National Academies Press: OpenBook

Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing (2014)

Chapter: 5 Fundamental and Engineering Limits of Active Electro-Optical Sensing

« Previous: 4 Active Electro-Optical Component Technologies
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

5

Fundamental and Engineering Limits of Active Electro-Optical Sensing

ILLUMINATION SOURCES

There are a small number of what might be called truly “fundamental” limits to the performance of sources, and these involve the basic physics concept of energy conservation. While physics is involved in all other types of limits, the concepts here are more in the engineering category, involving, say, properties of the materials involved and thermal management of the devices. As such, these are not hard “limits” and may change with further development efforts. Some have resisted the effort that has gone into overcoming them, such as the lack of suitable semiconductor materials to provide laser operation over a good portion of the visible wavelength range or the lack of nonlinear materials with high transparency in the ultraviolet to allow generation of power in that region. Others, such as electrical conversion efficiencies for diode lasers and limits to the (average and peak) power outputs for solid-state lasers are undergoing continuing improvement. Appendix C contains several tables that are a snapshot in time for many of these “soft limits” (Tables C-5 to C-12), summarized below.

Solid-State Lasers

Bulk Format

Solid-state lasers require optical pumping; the most efficient method would be to use a narrow-line source at a suitable absorption frequency larger than the laser’s, but as close as possible. The ratio of the energy of the laser photon to the pump photon is called the quantum defect. The theoretical efficiency for the 1,064 nm Nd:YAG laser line pumped by a laser diode at 808 nm is 76 percent. The theoretical best efficiency of the laser-pumped laser is then the product of the quantum defect and the pump laser efficiency. Table C-5 shows that pump diodes with 60 percent efficiency at 808 nm give a fundamental limit to the electrical efficiency for this Nd:YAG laser of 46 percent. In reality, such lasers demonstrate 20-25 percent efficiency, only half of the fundamental limit due to multiple engineering factors.

The maximum power out of solid-state lasers in the 1,000 nm wavelength region seems to have an engineering limit at the present time of ~ 2 kW in a single rod, 10 kW in a single disk, and 15 kW in a single slab. Higher powers are possible with multiple active media; the present record is >100 kW. Solid-state lasers at longer wavelengths are limited by multiphonon decay, but low-phonon hosts tend to be impractical for high-power operation.

The fundamental Schalow-Townes linewidth for lasers is only a few hertz and can be reached by active external stabilization. Otherwise, fluctuations in optical cavity length cause technical noise on the order of kilohertz owing to coupling between the pump power and gain medium refractive index, acoustic noise, and other cavity perturbations. If the laser is not stabilized, slow drifts of 10-50 MHz occur as a result of changing environmental conditions.

Mode-locked lasers have pulses that are given by the inverse bandwidth of the gain, with a repetition time equal to the round-trip time of the cavity. For Ti:sapphire, the theoretical pulse width is about 3.5 fs; experimentally the pulse width is about 25 percent larger because of dispersion in the optical

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

cavity, the finite spectral response of mirrors, and nonlinear effects. The calculated pulsewidth for the mode-locked, eye-safe Cr:ZnSe laser at 2,500 nm is 7 fs, while experiments report much longer pulsewidths (60 fs). Nd:YAG at 1,064 nm has a fundamental mode-locked pulsewidth of 2 ps.

Fiber Format

The quantum defect rate can be as low as 5 percent for a 1,030 nm Yb:fiber laser pumped at 976 nm. Practical limits to efficiency of diode-pumped fiber laser are due to poor spatial overlap of pumping and lasing regions in the core, incomplete absorption of the pump, losses in the laser material, excited state absorption of the pump or laser power, and upconversion from the upper laser level. In practice, 88 percent slope efficiency has been reported for Yb:fiber lasers. When the pump diodes have 65 percent efficiency, the Yb:fiber laser should have a fundamental limit of 62 percent electrical efficiency, but in reality it has an efficiency of 49 percent, mostly because pump light is lost in transport from diode facet to pump cladding.

Diode-pumped single fiber lasers have the ability to produce very high average powers in the 1000 nm wavelength region, limited mostly by stimulated Raman emission. A 1,030 nm Yb:fiber laser pumped by multiple 1,018 nm Yb:fiber lasers has produced 20 kW. In the eye-safe 2000 nm wavelength region, a single Tm-fiber laser can produce 1 kW when pumped by 790 nm diode lasers. A single-frequency, single fiber in the 1,000 nm region can put out 100 W with 5 kHz linewidth compared to 1 kW with 3 GHz linewidth. In both cases stimulated Brillouin scattering had to be overcome; the former used a thermal gradient along the length of the fiber; the latter used a frequency modulated source. In the eye-safe 2000 nm region, the average power in a single frequency from a single Tm-fiber laser is 600 W in a single frequency with < 5 MHz linewidth.

Q-switching a straight Yb:silica rod-type fiber can give 60 ns pulses of 0.45 MW peak power at 1030 nm (27 mJ) or a shorter pulse <1 ns with a peak power of 4.5 MW (4.3 mJ). Both regimes are limited by stimulated Raman scattering and optical breakdown. Flexible fibers typically generate less peak power: 25 kW in 400 ns pulses. The highest peak power comes from mode-locked pulses: Yb:silica laser pulses can be only 0.5 ps long, with a peak power from a single fiber of 3.8 GW (2.2 mJ). Such short pulses would be limited by self-phase modulation, as well as the mechanisms that affect Q-switched pulses, so to achieve this high power while avoiding nonlinear effects, the rod-type fiber system employs chirped-pulse amplification (CPA). Flexible fibers with CPA operate up to around 100 MW of peak power.

Nonlinear Optics

Harmonic generation can, in principle, be 100 percent effective for plane waves. In practice, many factors can reduce this efficiency: dephasing from beams of finite width; dephasing as a consequence of heating in the crystal from background absorption or multiphoton absorption and creation of color centers; losses at the entrance and exit faces of crystal; crystal or quasi-phase matching (QPM) material imperfections; and optical damage at material surfaces or in the bulk. All of these factors can limit second harmonic generation to 60-70 percent for Gaussian beams, though it is up to 90 percent for the high-energy, flat-top beams of the National Ignition Facility. Maximum third harmonic from Gaussian beams, theoretically, is 50 percent. The maximum amount of power generated is limited by the same phenomena that limit the efficiency. Lithium borate (LBO) has the lowest absorption of common nonlinear materials, and the powers are ultimately limited by damage in the coatings. Multiple hundreds of watts of second harmonic can be created from Nd- or Yb-doped lasers. The shortest wavelength achievable by a second harmonic is 176 nm, determined by phase-matching limits and vanishing nonlinear coefficients at short wavelength. This wavelength can be reached with KBBF crystals; the more readily available BBO crystals generate second harmonic out to about 205 nm.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

Optical parametric oscillators OPOs), in principle, have a fundamental limit for overall conversion efficiency from pump to signal plus idler of 100 percent. The OPO is, in principle, a lossless process. It breaks up one pump photon of energy hnp into two photons, hns and hni, where hnp = hns + hni. The maximum efficiency of conversion from the pump to the signal is hns/hnp. Similarly the maximum efficiency of conversion from the pump to the idler is hni/hnp. These are called the Manly-Rowe relations and assume plane waves. They show that the sum of the power in the signal plus idler should equal the pump power if there are no extraneous losses. The actual efficiency can be 90 percent for continuous wave (CW) OPOs, while 50 percent is more typical for pulsed lasers. For pulsed OPOs, the finite buildup time of the parametric process in the cavity leads to further loss in efficiency. Often loss in conversion efficiency results from some light not being converted away from the pump wavelength.

While theoretically there is no limit to the power than can be handled by an OPO, the actual power is limited by the same processes that limit the efficiency of second harmonic generation. Typical OPO outputs are presently below 100 W but have been limited primarily by the power available in pump lasers. Hundreds of watts should be possible in the near-IR with materials like LBO.

DETECTORS

The smallest light packets, photons, carry information that detectors attempt to extract. How much information carried is an active research topic.

The photon, the indivisible unit of electromagnetic energy, is a fundamental carrier of information. Numerous degrees of freedom are available for the conveyance of information on a photon including frequency, phase, arrival time, polarization, orbital angular momentum, linear momentum, superposition states, correlation, entanglement, etc.1

Active electromagnetic applications such as ladar extract more information than intensity from photons by including arrival time, wavelength (frequency) and sometimes polarization and phase in the classical regime.

Detector Fundamental Limits

Important parameters that characterize the fundamental properties of detectors are discussed next.

• Responsivity (Amperes/Watt). Responsivity is the ratio of the generated photocurrent to incident light power. The ideal would be for all the power to produce electron-hole pairs and for none to be wasted as heat or through other loss mechanisms. For detectors with gain, this parameter can be large, but gain also introduces excess noise (see discussion of gain and excess noise in Chapter 4).

• Signal to Noise (SNR). SNR is the ratio of the signal current to the root-mean-square average of the noise current. There is no ideal limit, but higher is better.

• Specific Detectivity (D*). Specific detectivity is used to compare detectors with different active areas and electronic bandwidths. It is defined as

image

where Ad is the detector area, Flux is the incident photon power, and Δf is the bandwidth. This expression assumes the noise is constant white noise across Δf. Some detectors approach ideal performance when the

____________________

1 See http://www.darpa.mil/Our_Work/DSO/Programs/Information_in_a_Photon.aspx.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

wavelength of the incident photons nears the detector cutoff. The ideal is when the photons have just enough energy to create carriers by exciting them over the bandgap and there is no lost excess energy.

• Quantum Efficiency (percent). For detectors without gain, quantum efficiency (QE) is the ratio of the number of carriers to the number of incident photons, with 100 percent being ideal. For detectors with gain, the ratio can be large and is usually limited by voltage controls.

• Dark Current (electrons/second). Dark current is the current in the absence of light when the detector is operated in photoconductive mode. Dark current has unavoidable shot noise. For semiconductors, dark current decreases as bandgap increases. The ideal is zero dark current, but that only occurs at 0 K.

• Response time. Photon absorption in detectors produces carriers (electron-hole pairs) that generate a current under influence of electric fields. The sum of the electron and hole transit times determines the minimum response time. The detector resistance, Rd, and capacitance, Cd, add to the response time, T = RdCd. External circuitry adds additional RC delay in the response time. The response time lengthens the impulse response of the detector. The ideal response time for a detector with zero resistance is the detector width divided by the speed of the carrier’s ballistic transport velocity, which is the speed of carrier transport without scattering and limited only by the carrier effective mass, a function of the band structure.

To highlight important detector fundamental parameters, a simplified analysis of passive detectors without detector gain is presented below. The model shows the limiting effects of several detector parameters but, as noted later, even with these constraints, systems with great utility are available. Better detectors will expand the EO system application options, but improved detectors are only one factor enabling improved systems.

To identify an object, a detector must distinguish the object’s signal from the noise that comes from multiple sources. A key measure of performance is therefore the detector SNR. SNR is the ratio of object signal to the uncertainty of all the other contributions to the signal measurement. Higher SNR is better. The desire for better SNR drives much of the development of electro-active imaging systems, components, and signal processing. The goal of this analysis is to elucidate the pure characteristics of detectors without the many application- and system-specific factors such as atmospheric transmission, collection apertures, external readout noise, scene spatial noise and clutter, and like factors that are important for application success but not directly related to detector performance. At the end of this analysis is an example of a complex model that includes many application-specific parameters.2 The more realistic complex model predicts performance (SNR) for the various combinations of detector focal plane arrays (FPAs), optics, and spectral target and scene characteristics. The model shows that detectors are only one of many aspects of overall system performance.

Signal

Detector signal is produced from the incident flux, F, (J/s) of photons multiplied by the photon to electron QE and integrated over time t.3 Quantitatively, the signal S is

image

where hν is the energy of the photon (Planck’s constant × frequency) and q is the electron charge. In addition to signal, there also is corrupting noise signal.

____________________

2 See Figure 5-4.

3 Not included are factors such as system aperture size and atmospheric transmission that affect system performance but are not exclusively related to the detector.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

Noise Sources4

Incident Flux Shot Noise

Usually photons are random in nature,5 described by Poisson statistics, and have arrival variations called shot noise. This noise in not related to any flaws in a detector and is unavoidable. Detector signal noise can never be less than flux shot noise, but some detectors operate near the shot noise limit6 (see Figure 5-2), and there are ways to operate detectors to reduce other noise for the photon sources.7 The performance for three commercial detectors compared to the ideal shot noise limit is seen in Figure 5-1, where it is apparent that the performance advantage produced by the high-gain Geiger-mode avalanche photodiode (GM-APD) detector8 approaches the shot noise limit over much of the irradiance range, while the intrinsic noise of the detector degrades SNR for low light levels.

image

FIGURE 5-1 Curves demonstrating the theoretical sensitivity of an ideal shot-noise-limited 50 µm pitch, 100 percent QE sensor (dashed line); a 50 µm pitch, GM-APD sensor; a 6 × 6 aggregated 8 µm pitch Kodak KA4021 sensor; and a 6 × 6 aggregated 8 µm pitch Fairchild 128-stage time delay and integration (TDI) sensor. SOURCE: Michael Gartley. Detectors, RIT Course Number 1051-465 Lecture Noise. Available at http://ridl.cfd.rit.edu/products/training/Detectors%20465%2020083/Lecture/Lecture%2011-Noise/Lecture%20Noise.pdf.

____________________

4 More detail on the following and additional noise sources is found in National Research Council, 2010, Seeing Photons: Progress and Limits of Visible and Infrared Sensor Arrays, National Academies Press, pages 33-36.

5 True for uncorrelated light photons from a source such as incandescence but with coherent light and proper selection, the light’s amplitude noise can be reduced by transfer or “squeezing” fluctuations into the light phase resulting in subclassical shot noise light.

6 Since for Poisson statistics the variance is equal to the mean, shot noise limited means the detector noise must be less than the shot noise = image where N is the number of incident photons.

7 Examples include processing procedures such as balanced mode detection that reduces laser noise by splitting the received signal onto two detectors and rejecting any imbalance between the photocurrents generated by the reference and signal detectors (http://www.newport.com/New-Focus-Application-Note14-A-Survey-of-Methods-/919636/1033/content.aspx).

8 See discussion of Geiger mode below and in Chapter 4.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

Dark Current Noise

In semiconductor detectors, additional noise signal that comes from thermally generated electrons called dark current. It appears even when there are no impinging photons. The dark current is from electrons interacting with thermally excited crystal vibrations called phonons. The excited electrons that exceed the semiconductor bandgap energy produce unwanted dark current.

Dark current is like shot noise and cannot be discriminated from incident photon shot noise or other shot noise sources. The dark current shot noise is dependent on bandwidth (BW) and detector area (A):

image

Another detrimental effect of dark current is that it can accumulate, reducing the detector charge storage capacity, and thereby limiting detector dynamic range and integration time. Most detector applications reduce detector dark current by cooling.

Other Noise Sources

The IR background light flux Fback, which depends on scene and target temperature, produces scattered photons that also contribute to detector shot noise.

image

The background noise contribution is also not directly detector related but can be significant for applications viewing dim targets. The readout of detectors adds several fundamental noise components such as Johnson, kTC (capacitor reset noise), persistence, and 1/f noise.9 Some components of this noise are detector-related and others are system- or application-related.

Johnson noise arises from the random motion of electron flow that results from scattering, from electron phonon interactions, and motion induced by thermal interaction; this adds a shot noise component to the readout noise.

image

where R is the detector resistance. For low-resistance detectors like HgCdTe and low-impedance readout integrated circuits (ROICs), Johnson noise is not a significant detector limit except at the very lowest signals.

Readout circuits and detectors both have capacitance, and during detector readout there is generally an associated capacitor reset noise, called kTC. The noise is characterized by

image

____________________

9 There are other contributions to noise in real systems that are not “fundamental.” Examples include pixel-to-pixel charge diffusion, electronic crosstalk, radiation interference from other components, and unstable power supplies. An operational, not fundamental, definition of readout noise, Nread, includes all noise sources. One way to measure this noise is to derive the standard deviation of multiple detector reads taken with minimum exposure time and under dark conditions at operating temperature.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

image

FIGURE 5-2 SNR gain is limited after about a dozen reads in MCDS. SOURCE: Massimo Robberto. Detectors: RIT Course Number 1051-465 Lecture Noise, available at http://ridl.cfd.rit.edu/products/training/Detectors%20465%2020083/Lecture/Lecture%2011-Noise/Lecture%20Noise.pdf.

where C is the capacitance and q the electronic charge. Within limits (see Figure 5-2), this noise can be reduced by using multiple correlated double sampling (MCDS) data processing before the signal is digitized.10

Persistence is noise signal caused by the slow release of trapped charge after the detector is read out. Impurity or defects cause charge traps to decay exponentially with time, so unless traps are eliminated or frozen by cooling there is persistent long tail of thermally activated charge, adding noise and slowing response time. This noise source and excess noise are also important in detectors with gain, as discussed in Chapter 4.

The principal sources of 1.f noise in electronic devices are almost invariably the slow fluctuations of properties of the condensed-matter materials of the devices. In many cases the specific sources of the fluctuations are known. These include fluctuating configurations of defects in metals, fluctuating occupancies of traps in semiconductors … 11

In general, 1/f noise is technology-dependent and the theory of 1/f noise12,13 is much disputed and remains a science and engineering mystery.14 In certain materials such as HgCdTe, however, detailed models seem to well describe the contribution of 1/f noise from defects and surface states.15 1/f and

____________________

10 For detectors that can be read without destroying the accumulated signal charge, MCDS and a related technique called Fowler sampling use several nondestructive reads from the detector while it is integrating signal. If N is the number of reads, by averaging the N reads the read noise is reduced by 1/image. MCDS processing decreases the system response time and adds to the system processing overhead and is therefore commonly used in astronomical imaging, where the image does not change and has a very low intensity, so that long integration times are unavoidable.

11http://en.wikipedia.org/wiki/Pink_noise#cite_note-Kogan-1996-2.

12 Sh. Kogan, 1996, Electronic Noise and Fluctuations in Solids, Cambridge University Press, ISBN 0-521-46034-4.

13 M. B. Weissman, 1988, "1/ƒ Noise and other slow non-exponential kinetics in condensed matter," Reviews of Modern Physics 60(2): 537.

14http://www.scholarpedia.org/article/1/f_noise.

15 M.A. Kinch, R.L. Strong, and C.A. Schade, 2013, “1/f Noise in HgCdTe Focal Plane Arrays,” J. Electronic Materials 42(11): 3243.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

Johnson noise are generally more important in thermal detectors than in photodetectors. 1/f noise is only important at lower frequencies.

Lumping all these readout noise sources into a single readout noise parameter, Nread, allows an estimate of the detector SNR that includes the signal and shot noise from the incident flux, background flux, Fback, dark current and total readout noise, and the effect of integration time, t:

image

where the highlights show the object signal (blue) and the background signal (red) while the dark current noise is green and the total readout noise is black.

The SNR is improved by increasing the numerator compared with the denominator. Fundamental detector parameters that can be improved are the QEv, idark and components of N2read such as 1/f, kTC, persistence, and Johnson noise from the detector. Other parameters such as t and the component of kTC introduced by external readout circuitry are controlled by the system design and engineering practices.

As seen in the equation above, the signal increases with exposure time, t. Noise also increases with time, but generally not by as much. Within limits, SNR improves with integration time as the dominant noise term changes from readout to shot noise (see Figure 5-3).

image

FIGURE 5-3 Signal with a minimum noise at t = 0 increases compared with noise as integration time increases, but the increase slows as the overall SNR becomes flux-, object-, and background-noise dominated. SOURCE: Don Figer, Detectors: RIT Course Number 1051-465 Lecture Noise, available at http://ridl.cfd.rit.edu/products/training/Detectors%20465%2020083/Lecture/Lecture%2011-Noise/Lecture%20Noise.pdf.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

image

FIGURE 5-4 SNCR model for missile detection that includes many factors in addition to the detector performance. SOURCE: R. Richwine, A.K. Sood, R.S. Balcerak, and K. Freyvogel, 2007, “EO/IR sensor model for evaluating multispectral imaging system performance,” Proc. SPIE 6543, Infrared Imaging Systems: Design, Analysis, Modeling, and Testing XVIII, 65430W, April 30.

In practical applications of detectors there is an inherent interplay between the fundamental, operational, and system parameters. For example, obscurant can be often be better penetrated by using longer wavelengths, holographic processing16 and range gated processing. Range gating may improve as the fundamental detector parameter response time improves. However, often there are other practical limits such as the RC time constant of the detector’s external readout circuitry, which overwhelms the fundamental detector performance. As another example, the background flux is greatly reduced by being in space or cooling the aperture, but these options are not often operationally practical. For illustration of the complexity of noise estimation in real applications, a system performance model is provided (Figure 5-4) that has many of the important detector and nondetector factors that determine the real-world application signal to noise and clutter ratio (SNCR) for missile detection:17

Fundamental detector properties are important constraints on applications but for practical systems there are many other factors that require time to manufacture, engineering attention, operational constraints, and—finally—the funds necessary to build, train, and use the system.

Detector Gain

The SNR estimate assumes the detector signal responds to incident photon linearly and without gain, but detectors can be very nonlinear when they have intrinsic gain from operating in the linear avalanche or Geiger mode. Avalanche gain comes with the cost of excess signal noise but with the advantage that the gain can overcome readout and other system noise sources. See the subsection “Comparison of Linear- and Geiger-mode Systems” in Chapter 2 and the subsection “Comparison of Linear-vs. Geiger-Mode APDs” in Chapter 4.

____________________

16 M. Locatelli, E. Pugliese, M. Paturzo, V. Bianco, A. Finizio, A. Pelagotti, P. Poggi, L. Miccio, R. Meucci, and P. Ferraro, 2013,” Imaging live humans through smoke and flames using far-infrared digital holography,” Optics Express 21(5): 5379.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

TABLE 5-1 Infrared Detector Parameter Limits with Comments on the Major Implications

Parameter Limit State of the art Implications
Dark current at 140 K, HgCdTe arrays, 2.5 µm cutoff 10-3 electrons/s/pixela 0.01 e/sec/pixel at 90 K Improves SNR in low background
kTC noise 0 electrons 3 electrons/64 reads Reduces SNR
Johnson noise 0 electrons Few electrons/s Important mostly for thermal detectors
1/f noise Generally a theoretical mystery but well understood in particular casesb Few electrons/s Important for low frequencies
Quantum efficiency 100 percent 70-80 percent Improves SNR

NOTE: See also Table 4-3, “Performance Parameters Comparison of Linear-, Geiger-, and Coherent-Mode Detectors” for more details on state-of-the-art APDs.

a J.W. Beletic, R. Blank, D. Gulbransen, D. Lee, M. Loose, E. C. Piquette, T. Sprafke, W. E. Tennant, M. Zandian, and J. Zino, 2008, “Teledyne imaging sensors: infrared imaging technologies for astronomy and civil space,” Proceedings of the SPIE Conference on Astronomical Instrumentation (Marseille, France).

b M.A. Kinch et al., op. cit.

Fundamental Parameters

The key fundamental detector characteristics, their theoretical limits, and estimates of the current state of the art for selected detectors are summarized in Table 5-1, along with comments on implications of achieving those limits.

Key Detector Technologies and Trends

Materials and Processing Trends

Since detectors are not perfect crystals with uniform boundaries and contacts, material and fabrication improvements tend to reduce noise and minimize signal loss. For example, a serious source of semiconductor detector dark current noise is current generated from bulk and surface states that have lower energy than the bandgap.18 There is some evidence that surface states may also contribute to 1/f noise. There are always efforts to fabricate detectors with lower impurities, fewer crystal defects, better contacts, and cleaner surfaces. Detectors are grown on lattice matched substrates that reduce the crystal strain physical imperfections. Lattice matched atomic layers with high purity materials and precise chemical compositions are formed with techniques such as HgCdTe grown by metal organic vapor phase epitaxy (MOVPE) on GaAs substrates. Special care is sometimes taken to anneal the detectors and add the treatment of hydrogen overpressure that is thought to create passivizing interstitial hydrogen. Many processing steps are proprietary and not published in detail but they reduce defects or deactivate them so the result is less thermally generated noise, better response time and lower signal losses. Better material perfection, uniformity and purity is a continuous quest.

____________________

18 Semiconductor dark current from electrons excited over the bandgap is unavoidable, but materials with a larger bandgap have smaller dark current at the cost of requiring higher energy (smaller wavelength) photons for sensing. There are, however, some materials, such as graphene, that have unusually small coupling of the crystal vibrations (phonons) to electrons, and these materials promise much smaller dark currents at ambient temperature.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

image

FIGURE 5-5 FPA variation within each chip and on the wafer. The pixel variations in dark count rate (DCR) are shown on first FPA and in photo detection efficiency (PDE) on the second FPA. Variations are shown by the color changes. The epitaxial growth process generally makes radial variations in DCR and PDE. SOURCE: Mark Itzler, Princeton Lightwave, presentation to the committee on Jan 30, 2013. Approved for public release, unlimited distribution.

Processing large focal plane arrays is more complicated than making single detectors. The material and processing improvements are especially difficult where precise atomic stoichiometry and layer thickness control is needed for a few tens of micron-size detector pixel over the relatively vast area of the multimillimeter size chips and even larger multicentimeter-sized production wafers. Achieving near-fundamental detector performance requires extremely precise and accurate process control over large areas. Typical state-of-the-art fabrication variations for InGaAsP/InP Geiger-mode avalanche photodiodes (GmAPDs) with high-efficiency single photon sensitivity at 1.06 µm are seen in Figure 5-5.

Pixel Size

Even if the specific detectivity D* does not improve, array pixel size is shrinking and will continue to do so because of technological advances. This trend is seen in Figure 5-6.

Smaller pixel size allows smaller system SWaP because smaller diffraction-limited optics and detector size reduction means the load on the cooling subsystem is reduced. There is some benefit from pixels even smaller than the incident wavelength from oversampling, but this benefit increases processing overhead.

Superlattice Detectors

The properties of superlattice detectors are being improved in several ways in attempts to achieve fundamental performance limits. Because the energy band structure in an n-type GaAs/AlGaAs quantum well infrared photodetector (QWIP) enforces directional photo absorption selection rules, normal incident radiation absorption is not possible.19 Optical structures that enhance normal incident absorption are seen in Figure 5-7.

____________________

19 Y. Fu1, M. Willander, W. Lu, and Wenlan Xu, 1998, “Optical coupling in quantum well infrared photodetector by diffraction grating,” J. Appl. Phys. 84: 5750.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

image

FIGURE 5-6 Pixel pitch for HgCdTe photodiodes (left) and amorphous silicon micro-bolometers (right) have continued to decrease due to technological advancements. SOURCE: Left, G. Destefanis, P. Tribolet, M. Vuillermet, and D.B. Lanfrey, 2011, “MCT IR detectors in France,” Proc. SPIE 8012: 801235-1; Right, A. Hoffman, 2006, “Semiconductor processing technology improves resolution of infrared arrays,” Laser Focus World, February, 81.

image

FIGURE 5-7 Grating light-coupling mechanisms used in QWIPs: (a) gratings with optical cavity, (b) random scattered reflector, and (c) corrugated quantum wells. SOURCE: A. Rogalski, 2012, “Progress in focal plane array technologies,” Progress in Quantum Electronics 36: 342, with permission from Elsevier.

QWIPs (n-type) have a gain of <1. The lower gain on QWIP detectors must be compensated by higher gain ROICs, which add to the readout noise. Better ROICs and improved quantum efficiency are recognized as important developments for progress in QWIPs.20

Quantum Dots

Researchers are pursuing quantum dot infrared photodetectors (QDIPs) because in 2002 a theoretical analysis predicted that “quantum dots are expected to have the potential to outperform quantum wells by several orders of magnitude and compete with HgCdTe.”21 QDIPs use a distribution of dots of material for photon absorption and carrier production, while QWIPs use a thin heterogeneous semiconductor for the same purpose.

____________________

20 Yetkin Arslan, Tahir Colakoglu, and Cengiz Besikci, 2013, “Diffraction-grating-coupled high quantum efficiency InP/InGaAs quantum well infrared photodetector focal plane array,” IEEE Journal Of Quantum Electronics, 49(2): February.

21 J. Phillips, 2002, “Evaluation of the fundamental properties of quantum dot infrared detectors,” Journal of Applied Physics 91(7): April.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

image

FIGURE 5-8 HgCdTe detector spectral range extended by removal of substrate.

SOURCE: J.W. Beletec, R. Blank, et al., 2008, “Teledyne imaging sensors: infrared imaging technologies for astronomy and civil space,” Proceedings of the SPIE, Conference on Astronomical Instrumentation (Marseille, France), Courtesy of Teledyne Imaging Sensors.

Currently, quantum dot infrared photodetectors have problems related to dot size and uniformity.22 QDIPs are promising, but there are as yet no commercial or military applications.23

HgCdTe Versus Other Materials

For the next decade, HgCdTe technology will likely continue to be the IR detector material of choice for MWIR and LWIR applications because of its excellent properties and its increasing manufacturability. Despite “competition from alternative technologies, HgCdTe is unlikely to be seriously challenged for high-performance applications—those requiring multispectral capability and fast response”24. HgCdTe arrays are sensitive over a broad spectral range and can incorporate low-noise gain, reducing the demands on ROICs. As seen in Figure 5-8, HgCdTe detectors without a substrate show good quantum efficiency into the near IR, and there are plans to extend this spectral range into the visible and, possibly, the ultra-violet.

However, the chemical stability of HgCdTe is not perfect, and nonuniformity over large areas is a challenge (see Figure 5-8). There is a steep learning curve for those seeking to fabricate high-uniformity HgCdTe focal plane arrays.

____________________

22 P. Bhattacharya, S. Ghosh, and A. D. Stiff-Roberts, 2004, “Quantum dot optoelectronic devices,” Ann. Rev. Mater. Res., 34: 1, August.

23 See additional discussion of QDs in Chapter 4.

24 A. Rogalski, 2012, “Progress in focal plane array technologies,” Progress in Quantum Electronics 36(2): 342-473.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

“Type II InAs/GaInSb superlattice structure is a relatively new alternative IR material system and has application potential for LWIR/VLWIR spectral ranges with performance comparable to HgCdTe, but it does not compete at the shorter wavelengths”25. Sb-based superlattices are fabricated using standard III-V technology that may be more competitive due to lower costs of series production.26

HgCdTe technology will continue to dominate military passive and advanced MWIR and LWIR active applications of IR detectors because of its excellent detector properties such as broad spectral response and low noise gain. Type II InAs/GaInSb and related superlattice structures are an “alternative IR material system that has great potential for dual LWIR/VLWIR spectral ranges with theoretical performance comparable to HgCdTe with the similar cutoff wavelength”27. Based on what appeared to be more mature material and fabrication processes, hopes were high, but actual progress has been slow. More research is needed to understand avalanche gain limitations, gain noise, quantum efficiency, and choice of optimal semiconductor alloys.

SIGNAL PROCESSING

Limits of Signal Processing

Signal processing consists of a series of computational steps in the transformation of data into information. In the case of a ladar, the steps might begin with a ROIC attached (typically bump-bonded) to a detector array. The ROIC puts out multiple concurrent data signals that must be “processed” by a computational method (“algorithm”). Figure 5-9 illustrates a set of steps mapped onto some hardware components.

The computational method is typically decomposed into steps. The decomposition is based on either or both of the following: (1) orderings of operations required by the mathematics of the method (e.g., a series of transforms performed in order) and (2) current technology limits on storage, communication and computation. These technology limits can be component limitations (e.g., the clock rate of a processor or the cycle times of storage devices) or architectural limitations (e.g., mismatches among components, such as wireless communications links and airborne sensor collection capabilities or limitations on parallelism inherent in the algorithm).

Each step will have an associated delay (measured in seconds) and throughput (measured in bits per second). The delay through the signal processing system as a whole is the sum of the delays for each step, plus the time needed to communicate data between each step. Several schemes can be used to improve performance, some of which are enabled by improvements in algorithms and their implementations, and others of which are enabled by improvements in hardware technologies such as semiconductor manufacturing.

An example of an algorithmic improvement is one that allows multiple data items to be processed at once, in a mode of operation called parallel processing. When n operations can be performed at the same time, the delay per operation is reduced by a factor of n and the throughput is increased by a factor of n. So, if a signal processing step allows for 10 operations to be processed at once and 10 hardware units are available, it takes 1 unit of delay to process 10 operations, so each operation accumulates 1/10 of the delay required for nonparallel operation. Since 10 operations on b bits of data are performed at once, the

____________________

25 A. Rogalski, 2009 “Infrared Detectors for the Future” ActA PhysicA Polonica A 116: 389-406.

26 A. Rogalski, Material considerations for third generation photon detectors, http://www.phy-astr.gsu.edu/qwip2006/Presentations/Material%20considerations%20for%20third%20generation%20detectors%20-%20A%20R.pdf. Accessed on March 14, 2014.

27 A. Rogalski, 2010, Infrared Detectors, Second Edition, CRC Press, pp. 846–847.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

image

FIGURE 5-9 Components of a signal processing algorithm for an FPA. SOURCE: Figure 4-10 from National Research Council, 2010, Seeing Photons: Progre ss and Limits of Visible and Infrared Sensor Arrays, The National Academies Press, Washington, D.C.

throughput is 10 × b, or 10 times as much. Unfortunately, not all computation is parallelizable, and the gains for parallelism are limited to the portions of the computation that can be parallelized. Amdahl’s Law28 states this succinctly, as:

Speedup for n processors = 1/(serial + ([1-serial ]/n))

where serial, a number between 0 and 1, is the fraction of the computation that must be executed one operation at a time. Fortunately, much of the signal processing can be parallelized and the major limits are technological in nature (e.g., the number n of operations that can be carried out in parallel). These limits are being addressed aggressively—for example, by companies such as NVidia that develop graphics processor units (GPUs) that carry out thousands of operations simultaneously. For example, an NVidia Tesla K20X offers29 2688 cores and peak performance figures of 1.31 double-precision TFLOPS30 and 3.95 single-precision TFLOPS.

A technology improvement such as an improvement in semiconductor fabrication can affect the ability to integrate multiple steps of an algorithm on a single chip. This reduces the delay for communicating data between steps of an algorithm and reduces the potential for errors and noise in a communications path. Further, integration may permit further processing optimizations when data are

____________________

28 Gene Amdahl, 1967, “Validity of the single processor approach to achieving large-scale computing capabilities,” AFIPS Spring Joint Computer Conference Proceedings, 483.

29http://www.nvidia.com/object/tesla-servers.html.

30 1012 floating point operations per second.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

TABLE 5-2 Limitations on Various Signal Processing Parameters

Parameter Limit State-of-the-Art Implications
Digital process energy kTLn(2) = 0.0178 eV 1 × 104 eV Battery life, system size, and Moore's Law seems to prevent improvement beyond 100 ×.
Parallelized speedup 1/(serial+([1-serial]/n)) (See discussion in text.) 64 (problem dependent, technology dependent) Faster processing for certain problems
Power efficiency in multiply accumulates per second Assuming digital limit kTln(2) ~1 nw 10 /mW Analog processing might improve the state of affairs by 1000×.

accessible to multiple steps of a computation—for example, processing it in place or exploiting the principle of locality of reference.31

The ultimate limits of parallelism and integration are achieved when all of the digital functions are combined with the ROIC. While seemingly far-fetched, technology and algorithmic directions pointing towards this outcome are being explored by, among others, the Defense Advanced Research Projects Agency (DARPA) in the RAPID UPSIDE32 Program. These limits are summarized in Table 5-2.

Another important system consideration is energy. In the shorter term, DARPA has a program called Power Efficiency Revolution for Embedded Computing Technologies (PERFECT)33 focused on gaining higher performance processing on a lower energy budget, with a program target of at least 75 GFLOP/W. While it is not clear how to achieve it, there is a lower bound on the energy required for computing derived from thermodynamics, called Landauer’s principle,34 which states that there is a minimum amount of energy, kT ln(2), required to change a bit. This prediction has recently been verified experimentally.35 Landauer’s principle applies to irreversible computations (those that increase entropy) but arguments have been made36 that reversible computing37 might be able to overcome this lower limit, at least for portions of the computation. The current state-of-the-art processor’s power consumption is several orders of magnitude higher than that predicted by Landauer’s principle, so this fundamental limit is not a practical barrier.

Finally, whenever practical, adding more signal processing functionality to the ROIC is a preferred option. This contributes to making the systems smaller and lighter and may even result in lower overall power consumption. Adding functionality on the ROIC however, often comes at the expense of added thermal management due to the added thermal dissipation, and more significantly as the bandwidth

____________________

31 Locality of reference is observed in multiple forms in computation, but the important idea is that data nearby to each other in time (referenced recently) and space (nearby in the data set) are more likely to be referenced.

32 RAPID UPSIDE, Solicitation Number: DARPA-BAA-12-53.

33http://www.darpa.mil/Our_Work/MTO/Programs/Power_Efficiency_Revolution_for_Embedded_Computing_Technologies_(PERFECT).aspx.

34 Rolf Landauer, 1961, “Irreversibility and heat generation in the computing process,” IBM Journal of Research and Development, 5: 183.

35 Antoine Berut, et al., 2012, “Experimental verification of Landauer’s principle linking information and thermodynamics,” Nature, 483: 187.

36 Graham P. Boechler, Jean M. Whitney, Craig S. Lent, Alexei O. Orlov and Gregory L. Snider, 2010, “Fundamental limits of energy dissipation in charge-based computing’” Applied Physics Letters, 97(10).

37 Tommaso Toffoli, 1980, “Reversible computing,” Springer: Berlin, Heidelberg.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

of the circuitry goes up. This aspect of adding functionality on the ROIC is common to passive detectors as well and has been extensively addressed in a recent NRC report.38

In summary, there are a few unique efforts to develop more efficient image processing algorithms and specialized IR image processing hardware. More importantly, EO IR systems will continue to benefit from spinoffs from the imaging processing hardware developed for games as well as the low-power processors developed for mobile phones and cameras. These will make smaller, lighter, and lower-power IR systems available worldwide.

There are multiple axes along which processing performance is measured, including throughput, latency, and SWaP requirements. Technological improvements in parallel processing and integration are promising but may come at the expense of SWaP. The technology goal can be expressed as maximizing the ratio of performance/SWaP. Evidence from biological systems (e.g., the human brain) and physics (e.g., Landauer's principle) indicates that substantial opportunities exist.

Recommendation 5-1: Novel approaches to signal processing based on biological systems and targeting physical limits should be explored.

PROPAGATION EFFECTS

Atmospheric Propagation

The atmosphere affects active EO sensors in three main ways: absorption, scattering, and refractive index variations (the latter causing beam spreading or fluctuations). For active EO sensors, these effects occur both for the illumination beam and for backscatter from the object to the sensor. Obviously absorption and scattering are negative influences. Normally, refractive index variations are also negative, but there are occasions where something called lucky imaging can turn these variations into a positive influence.39 In lucky imaging, portions of the image that are improved by the index of refraction variations are kept and combined over time with other improved portions of the image. Angular resolution beyond the diffraction limit can occur using lucky imaging. Index of refraction variations can cause beam spreading, beam wander, image dancing, and beam scintillation.

Absorption is a process whereby atmospheric molecules absorb energy from photons. This is a quantum process, so photons with a certain amount of energy, at certain wavelengths, are preferentially absorbed. Figure 5-10 shows atmospheric transmission vs. wavelength for the visible and infrared regions.

Scattering is another significant effect of the atmosphere. Rayleigh scattering occurs when the photons scatter off of particles that are small compared to the wavelength of light. Rayleigh scattering varies inversely with the fourth power of the wavelength and is isotropic.40 When the wavelength of light is very much larger than the particles, very little scattering occurs. This is why microwave radar usually has excellent transmission through the atmosphere, even through clouds, fog, and often rain. Infrared radiation is not scattered as much as visible radiation. Mie scattering occurs when particles are about the same size as the wavelength of the light and is biased toward forward scattering. Mie scattering tends to scatter more than Rayleigh scattering. Particles much larger than the wavelength of electromagnetic radiation exhibit reflection and refraction, similar to hitting a large object. Total atmospheric transmission, combining losses due to absorption and due to scattering, is characterized at a given wavelength by Beer’s law:

____________________

38 National Research Council, 2010, Seeing Photons: Progress and Limits of Visible and Infrared Sensor Arrays, National Academies Press.

39 David L. Fried, 1978, "Probability of getting a lucky short-exposure image through turbulence," Optical Society of America 68(12): 1651.

40 C.E. Barnett, 1942, "Some application of wavelength turbidimetry in the infrared," J. Phys. Chem 46(1): 69.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

image

FIGURE 5-10 Atmospheric absorption. Figure created by R. Pogge from ATRAN model data of Lord, S.D. (1992, NASA Technical Memor. 103957) from the Gemini Observatory website (http://www.gemini.edu).

image

where τ is the absorption over some distance and α is the absorption per unit distance for a given wavelength.41 Beer’s law provides an exponential decay of the signal level vs. distance. For large values of α, Beer’s law can dramatically reduce signal level after some distances. There is a phenomenon called a precursor wave, which exhibits only linear decay vs. wavelength, but the explanation is thought to be that the decay is the sum of exponential decays with different decay rates at different wavelengths. The sum of the wavelengths exhibits linear decay, but the mix of wavelengths transmitted shifts to those wavelengths with lower exponential decay.

Turbulence can be described in two limiting cases: thin turbulence or distributed turbulence, as shown in Figure 5-11.

Thin turbulence is what is usually considered, because it is easier to compensate. With thin turbulence it is assumed that all the phase distortions are at, or near, the pupil plane. As a result, phase change that occurs in the pupil plane will completely compensate for the thin turbulence optical distortions. Adaptive optics systems can provide phase compensation for thin turbulence.

Aero-optical effects occur near an air vehicle, and therefore fit into the region of pupil plane disturbances. The difficulty with compensating for aero-optical effects is that they tend to be much more rapid than other sources of turbulence, requiring perhaps an order of magnitude faster compensation (on the order of ten kHz instead of 1 kHz).

Due to the diffraction limit, when an illuminator beam is emitted from the active EO system and sent to the target, the beam can be made narrower by using larger optics. Large optics mean high antenna gain and a narrow beam. This puts high flux (watts per square meter) on a target. The atmosphere, however, can limit how narrow the illumination beam is even with large optics. David Fried defined the Fried parameter, r0, which is the largest effective diameter that can be used under given atmospheric conditions.42 When calculating the beam divergence of the illuminator beam, the estimated full width, half max, beam divergence of the illuminator beam in a vacuum is:

image

____________________

41 Larry Andrews, 2004, Field Guide to Atmospheric Optics, SPIE FG04: 4.

42 D. L. Fried, 1966, “optical resolution through a randomly inhomogeneous medium for very long and very short exposures,̵ Journal of the Optical Society of America 56(10): 1372.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

image

FIGURE 5-11 Thin turbulence is shown on the left, and distributed turbulence—sometimes called deep turbulence—is shown on the right.

where D is the diameter of the illuminator aperture. In the atmosphere, however, the beam divergence calculation will substitute the Fried parameter, r0, for D, because the atmosphere will not allow an area larger in diameter than r0 to have constant phase across it (to act like a single aperture). The Fried parameter depends on the structure parameter, or strength of the atmospheric turbulence, Cn2, of the atmosphere for that day, time, and location.43 It also depends on the path length. The Fried parameter44 may be calculated as

image

where k = 2π/λ.

One major assumption is usually the frozen atmosphere assumption—that is, the atmosphere stays frozen but just moves past at some velocity. The faster it moves, the more turbulent the atmosphere. The velocity of the wind will influence the Greenwood time constant,45 the time interval over which turbulence remains essentially unchanged.

Scintillation refers to the twinkle of a star, the spatial or temporal fluctuations in irradiance caused by small random index of refraction fluctuations. The Rytov variance characterizes scintillation.

The illuminator beam also suffers from Beer’s law transmission. Light scatters off the object of interest in either a specular or Lambertian fashion. Lambertian scatter provides scattering over a wide angle of return, again encountering Beer’s law as light returns to the active EO sensor receiver. The index of refraction variations in the atmosphere make the return light come back in a varying, crooked, path. Detail in the object being viewed moves around as the atmosphere distorts its image. When this moving detail is averaged, it reduces angular resolution.

One limit is how much loss might be acceptable. Beer’s law can be used to calculate how many absorption lengths would yield an acceptable amount of atmospheric loss, or an acceptable allowance in the link budget. Table 5-3 shows the loss as a function of the number of absorption lengths.

Active EO sensors require two-way transmission, so a 15 db one-way loss means a 30 db loss both ways, or a loss in link budget by a factor of 1,000. From this table, it is obvious that no more than 4 or 5 absorption lengths loss can be tolerated for any reasonable-power ladar. A loss of 1,000 or 10,000 is a significant link budget loss.

Another limit is the amount of beam spreading that can be tolerated for the illuminator beam. If flash illumination is used, then this is not much of a limit. Assuming a 32 × 32 pixel FPA on receive, the illuminator beam can be 32 times larger than a pixel, and the illuminator aperture can be 32 times smaller than the receive aperture. For a 20-cm receive aperture the transmit aperture can be less than 1 cm. Another way to think of this is that the Fried parameter can be less than 1 cm without interfering with the illuminator beam.

____________________

43 Larry Andrews, 2004, Field Guide to Atmospheric Optics, SPIE FG04: 11.

44 John W. Hardy, 1998, Adaptive Optics for Astronomical Telescopes, Oxford University Press, 92.

45 See http://en.wikipedia.org/wiki/Greenwood_frequency.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

TABLE 5-3 Atmospheric Loss as a Function of Absorption Lengths

Number of Absorption Lengths

Transmission τ

Loss (db)

1

0.367879

4.3

2

0.135335

8.7

3

0.049787

13   

4

0.018316

17.4

5

0.006738

21.7

6

0.002479

26.1

7

0.000912

30.4

8

0.000335

34.7

Another limit is the turbulence on receive, which will limit angular resolution. The diffraction limit on receive will be limited to the diffraction associated with an aperture diameter equal to the Fried parameter unless some form of compensation is employed. Methods of compensation were discussed in Chapter 4.

Atmospheric compensation techniques are key for correcting aberrations caused by the atmosphere. Traditional adaptive optics only correct for phase variations at the pupil plane. There is no effective approach to reduce exponential loss except going to longer wavelengths. Longer wavelengths will significantly reduce exponential loss from scattering for small particles.

Conclusion 5-1: Any long-range active EO sensor operating in the atmosphere will be limited by the atmosphere. The amount of limitation will vary from hour to hour and day to day.

Conclusion 5-2: The ability to compensate for extended turbulence will greatly aid long-range active EO sensors. If some method could be found to reduce the exponential decay, such as a soliton wave through highly scattering or absorptive media, this would be immensely beneficial for active EO sensors.

Propagation Underwater

“Active EO sensors transmit energy that is elastically or inelastically returned from the probed region of interest and analyzed for target signatures”46. Whether or not active EO sensing is practical underwater depends on whether the target is visible. Two cases can be considered: (1) active illumination from below the surface of the water—that is, the source and detector are immersed, and (2), active illumination and detection from above the surface of the water.

The ultimate limit to observability is the transparency of the water. Besides the natural processes of absorption and scattering in pure water, there may be further degradation of visibility due to suspended particles, turbidity and turbulence, and water color. The final limits will depend on the reflectance and contrast of the target, the character of the optical image, and the detection sensitivity.

____________________

46 Arete Associates, Active EO (electro-optical), http://www.arete.com/operational_customers/active_eo.aspx. Accessed on March 14, 2014

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

The degree to which seawater is transparent is a function of the combined effects of scattering and absorption of light by the water. Both absorption and scattering cause the amount of light to decrease exponentially with distance z : I = Ioexp(αz), where α is the fraction of light absorbed per unit length, called the loss coefficient. The loss coefficient will be the sum of the coefficients for all loss mechanisms. For example, the loss coefficient α = αwater + αdissolved + αparticle, assumes that the ocean or lake contains losses inherent in pure water plus those due to dissolved substances and those due to suspended particles. Each of these many coefficients have absorption and scattering components.47

Transparency of Pure Water

The absorption of pure water has a strong wavelength dependence and is found to be smallest in the blue-green wavelengths (see Figure 2-23). The minimum absorption is near the blue wavelength 418 nm, where the absorption coefficient αabs = 0.005 m-1. The water absorption increases rapidly at red wavelengths because of overtones of infrared vibrational resonances.

In addition to absorption, even pure water exhibits scattering losses. Rayleigh scattering is scattering from molecules, particles, and density fluctuations that are small with respect to the wavelength of the light. This scattering is much larger for shorter wavelengths in the blue because the scattered intensity is proportional to 1/λ4, where λ is the wavelength of light. This increased scattering in the blue pushes the wavelength of minimum total attenuation more to the green. The optimum operating wavelength would be where the sum of the two absorption coefficients is a minimum, close to the argon laser wavelength of 488 nm. This encouraged early researchers to investigate active underwater sensing using argon lasers. More recently frequency-doubled Nd:YAG lasers, operating at 532 nm, have become more practical. According to Figure 5-12, at this wavelength αabs = 0.046 m-1 and αscat = 0.026, for a total attenuation coefficient of 0.072 m-1. This means that in the purest water, the laser signal at 532 nm will drop to 1/e in a distance of 14 meters. If used for active sensing, the signal must make a round trip and still be detectable. This loss must be figured in the link budget. In addition, any scattering must be gated out or minimized in some fashion.

A range-gated system was built in China using time resolution to minimize the signal from scattering.48 It was tested in a swimming pool and apparently was able to detect objects at 26 meters. Given the data from Figure 5-12, if the water were ideal, with a loss coefficient of α = 0.072 m-1, at 26 m there should still be 15 percent of the incident light. In fact, the signal was very weak at this distance. Images were reported only at 15 m distance. Clearly the loss in real water is much larger than the ideal, even in a quiet pool.

Real water is rarely pure, and additional factors are much more important than intrinsic water absorption and Rayleigh scattering. The other major concern for water transparency is Mie scattering, which is due to particles that are larger than, or on the same order of the size as, the wavelength of the incident light. Mie scattering has no strong wavelength dependence. The direction of scattered light peaks forward. Particles in water such as algae and mud usually introduce both scattering and absorption. In water, Mie scattering from particles larger than a wavelength of light is usually the most predominant. It has these general characteristics: (1) all wavelengths are scattered by roughly similar amounts, (2) most of the light is scattered in a forward direction, much less is scattered backwards or sideways, and (3) Mie scattering is very much stronger than Rayleigh scattering and increases as the concentration of particles increases.

____________________

47http://www.deepocean.net/deepocean/index.php?science07.php.

48 Jin Weiqi, Cao Fengmei, Wang Xia, Liu Guangrong, Huang Youwei, Qi Huaichuan, Shen Fei, 2008, “Range-gated underwater laser imaging system based on intensified gate imaging technology,” Proc. of SPIE 6621: 66210L.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

image

FIGURE 5-12 Spectral absorption coefficient αabs and total scattering coefficient αscat for pure water. The two losses at wavelength 488 nm sum to a total minimum attenuation. SOURCE: Adapted from Figure 5 in R.A. Maffione, 2001, “Evolution and revolution in measuring ocean optical properties,” Oceanography 14(3): 9.

CONCLUDING THOUGHTS AND OVERARCHING CONCLUSION AND RECOMMENDATION

Since the game-changing invention of microwave radar technology in the Second World War, radar has developed a mature technology and exploitation base and has proliferated to a wide variety of military and commercial applications. Similarly, passive EO sensing systems have developed widespread commercial applications and have been exploited for military use in space, in the air on intelligence, surveillance, and reconnaissance (ISR) missions, and on the ground in night vision systems. Active EO sensing systems enjoy the best of both worlds in that they have resolution advantages over microwave radar and illumination advantages over passive EO systems. Recent advances in laser illuminators, sensitive broadband detectors, and rapid data processing are combining to enable active EO sensing to bring revolutionary advances in areas such as mapping, targeting, autonomous robotics, environmental monitoring, weather prediction, and intelligence gathering. In the context of this committee’s task to focus on those areas of active EO sensing that could produce technological surprise adverse to U.S. national security, the committee arrived at the following overarching conclusion and recommendation:

Overarching Conclusion: Active electro-optical sensing is a rapidly emerging technology with many applications across intelligence, military, scientific, and commercial domains. It has the potential to alter the balance in many of these areas. Interest in developing and applying active electro-optical

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

sensing has consequently risen in other nations, putting U.S. leadership at risk, to the degree that in some instances the United States no longer leads.

Overarching Recommendation: The uses and development worldwide of active electro-optical (EO) sensing should be tracked aggressively by the U.S. intelligence community, and active EO sensing should be aggressively developed by the United States.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

This page intentionally left blank.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×

This page intentionally left blank.

Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 255
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 256
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 257
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 258
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 259
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 260
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 261
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 262
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 263
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 264
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 265
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 266
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 267
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 268
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 269
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 270
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 271
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 272
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 273
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 274
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 275
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 276
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 277
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 278
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 279
Suggested Citation:"5 Fundamental and Engineering Limits of Active Electro-Optical Sensing." National Research Council. 2014. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing. Washington, DC: The National Academies Press. doi: 10.17226/18733.
×
Page 280
Next: Appendix A: Committee Biographies »
Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing Get This Book
×
Buy Paperback | $75.00 Buy Ebook | $59.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

In today's world, the range of technologies with the potential to threaten the security of U.S. military forces is extremely broad. These include developments in explosive materials, sensors, control systems, robotics, satellite systems, and computing power, to name just a few. Such technologies have not only enhanced the capabilities of U.S. military forces, but also offer enhanced offensive capabilities to potential adversaries - either directly through the development of more sophisticated weapons, or more indirectly through opportunities for interrupting the function of defensive U.S. military systems. Passive and active electro-optical (EO) sensing technologies are prime examples.

Laser Radar considers the potential of active EO technologies to create surprise; i.e., systems that use a source of visible or infrared light to interrogate a target in combination with sensitive detectors and processors to analyze the returned light. The addition of an interrogating light source to the system adds rich new phenomenologies that enable new capabilities to be explored. This report evaluates the fundamental, physical limits to active EO sensor technologies with potential military utility; identifies key technologies that may help overcome the impediments within a 5-10 year timeframe; considers the pros and cons of implementing each existing or emerging technology; and evaluates the potential uses of active EO sensing technologies, including 3D mapping and multi-discriminate laser radar technologies.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    Switch between the Original Pages, where you can read the report as it appeared in print, and Text Pages for the web version, where you can highlight and search the text.

    « Back Next »
  6. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  7. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  8. ×

    View our suggested citation for this chapter.

    « Back Next »
  9. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!