Description

The present invention relates to a system for the measurement of electrical signals with picosecond resolution, and particularly to a signal analyzer and display system with picosecond electrooptic sampling of the signal being analyzed.

The invention is especially suitable for meas- urement of small (millivolt amplitude) transient sig¬ nals, such as are produced by picosecond photodetectors, photoconductive switches and other ultrafast devices and enables the characterization of the response of such signals with picosecond accuracy. The measurement of ultrafast electrical signals has been difficult to accomplish. Sampling oscillo¬ scopes have been used, but are limited in temporal reso¬ lution by their electronic sampling window which is typ¬ ically, approximately 25 picoseconds. It has been sug- gested that amorphous semiconductors may be switched fast enough to resolve electrical transients as short as 5 to 10 picoseconds (see D.H. Auston, et al., Appl. Phys. Lett. ___, 371, 1980). The resolution of the meas¬ urement is constrained by the material recovery time which is approximately 10 picoseconds. The response of electroopt c switch/modulators themselves has been meas¬ ured with picosecond optical sampling pulses according to an article by Alferness, et al., Appl. Phys. Lett. ___{ ! ) , 597, 1980, but not for the display and analysis of electrical signals generally.

It is an object of the present invention to provide an improved system for the measurement of elec¬ trical signals electrooptically with high resolution both temporally and in amplitude (e.g., temporal resolu- tion in the picosecond range and sensitivity less than 100 microvolts minimum).

It is another object of the present invention to provide an improved system which enables the analysis and display of signals having very broad bandwidths (e.g., 20 GHz and higher) such that changes in amplitude (e.g., rise times in the picosecond range - such as less than 20 picoseconds) can be displayed and analyzed. It is still another object of the present invention to provide a system having extremely high sen- sitivity (e.g., less than 100 microvolts minimum sensi¬ tivity) for the analysis and display of electrical sig¬ nals, with high temporal resolution (e.g., in the picosecond range).

It is a further object of the present invention to provide an improved system for the characterization of the response of optically activated devices such as photodetectors and photoconductive switches which are ultrafast .in operation (e.g., are activated in picoseconds). is a still further object of the present invention to provide an improved system for the measure¬ ment of electrical signals by electrooptic sampling using the Pockels effect.

It is a still further object of the present invention to provide an improved signal analysis and display system using ^" electrooptic sampling wherein limi¬ tations in resolution, both temporally and in amplitude, due to the. electrooptic sampling device are reduced.

Briefly described, a system for measurement of an electrical signal with picosecond resolution, which embodies the invention, makes use of means including a traveling wave Pockels cell for electrooptically sam- 5 pling successively occurring portions of the signal.

Means for processing the samples are used to provide a display of the signal on a time base which is propor¬ tional to, but much longer than, the duration of the signal so that it can be analyzed and otherwise measured, 10 The foregoing and other objects, features and advantages of the invention, as well as a presently pre¬ ferred embodiment thereof, will become more apparent from a reading of the following description in connec¬ tion with the accompanying drawings in which: 15 FIG. 1 is a block diagram schematically illus¬ trating an electrooptic system for the measurement of electrical signals with picosecond resolution which embodies the invention;

FIG. 2 is a transmission curve of the electro- 20 optic sampler which includes the polarization means and the traveling wave Pockels cell shown in FIG. 1; and FIG. 3 is a display of a typical electrical signal; namely the response of a chromium doped gallium arsenide photoconductive solid state switch detector of 25 the type described in U.S. Patent No. 4,218,618 issued August 19, 1980 to G. Mourou.

Referring more particularly to FIG. 1, an opti- ∑ cal pulse generator 10 generates and transmits separate trains of simultaneously occurring pulses along separate 30 paths 12 and 14 defined by mirrors 16, 18, 20 and 22. ^ The generator 10 is .a laser, preferably a CW mode-locked - ring dye laser which is pumped by a CW Argon laser. The ring laser operates as a colliding pulse mode-locked

(CPM) laser which produces its pulses and propagates them along the paths 12 and 14. Any laser which pro¬ duces subpicosecond pulses may be used. The pulses in this embodiment may be at a wave length of 6150 Ang- stroms; the laser pump being at 5145 Angstroms. The pulses are repetitive at a rate of 100 MHz and are of subpicosecond duration, e.g., 120 femtoseconds.

The path 14 is through an electrooptic sampling channel 24 and is for the optical sampling pulses. The pulses which propagate along the path 12 are used to synchronize the generation and propagation of the sig¬ nals which are to be measured. These pulses drive a signal source 26 which, in this embodiment, is a photo¬ conductive device. The system is operative to charac- terize the response of that device. Any signals which are produced or switched for electrooptic sampling in synchronism with the optical sampling pulses which are transmitted through the channel 24 may be measured by means of the system. When the signal source 26 is a photoconductive device such as a solid state light actu¬ ated switch, it is directly activated by the optical pulses. The signals which are measured are therefore automatically synchronized with the optical sampling pulses. For the sampling channel 24, a traveling wave

Pockels cell 28 is used. This cell is a crystal of birefringent material which has an index of refraction proportional to the electrical field established therein. The crystal is preferably composed of Lithium Niobate. Lithium tantalate (LiTaO.-,) crystals may also be used. The crystal is rectangular and has a length in the direction of propagation of the electrical signal therein,.!; a width w in the direction of transmission

of the optical pulses therethrough, and a thickness t. Using a crystal where 1, w, and t are approximately 10 millimeters,1.5millimeter, and .7 millimeter, respec¬ tively, signals having a bandwidth of at least 17 GHz 5 have been measured with the minimum sensitivity of

100 microvolts. Since the sensitivity is proportional to the ratio of the width to the thickness, and the fun¬ damental bandwidth is inversely proportional to the width (the time the optical sampling pulse is in the 10 cell), the reduction of the width and thickness dimen¬ sions can provide a bandwidth over 50 GHz without degra¬ dation of the sensitivity. It will be seen therefore that the temporal resolution of the system is very high and is not limited by the materials used in the cell. 15 The electrical field from the signal source propagates along the length of the cell on a balanced strip transmission line provided by a strip electrode 30 and a ground strip electrode 32 on opposite faces of the cell across the thickness thereof. The signal source 26 20 is coupled to the strip line by means of a coaxial cable 34 which desirably provides an impedance match between the source and the strip line. Other broadband coupling means may be used. The strip line is terminated by a resistance Z _n which is matched to the strip line impe- 25 dance and is connected thereto by another coaxial cable 36. A signal is launched and propagates along the length of the cell each time the source is activated by an optical pulse.

The sampling pulses pass through a polarizer 38 30. to produce linearly polarized light as shown by the E vector 40. This light is focused into a narrow beam by a lens 42. The optical pulses of polarized light sample the amplitude of the electrical field in the

OMPI

crystal due to the propagating signal and change its polarization. In other words, the change in polariza¬ tion varies as a function of the amplitude of the signal. A second crystal 31 identical to the first, but without electrodes, is placed adjacent to the first crystal so that its ordinary and extraordinary axes are interchanged with those of the first crystal. This eliminates static birefringent effects. The beam trans¬ mitted through the crystals 28 and 31 is collimated by 0 another lens 44 and passes through a compensator 46 and an analyzer 48. The compensator 46 interposes an addi¬ tional variable birefringence such that in the absence of a signal on the cell 28 circularly polarized light reaches the analyzer 48. This is indicated by the E 5 vector 50 ahead of the analyzer.

The compensator thus biases the modulator arrangement at its quarter wave point. This is shown in FIG. 2 where the curve represents the transmission response of the system as a function of the voltage

20 which produces the electrical field across the thickness of the cell. The compensator causes the system to oper¬ ate on the steep slope of the transmission characteris¬ tic. It is possible to use DC bias on the strip line to provide the biasing of the system or the combination of

25 DC bias and a compensator which, together with the bias, results in a change of polarization in the cell 28 and in the compensator 46 such that circularly polarized light reaches the analyzer 48.

Separate photodetectors 52 and 56 obtain sepa-

30. rate outputs corresponding to orthogonal, linear compo¬ nents of the polarized light which reach the analyzer. The transmitted component (E vector perpendicular to E vector 40) reaches the photodetector 52 and the rejected

.own

component (E vector parallel to E vector 40) reaches the photodetector 56. In the absence of an electrical sig¬ nal, the outputs of the photodetectors will be equal. A difference amplifier 58 which responds to the difference between the photodetector outputs will then produce an essentially zero output. When the signal propagates, this difference output changes, since the components of the polarized light reaching the analyzer change in opposite directions with a single component either increasing or decreasing depending upon the polarity of the signal as it propagates along the strip line.

A variable delay 60 enables the optical pulses transmitted through the channel 24 to sample successive portions of the signal. The variable delay 60 includes a delay stage 64 through which pass the optical pulses synchronous with the sampling pulses in the channel 24 and in the cell 28.

The delay stage is reciprocated along the lin¬ ear path at a constant rate, for example, one centimeter per second. The delay stage uses a corner reflector 70 which provides a double pass of the optical pulses before reaching the signal source 26. Since a travel of 1 millimeter of the delay stage 64 produces approxi¬ mately a 6 picosecond delay, the pulses reaching the signal source 26 will be variably delayed. The signal is generated repeatedly each time a pulse reaches the signal source. Since the optical pulses are variably delayed, different portions of the signal will be coin¬ cident with the optical sampling pulse which is trans- mitted through the cell 28 as the delay stage travels, say from left to right, as viewed in FIG. 1. The rate of travel of the delay stage 60 is synchronous with the time base in the display 72, since the delay stage is

V ^' l

coupled to the display as indicated schematically by the dash line 74. In the event that the display is an oscilloscope, the horizontal sweep which defines the time base is locked to the displacement of the stage 64. For example, a ramp voltage which drives the motor or electromechnical actuator which moves the delay stage may be used to control the horizontal sweep. Accord¬ ingly, the time of the sweep is proportional to the var¬ iable delay of the signals which are launched into the cell 28. The display is therefore a temporally enlarged replica of the signal.

A chopper 62 is used to encode the electrical signal applied to the crystal by effectively turning the signal on and off at typically a 1 KHz rate. The chopper 62 may be a commercially available rotating disc chopper which chops the optical pulses into successive groups at a 1 KHz rate. The train of optical pulses which propagates along the path 12 is repetitive at 100 MHz or 10 nanosecond intervals. This train is shown at 66. It is chopped into groups which are repetitive at a 1 KHz rate. Each group is 1/2 millisecond long as shown at 68. The chopper produces a rate signal synchronous with these groups of optical pulses. A lock-in amplifier, 76 tuned to the same frequency, then extracts that 1 KHz signal proportional to the signal amplitude on the crystal, thus increasing the signal to noise ratio very effectively.

The signal is applied to the display by a sig¬ nal averager 78 which is operative to further reduce the noise and increase the sensitivity of the measurement. A lock-in amplifie -suitable for use in the system may include the difference amplifier 58 as the preamplifier thereof. Such a lock-in amplifier is commercially

OMP

available from EG&G., Princeton Applied Research of Princeton, New Jersey (for example, their model 124A) . The signal averager is also available from the same source, for example, their model 4202 or 4203. FIG. 3 shows the response characteristic of a chromium doped gallium arsenide photoconductive switch, when such a switch has a 90 micrometer gap biased at 40 volts. The switch is activated by the optical pulses transmitted through the chopper and variable delay 60. The optical energy of each pulse may be 0.1 nano-Joule. FIG. 3 shows that the rise time of the signal is 21 picoseconds and the full width half magnitude is 65 picoseconds. The displayed rise time is due to the convolved bandwith limiting effects of the striplines, coaxial connectors, and optical sampling time of the geometry in the herein illustrated embodiment. These can readily be improved to enhance bandwith and sensitivity.

From the foregoing description it will be apparent that there has been provided an improved system for measurement of electrical pulses with a temporal response in the picosecond range at a sensitivity in the microvolt range. Variations and modifications in the herein described system, within the scope of the inven- tion, will undoubtedly suggest themselves to those skilled in the art. For example, higher resolutions (bandwidths over 50 GHz) and sensitivities better than 100 microvolts may be obtained by reducing the w and t dimensions of the crystal in the cell 28. Accordingly, the foregoing description should be taken as illustra¬ tive and not in a limiting sense.