1: \documentclass[12pt,preprint]{aastex}
2:
3: \newcommand{\myemail}{jbp@cmu.edu}
4:
5: \begin{document}
6:
7: \title{Retroactive Coherence}
8:
9: \author{J. B. Peterson\altaffilmark{1}}
10: \affil{Department of Physics, Carnegie Mellon University, Pittsburgh, PA 15213}
11:
12: \altaffiltext{1}{email:\myemail}
13:
14:
15: \begin{abstract}
16: In gravitational microlensing, interference visibility is established
17: when the spectral resolution of the observation
18: is less than the reciprocal of the time delay difference. Using high resolution
19: spectroscopy fringes will be visible at large path differences,
20: even when the source emission is incoherent. This is possible because
21: coherence can be established after interference has already occurred.
22: Results of a
23: laboratory test are presented which demonstrate
24: this retroactive establishment of coherence.
25: \end{abstract}
26:
27: \keywords{gravitational lensing, microlensing, interference, coherence}
28:
29: \section{Introduction}
30:
31:
32: Interference can occur when there are two paths for light propagation
33: from a source to an observer.
34: In gravitational microlensing (\cite{a00}, \cite{w01}, \cite{d99})
35: wavetrains from a point on the source star
36: travel on two paths,
37: and the wave amplitudes are combined before the light is
38: detected. The lensing geometry has the potential to produce interference,
39: and applications of gravitational lensing interference have been
40: considered for wavelength regimes from radio waves to gamma rays
41: (\cite{man81}, \cite{ss85}, \cite{dw88}, \cite{pf91}, \cite{g92}, \cite{spg93},
42: \cite{ug95}).
43: As many of these authors note, interference visibility will
44: be diminished if the difference of the length of the two paths
45: exceeds the coherence length of the light.
46:
47: The light sources in gravitational microlensing are incoherent.
48: In many microlensing examples the light sources are stars and stellar
49: photospheres emit primarily blackbody radiation.
50: The temporal coherence length of this light is $L_{coh}= c~t_{coh} = c / \Delta \nu$,
51: where $c$ is the speed of light, $t_{coh}$
52: is the coherence time and $\Delta \nu$ is the half-power bandwidth
53: of the source emission spectrum.
54: Because of the large bandwidth of the
55: blackbody spectrum, the coherence length $L_{coh}$ is about equal to
56: the peak wavelength. Blackbody emission is often labeled ``incoherent''.
57:
58: Blackbody emission can be modeled as a continuous
59: stream of short wavepackets.
60: Figure 1 shows wavepackets of length $L_{coh}$ traveling on two
61: paths from source to observer around a lensing mass.
62: Because the wavepackets do not overlap at arrival,
63: the short coherence length of the light would seem to
64: indicate that gravitational lensing
65: interference is undetectable except in cases where the path length
66: difference amounts to just a few wavelengths. However, this is not
67: correct, as the experiment described below demonstrates.
68:
69: Coherence can be established retroactively. {\it After}
70: incoherent light has passed through the interference apparatus
71: (in the lab or in the Universe), it can be filtered to a narrow bandwidth and
72: that filtered light will have a new, longer coherence length. This allows
73: gravitational microlensing interference to be observed for large path
74: length differences.
75:
76: In this Letter I present results of a laboratory experiment, carried out
77: using a Michelson interferometer, which demonstrates retroactive coherence.
78:
79: \section{Two Path Coherence}
80:
81: An analysis of partial coherence for two path geometries is presented in
82: \cite{bw80}. I adapt their analysis to the present purpose.
83:
84: To describe the degree of interference seen in the patterns formed by his
85: interferometers
86: Michelson defined a quantity called the
87: interference visibility
88: $V(\Delta \ell) = (I_{max}-I_{min})/(I_{max} + I_{min})$.
89: Here the maximum and minimum of the intensity $I_{max}$
90: and $I_{min}$
91: are taken as the
92: path length difference $\Delta \ell$
93: varies. $V$ ranges from 0 to 1, and generally has a maximum
94: at zero path difference.
95:
96: For two-path gravitational microlensing, or for light that passes through a
97: Michelson interferometer, the detected intensity is
98:
99: \begin{equation}
100: I(\Delta \ell)= <|\sqrt{I_1} f(t) + \sqrt{I_2} f(t+\Delta \ell/c)|^2 >_T
101: \end{equation}
102:
103: where $I_1$ and $I_2$ are the intensities on the two paths, $f(t)$
104: is
105: a complex function describing the time variation of the arriving field strength
106: on the shorter path, and
107: $f(t + \Delta \ell/c)$ is the same function shifted in time by the path delay
108: difference.
109: $f$ is normalized, i.e. $<f^*(t)f(t)>=1$, and assumed to be stationary.
110: $T$ is the duration of the observation, the time over which the
111: intensity is averaged, and is taken to be much longer than
112: $t_{coh}$ and $\Delta \ell/c$.
113:
114:
115: The total intensity is then
116:
117: \begin{equation}
118: I(\Delta \ell) = I_1 + I_2 + 2 \sqrt{I_1 I_2} Re (<f^*(t)f(t+ \Delta \ell/c) >_T)
119: \end{equation}
120:
121: The last term describes the interference is also called the
122: real part of the complex degree of coherence
123: $\gamma(\Delta \ell) = <f(t)^*f(t+\Delta \ell/c)>$.
124:
125: For an optical frequency band centered at $\nu_c$ the
126: magnitude of the degree of coherence
127: is related to the interference visibility by
128: $Re (\gamma(\Delta \ell)) = V(\Delta \ell) cos (2 \pi \nu_c \Delta \ell /c)$.
129:
130: The autocorrelation of the wave amplitude $\gamma$ is the
131: fourier transform of the power spectral density $P(\nu)$
132: of the light and the coherence length is the width of the function
133: $|\gamma_{12}(\Delta \ell)|$.
134: This means that decreasing the bandwidth of
135: the observation, i.e. forcing $P(\nu)$ to zero outside a narrow range of
136: frequencies, increases the coherence length and makes interference visible at
137: large path length differences. The experiment described below shows that this
138: is true even when the bandwidth restriction occurs after the light has
139: passed through the interference apparatus.
140:
141: \section{Laboratory Analog of Lensing Interference}
142:
143: I set up a laboratory analog of gravitational lensing interference using a
144: Michelson interferometer, as shown in figure 2. In the experiment
145: light from a
146: laboratory thermal source
147: arrives at a partially
148: reflecting beam splitter which divides the light among two paths.
149: These paths are the two orthogonal arms of the interferometer.
150: At the ends of the arms mirrors return the light to the
151: beam splitter with a path
152: length difference that can be adjusted.
153: The beam splitter recombines the light before it passes through
154: the exit aperture.
155:
156: Filtering the light restores interference visibility.
157: In gravitational lensing the path length varies with time because of
158: relative motion among the source, lensing masses, and the observer.
159: In the Michelson interferometer, the path lengths
160: are stable
161: but the interferometer can be adjusted to give various path lengths through
162: various parts of the instrument.
163: I tilted one mirror slightly so zero path length difference
164: occurs only at a line dividing the pattern vertically.
165: The pattern is also divided horizontally.
166: The top half of the exit aperture of the interferometer is unblocked and here
167: the full bandwidth of the thermal source contributes to the
168: interference pattern.
169: The bottom half of the exit aperture is covered by a narrow-band
170: ($486 nm \pm 15 nm$) filter. In figure 3 the results of the experiment are
171: displayed. For white light the visibility falls almost to zero within a few
172: fringes. For filtered light the visibility remains near one for
173: the entire width of the pattern.
174:
175: Considering the frequency domain it is easy to see
176: how coherence is restored through bandwidth restriction: the white light
177: interference pattern
178: is composed of many narrow-band patterns overlaid. The spacing of
179: the fringes is proportional to wavelength.
180: For all wavelengths, the patterns are aligned in the center of the image,
181: at zero path difference.
182: The central few fringes have high visibility but, away from the
183: center of the image, the pattern develops colored stripes, as red and
184: blue fringe patterns separate in phase.
185: At the edge of the pattern the visibility has
186: fallen nearly to zero because of cancellation,
187: among the various wavelengths, of constructive and destructive
188: interference. By selecting a narrow range of frequencies
189: cancellation is avoided and the interference becomes
190: visible across the entire field. This is shown
191: in the right panel of figure 3.
192:
193: In the time domain a puzzle remains: how can short coherence length
194: wave packets,
195: which don't overlap as they exit the interferometer, create visible
196: interference at large path length differences?
197: The answer is that the filter stores multiply delayed
198: copies of the wave train incident upon it.
199: Interference occurs among these copies. Light storage in a filter is
200: illustrated in figure 4.
201:
202: \section{Discussion}
203:
204: In the experiment described here a single narrow filter is used to
205: restrict the passband of the observation. Most of the optical power
206: from the source is
207: blocked by this filter. In a gravitational microlensing observation a
208: spectrophotometer could be used instead of a filter. This would
209: divide the spectrum into many
210: narrow passbands, each acting as a narrow filter.
211: This allows all the optical power to be used in the observation.
212: For a spectrophotometer the coherence time $t_{coh}$ is the reciprocal of the
213: frequency resolution.
214:
215: Although temporal coherence can be increased by using high spectral
216: resolution, limited {\it lateral} spatial coherence remains a factor reducing
217: lensing interference visibility for most geometries studied so far.
218: As described in the references above, the path length
219: difference varies from point to point across
220: the disk of a source star. The total variation is generally many wavelengths,
221: and this substantially reduces interference visibility.
222:
223: Occasionally, microlensing events include caustic crossings.
224: At the first crossing two new, high magnification paths appear. These
225: paths have the same length at first so
226: the size of the laterally coherent region (\cite{jp95}, \cite{gg97}
227: %, \cite{zp02}
228: ) can cover a much larger part of the source disk than in the more common
229: Schwarzschild case.
230:
231: A Michelson interferometer is useful, not just as a laboratory analog
232: of gravitational lensing interference, but also as a tool for observation
233: of the effect. Microlensed light, collected with a telescope, can be
234: passed through a Michelson interferometer before detection. Then,
235: if the interferometer path length difference matches the lensing
236: path length difference, the two arriving wavepackets seen in figure 1
237: can be realigned, and bandwidth reduction of visibility can be avoided.
238: For slight mismatches of the two path length difference the
239: required bandwidth for full visibility is
240: $\Delta \nu < c/ (\Delta \ell_{interferometer} - \Delta \ell_{lensing})$.
241:
242: Even though a complex representation of the wave amplitude is used
243: in this discussion of interference, and even
244: though the uncertainty principle $\Delta t_{coh} ~ \Delta \nu \sim 1$
245: applies to the coherence time,
246: the effects described here are classical rather than quantum
247: mechanical. No use is made, in this analysis, of photon quanta.
248:
249: The use of high resolution spectroscopy establishes coherence
250: retroactively, allowing gravitational microlensing interference
251: to be visible for path length differences many times larger than the
252: wavelength of the light.
253:
254:
255: \begin{thebibliography}{}
256: \bibitem[Alcock, et.al.(2000)]{a00}Alcock, C., et.al 2000, \apj, 542, 281
257: \bibitem[Born and Wolfe(1980)]{bw80}Born and Wolfe
258: \bibitem[Deguchi and Watson(1988)]{dw88}Deguchi, S. and Watson, W. D. 1986, \apj, 307, 30
259: \bibitem[Derue, et.al.(1999)]{d99}Derue, F., et.al. 1999, \aap, 351, 87
260: \bibitem[Gould(1992)]{g92}Gould, A. 1992, \apj, 386, 5
261: \bibitem[Gould and Gaudi(1997)]{gg97} Gould, A. and Gaudi, B.S. 1997, \apj, 486, 687
262: \bibitem[Jaroszynski and Paczynski(1995)]{jp95}Jaroszynski, M. and Paczynski, B. 1995, \apj, 455, 443
263: \bibitem[Mandzhos(1981)]{man81} Mandzhos, A. V. 1981, Soviet Astron. Lett., 7, 213
264: \bibitem[Peterson and Falk(1991)]{pf91}
265: Peterson, J. B. and Falk, T. 1991, \apj, 374, 5
266: \bibitem[Schneider and Schmid-Burgk(1985)]{ss85} Schneider, P. and Schmid-Burgk, J. 1985, \aap, 148, 369
267: \bibitem[Stanek, Paczynski and Goodman(1993)]{spg93}
268: Stanek, K. Z., Paczynski, B. and Goodman, J. 1993 \apj, 413, 7
269: \bibitem[Ulmer and Goodman(1995)]{ug95} Ulmer, A. and Goodman, J. 1995, \apj, 442, 67
270: \bibitem[Wozniak, et.al.(2001)]{w01} Wozniak, P. R., et.al. 2001,
271: Acta Astron., 51, 175
272: %\bibitem[Zabel and Peterson(2002)]{zp02} Zabel, S. and Peterson, J. B. 2002,
273: %{\it Diffraction and interference at gravitational microlensing caustic crossings}
274: %in preparation
275: \end{thebibliography}
276:
277:
278: \clearpage
279:
280: \begin{figure}
281: \plotone{f1.eps}
282: \caption{{\bf Gravitational microlensing time delay.}
283: In gravitational microlensing light travels on two
284: paths, and is recombined before detection.
285: The spike waveforms shown represent the short coherence length of a
286: wavepacket from a blackbody source. If the lengths of the two paths differ
287: by more than the coherence length, and the detector is sensitive to the
288: full bandwidth of the arriving light, there will be strong attenuation of the
289: interference visibility.
290: \label{lens}}
291: \end{figure}
292:
293: \clearpage
294:
295: \begin{figure}
296: \plotone{f2.eps}
297: \caption{{\bf Michelson interferometer used to demonstrate retroactive
298: coherence.}
299: A thermal light source illuminates a ground glass diffuser.
300: Light from each point on the diffuser
301: is divided by the beam splitter, reflects off the
302: flat mirrors and returns to the beam splitter. Light amplitudes are
303: recombined at the beam splitter and the composite waves pass out the
304: exit aperture. A filter processes the wave, after it has passed through
305: the interferometer, and the interference pattern is recorded with a camera
306: focused to image the diffuser.
307: \label{mich}}
308: \end{figure}
309:
310:
311: \clearpage
312:
313: \begin{figure}
314: \plottwo{f3.eps}{f4.eps}
315: \caption{{\bf Interference Patterns.}
316: Both panels show the diffuser photographed through the Michelson interferometer.
317: The vertical stripes are due to interference. The interferometer is set for
318: zero path difference along a vertical line through the center of the
319: image. One of the interferometer mirrors has been tilted so
320: the path length difference increases to the right and left of the centerline.
321: The bottom half of the exit aperture is
322: covered by a narrow-band filter.
323: In the left panel the exposure is set so detail in the white light
324: interference pattern can be seen in the top half of the image.
325: The visibility of the pattern falls nearly to zero in just a few fringes
326: because the coherence length of the white light is just a few times the
327: peak wavelength.
328: %The dark line near the bottom of the white light pattern is
329: %due to the edge of the glass encapsulating the filter.
330: In the right panel the exposure
331: has been increased so that the narrow band interference pattern, photographed
332: through the filter, can be seen. Interference visibility is
333: maintained across the entire width of the pattern because the filter has
334: increased the coherence length of the light. This shows that coherence
335: can be established in the light waveform after the light has passed
336: through the interference apparatus.
337: \label{fringes}}
338: \end{figure}
339:
340: \clearpage
341:
342: \begin{figure}
343: \plotone{f5.eps}
344: \caption{{\bf Light Storage in a Bandpass Filter.}
345: A narrow passband filter stores light for a long time,
346: increasing the coherence length of the stored waves. Here multiple
347: reflections in a Fabre-Perot Etalon
348: are shown.
349: \label{fabre-perot}}
350: \end{figure}
351:
352:
353: \end{document}
354:
355:
356: