Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
HIGH SPEED OPTICAL SPECTROMETER
Document Type and Number:
WIPO Patent Application WO/2007/115312
Kind Code:
A2
Abstract:
A subject light (12) is optically amplitude modulated (14) to generate a sequence of light pulses, each having wavelength components of the subject light. The light pulses propagate through a wavelength dispersive element (18) that spreads the different wavelength components of the pulse in time with a spacing corresponding to the frequency spacing of the components. The time spread light is converted to an electrical signal (20, 22). The time domain form of the electrical signal represents the frequency spectrum of the subject light.

Inventors:
WANG YONGXIN (US)
WANG ANBO (US)
HAN MING (US)
Application Number:
PCT/US2007/065983
Publication Date:
October 11, 2007
Filing Date:
April 04, 2007
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
VIRGINIA TECH INTELL PROP (US)
WANG YONGXIN (US)
WANG ANBO (US)
HAN MING (US)
International Classes:
G01J3/46
Foreign References:
US20020101592A1
Attorney, Agent or Firm:
WHITHAM, Michael, E. et al. (Curtis Christofferson & Cook, P.C.,11491 Sunset Hills Road, Suite 34, Reston VA, US)
Download PDF:
Claims:
We hereby claim:

1. A spectrometer, comprising: an optical modulator, having an optical in port and an optical out port, to receive a light at the optical in port, and modulate the light to output a succession of pulses, each pulse having a length and form to having a spectral content substantially equal to the spectral content of the input light; an optical dispersion element having an optical input optically connected to the optical out port of the optical switch, and an optical out port constructed and arranged to a wavelength-dependent propagation time from said input to said output; and a sampler, constructed and arranged to sequentially sample light from the optical out port of the optical dispersion element, and to generate a corresponding sequence of time samples.

2. The spectrometer of claim 1 , further comprising a processing circuit constructed and arranged to generate a spectrum data based on the corresponding sequence of time samples.

3. The spectrometer of claim 2, wherein the dispersion element is constructed and arranged to have a first propagation time at a first given wavelength and a second propagation time at a second given wavelength shorter than the first given wavelength, the second propagation time being longer that the first propagation time.

4. The spectrometer of claim 2, wherein the dispersion element is constructed and arranged such the propagation time is proportional to the wavelength of the light.

5. The spectrometer of claim 1 , further comprising: a light source for generating a light having a first spectrum; an interferometer, optically connected to the light source, constructed and arranged to receive the light having the first spectrum and to output an interference light having a second spectrum; and

an optical transmission path having a receiving end and a terminal end, the receiving end arranged to receive the interference light, the terminal end connected to an input of the optical modulator.

7. A method for detecting a spectrum of a light having a plurality of wavelengths, comprising: modulating the light to output a succession of pulses, each pulse having said wavelengths; spreading the wavelengths apart from one in time to generate a time domain light having an amplitude versus time form corresponding to the wavelength versus amplitude of said spectrum; and generating a spectral measurement based on the amplitude versus time form of the time domain light.

8. The method of claim 7, wherein said spreading includes propagating the pulses through a dispersive element.

9. The method of claim 7, wherein said spreading includes injecting the pulses into a fiber grating to generate the time domain light as a reflected dispersed signal.

10. The method of claim 7, wherein said generating a spectral measurement includes converting the time domain light to an electrical detection signal.

11. The method of claim 10, wherein said generating a spectral measurement further includes detecting amplitude versus time characteristics of the electrical detection signal.

12. The method of claim 11 , wherein said detecting amplitude versus time characteristics of the electrical detection signal includes identifying predetermined forms of the electrical detection signal.

13. The method of claim 12, wherein said identifying predetermined forms of the electrical detection signal includes detection of edges.

14. The method of claim 13, wherein said detection of edges includes sampling said electrical detection signal at a sample rate higher than twice the reciprocal the distance between peaks and valleys of the electrical detection signal.

15. The method of claim 14, wherein the sample rate is higher than approximately one gigasamples per second.

Description:

HIGH SPEED OPTICAL SPECTROMETER

TECHNICAL FIELD

[0001] The present invention relates to optical spectrometry and, more particularly, high speed optical spectrometry.

BACKGROUND

[0002] Optical spectrum measurement is used across a wide range of industries and sciences. Various devices, systems and methods of spectral measurement are known. One example is a diffractive grating and optical detector; the diffractive grating disperses different optical wavelengths light and different angles, causing the light to have different positions of incidence on the optical detector, which indicates the spectral content.

[0003] Optical spectrum measurement may identify an emission or absorption spectra, to measure chemical reactions in a laboratory or identify the constituent elements of stellar objects. Optical spectrum measurement may be used with interferometric sensors, to detect almost any physical quantity that can impart a strain on the sensor including, for example, pressure, deflection, and temperature. Example types of interferometric sensors are Michelson, Mach-Zhender, Fabry- Perot, and Sagnac interferometers (referenced generically in the disclosure as "interferometer.")

[0004] Interferometers are typically illuminated by either a monochromatic light source or a broadband light source. When illuminated by a monochromatic light, such as a laser, the laser light is propagated or reflected to form multiple light paths, to cause constructive or destructive interference patterns of fringes, spatial or temporal. Physical strain and distortion of the interferometer by, for example, pressure, bending, or temperature expansion, changes the length of the optical paths, causing a change in the interference fringes. Therefore, if the movement of the fringes is monitored a quantity of interest, e.g., pressure, bending, or temperature expansion, may be measured. However, only the changes in the interference patterns are measured. The measurement is therefore relative, not absolute. Further, since only changes in the fringe pattern are measured, the fringe tracking must be continuous. If the tracking is interrupted, the historical information

on fringe changes is lost. In addition to being relative, a calibration must be performed before each measurement. For these and other reasons, industrial applications of monochromatic interferometers are very limited. [0005] "White-light interferometers," also called "broadband interferometers," illuminate the interferometer with a broadband light source, ad are another major category of interferometer sensor. Broadband light interferometers have a much wider field of use, as they provide absolute measurement of an optical spectrum, and can be switched on and off without loss of information and without requiring recalibration.

[0006] A broadband light interferometer apparatus typically comprises a broadband light source, a sensing interferometer and a spectrometer. The sensing interferometer is typically a Michelson, Mach-Zehnder, Sagnac, or Fabry-Perot (FP) type. The general principle is that the different wavelengths in the broadband light, after being split and recombined along different paths through the interferometer, interact to produce interference patterns of fringes. The patterns are spatial and temporal, having characteristic peaks and valleys. The relative spacing of the peaks and valleys changes when the interferometer sensor is physically distorted, as this changes the optical path lengths formed by the interferometer. Unlike the monochromatic interferometer, the relative spacing of the fringes unambiguously indicates the physical distortion of the sensor. The movement history of the fringes does not have to be continuously monitored.

[0007] However, since the fringe pattern changes as the physical parameters of the interferometer change (e.g., from pressure, flexion, temperature-induced stress) the fringe pattern must be sampled at a "frame rate" high enough to detect the fastest changes in the event being monitored. Each frame of the spectrum from the sensor is typically recorded and processed.

[0008] At present, linear coupled-charge-device (CCD) detectors measure the fringe patterns of interferometer outputs. However, CCD-based spectrometers have shortcomings. One is low speed, typically lower than 100 frames-per-second, because CCDs require a long charge accumulation time. [0009] Another shortcoming, caused by current CCD spectrometers being silicon based, is frequency response, typically limited to wavelengths below 1.1mm. The present inventors observe that this presents a significant problem, because

reasonably priced broadband high power fiber pigtailed light sources having a wavelength below 1.1mm are not available. One reason is that communication wavelengths are at 1.3 or 1.55mm and, therefore, light sources below 1.1mm are not in mass production. Therefore, silicon CCD technology has not been available because it does not work well at the popular 1.3 or 1.55mm wavelengths. The CCD industry has attempted a two-layer semiconductor as a solution. The first layer absorbs the incident longer wavelengths and re-emits photons at a shorter wavelength to which the silicon detectors are sensitive. This new technology has not been widely accepted, however, for at least two reasons. First, the technology is difficult and costly, such that a CCD spectrometer having it is approximately six times the price of a silicon CCD-based spectrometer. Secondly, the two stages of light detection render a significantly reduced sensitivity.

SUMMARY OF THE INVENTION

[0010] In view of these and other shortcomings of the prior art, one object of the invention is to provide a high-speed spectrometer, having a sample rate significantly higher than provided by CCD-based spectrometers.

[0011] A further object is to provide a high-speed spectrometer, providing high frequency resolution over a wide range of frequencies.

[0012] A further object is to provide a low cost, rugged fiber grating based sensor, having a low cost spectrometer with sufficient frequency resolution to accurately measure the fiber grating fringes.

[0013] According to one aspect, a subject light is received by an optical modulator. The optical modulator amplitude modulates the subject light to generate an optical pulse train. The amplitude modulation is such that the frequency domain spectrum of each generated optical pulse is approximately the same as the spectrum of the subject light. The optical pulse propagates through a wavelength dispersive element that spreads the different wavelength components of the pulse in time. The dispersive element is constructed and arranged such that, upon exit from the dispersive element, the time spread between the various wavelength components of the optical pulse represents the frequency (or wavelength) spacing of the components in the frequency domain. The time spread light is converted to an

electrical signal. Peaks and valleys of the electrical signal may be detected. The detected peaks and valleys correspond to peaks and valleys of the frequency spectrum of the subject light.

[0014] According to one aspect the electrical signal is sampled by an analog to digital converter, to generate digital samples. The digital samples may be processed by, for example, a programmable general-purpose computer, to detect centers of the peaks and valleys, edges, and other parameters of interest in the electrical signal conversion of the dispersed pulse light.

[0015] According to one aspect, the detected peaks and valleys of the electrical signal are averaged over a plurality of pulses.

[0016] According to one aspect, the bit resolution of the analog to digital converter is high, and the digital samples are processed using, for example, curve-fitting processes to detect or estimate, for example, positions of peaks and valleys, edges, and other parameters of interest in the electrical signal conversion of the dispersed pulse light.

[0017] According to one aspect, a sampling rate of the analog to digital converter is high, and the digital samples may be averaged to perform smoothing, and may be processed to detect positions of peaks and valleys, edges, and other parameters of interest in the electrical signal conversion of the dispersed pulse light.

[0018] According to one aspect, a broadband light source illuminates an interferometer sensor. Light reflected from the interferometer, having a fringe pattern corresponding to a physical state of the interferometer, is input to the modulator as the subject light.

[0019] According to one aspect, the dispersive element is arranged in a transmissive mode such that light pulses from the optical modulator propagate through the transmissive element.

[0020] According to one aspect, the dispersive element is arranged in a reflective mode, such that light pulses from the modulator pass through a circulator, or equivalent directional optical coupler, are injected into the dispersive element, and reflections have a time spread between the various wavelength components of the pulse representing the frequency (or wavelength) spacing of the components in the frequency domain.

BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The foregoing and other objects, aspects and advantages will be better understood from the following detailed description of a preferred embodiment of the invention with reference to the drawings, in which:

[0022] FIG. 1 shows a high-speed spectrometer;

[0023] FIG. 2A graphically depicts an example spectrum of one light input to the

FIG. 1 high speed spectrometer; 2B graphically depicts an example modulated pulse, and FIG. 2C graphically depicts a time domain spectral detection signal;

[0024] FIG. 3 is a functional flow chart of an example high speed spectral measurement;

[0025] FIG. 4 graphically illustrates a diffraction pattern of a Bragg dispersion element;

[0026] FIG. 5 graphically illustrates a diffraction pattern of a chirped grated fiber dispersion element;

[0027] FIG. 6 is a functional schematic of one embodiment of a high speed spectrometer with an interferometer sensor;

[0028] FIG. 7 is a functional schematic of another example embodiment of a high speed spectrometer with an interferometer sensor;

[0029] FIG. 8 is a functional schematic of one alternative of the FIG. 7 example, having a reflective mode chirped fiber element, connected by a circulator;

[0030] FIG. 9 is a functional schematic of one embodiment of an over-sampling high speed spectrometer, having a fiber interferometer using a chirped fiber Bragg grating;

[0031] FIG. 10 is a functional schematic of one alternative of the FIG. 10 example, having an optical light amplifier;

[0032] FIG. 11 graphically illustrates spectral fringe patterns, and sampling corresponding dispersed time domain signals;

[0033] FIGS. 12A and 12B graphically illustrate, respectively, spectrum from a fiber grating sensor and a measurement of the spectrum; and

[0034] FIGS. 13A through 13C graphically illustrate three sampling rates and resolutions.

DETAILED DESCRIPTION OF EXEMPLARY EMBODIMENTS

[0035] The following detailed description of the invention is in reference to accompanying drawings, which form a part of this description. The description is of illustrative examples of various embodiments in which the invention may be practiced. The invention is not limited to the specific illustrative examples. Other configurations and arrangements embodying or practicing the present invention can be readily implemented by persons skilled in the arts, upon reading this description. [0036] In the drawings, like numerals appearing in different drawings, either of the same or different embodiments of the invention, reference functional or system blocks that are, or may be, identical or substantially identical between the different drawings.

[0037] Various embodiments of the invention, although different, are not necessarily mutually exclusive. For example, a particular feature, function, act or characteristic described in one embodiment may, within the scope of the invention, be included in other embodiments. Further, various instances of the phrase "in one embodiment" do not necessarily refer to the same embodiment. [0038] Unless otherwise stated or made clear from its context, the terminology and labeling used herein is not limiting and, instead, is only for purposes of internal consistency in referencing the examples.

[0039] Fig. 1 shows a functional schematic of one example high-speed spectrometer 10. Referring to FIG. 1 , the example 10 includes an optical gate or modulator 14, which modulates light based on a modulating pulse generator 16. The optical modulator 14 and modulating pulse generator 16 are shown as separate blocks to illustrate functions; a single hardware unit (not shown in FIG. 1) may perform both of these functions. Alternatively, an external generator (not shown in FIG. 1) may substitute for the modulating pulse generator 16. [0040] With continuing reference to FIG. 1 , a dispersion element 18 connects the modulator 14 to a photodetector 20, which connects to a data acquisition device 22. The dispersion element 18 propagates different wavelengths of light at different speeds, i.e., the dispersion element 18 causes different wavelengths of light to incur a different delay traversing from the input to the output of the dispersion element18. The purpose of the different delays will be described in greater detail below. The element 18 may be constructed to provide more delay for shorter wavelengths than

for longer wavelengths, or the reverse, i.e., the element 18 may exhibit a light propagation speed that is approximately proportional or inversely proportional to the wavelength, at least over the wavelengths of interest. [0041] One example operation of the FIG.1 high speed spectrometer will be described in reference to FIGS. 2A - 2C. FIG. 2A shows the P(ω) spectrum of an arbitrary input light, labeled 12 on FIG. 1, input to the modulator 14. It is assumed that the light 12 is continuous. The light 12 may be directly input, or transmitted to the modulator 14 through an input path 14A that may, for example, be an optical fiber. Optional, but not necessary, a light amplifier (not shown in FIG. 1) may be inserted prior to the modulator 14.

[0042] Referring to FIG. 2B, the controller 16 and modulator 14 are preferably constructed such that the output of the modulator 14 is a sequence of light pulses So(O, each having a slightly rounded form with a width of approximately WC. Referring to FIGS. 2A and 2B, the width WC and shape of each pulse S 0 (O are such that the bandwidth S 0 (O is sufficiently narrow, compared to the P(ω) spectrum of the input light 12, that the spectrum of S 0 (O is approximately the same as the P(ω) spectrum. Each of the pulses So(O can therefore be approximated as a superposition of pulses with continuous spectral components ω, and power P{ω). The form of the pulse So(O therefore cannot be overly narrow or rectangular, as the spectrum of So(O would then be a SinX/X form (the Fourier transform of a rectangular pulse) convolved with the P(ω) spectrum, which would be very wide, and would not be a replica of the P(ω) spectrum.

[0043] Referring to FIG. 1 , the series of light pulses S 0 (O are transmitted through, for example, an optical fiber to the dispersion element 18. As described above, the dispersion element 18 has a different propagation speed, and therefore delay, for each of the different wavelengths in So(O, e.g., ω1 , ω2, ω3, cu4 and ω5. Each of the So(O components or bands ω1 , ω2, ω3, ω4 and ω5 therefore propagates at a respectively different speed through the dispersion element 18. Therefore, although the signals within the pulse S 0 (O at the different bands are concurrent in time upon entering the dispersion element, they incur a time spread from one another as they progress through the dispersion element 18.

[0044] FIG. 2C is a time domain plot of S(O at the exit of dispersion element 18, showing the result of the different delay incurred by the bands ω1 through ω5 within

S 0 (O passing through the element 18 - provided the bandwidth of the pulse S 0 (O is much smaller than the bandwidth of the incoming light spectrum P(ω) being measured. As seen in FIG. 2C, the S(O output from the dispersion element 18 has the same form, along its time axis, as the light spectrum P(ω) of FIG. 2A along the frequency axis ω. More specifically, time instants t1 , t2, t3, t4, and t5 on the time axis of FIG, 2C correspond to frequency values ω1 , ω2, ω3, ωλ and ω5 along the frequency axis of FIG. 2A.

[0045] Referring to FIG. 1 , the S(t) output from the dispersion element 18 is detected by, for example, the photodetector 20. If desired, a processor or data recorder such as, for example, the data acquisition device 22. [0046] Therefore, by measuring (e.g., sampling by photodetector 20 and a data acquisition device 22) the pulse profile S(O from the dispersion element 18 in the time domain, the spectrum P{ω) of the subject light 12 is determined. Since time domain measurement can be performed at a very high rate by, for example, commercially available high-speed optical data acquisition devices, a spectrometer according to FIG. 1 provides very high speed spectral measurement high speed. [0047] Various alternative implementations of the dispersion element 18 are contemplated. For example, the dispersion element 18 may be a bulk optical element or an optical waveguide, such as a particular construction of an optical fiber, providing a large dispersion of by effecting different time delays for different wavelengths of light. Various constructions of optical fibers and other light propagation structures are contemplated to obtain the desired dispersion characteristic in the dispersion element 18. Referring to FIGS. 3 and 4, one specific example implementation of element 18 is a chirped longitudinal grating, having a refractive index that changes periodically along its longitudinal direction LD as shown in FIG. 4. Referring to FIG. 3, the change in refractive index reflects light satisfying the so called Bragg condition, given by

ls=2λn (Equation No. 1)

where λ is the spatial period and n is the average refractive index of the grating material.

[0048] Referring to FIG. 4, a chirped grating has varying grating periods along the propagation direction, so light beams with different wavelengths are reflected at different positions so different delays are generated. A chirped grating can be created in a bulk optical material or in an optical waveguide. Upon reading this disclosure, known methods and processes for fabricating a chirped grating in a bulk optical material may be utilized and, combined with this disclosure, used to implement a dispersion element 18 and, therefore, further detailed description of a bulk optical implementation is not necessary.

[0049] Another example implementation of the dispersion element 18 is a waveguide structure, such as an optical fiber. An optical fiber of sufficient length can exploit the fiber's material dispersion and the different delay imparted to different wavelengths that is inherent to optical fiber, to obtain a required dispersion. This is readily implemented because optical fiber is easily made very long. As will be understood from this disclosure, a dispersive element 18 having a long differential delay time (e.g., referring to FIG. 2C, a substantial difference between f1 and t5) by selecting an optical fiber having a proper (e.g., linear) chromatic dispersion over the wavelengths of interest, and simply calculating the required length of the fiber. As can also be understood, based on the sample rate of the data acquisition unit 22 or equivalent, and the sample rate of the pulse generator 16 (i.e., the rate of the pulses So(O injected into the dispersive element 18), a person of ordinary skill in the art can readily implement a high speed spectrometer according to the example 10 of FIG. 10, having a any desired wavelength measurement resolution. [0050] FIG. 5 is a functional flow diagram of one example method for measuring the spectrum of a light that may be performed on, for example, the high speed spectrometer 10 of FIG. 1.

[0051] Referring to FIG. 5, at 50 an incoming light of arbitrary spectrum is received. Graph plot 5OA graphically depicts an example spectrum, in a wavelength/amplitude form, having three bands, labeled A 1 A 2, A 3 , with respective amplitudes / 0 (A 1 ), / 0 (A 2 ) and lo(λ 3 ). At 52 the received light is modulated to output a series of light pulses, each of the form S 0 (O shown in graph plot 52A. As described previously, the pulses S 0 (O are preferably formed such that the bandwidth of S 0 (O is substantially narrower than the bandwidth of the light received at 50. With continuing reference to FIG. 5, at 54 the light pulses S 0 (O are passed through a dispersive element to cause a time

spread between the difference wavelength (or frequency) components of S 0 (O- Graph plot 54A shows an example time spread, with t1 , t2, and t3 corresponding to the frequency components labeled in graph plot 5OA as A 1, A 2 and A 3 . At 56, the dispersed signal shown in graph plot 54A is converted to an electrical signal by an optical-to-electrical converter, at 57 the electrical signal is sampled and, at 58, the samples are processed to convert the time domain form shown in 54A to a spectral plot such as 58A.

[0052] FIG. 1 shows only one example arrangement and configuration for the example 10. Various alternative implementations will be apparent upon reading this disclosure. For example, FIG. 1 shows the dispersion element 18 connected in a transmission mode, where the gated light pulse So(O passes through the dispersion element 18. One contemplated variation, as shown in FIG. 6, connects a dispersion element equivalent to element 18 of FIG. 1 in a reflective mode instead a transmission mode.

[0053] Referring to FIG. 6, the example 60 has a light modulator unit 64, controlled by a pulse generator 66. The light modulator unit 64 and the pulse generator 66 may be in accordance with as the modulator unit 14 and gate control generator 16 described in reference to FIG. 1. An optical coupler or optical circulator 68, hereafter reference as an "optical circulator", has three ports: in port 68a, in/out port 68b, and out port 68c. Various optical circulators are available from various commercial vendors and, based on this disclosure, a person of ordinary skill in the art can readily select an appropriate model to construct an apparatus or perform a method embodying or practicing one or more of the appended claims. [0054] The optical circulator 68 is arranged in the schematic of FIG. 6 to function as follows: light propagation with low attenuation from in port 68a to in/out port 68b; light propagation with low attenuation from in/out port 68b to out port 68c; no substantial light propagation (i.e. low crosstalk) from in port 68a to out port 68c; and no substantial propagation from in/out port 68b to a in port 68a. The phrases "with low attenuation" and "no substantial propagation" are, as can be understood by persons skilled in the art, relative terms for which specific numerical values are dependent on particular system requirements, but can be readily determined by combining this disclosure with standard engineering design methods known to persons of ordinary skill in the art.

[0055] With continuing reference to FIG. 6, a reflective-type dispersive element 70 such as, for example, a chirped fiber Bragg grating is connected to in/out port 68b of the optical circulator 68. The chirped Bragg grating 70 may be a bulk optical material or an optical fiber. An optical-to-electric converter 72 is optically connected to the out port 68c of the optical circulator 68, and a data acquisition unit 74 and processor 76 sample and process the output of the optical-to-electric converter 72. The optical-to-electric converter 72, data acquisition unit 74 and processor 76 are only functional blocks, and these functions may be incorporated into a single hardware unit (not shown).

[0056] Referring to FIG. 6, an example spectral measurement is performed as follows: a light having a wavelength spectrum such as A 1 , A 2 and A 3 of the illustrative example of 5OA of FIG. 5, is modulated by the light modulator unit 64 to produce a series of pulses such as S 0 (O illustrated in the graph plot 54A of FIG. 5. The light pulses S 0 (O are injected into port 68a of the optical circulator 68. According to the above-described arrangement of the optical circulator 68, the light pulses S 0 (O exit from port 68b. The amount of s o (f) exiting port 68c, determined by the cross-talk specification of the particular implementation of the optical circulator 68, should be very small.

[0057] With continuing reference to FIG. 6, the light pulses S 0 (O travel in the XMIT direction, enter the chirped fiber Bragg grating 70 dispersion element, and the different wavelength components (e.g. A 1 , A 2 and A 3 ) are spread in time and reflected, in the RFLT direction, back to the in/out port 68b of the optical circulator. The reflected, dispersed signal may, for example, be in accordance with the s(t) signal depicted at 54A of FIG. 5. The reflected, dispersed signal s(t) enters the in/out port 68b and, in accordance with the above-described arrangement of the optical circulator 68, exits through the out port 68c. The signal s(t) is then sampled, and information representing, for example, the frequency spectra graphically depicted at 58A is generated.

[0058] FIG. 7 shows a schematic of one example high speed spectrometer 80, employing a dispersion element 82 in a transmission mode, arranged to illuminate and measure an interferometer sensor 84. As will be described, the dispersion element 82 disperses the light reflected from the interferometer sensor 84 to spread

its frequency components in time. The spectrum of the interferometer sensor output is measured at a very fast sample rate.

[0059] Referring to FIG. 7, the example 80 includes a broadband light source 84 outputting a broadband light of spectra 84A, centered at, for illustrative example, nanometers 1500 (nm). The broadband light connects to the in port 86A of an optical coupler or circulator 86. An interferometer sensor 88 connects to the in/out port 86B of the optical circulator 86. Various implementations of the interferometer sensor 88 are contemplated such as, for example, Michelson, Mach-Zehnder, Sagnac, or Fabry-Perot (FP) interferometers. Various kinds of devices and methods to implement broadband light source 84 are contemplated. One example broadband light source 84 is an off-the-shelf superluminescent light emitting diode (SLED). These are available at various center frequencies such as, for example, 1500 nm. This is only one illustrative example. Various other implementations will be apparent to persons skilled in the art upon reading this disclosure.

[0060] With continuing reference to FIG. 7, the out port 86c of the optical circulator 86 connects to a modulator 90. The optical modulator 90 performs an optical gate function such as, for example, the gate function of the optical modulator 14 and associated modulator pulse generator 16 described in reference to FIG. 1. Each light pulse output SP(O of the modulator 90 is a narrow pulse, as depicted, having an envelope such that the frequency spectrum of SP(O is substantially narrower than the spectrum of the interference light returned from the interferometer sensor 88, as described in further detail below. One illustrative example rate of the light pulses SP(O is 1 x 10 6 pulses per second. One illustrative pulse width PW is approximately 6 nanoseconds (ns).

[0061] Referring to FIG. 7, the dispersion element 82 is arranged between the output of modulator 90 and a high-speed light detection unit 92. One illustrative example implementation of the dispersion element 82 is a 5 kilometer (km) dispersion compensation fiber (DCF). The high speed light detection unit 92 may have an optical-to-electrical converter (not separately shown) and an analog-to- digital converter (ADC) (not separately shown). The high speed light detection unit 92 may output digital samples to, for example, a signal processor 94 or equivalent. The signal processor 94 may include various display, user interface and control, and data manipulation functions readily apparent to persons skilled in art upon reading

this disclosure. Alternately, the signal processor may be omitted; it is not required to practice the invention.

[0062] With continuing reference to FIG. 7, the high-speed light detection unit 92 may, for example, include a high-speed data buffer (not shown) to receive high sample rates from its ADC (not shown) in a burst mode and transfer the samples at a lower rate to the signal processor. The design and use of such data buffers is known in the electrical engineering arts. As will be understood from this disclosure various configurations and implementations of the data processor 94 may be selected and/or constructed such as, for example, a general purpose programmable computer (not separately shown) or, alternatively, a single chip programmable microprocessor or controller.

[0063] Alternatively, the high speed light detection unit 92 may include an optical- to-electrical converter (not separately shown) that generates an analog signal corresponding to a detected light and, without ADC conversion, the signal may feed a real-time observation device such as, for example, an oscilloscope. The oscilloscope may be triggered by a signal (not separately shown) output from, or input to the modulator 88, enabling real-time observation of a waveform such as, for example, the dispersed signal 54A shown at FIG. 5.

[0064] Referring to FIG. 7, one illustrative example of one method of measuring an event using the example 80 will be described. For this method, the broadband light source 84 is implemented as an SLED having a center frequency of approximately 1500 nanometers (nm), operated in a continuous (cw) mode. An example spectrum of the broadband light from 84 is shown as 84A. A Fabry-Perot (FP) interferometer or equivalent implements the interferometer 88. The light 84A enters port 86A of the optical circulator, exits port 86B and is injected into the interferometer 88. A light having an interference spectrum Sl(ω) reflects from the interferometer 88 in the "Returned" direction. As known in the art, spectrum Sl(ω) may indicate a state (e.g., pressure, stress, temperature) of the interferometer 88. The reflected interference light Sl(ω) re-enters the port 86b of the optical circulator 86, exits through the out port 86c, to the modulator 90. The modulator 90 modulates the Sl(ω) signal at approximately 1 x 10 6 cycles per second with to generate 1 x 10 6 pulses per second, each pulse having a general shape of SP(f) and a width PW of, for example,

approximately 6 ns. At this width and shape, the frequency bandwidth of each pulse SP(O is narrow compared to the bandwidth of Sl(ω).

[0065] Referring to FIG. 7, the SP(O pulses pass through the dispersive element 82. Each spectral band or component of the reflected signal SP(O propagates through the dispersion element 82 at a different speed, thereby spreading the spectral components in time. At the exit (not separately numbered) of the dispersive element the dispersion produces a signal such as DS(O, which corresponds to the Sl(ω) spectrum received from the interferometer 88. It will be understood that DS(O and Sl(ω) are not necessarily drawn to scale. It will also be understood that the spacing of the peaks and valleys of DS(O ' s based, at least in part, on the total differential dispersion provide by the dispersive element 82. The dispersed signal DS(O ma y be converted to an electrical signal by, for example, the high speed detection unit 92. A user may observe the peaks and valleys of the time domain signal DS(O on an oscilloscope (not show). Alternatively, a processor such as 94 may convert the time domain signal to a spectral plot such as, for example, the plot illustrated by FIG. 2C and by 58A of FIG. 5.

[0066] To reduce the effects of noise, the time domain dispersed signals DS(O may be averaged. For example, averaging DS(O from over 100 pulses provides 1 x 10 4 frames per second of spectral measurement of the interferometer 88. The wavelength (or frequency) resolution depends, in significant part, on the dispersion provided by the dispersion element 82, and on the accuracy (in terms of time and amplitude) in measuring the time domain dispersed signal at the exit of the element 82. An example wavelength resolution, achieved with an implementation according to FIG. 7 (having a dispersion element 82 implemented as a 5 km DCF as described above, and having an EDFA inserted prior to the modulator 88), is approximately 15 nm.

[0067] FIG. 8 shows a variation 100, similar to the FIG. 7 example, with a chirped fiber Bragg dispersive element 102 arranged in a reflective mode, by way of optical circulator 104, as opposed to the transmissive mode of the dispersive element 82 of the FIG. 7 example. The FIG. 8 example 100 may otherwise be according to the FIG. 7 and, accordingly, like functioned blocks are labeled the same. Referring to FIG. 8, one illustrative example of one method of measuring an event using the example 100 will be described. As described in the example method on the FIG. 7

system 80, the broadband light source 84 may be an SLED having a center frequency of approximately 1500 nm operated in a continuous (cw) mode, with Fabry-Perot (FP) interferometer 88. The light enters port 8A of the optical circulator, exits port 86B and is injected into the interferometer 88. Light reflected from the interferometer, having interference spectrum Sl(ω), propagates back in the "Returned" direction, re-enters port 86B of the optical circulator 86, exits through the out port 86c, and into the modulator 90. Modulator 90 modulates the Sl(ω) signal at, for example, approximately 1 x 10 6 cycles per second to generate 1 x 10 6 PS(O pulses per second, each pulse having a general shape as shown in SP(O, with a PW width of, for example, approximately 6 ns. As described above, the width PW and shape of SP(O are such that the frequency bandwidth of each pulse SP(O is narrow compared to the bandwidth of Sl(ω).

[0068] Referring to FIG. 8, the SP(O pulses pass enter port 104A of the circulator 104, exit through port 104B and are injected into the chirped fiber Bragg dispersive element 102. Each wavelength of the pulsed signal SP(O is reflected back from the chirped fiber Bragg dispersive element 102 with a different time delay, thereby spreading the spectral components of SP(t) in time. The dispersed signal has a time domain form DS(O, which corresponds to the Sl(ω) spectrum received from the interferometer 88.

[0069] A frequency resolution of, for example, approximately 10-15 nm is generally sufficient to provide useful measurement of, for example, a Fabry-Perot interferometer sensor. However, Fabry-Perot interferometer sensors have structural shortcomings, rendering them impractical or unsuitable for many environments. Also, Fabry-Perot interferometer sensors are generally expensive. Similar characteristics apply, generally, to Michelson, Mach-Zehnder and Sagnac interferometer sensors.

[0070] A fiber grating sensor, itself, is typically much less expensive than a Fabry- Perot interferometer sensor. However, when used to detect physical quantities such as pressure and strain, a fiber grating sensor typically exhibits a much smaller change in its spectrum than is exhibited by a Fabry-Perot interferometer sensor. Fiber-grating sensors therefore typically require a spectral measurement apparatus having much better frequency resolution than required for measuring a Fabry-Perot interferometer sensor. Stated more specifically, the spectrum width of the reflection

peak generated by a fiber grating may be much smaller than 1nm. Spectrometers or optical spectra analyzers (OSA) such as, for example, CCD-based spectrometers, having such a resolution are available, but are typically very expensive. Therefore, in the related art, fiber grating sensors are often impractical. [0071] As described above, spectral measurement devices according to FIGS. 6 and 7, which convert the frequency spectrum of interferometers (such as the FP interferometer 88) to a time domain signal, instead of detecting the spatial fringe pattern as CCD-based devices, are inherently faster and less expensive than CCD devices. Further, spectral measurement devices according to, for example, FIGS. 6 and 7 with a frequency resolution better than 1 nm are contemplated. Therefore, spectral measurement devices according to FIGS. 6 and 7, which substitute a fiber grating sensor for the Fabry-Perot interferometer 88, are contemplated. [0072] However, a minor addition to FIG. 1 , or of FIGS. 6 or 7 is one alternative, that provides a significantly higher frequency resolution without requiring a substantially longer dispersive element, will be described. The addition is a high speed ADC unit, having a very high sampling rate and a high sampling resolution that much more accurately measures the form of the pulse, e.g., So(t) or DS(O, exiting the dispersion element.

[0073] FIG. 9 shows, in schematic form, an example 200 according to one over- sampling alternative embodiment. The FIG. 9 example is depicted as having certain functional blocks and components common with FIG. 7, which are like numbered accordingly. Referring to FIG. 9, wideband light source 84 connects to the in port 86A of circulator 86. The in/out port 86B of circulator 86 connects via, for example, optical fiber path 202, to a fiber grating sensor 204. The optical fiber path 202 is not necessarily different from the path in FIG. 7 connecting the in/out port 86B of the circulator to the Fabry-Perot interferometer 88. Referring to FIG. 9, the out port 86C of the circulator 86 connects to the optical modulator 206, which is controlled by the modulation pulse generator 208. The output 206A of the optical modulator connects to dispersive fiber 210, or equivalent dispersive element, which connects to an optical-to-electrical (O/E) converter 212. A digital acquisition unit (DAQ) 214 samples the output of the O/E converter with, for example, a high-speed analog-to- digital converter (ADC) (not separately shown). The ADC of the DAQ 214 may have

a sample rate of, for example, 10 Gigasamples/second (GS/s), at a resolution of, for example 6 to 8 bits.

[0074] In the depicted FIG. 9 example, a processor 216 receives the ADC samples from the DAQ 214. The processor 216 provides a computational resource for extracting time and amplitude information from the ADC samples from the DAQ 214 and, for example, converting these to high-resolution spectral data, as described in further detail below. The processor 216 is not required, though, to practice the invention or its aspects. The FIG. 9 example shows control lines (not separately numbered) connecting between the modulating pulse generator 20, the DAQ 214, and the processor 216. The control lines are shown as a functional aspect, which may be implemented in various ways, to provide synchronization of the modulating pulse generator 20, the DAQ 214, and the processor 216. The synchronization may be preferred because, for example, of the high sample rate of the DAQ 214, to avoid collecting a large quantity of samples between pulses. Synchronization may also provide a time reference for edges detected in the time domain waveform sampled by the DAQ 214, as described in greater detail below. The specific synchronization scheme, and physical implementation, depends on the specific construction of the spectrum measurement system according to FIG. 9, and is easily implemented by a person of ordinary skill in the art upon reading this disclosure. [0075] FIG. 10 shows one example variation according to FIG. 9, having an erbium doped fiber amplifiers (EDFA) 220 inserted prior to the optical modulator 206. In the depicted example FIG. 10, the EDFA 220 is shown as controlled by the processor 216. This is optional, as the EDFA 220 may have a fixed gain or, for example, may be directly manually controllable.

[0076] One method for high-resolution spectral measurement will be described in reference to FIG. 9, and to graphical illustrations of the oversampling shown in FIGS. 11 , 12A, 12B, 13A and 13B. The method is generally identical to the example method described in reference to FIG. 7, with an additional high-speed oversampling to provide much greater spectral resolution.

[0077] Referring to FIG. 9, the broadband light source 84 is as described in reference to FIG. 7, operated in a continuous (cw) mode. The light (e.g., 84A of FIG. 7) enters port 86A of the optical circulator, exits port 86B, propagates through the optical connection fiber 202, and is injected into fiber grating sensor 204. A light

having an interference spectrum Sl'(ω) reflects back from the fiber grating sensor 204. Because of using the fiber grating sensor 204 instead of the Fabry-Perot interferometer, though, the bandwidth of Sl'(ω) is much narrower than the bandwidth Sl(cϋ) described in reference to FIG. 7. The reflected signal having the Sl'(ω) spectrum re-enters the port 86b of the optical circulator 86, exits through the out port 86c, and enters the modulator 206. The modulator 206 and modulation pulse generator 208 may be in general accordance with the modulator 90 of FIG. 7. For this example, modulator 90 will modulates the reflection signal from the fiber grating sensor 204 at approximately 1 x 10 6 cycles per second, to generate 1 x 10 6 pulses per second, each pulse SP'(f) having a general shape and width of, for example SP(O and PW described in reference to FIG. 7. Regardless, in selecting the implementation of the optical modulator 206 and its associated modulation pulse generator 208, the following criterion must be maintained: the frequency bandwidth of each pulse SP'(f) must be narrow compared to the bandwidth of Sl'(ω) reflected from the fiber grating sensor 204, so that the frequency spectrum of SP'(f) is approximately the same as of Sl'(ω).

[0078] Referring to FIGS. 9 and 10, the SP'(f) pulses from the optical modulator 206 pass through the dispersive element 210 which, for this example, is a dispersive fiber. Each wavelength component of the reflected signal SP'(f) propagates through the dispersive fiber 210 at a different speed, thereby spreading the spectral components in time. At the exit (not separately numbered) of the dispersive fiber 210, the dispersion produces a signal DS'(t). DS'(t) corresponds in form to the Sl'(ω) spectrum reflected from the fiber grating sensor 204. It may be assumed, as one kind of example, that the dispersive fiber 210 is constructed to provide a total dispersion comparable to the total dispersion provided by the fiber 82 of FIG. 7. Therefore, DS'(0 will be appear as a shortened version of DS(t) in FIG. 7, because the bandwidth Sl'(ω) of the reflection from the fiber grating sensor 204 is considerably less than the bandwidth Sl(ω) of the reflection from the Fabry-Perot interferometer 88. As described in further detail below, DAQ 214 oversamples the DS'(0 signal to extract the peaks and valleys to a much higher resolution, and, accordingly, measure Sl'(ω) sufficient to identify changes of interest in the fiber grating sensor 204.

[0079] Referring to FIGS. 7 and 11 , in the FIG. 7 example the signal processor 94 samples the electronic signal from the high-speed light detection circuit 92 at a speed sufficient enough to resolve the valleys and peaks in the DS(O signal. According to standard sampling theory, the sample rate is determined by the highest frequency component DS(O- Graph 300 of FIG. 11 depicts a typical fringe from a Fabry-Perot interferometer, such as Sl(ω) from the interferometer 88 of FIG. 7, and graph 302 depicts a corresponding time domain dispersed signal, such as DS(O °f FIG. 7. This corresponds, for example, to Sl(ω) of FIG. 7. Samples 306 of the 302 signal are at a rate SR, which can be seen, is more than twice the frequency of the peaks and valleys. Graph 304 shows sampling of the dispersed signal at a rate OSR, that is many times greater than SR.

[0080] Referring to FIGS. 9 and 10, the DAQ 214 samples the dispersed reflection signal DS'(0 at, for example, 10 GS/s. The high sampling enables detection of peaks and valleys in DS'(0 sufficient to resolve down to, for example, approximately 0.1 nm. This is performed as follows. FIG. 12A illustrates an actual spectrum from a fiber grating sensor such as 102, and FIG. 12A illustrates a corresponding spectral measurement shown by the dispersed signal DS'(0- As seen, a spectrometer according to the FIG. 9 example may not resolve the width of grating signal FIG. 12A because its spectral resolution may be lower. However, the center of FIG. 12A and FIG. 12B peaks are the same. Further, when the center of the FIG. 12A spectrum moves then the center position of the FIG. 12B will move identically. The high sampling date of the DAQ 214 enables accurate estimation of the center of the FIG. 12B peak, using conventional, straightforward curve analysis techniques. Therefore, the high sampling performed of the DS'(0 dispersed signal by DAQ 214 or an equivalent enables accurate enough detection of the peaks and valleys in DS'(t) to sufficiently measure the spectrum of the fiber grating sensor. Based on the present disclosure, a spectral measurement apparatus according to FIGS. 9 or 10 can be readily practiced, resolution better than approximately 0.1 nm. [0081] Referring to FIGS. 13A through 13C, oversampling by the DAQ 214 arranged according to FIGS. 9 and 10 provides, in addition to high spectral positioning resolution, amplitude information of the signal DS'(0 for even better spectral positioning resolution. Referring to FIG. 13A, the position of a peak or a pulse in the dispersed signal DS'(0 ma y °e measured by detecting, or estimating the

edge of the peak, such as edges 402 and 404. For ease of explanation, edges 402 and 404 are assumed to be straight lines. This description can be readily adapted to jagged or curved lines, using conventional edge estimation and analysis techniques. [0082] Referring to FIG. 13A, the vertical axis U is shown with a one-bit resolution. Assuming a threshold-based edge detection (anywhere between "0" and "1") the sample rate SRX in the time domain therefore directly determines the positioning resolution. Referring to FIG. 13B, by increasing the vertical resolution to two bits, or four levels, it is seen that the positioning resolution is four times better than FIG. 13A, at the same sample rate SRX. Referring to FIG. 13C, if the vertical resolution is kept at two bits, and the sample rate doubled to 2SRX, the positioning resolution may not necessarily be significantly improved, but the effects of noise on the positioning accuracy may be reduced.

[0083] Referring to FIGS. 9 and 10, it is contemplated that a DAQ 214 vertical axis resolution may be approximately eight or more bits. Since the typical signal amplitude (e.g., the amplitude of DS'(f) exiting the dispersive element 210) only fills half of the full scale of the DAQ 214, the effective vertical axis resolution is will be four or more bits. If the total amount of dispersion of the dispersive fiber 214 is, for example, 1000ps/nm at 1550 nm, to achieve a spectral positioning resolution of 0.01nm, a 10 Gs/s sample rate is required. Various implementations of DAQ 214 meeting this example requirement will be apparent to persons skilled in the arts, upon reading this disclosure, such as, for example, various off-the-shelf data acquisition devices, available from various commercial vendors. [0084] While certain embodiments and features of the invention have been illustrated and described herein, many modifications, substitutions, changes, and equivalents will occur to those of ordinary skill in the art. For example, FIGS. 9 and 10 show a dispersive fiber 210 connected in a transmission arrangement, between the modulator 206 and the O/E converter 212. One alternative is to connect a dispersive fiber, or equivalent, in a reflective arrangement, using a circulator such as shown in the FIG. 8 example. It is therefore to be understood that the appended claims are intended to cover all such modifications and changes as fall within the spirit of the invention.