Deep-spaceoptical communications

Recent investigations have shown that laser systems, particularlythe incoherent direct detection and transmitted reference systems, haveimportant potential advantages over local heterodyning techniquesfor achieving effective deep-space communications

E. Brookner, M. Kolker Raytheon CompanyR. M. Wilmotte Consultant

A major problem in deep-space communication sys- days to transmit 1010 bits. At the same conference, it wastems is that of obtaining high data rates (of the order suggested that 1000 b/s may be achieved in time for theof 107 bits per second). This article proposes some early Voyager program. As many as 115 days would bedesign concepts that indicate the probable feasibility required, even at this data rate. At the 10-Mb/s data rateof achieving wide-band communications by means of suggested as a minimum requirement, the transmissionthe laser. The example selected here is a hypothetical time assumes values in seconds.mission to Venus, chosen because of its great bright- To increase the information data rate capability sig-ness and, hence, high background-noise level. Since nificantly at radio frequencies implies consideration ofno earth satellite relay is assumed, the communica- larger antennas in the spacecraft and on the ground,tion channel includes the atmosphere. The down-link increased power in the spacecraft, and use of higheris the one considered because of its high-information- frequencies (for example, EHF) with the commensuraterate requirement. development requirements and cost to change from the

current NASA deep-space instrument facility S-band sys-The primary goals of any space subsystem are generally tem. The laser-with its extremely narrow beam due to its

taken to be low power and low weight. For deep-space, short wavelength, notwithstanding its high quantum andwide-band communication, however, another factor may background noise-offers the possibility of surpassingRFbe equally important-namely, the size of the transmit- techniques in its ability to satisfy deep-space require-ting aperture. A very large aperture, as would undoubt- ments. Should it prove superior to RF at data rates of theedly be required by a microwave channel, is likely to prove order of 107 b/s, its growth capability to higher data ratesan obstruction to the sensors of the spacecraft and will, will be much greater than that ofRF systems. We can ex-therefore, reduce the time available for collecting infor- pect the laser communication art to develop in all itsmation or transmitting it. In this respect, the laser has an component areas, as has been historically achieved in allimportant advantage over microwave, as will be seen new technologies.later.The magnitude of the deep-space communication prob-

lem is shown in Table I. At the August 23-27, 1965, con- I. Space communication problemference at Virginia Polytechnic Institute on "The Explora- Transmission Timetion of Mars and Venus," the total information required Earlyin the early Voyager program was suggested to be of the Voyager at at at atorder of l08 to 10'° bits. A typical imagery requirement of Requirements, 10 b/s, 100 b/s, 1000 b/s, 107 b/s,the geologist interested in rough terrain characteristics of toabis dy dys as sensMars or Venus is 20-cm X 20-cm photography, by means *108 115 11.5 1.15 10of whichasingle picture could contain asmanyas 108 bits. 1010 11 500 1150 115 1000Based on current capability of the order of 8 bits per sec- * Approximate requirements for one 20-cm X 20-cm photo withond (b/s) of Mariner IV, it could take as long as 11 500 15 shades of gray at 10 lines/mm.

IEEE spectrum JANUARY 1967 75

Candidate optical systems must be maintained accurately at a specified frequencyThree types of systems have been considered: difference from that ofthe received signal light, and thus it1. Local heterodyne system (LHS) is necessary that the local laser be continuously corrected2. Direct detection system (DDS) for the Doppler shift. For some missions, the Doppler3. Transmitted reference system (TRS)' shift can be very large, typically well over 10 GHz.

Block diagrams of these systems are shown in Fig. 1. The direct detection system is simply a straightforwardThe local heterodyne system often provides the highest transmission and detection system, with a single modu-

signal-to-noise ratio (SNR) of the three systems because lated carrier providing video detection. It has a limitationthe local heterodyne laser can be made sufficiently strong in the loss of phase information of the carrier. Neverthe-that the shot noise from it dominates all other sources of less, DDS appears to be the most attractive choice at thisnoise. However, as shall be indicated later, the DDS and time.TRS can be designed so that the power efficiency is nearly The transmitted reference system is a heterodyne sys-as high as that of the LHS. Moreover, the LHS system tem in which the reference is transmitted with the signalsuffers from the serious disadvantage of receiver SNR from the spacecraft. This technique avoids the Dopplerdegradation due to spatial dispersive effects of the at- shift problem of LHS, but its SNR is lower than that ofmosphere. either the DDS or LHS. It is lower than that of the DDSFor LHS to operate properly, the local laser radiation principally because only half the power transmitted

should maintain spatial coherence with the received signal from the spacecraft is signal power. The successful per-light over the whole receiving optical aperture. The at- formance of this system depends on the assumptionmosphere disperses the signal beam so that coherence is that the atmosphere will not disperse the very close fre-lost for much of the time except for very small apertures. quencies of the signal and the reference sufficiently toFor example, based on an atmospheric transmission ex- damage their spatial coherence over the receiver aper-periment by Goldstein et al.2 from noon to midnight over ture. (The frequency separation is of the order of 0.2-10a 4-km path at a wavelength of 0.63 ,um, a 3-cm dish would GHz.) This system can use almost any form of modula-experience 70 percent of the time a loss of at least 15 dB tion including phase-shift keying (PSK).greater than that experienced under atmospheric condi- Diffraction-limited optics do not provide increasedtions that permit perfect coherence. During the experi- SNR for the DDS and TRS. Since nondiffraction-limitedment reception was found to be generally poor except optics will simply have the effect of producing a focalshortly after sunset. Smaller losses may be expected for in- area larger than the Airy disk, the detec-or is requiredstallations on high mountains and with longer wave- to have a larger area. It is, therefore, possible to receivelengths, but the prospects are not promising for maintain- light signals through the atmosphere on extremely largeing reliably the spatial coherence across the aperture nondiffraction-limited apertures with high efficiency byneeded for an LHS. Another alternative for the LHS is to providing adequate detector area at the focus of thehave a receiver system that consists of a large number of optics. For the greatest accuracy, it is desirablesmall diffraction-limited dishes. The randomness of the 1. That the optics do not enlarge the focal areaphases of the signals from each dish would then be com- beyond a diameter for which the collection of photonspensated for by some adaptive scheme that adds the by the detector becomes difficult.signals in phase. 2. That the highest frequency of the modulationLHS suffers from one other problem. The local laser carried by the light beam remain coherent over the area

Fig. 1. Optical communication system types. A-Direct detection system.B-Local heterodyne system. C-Transmitted reference system.

A Diffraction- PhotonInformation;, limited collector

B Information Photon collector

Diffraction-limited mirror Dp H

C Diffraction-limited mirror

Laser Modu1atd~~~ 'OpticsOptical Photon Current I I~~~lnformation

Photonchanger collector

76 IEEE spectrum JANUARY 1967

LHFig. 2. Multinondiffraction-limited dish system.

of the detector; that is, the mechanical irregularities noise the performances of all three systems are dependentshould not exceed the equivalent of about 1/10 of the on the receiver collecting area and not on the number ofwavelength of the highest modulation frequency. dishes involved. This important property applies theoreti-When it is desired to have photon collectors of very cally to the TRS, LHS, and DDS systems, regardless of

large dimensions, the structural requirements may lead the output SNR per dish. Letto the use of a number of nondiffraction-limited dishes, PLii1 = transmitted power required for the LHSeach with its own detector system; see Fig. 2. The P1,)1) = transmitted power required for the DDScontinuously variable delays are introduced to compen- P'ris = transmitted power required for the TRSsate for the changes in path length with change in direc- Nb = background noise received by each dish (aftertion of the received beam. = filterindnoise per secondThe configuration of the several receiving apertures optical filtering), photons per second

must be such that they will not interfere with one another M = number of dishesNhr = total background noise received by the Min the directions of interest and such that the compensat- dishes (after optical filtering), photons pering delays can be accurately established. In the simplest secondconfiguration, all the dishes are installed on a single, secondvery large structure. If it is sufficiently rigid, the whole Ba = signal bandwidthstructure can be moved normal to the direction of the aS = quantum efficiency of the receiverreceived beam and compensating delays are not required XsT = power SNR at output of receiver sum pointexcept to correct for the effects of temperature and me- XbT = BTpowNR at= BrIouMNb

chanical stresses. -~~~~~~power SNR at output of receiver sum point,if there were collected by the receiver antenna

Comparison of the LHS, DDS, and TRS complex one photoelectron per hertz of trans-In order to compare the performances of the LHS, mitted signal bandwidth

DDS, and TRS, an analysis was made of the transmitter Figures 3 and 4 give plots of Pi)i)sIPij,is and P'i'nslpower required by the three systems for shot-noise-limited 2PLIIS versus XbT for X,SN = 10. It is noted that theconditions. The systems were put on an equal footing for curves are independent of M in accordance with thethe analysis, by assuming that all three systems have the results already given above. The curves indicate thatsame parameters-in particular, the same transmitter for high XbT, the incoherent DDS has a power efficiencyfrequency and area, the same total receiver background as high as that of the LHS. The TRS requires four timesnoise, and the same detector efficiency. Multiple-dish as much power as the LHS for high values of Xb7 be-receiver systems are assumed, with the total receiving cause half the transmitter power is in the information-area (not necessarily the number of dishes) the same for carrying part of the signal, which results in a fourfoldthe systems. For the LHS, it was also assumed that the decrease in the power SNR after detection. By proper de-atmosphere does not degrade the coherence of the in- sign of the system (that is, by the use of a wide bandwidthcoming field. Equivalently, it was assumed that the for the signal when necessary and a narrow-band opticalspatial incoherence of the received signal is compensated filter and small field of view, the quantity XhT can befor by an adaptive technique for adding in phase the made large.signals from the various dishes. For the TRS the signals It is found that one generally can achieve, or come closefrom the outputs of the detectors are added in phase, to achieving, shot-noise-limited conditions for the DDSwhereas for the DDS they are added linearly, as indi- and TRS. A sufficient condition for the DDS and TRScated in Fig. 2. The pessimistic assumption is made that to be shot-noise limited is thatthe background noise is spatially coherent over the total BT I aNb',collecting area of the systems. And it is assumed that the BT =0

For all these systems aNbT/B0 < 0.1. It is important to The DDS is assumed to have the same receiver dish con-point out, however, that if the systems are partially or figuration as specified for the Venus mission given incompletely limited by classical background noise (instead Table II; that is, it consists of 25 ten-meter dishes. Theof shot noise), the power performance of the systems is assumption is again made that the degrading effects ofeven closer than is indicated in Figs. 3 and 4. The systems the atmosphere can be ignored for the LHS.become classical background-noise limited when the What was allowed to vary for the systems and serve asdirection of inequality is reversed in the above equation. a parameter for comparison is the transmitter dish di-Using the results of Fig. 3 one finds then that when ameter. The transmitter dish diameters were set so as

to give the same receiver SNR in all the systems. Also1 - NbT < 2 used as a basis of comparison is the number of receiver

XbT BT dishes required for the systems. The sizes of the trans-

and XSN = 10, the DDS will require less than 1 dB more mitter dishes required are given in Table III in terms oftransmitter power than an equivalent diffraction-limited the dish diameter DTO required for the LHS operatedLHS if the DDS is limited by shot or background noise at 0.84 ,um. It can be seen that at 0.84 ,m the transmitterinstead of detector- or receiver-generated noise. These dish needed for the DDS is the same as for the LHSconditions are met for the GaAs DDS using a photo- and hence, on this basis, the systems are equivalent.multiplier detector. Moreover, it is found that this DDS However, when the systems are compared on the basisrequires the same power as the comparable LHS. of the number of dishes required, this is not the case.So far in the foregoing discussion the comparison has In particular, the LHS requires 2.8 million dishes if one

been on the basis that the three systems are operating at uses a dish size of 3 cm in order to attempt to eliminatethe same frequency with no concern being given to the atmospheric degradation. This is in contrast to requir-number of receiver dishes required by each system. Now ing only 25 dishes for the DDS system. As was noteda comparison is made for different frequencies of opera- previously,2 the dish size of 3 cm actually is too large,tion. Table III gives a comparison of the DDS and providing serious degradation a large percentage of thethe LHS, both operating at 0.84 Am and 10 ,im. In the time. Hence, the performance of the LHS will actuallyderivation of the table, it was assumed that the systems be worse than indicated.are signal shot-noise limited with XbT large so the DDShas as high a power efficiency as the LHS. To put thesystems on the same basis for comparison, they were 11. Mission to Venusspecified this time to have the same transmitter power, Transmitterthe same receiving area, and the same detector efficiency. Aperture

System Diameter, cmNumber Laser Detector DDS TRS

-rlmrnmill vFninlf irninipvmm-rrlilill iiml -rlimrr 1 GaAs S-1 photo- 29 5620 multiplier

XSN 10 2 GaAs Diode 209 1163 GaAs Avalanche diode 29 31

10 \ 4 Semiconductor S-20 photo- 3.5 68 _ in visible multiplier

OI_6 \ (unavailable) (X = 0.42 Ain)X4_\ 5 Argon II S-20 photo- 78 154

ltip lie6 N2-CO2 Cu-doped 30 000 42 300

2 \ germanium7 N2-CO2 Ideal (not 125 221

Xl1|11111- 11111 H11111 -| W1 11111available)0.0001 o.oi1 0.01 0.1 1 10 8 Ho-doped YAG Ideal (not 7 15

Xb.T available)Fig. 3. PDDS/PLHS as a function of XbT. Microwave S-band system: diameter = 2000 cm

Notes:Distance = 180 million kmPower input to transmitter = 30 watts

Fig. 4. PTRS/2PLHS as a function of XbT- Information rate = 107 b/s Error rate = 10damLaser receiver: 25 apertures, each 10 meters in diameterMicrowave receiver: one paraboloid, 50 meters in diameter

2I01 1 1il IY1l 1[F IITil17 IIT 111 FIlli FF lIil Modulator: PCM/PPM, alphabet size of 32, BT = 108 Hz

XSN = 10Ill. Comparison of DDS and LHS

Q _ Characteristic DDS LHSavelength (X),

4 \___ micrometers 0.84 10 0.84 10+ ~~~~~~~~~~~Transmitter dish

damneter required DTO 3.5 DTO DTO 3.5 DTO______ Receiver dish

diameter, meters 10 10 0.03 0.3t ±il ||llll il ZXXllill f i[JJJ ILLl'llll Number of receiver

0.0001 0.001 0.01 0.1T 1 10 dishes (M) 25 25 2.8>X 106 27 800

78 IEEE spectrum JANUARY 1967

Table III also indicates that the transmitter dish about 12 ,rad), which is also planned in the OAO pro-diameter for the 10-gm system has to be about 3.5 times gram, and maintaining alignment of optical system ele-that for the 0.84-ym system. Use of a CO2 laser operating ments and diffraction-limited characteristics after with-at 10 ,um for the LHS will offer considerable improve- standing the launch and space environment during ament as far as the number of dishes necessary; however, long mission to the planets. These problems increaseas the table shows, an excessively large number still can rapidly with the size of the aperture. It is, therefore, im-be expected to be required. The spatial correlation dis- portant to minimize the aperture size.tance is approximately proportional to the signal wave- Optical filter. An important component of the receiv-length; hence, for the 10-,um system, a receiver dish di- ing optics is the optical filter, which is incorporated toameter of 0.3 meter should be used. The table indicates reduce the background noise. The bandwidth of thethat for this dish diameter 27 800 dishes are required. filter will usually be large compared with the modulationEven if one optimistically assumes that a receiver dish bandwidth.diameter of 2 meters can be used, 625 dishes would be Sharp filters operate on the light interference and areneeded for the ground complex. sensitive to the angle of incidence. It is, therefore, im-

portant to insure that all the signal light is incident onSystem--analysis the filter within its angular field of view, and no back-

Sources of noise. The four sources of noise in an opti- ground light which is incident at larger angles reachescal detector are as follows: the detector.

1. Thermal. Similar to that at microwave frequencies. Lyott filters are attractive because they provide a wide2. Quantum or shot. Very high compared with that field of view with narrow bandwidth. A filter of 0.5 A

at microwave frequencies. This category includes the (0.05 nm) at a wavelength of 0.84 gim, about 5 cm inshot noise due to the signal photons, the background diameter and 40 cm long, can be made, using calcitephotons, and the photons that are equivalent to the dark and quartz, with a field of view of 50 (0.1 rad). The fil-current. ter, which could be tuned through ±0.5 A, is expected

3. Dark current. Adds to shot noise. There is no to have a transmissivity of 0.15. It is sensitive to tempera-equivalent at microwave frequencies. ture changes, which should be maintained within 0.1°K.

4. Background. Very high compared with normal Transmitter laser choice. Of the three types of lasers-operations in the microwave region. The background gas, solid-state, and semiconductor-the most desirablecalculated for the systems listed in Table II was based for ultimate development for deep-space communica-on irradiance of 10-10 W/cm2 - gm for Venus, and a radi- tions is the semiconductor type (currently GaAs) becauseance from the sun-illuminated atmosphere of 1.3 X of its small size and weight, its promise of ready capability10-4 W/cm2sr-,m.gm for wide-band pulsed internal modulation with simpleOther atmospheric effects. Because of the earth's rota- techniques, and its potential for high efficiency (between

tion, at least three ground sites are required. In addition, 0.3 and 0.6) at reasonable temperatures.to minimize the attenuation that may occur due to bad In the visible region, gas lasers having a single mode,weather, these sites must be selected for their high proba- very narrow bandwidth, and high power can be made;bility of clear weather. There are such areas on the earth. 4 however, the efficiency is low-about 0.1 percent or lessThe probability of clear weather can be further increased for the narrow-band, single-mode operation. In the far-by providing redundancy with additional sites. The values infrared region, molecular gas lasers have recently ap-used for atmospheric attenuation in the systems listed peared. The N2-CO2 laser (X = 10.55 gm), which isin Table II are for clear weather.5 receiving much attention at present, has been made

Transmitter optics. Spacecraft transmitter optics must with an efficiency of about 10 percent. This wavelengthbe small and light. A laser should be excellent for this falls within a wide atmospheric window. However, com-purpose. If the beam from it is perfectly coherent, it can pared with GaAs, it has the following disadvantages:in theory be focused to a point of dimension of the order 1. It is not possible to obtain high data rate and veryof a few wavelengths of light, so that it is possible to narrow pulsed operation. At present, fundamental limi-make full use of the collimating capability of diffraction- tations rule out wide-band internal modulation.limited optics. The limitations for achieving this are 2. For a given diameter of the diffraction-limited(1) the degree to which the laser beam is truly coherent dish in the spacecraft, the gain is -22 dB relative toand (2) the stability of the lasing area. GaAs (because of longer wavelength).Gas lasers are currently the best lasers for meeting It remains also to develop wide-band detectors that are

these two conditions. Semiconductor lasers are at present not detector-noise limited for the direct detection andpoor in this respect. Although the gallium arsenide laser transmitted reference system. The transmissivity of themay not be steady at present (no measurements are atmosphere is about the same in clear weather as atknown to have been made to determine the lasing area a 0.84-gm/wavelength. Although the N2-CO2 laser willstability), it is expected that it will be when single-mode operate satisfactory in worse weather conditions thanoperation is achieved. It may be necessary, however, to the GaAs laser will, there are weather conditions in whichmaintain the temperature very constant. neither laser can operate.As to the optical mirror, diffraction-limited dishes up Solid-state lasers, such as ruby, that radiate in the

to one meter are cur-rently being discussed for deep-space visible region have low power efficiencies (less than 1communication transmitters. Such a size appears feasible percent) and are useful mainly for high-peak power pulsesin view of plans for the Orbiting Astronomical Observa- at low repetition rates. A holmium-doped yttrium alu-tory (OAO) program to orbit a telescope of this diameter. minum garnet (YAG) laser has been made to lase CW inSuch large dishes lead to difficult design problems, in- the infrared region at 2.3 gum with a power efficiency ofcluding an extreme tracking requirement (for example, 5 percent at liquid nitrogen temperatures (77°K); a

Brookner, K(olker, Wilmotte-Deep-space optical communications 79

realizable efficiency of 10 percent seems reasonable. of 10-s or 10-2 can produce pictures of satisfactoryOutputs of the order of a few watts were achieved, with quality. Of course, engineering data, which would bemuch higher outputs being anticipated. As in the case of transmitted at a lower data rate than video data (a rateN2-CO2, there is the problem of developing an efficient of 105 b/s as compared with 107 b/s), need a bit-errormodulator for obtaining narrow (nanosecond) pulses probability of about 10-5 and hence require error-cor-at a high data rate of the type desired. High-efficiency recting codes. In Table II, the conservative bit-errorpulsed operation with higher repetition rates than tens probability of 10-4 was used for comparison purposesof kilohertz appears unfeasible because of the lack for the high data rate transmission.of a flash lamp that can operate at these higher rates. Detectors. Photodetectors are in effect photon-to-elec-Also, there remains the equally important problem of tron converters. In the course of conversion, they providedeveloping a good detector for receiver systems that gain and noise in varying degrees. With the weak lightdo not use local heterodyning. intensity of deep-space communication, the possibility ofSemiconductor lasers (for example, GaAs at 0.84 gain in the photodetector is of prime importance to

Mm) are at present the most promising, but require minimize the effect of thermal noise generated in the out-considerable development before they can be effectively put resistor. Two detectors are of special interest in thisused for deep-space communication. The present prob- respect. The first, the photomultiplier, is excellent be-lems with GaAs lasers are concerned with cause of its high gain (of the order of 106) with little

1. Multimode operation. The fluorescence bandwidth generation of noise; however, it has only a fair quantumis very large, about 200 A wide; most of the power is efficiency, which rapidly becomes poor beyond a wave-within a band of about 20 A (860 GHz). It comprises length of about 0.7 ,im. The second detector of interesta large number of equispaced lines, each about 10 is the microplasma free avalanche diode, recently de-MHz wide. After single-mode operation is achieved, veloped by Bell Telephone Laboratories.78the type of modulation selected must, therefore, be In an avalanche diode recently tested, the shot noise isable to operate with a carrier having this bandwidth. proportional to the cube of the current gain of the de-

2. Stability. The lasing area may shift in positionwhen it is internally modulated. The lasing area mustbe extremely stable if we are to make full use of the gain IV. Modulation and codingcapability of the diffraction-limited optics at the trans- for bit-error probability Pe G i0'mitter.

3. Power. Continuous-wave power of the order of SNR per Bit Signal10 watts requires operation at 4°K, a temperature diffi- Modulation and coding dB Required, Hzcult to reach in a spacecraft. However, it is expected thattemperatures of 77°K might be achieved in a spacecraft PAM-1SBSC* Incoherent 19 2 X 107and that the GaAs laser can be developed to powers in Coherent 13 2 X107excess of 10 watts at such temperatures or higher. One- PAM-PPM, PAM-FM 6 6 X 107watt CW power has been achieved to date at 77°K. PCM-Binary (incoherent) 13 107watt CW power ~~~~~~~~~~~~Orthogonal(DPSKt) 10 107

It is anticipated that solving the first two problems PCM-32 Orthogonal 7 6 X 107will permit optical collimation down to a narrow beam. PCM-1024 Orthogonal 5 109Although gas lasers currently have the desirable char- Sequential decoding

acteristics that the semiconductor laser has yet to achieve Binary PSK, 3-bit(single-mode operation with spatially coherent, stable quantized detector 2.4 5 X 107output), future developments in the problem areas just 32 orthogonal, list of 8mentioned may lead to the choice of a semiconductor decoding 5.4 3X( 107laser type for deep-space communications. Shannon channel limit -1.6

Coding and modulation. Table IV6 is based on Gaussian * Pulse amplitude modulation, double-sideband suppressedcarrier.noise statistics that apply to microwave communications. t Differential phase-shift keying.With the quantum nature of light and the unknownstatistics of atmospheric effects, comparable figures foroptical communication must still be derived. However,the relative order of magnitude of the required SNR Fig. 5. Signal-to-noise ratio for avalanche diode.per bit for each type shown is expected to hold for op-tical communications. _ isAn optical beam can be modulated in phase, amplitude, -. _

and polarization. The last has some valuable character-F - -istics. The relative merits of these are not discussed OF A

herein. The modulation type selected to compare the m 1systems listed in Table II is PCM/PPM with an al- SNR Shot noisephabet size of 32 (PCM-32 orthogonal). A PCM sys- - - -1tem of larger alphabet is not used because of the L~ _ _ 7 ' SNoR mI_Acomplexities involved. Sequential decoding is not chosen /

Moreover, the small bit-error rates (of the order of 10-6 - _ _r__ =or less) that can be achieved by using sophisticated I T I I1 T Tcodes is not necessary for the high-data-rate video pic- -__- - -ture communication. Bit-error probability of the order Log diode current

80 IEEE spectrum JANUARY 1967

tector, whereas the signal is proportional to the square 3. Solid-state lasers in the visible region are notof the gain; see Fig. 5. The ordinate distance between efficient and operate at low repetition rates. Holmium-the signal and noise curves is the SNR. It is clear that doped YAG in the infrared region has a high efficiencythe smaller the gain, the greater the SNR down to the but requires the development of efficient wide-band pulsedpoint where the constant noise-that is, the thermal modulators and detectors for systems that do not usenoise-takes over. For maximum SNR, therefore, the local heterodyning. In addition, it requires cooling togain should be adjusted so that the shot noise is approxi- 770K.mately equal to the thermal noise, and the thermal 4. Semiconductors are most desirable because ofnoise should be kept to a minimum by cooling. their efficiency, size, weight, and potential ease of modu-The characteristics obtained to date and predicted lation; however, the most efficient to date, GaAs, oper-

on the silicon diode for 0.75 ,um (gallium arsenide phos- ating in the near infrared, suffers from multimode andphide laser) are: possibly spatial instability and, as yet, only one watt

1. The shot noise increases as GI, where m generally has been achieved at 770K. Developed detectors at thislies between 2.5 and 3.0. To date the figure is 3.0. This wavelength (photomultipliers having poor quantum effi-figure will probably be reduced. ciency and diodes with too small gains) are of only fair

2. Gain-bandwidth product is 10". It is not likely quality.to be increased materially, possibly up to 2 X 1011.

3. Avalanche operation is stable. Size of optics4. The sensitive area is very small, about 0.005 cm in It is believed that an important criterion of the ac-

diameter. ceptability of a system will be the size of the aperture in5. Quantum efficiency is high, over 0.5. the spacecraft, to minimize obstruction to view. In this6. Dark current between 10-10 and 10-11 ampere at respect, the laser appears far superior to other communi-

room temperature. It varies with the gain. cation systems. For laser operation, small optics in theThe germanium diode has similar avalanche character- spacecraft are desirable also for four other reasons:

istics at 0.84 ,um. At room temperature, however, its dark 1. Simpler pointing and tracking equipment.current noise is too high, although at low temperatures it 2. Reduced sensitivity to vibration and distortionis likely to be acceptably low. due to temperature gradients.

3. Simpler optics in the spacecraft for illuminatingLaser-detector combination the diffraction-limited mirror with the laser.At this stage of laser and detector development, there 4. Higher antenna efficiency.

is no combination that could be said to be unquestion-ably superior to microwave. In broad summary, one can Quantitative comparison of the DDS and TRSsay: In order to compare possible future systems, a specific

1. He-Ne and argon gas lasers are too inefficient. communications mission to Venus was considered.2. Although the molecular gas lasers have the ad- Its principal components are listed in Table II. The

vantage of reasonably high efficiency, they have a criterion for the comparison of the several systems shownnumber of disadvantages that must be overcome if they is the diameter of a diffraction-limited dish in the space-are to be used most effectively for deep-space high-data- craft that will provide the communication performancerate links. For one thing, it is not possible at present to specified for an input power of 30 watts to the laser orobtain very wide-band pulse modulation (for example, microwave transmitter. PCM/PPM with an alphabetpulse widths of the order of nanoseconds for a PCM size of 32 was chosen, as indicated previously. The resultsincoherent direct-detection communication link). Also, listed below were calculated from the quantities shownit is necessary to develop exceptionally good detectors. in the Appendix and in Table V. Analysis of Table VFinally, the long wavelength has the disadvantage of shows that maximizing the ratio of the transmitted band-requiring a larger transmitter aperture, all other things width to the information bandwidth significantly affectsbeing equal. the system SNR. For the system examples presented,

V. Expressions for output signal and noise

DDS LHS TRS

Signal a2q2G2RONS2 a2q2G2R,NhN a2q2G RoNs2NoiseThermal kTBT kTBT kTBTShot 2aq2GmRo(Ns + Nb + Nd)BT 2aq2G-R0NhBT 2aq2GmRO(2N, + Nb + Nd)BT

BT BT BTBackground a2q2G21R.Nb(2N , + Nb)B 2a2q2G2RNbNb B a2q2G21RONb(4N0 + Nb) B

Noise-signal5.4 X 1014 TBT 5.4 X 1014 TBT 5.4 X 1014 TBT

Thermal 2GIRoNI2 a'G2RQN0N, GRoS

Shot ~~~2Gm-l(N, + Nb + Nd)BT 2GmIBT 2Gm-l(2N0 + Nb + Nd)BTShot ~~~~aN0' atN8 aNS'Nb(2N0 + Nb)BT 2NbBT Nb(4N0 + Nb)BT

Background NS'BO NsBo Ns'Bo

Notes:1. N8 BT/Bi 2. Nb b B0 3. m = 2, except for avalanche diode, where 2.5 < m < 5. Chosen as 3 in calculations.

Brookner, Kolker, Wilmotte-Deep-space optical communications 81

a 0.1-GHz bandwidth for pulses of the order of 10 ns of shot noise = thermal noise.was chosen for the PPM. It should be recognized that 8. H1, irradiance of Venus = 10-10 W/cm2. ,um;such a high modulation bandwidth may be difficult Ha, radiance of sunlit atmosphere = 1.3 X 10-4to achieve because of modulator limitations, and even if W/cm2. sr*,m.achieved it may reduce the efficiency of the laser. 9. 1, optical transmissivity; lo for whole system,The following comments refer to the DDS listed in 0.05; It for transmitter, 0.5; 4, for atmosphere,

Table II. 0.7; Ir for receiver, 0.15 (includes optical filter).1. The GaAs laser with S-1 photomultiplier is at- 10. N, number of effective photons incident per second

tractive despite the low quantum efficiency' of the S-1 on the detector; N, for signal, Nh for heterodynephosphor. In view of the high gain possible with the reference, Nb for background, Nd for equivalentphotomultiplier, the thermal noise is small relative to dark current (negligible in systems of Table IL).shot noise. 11. m, exponent of G for shot effect; equals 2 except

2. The GaAs laser with diode detector is worse than for avalanche diode when it is between 2.5 and 3.System 1, because the gain is unity and thus the thermal 12. PL, power of laser radiation, watts; Psi = powernoise becomes dominant. The quantum efficiency of incident in detector, Pbi = background power on0.5 is excellent. detector.

3. In the case of the GaAs laser with an avalanche 13. R, range = 1.8 X 101 km (108 nmi).diode detector, the results in the table indicate that at the 14. R0, output impedance = 50 ohms.0.84-jAm wavelength, this detector gives efficiencies as 15. T, temperature of output impedance = 20°K.great as the photomultiplier detector. 16. a, quantum efficiency = 3.6 X 10-3 for System 1,

4. The use of a semiconductor laser in the visible 0.5 for diodes, 0.18 for Systems 4 and 5.region, with an S-20 photomultiplier detector, shows 17. X, light wavelength = 0.48 X 10-4 cm for argon.excellent performance potential; however, there are no 18. h, Planck's constant = 6.62 X 10-34 J. s.signs at this time of the possibility of developing a high- 19. k, Boltzmann's constant = 1.38 X 10-23 J/'K.power semiconductor laser in the visible part of the 20. q, electron charge = 1.6 X 10-19 coulomb.spectrum. 21. The SNR required for PCM/PPM with an alphabet

5. The argon laser with an S-20 photomultiplier size of 32 for laser communication is 10.is a bit poorer than System 1 because of the very low 22. Range equation:efficiency of its laser.

6. The N2-CO2 system with Cu-doped germanium is psi 2 DT2DR21OPLpoor, because it is detector-noise limited. '4l X2R2

7. When N2-CO2 is used with an ideal detector, the 3 X