et al. 2000; Petitjean et al. 2000; Levshakov et al. 2002; Ledoux et al. 2002). In the low metallicity environ- ments typical of DLAs ([Zn/H] ≈ −2 to −1, i.e. ∼ 0.01 to 0.1 solar, e.g. Prochaska et al. 2001), Liszt (2002) argues that even if the gas is cool, the ambient ionization is suf- ficient to suppress the formation of H2. This may offer a potential explanation for our non-detections. |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | and, from an upper limit of NCO from the A–X molecular −8 bands, Srianand & Petitjean (1998) find NCO/NHI < 10 in DLAs, which is indeed only 10% of the local ratio. In addi- tion, at earlier epochs of chemical enrichment we would ex- pect that there is simply a lack of raw materials to form sig- nificant amounts of tracer molecules (e.g. CO, HCO+, HCN and CS) even in cold dark clouds where H2 readily forms. A reliable conversion ratio could be obtained by searching for CO absorption towards the DLAs in which H2 has been de- tected. However, the detection of H2 biases towards optically bright sources and hence an under-abundance of dust. Fur- thermore, none of the quasars illuminating the 7 known H2 absorbing DLAs have appreciable 12 or 3 mm fluxes (Curran et al., in preparation). |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | In summary, not only do the low metallicities reduce the molecular hydrogen content in DLAs, but they may also further lower the column densities of other molecular trac- ers, making the detection of these much more difficult than expected. It is clear that further statistics are required in or- der to identify the factors most crucial (e.g. column density, visual magnitude or metallicity) in determining the molec- ular content of DLAs. It would be of interest to compare the metallicities of the four known molecular absorbers, es- pecially in the highest redshift case (zabs = 0.886, Wiklind & Combes 1998). However, these lie along the lines-of-sight to quasars which, although bright in the millimetre regime, are too optically dim (V> 20 in all cases) to estimate the metal-to-hydrogen ratio. |
Processed_A_22GHz_search_for_molecular_absorption_at_z~3_wit.txt | We wish to thank the John Templeton Foundation for sup- porting this work. Also Tony Wong and Lister Staveley- Smith at the Australia Telescope National Facility for their support in using the upgraded ATCA and Alberto Bolatto at the Berkeley Radio Astronomy Laboratory for his prelimi- nary results. Steve Curran acknowledges receipt of a UNSW NS Global Fellowship. This research has made use of the NASA/IPAC Extragalactic Database (NED) which is oper- ated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | pΛK + total cross section data which were recently measured by the COSY-11 collaboration at an energy 2 MeV above the reaction threshold. Our analysis suggests that the measurement of the invariant mass spectrum for the ΛK + at energies around 100 MeV above the threshold can provide a new constraint on the theoretical calculations for this reaction. In par- ticular, the measurement can give a clue as to whether the reaction is dominated by resonance excitation or not. Thus, it should contribute to further under- standing of the strangeness production mechanism. We show the invariant mass spectra for the ΛK + system in these optimal kinematic conditions calculated by several approaches, and demonstrate that our suggestion is experimentally feasible. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | Recently the COSY-11 Collaboration [1] measured the total cross section for the pΛK +, at energy 2 MeV above the reaction threshold. The data were reaction, pp compared with the theoretical models which are presently available for this reaction [2, 3, 4, 5]. Among them, both the predictions of Sibirtsev [2] and Li and Ko [3] could reproduce the experimental data at this energy well only with the inclusion of pion and kaon exchanges, but without the inclusion of any resonances which are observed to decay to the ΛK + channel. Moreover, it was found [2, 3] that the dominant contribution pΛK + reaction comes from the K-exchange mechanism. Alternatively, the to the pp study of F¨aldt and Wilkin [5] which was made analogous to the pp ppη reaction with a nonrelativistic treatment including the final state interaction, suggests that the total pΛK + reactions at energies only slightly above threshold can cross sections for the pp possibly be explained by one pion-exchange followed by N ⋆(1650) resonance excitation. Furthermore, they claim that these two ingredients alone are almost enough to describe the data. One of their motivations was to study whether the same type of meson exchange model could be capable of explaining simultaneously the near-threshold η and K production data – one pion exchange followed by N ⋆(1650) resonance excitation for the present case. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | In Ref. [4], within the resonance model, we performed the most detailed study of kaon production in proton-proton collisions with a full relativistic treatment. This included the mechanism adopted by F¨aldt and Wilkin [5] as just one of the processes in evaluating the total cross sections. However, it has turned out that our calculation underestimates the data [1] at energy 2 MeV above the threshold – as shown in Fig. 1. It underestimates the experimental data from COSY by about a factor of 2, although it reproduces the existing data fairly well at energies more than 100 MeV above the threshold. One of the reasons for this underestimation is that our calculation [4] involves an incoherent sum of the contributions from each resonance, and thus would not usually be expected to reproduce the data well at energies very slightly above the threshold. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | In Fig. 1 we also show the energy dependence of the total cross sections calcu- lated using a constant matrix element, adequately normalized so as to fit the data (or adequately scaled phase space distrubutions multiplied by a constant factor) for the reaction – denoted by phase space. It implies that the experimental data at low energies can be described solely by phase space with a constant matrix element nor- malized to fit the data, because a similar energy dependence of the production cross section is also obtained by both the K-meson exchange model of Refs. [2, 3] and the resonance model [4], except for the absolute values. Thus, the total cross section data for this reaction at very low energies might not be sufficient to constrain the theoretical calculations even at energies slightly above the threshold – that is, to decide whether the reaction is dominated by kaon exchange without resonance excitation [2, 3], or by meson exchange followed by resonance excitation [4, 5]. Furthermore, it is difficult to decide what kind of meson exchange mechanism is dominant for the reaction as was questioned in Ref. [5]. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | marized in Table 1. Most of the coupling constants, cut-off parameters and form factors Y K (Y = Σ, Λ) reactions [6, 8]. have been fixed by the previous studies for the πN pΛK + reaction [4], we Together with the quantities newly appearing in the pp summarize the parameters of our model used for this study in Table 2. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | Here, we should comment that the main features of our results are very similar to In particular, the π-meson exchange is dominant at those of F¨aldt and Wilkin [5]. energies close to the reaction threshold. Moreover, the N ⋆(1650) resonance is strongly coupled to the pion, and thus gives the main contribution for the reaction, which can also be understood from Table 1. Furthermore, it should be emphasized that we are able to reproduce the data of COSY-11 [1] if we adopt the upper limits for the branching ratios for the N ⋆(1650) ΛK + channels in Table 1. Indeed, the result was increased by a factor of about 1.8 when we adopted the upper limits in the calculation. Moreover, if we vary these branching ratios as free parameters, as was done by F¨aldt and Wilkin [5], we will be certainly able to reproduce the experimental data from COSY [1]. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | However, such approaches cause two problems. First, with the same set of param- eters, one should reproduce simultaneously the experimental data for the πN ΛK reaction. Indeed, the approach of the resonance model made previously [4] and again here uses the same parameter set consistently with that used for the study of the πN ΣK reactions [6, 8]. Thus, our study for → this reaction does not have much freedom to vary the parameters, nor to introduce new parameters. Second, the model should reproduce not only the recent data [1], but should also explain the data available at energies larger than 100 MeV above the threshold [9]. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | Keeping in mind the discussions made above, as well as the theoretical and exper- imental uncertainties for the N ⋆ resonances, we may conclude that in principle both the resonance model and the model of F¨aldt and Wilkin [5] can reproduce the data pΛK + at low energies, although this was not our motivation, nor the for the pp intention of the calculation [4]. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | Here, we point out further that a more definite conclusion can be drawn from the measurement of the ΛK + invariant mass spectrum. This might also clarify whether the other resonances, N ⋆(1710) and N ⋆(1720), may play a role in the reaction as well as the N ⋆(1650) resonance at relatively low energies near the threshold. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | normalization. It is easily seen that the shapes of all spectra for the three cases at this energy are similar to that of simple phase space. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | The reason is that the width of each N ⋆ resonance is larger than the maximal total energy (when the proton in the final state has zero momentum) available for the ΛK + pair, ǫ = 10 MeV. Because the upper limit of the spectrum is given by, √s0 + ǫ mp = mΛ + mK + ǫ, the range of the invariant mass distribution (i.e. the difference between the upper and lower limit for the spectrum) is equal to the excess energy, ǫ, above threshold. Thus, for a small value of the excess energy, ǫ, the structure of the corresponding resonance – in the present case the lowest baryonic resonance, N ⋆(1650) – cannot affect the shape of the ΛK + invariant mass spectrum noticeably, because the width is much larger than the excess energy. Obviously when the quantity, mΛ + mK + ǫ, is larger than the value, mN ⋆(1650) + ΓN ⋆(1650)/2, the mass spectrum will be certainly affected by this resonance – where, mN ⋆(1650) and ΓN ⋆(1650) are respectively the mass and width of the N ⋆(1650) resonance. Then, if resonance pΛK + reaction, we expect that excitation is the dominant mechanism for the pp the ΛK + invariant mass spectrum will be certainly affected by the baryonic resonances under the condition, ǫ 100 MeV. (The N ⋆(1650) resonance will at first start to give a contribution as the energy increases.) In order to demonstrate that our argument is correct, we illustrate the ΛK + invariant mass spectrum calculated at an excess energy of 100 MeV in Fig. 3, as an example. As can be seen from Fig. 3, the contributions of π and η-meson exchange become important when all the resonances, N ⋆(1650), N ⋆(1710) and N ⋆(1720) are included together. Furthermore, we emphasize that the result of the resonance model can be now distinguished from those of the simple phase space spectrum with an arbitrary normalization, and the pion exchange with the N ⋆(1650) resonance alone in our model. Thus, the measurement of the ΛK + invariant mass spectrum may serve as a decisive test for the theoretical models and the mechanism for strangeness production. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | It is of great interest that a measurement of assosiated strangeness production in proton-proton collisions is planned by the COSY-TOF Collaboration [11] in the near future. It is also crucial, to test our suggestion, that the experiment can be performed with a clear identification of the Λ and Σ hyperons in the invariant mass spectra – as was emphasized by Laget [12]. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | Acknowledgement A.S. would like to thank L. Jarczyk, W. Cassing, B. Kamys and U. Mosel for productive discussions. This work was supported in part by the Forschungszentrum J¨ulich, and the Australian Research Council. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | See, G.Q. Li, C.-H. Lee, and G.E. Brown, nucl-th/9706057, for comparion of their calculation and the experimental data. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | Table 1: Summary of the resonance properties used in the present study. The confi- dence levels of the resonances are, N ⋆(1650) [13]. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | Table 2: Coupling constants and cut-off parameters. κ = fρN N /gρN N = 6.1 for the ρNN tensor coupling is used. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | pΛK + reaction. Figure 1: Energy dependence of the total cross section for the pp The square indicates the experimental result of Ref. [1], while the circles are data from Ref. [9]. In the upper part of figure, the solid lines indicate the results of the resonance model, and the phase space considerations using the constant matrix element normalized so as to fit the data. The lower part in figure illustrates the separate contributions from π, ρ and η-meson exchanges. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | Figure 2: The ΛK + invariant mass spectrum for the pp pΛK + reaction. Results are shown for the resonance model, and a simple phase space consideration with an arbitrary normalization. Here the reaction energy is 10 MeV above the threshold. |
Processed_A_clue_to_the_mechanism_of_$ΛK^+$_production_in_$p.txt | Figure 3: Same as Fig. 2, but the reaction energy is 100 MeV above the threshold. |
Processed_Nonlinear_Self-Duality_and_Supergravity.txt | The concept of self-dual supersymmetric nonlinear electrodynamics is generalized to a curved superspace of N = 1 supergravity, for both the old minimal and the new minimal versions of N = 1 supergravity. We derive the self-duality equation, which has to be satisfied by the action functional of any U(1) duality invariant model of a massless vector multiplet, and construct a family of self-dual nonlinear models. This family includes a curved superspace extension of the N = 1 super Born-Infeld action. The supercurrent and supertrace in such models are proved to be duality invariant. The most interesting and unexpected result is that the requirement of nonlinear self-duality yields nontrivial couplings of the vector multiplet to K¨ahler sigma models. We explicitly derive the couplings to general K¨ahler sigma models in the case when the matter chiral multiplets are inert under the duality rotations, and more specifically to the dilaton-axion chiral multiplet when the group of duality rotations is enhanced to SL(2, R). |
Processed_Nonlinear_Self-Duality_and_Supergravity.txt | In 1935, Schr¨odinger [1] showed that the nonlinear electrodynamics of Born and Infeld [2], proposed as a new fundamental theory of the electromagnetic field, possessed a remark- able property – invariance under U(1) duality rotations. Although the great expectations, which originally led the authors of [2] to put forward their model, never came true, the Born-Infeld action has re-appeared in the spotlight since the 1980’s as a low energy effec- tive action in string theory [3, 4]. Along with patterns of duality in extended supergravity [5, 6], this motivated Gaillard and Zumino, Gibbons and others [7, 8, 9, 10, 11, 12, 13] to develop a general theory of (nonlinear) self-duality in four and higher space-time di- mensions for non-supersymmetric theories1. Extension to 4D N = 1, 2 globally super- symmetric theories was given in [15, 16]. As a final step in developing the formalism, it was also shown in [17] how to re-formulate the requirement of nonlinear self-duality (i.e. a nonlinear equation which the action functional has to satisfy in order for the theory to be duality invariant) as a condition of manifest invariance of the interaction; in a sense, this is a nice extension of Schr¨odinger’s ideas [1]. |
Processed_Nonlinear_Self-Duality_and_Supergravity.txt | The above features provide enough evidence for considering supersymmetric self-dual systems to be quite interesting and their properties worth studying. In the present note, self-dual nonlinear supersymmetric electrodynamics [15, 16] is coupled to N = 1 super- gravity in the presence of nonlinear K¨ahler sigma models. To describe N = 1 supergravity, we separately consider its two off-shell realizations3: the old minimal formulation [25, 26] and the new minimal formulation [27]. By now, 4D N = 1 superfield supergravity is a subject of several textbooks [29, 30, 31] to which the interested reader is referred for further details and references. |
Processed_Nonlinear_Self-Duality_and_Supergravity.txt | The structure of the paper is as follows. Working with the old minimal version of N = 1 supergravity, in section 2 we derive the self-duality equation as the condition for a vector multiplet model to be invariant under U(1) duality rotations. We then demonstrate self-duality under a superfield Legendre transformation, introduce a family of self-dual nonlinear models and argue that the supercurrent and supertrace in such models are duality invariant. To a large extent, these results are just a minimal curved superspace extension of the globally supersymmetric results presented in [16]. However, the game will become more interesting when we turn to the couplings to new minimal supergravity and K¨ahler sigma models. Section 3 is devoted to the coupling of self- dual nonlinear supersymmetric electrodynamics to the dilaton-axion multiplet in curved superspace; this is of interest from the point of view of string theory. |
Processed_Nonlinear_Self-Duality_and_Supergravity.txt | is the full flat superspace measure, and d6z = d4x d2θ is the measure in the chiral subspace. |
Processed_Nonlinear_Self-Duality_and_Supergravity.txt | Z where the equality in the last line takes place in the so-called chiral representation (see [29, 31] for more details). This result is especially simple in the chiral case, L = Lc/R, with Lc a covariantly chiral scalar, ¯D ˙αLc = 0. |
Processed_Nonlinear_Self-Duality_and_Supergravity.txt | One of the nice properties of all models of self-dual electrodynamics is invariance under Legendre transformation [12]. It was shown in [15] that this property also holds for any globally supersymmetric model of the massless vector multiplet that is invariant under U(1) duality rotations. We will demonstrate that this property also naturally extends to curved superspace. |
Processed_Nonlinear_Self-Duality_and_Supergravity.txt | with the Lagrange multiplier VD a real scalar superfield. This model is equivalent to the original model, since upon elimination of WD by its equation of motion we regain S[W, ¯W], with the condition that W satisfy the Bianchi identity (2.8). However, we may instead eliminate W by its equation of motion, in which case we obtain a dual action SD[WD, ¯WD], the Legendre transform of S[W, ¯W ]. Now, let S[W, ¯W ] be a solution of the self-duality equation (2.14), and hence it satisfies the invariance condition (2.15) under arbitrary U(1) duality rotations (2.12). Then, eq. (2.15) for a finite duality rotation τ = π/2 is easily seen to be equivalent to the statement that SD = S. |
Processed_Nonlinear_Self-Duality_and_Supergravity.txt | This action is of the form (2.18), and one can readily check that the differential equation (2.20) is satisfied. Therefore, the minimal curved superspace extension of the N = 1 super Born-Infeld action is self-dual. |
Processed_Nonlinear_Self-Duality_and_Supergravity.txt | Gaillard and Zumino [7, 12] developed an elegant, model-independent proof of the fact that the energy-momentum tensor of any self-dual bosonic system is invariant under It is not quite trivial however to generalize this proof to the U(1) duality rotations. supersymmetric case, and this is why we will follow a brute-force approach, similar to [1, 9, 10], and directly check duality invariance of the supercurrent and supertrace. |
Processed_Nonlinear_Self-Duality_and_Supergravity.txt | It is an instructive exercise to show that δT vanishes for Λ 6= 0 only if the self-duality equation (2.20) is taken into account. Now, the conservation equation (2.26) is to be satisfied both before and after applying the duality rotation. Since T is duality invariant, the left hand side of (2.26) should also be invariant. This essentially implies duality invariance of the supercurrent. |