FIELD OF THE INVENTION

The present invention relates to a device, which includes a cryogenic sensor housed

within a vacuum space, for microscopy of physical properties of a room-temperature object located outside the vacuum space.

BACKGROUND OF THE INVENTION

In recent years, with the advent of microelectronics circuitry and related advances in electrical engineering, many industries have found a greater need to non-invasively measure the

electrical and magnetic properties of materials and devices. The process of magnetic imaging

at high spatial resolution and high sensitivity has been impractical, while low sensitivity or low spatial measurements have been unable to resolve crucial electrical properties.

In the field of semiconductors/microelectronics testing, there is a need to measure the

current flow and image the data relating to the operation of semiconductor/microelectronic devices and their related current paths.

With the advent of magnetic resonance imaging in the field of biology, many new discoveries have been made regarding biological and biochemical subjects. Unfortunately, none

of the current technologies applied in this field provide a very sensitive reading in the picotesla

range at low frequencies, or provide good spatial resolution at high frequencies.

A number of techniques have been developed to image magnetic fields at length scales

of a few μm or relatively smaller. These include decoration techniques, magnetoresistive or

Hall probe sensors, magneto-optic thin films, magnetic force microscopy, and electron beam

interferometry. These techniques have provided limited success and are not practical for high resolution and high sensitivity imaging of fields and flux lines.

Additionally, a number of susceptometers and magnetometers have been proposed using Superconducting Quantum Interference Devices, or SQUIDS. Though previous SQUID

systems have been developed to provide high magnetic field resolution, they are impractical to

implement in a microscope imaging device. The prior art magnetic imaging devices using SQUIDS have had spatial resolution on the scale of a mm or larger which is to crude for microscopically resolving images. These devices may also require placing samples in a

vacuum. Of course, many samples such as liquids and biological specimens cannot tolerate a vacuum. Thus it is not practical to measure sources of biomagnetism which are currently the

focus of much of the existing low-spatial-resolution SQUID imaging work.

U.S. Patent Application Serial No. 08/061 ,102, entitled "Method and Apparatus for Imaging Microscopic Spatial Variations in Small Currents and Magnetic Fields," by Wellstood et al., herein incorporated in its entirety by reference, discloses one such apparatus capable of

providing all of the above discussed measurements with enhanced spatial resolution and

magnetic field sensitivity. However, the device still requires placing a sample within a dewar,

which may result in the unwanted destruction of the sample when it is exposed to the cryogenic liquid or vacuum. Even if the sample could tolerate the vacuum environment or cryogenic environment, it is time consuming and cumbersome to introduce a sample into a vacuum space

for imaging. Another shortcoming is the limited size of the samples that can be imaged.

SUMMARY OF THE INVENTION

Therefore, an object of the present invention is to conveniently measure the physical properties, such as the electrical and magnetic properties, of a sample.

Another object of the invention is to measure microscopic physical properties of a sample without destroying the sample.

Another object of the invention is to allow for magnetic and electrical imaging of objects which are bigger than objects now measured by conventional devices.

Another object of the invention is to allow the use of cryogenic sensors for obtaining microscopically spatially resolved images of physical objects of room temperature samples.

Another object of the invention is to generate microscopic spatially resolved images of

the magnetic and electrical properties samples at room temperature.

These and other objects of the invention are obtained by including a thin, stiff, transparent substrate or window within the outer wall of the vacuum space of a dewar and a

cryogenic sensor within the vacuum space and spaced very close distances to the window. This

construction allows for positioning a sample for measurement outside of the vacuum space, at room temperature or higher and for microscopy of physical properties of the sample by monitoring the output from the cryogenic sensor as it is scanned along the surface of the sample.

BRIEF DESCRIPTION OF SEVERAL VIEWS OF THE DRAWINGS

Fig. la is a schematic diagram of the device of invention.

Fig. lb is a fragmentary schematic view of features of the device of the invention,

including a cryogenic sensor within the vacuum space of a dewar and a thin transparent window in the outer wall of the dewar.

Fig. 2 is a magnified plan view of the encircled portion of the device of Fig. lb. Fig. 3a and 3b are schematic views of a SQUID;

Figs. 4a through 4c show, in sequence, stages in the manufacture of a SQUID and sapphire point cold-finger, which is an example of the cryogenic measuring device used in the

present invention.

Fig. 5 shows a top view of the stage and mechanism for moving the stage upon which

the sample is placed for measurement.

Fig. 6 is a schematic representation of a preferred embodiment of the invention. Fig. 7 is a diagram showing the scanning pattern used by the control program. Dashed

lines show the paths of the SQUID (relative to the sample) during a scan.

Fig. 8 is a hierarchical format for data set.

Fig. 9a is a photomicroprint of the fine ink pattern around the portrait on a $100 bill. Fig. 9b shows the magnetic-field image with fields ranging from 500nT (black) to

500nT (white). Fig. 10 is a vertical slice through the magnetic image shown in Fig. 9B indicating a

spatial resolution of 50 μm.

Fig. 11 is a photograph of printed circuit board with arrows indicating current flow

(100 μA) in wires.

Fig. 12 shows the static magnetic field image of current flow in printed circuit board

of Fig. 11.

Fig. 13 shows a 49 KHz eddy -current image of a lap-joint sample.

Fig. 14a shows a drive coil arrangement used for the eddy-current detection of

subsurface cracks in conductors.

Fig. 14b is a side view of 14a.

Fig. 15 is a photograph of wire bent into a meander pattern carrying an alternating

current at 400 MHZ.

Fig. 16 shows the radio frequency image of a 400 MHZ current flowing in the wire

meander of Fig. 15.

DETAILED DESCRIPTION OF THE INVENTION

In a preferred embodiment of the invention as shown in Fig. 6, the apparams of the

invention is composed of a modified dewar, a cryogenic sensor, a stage, and a computer and

its associated software and electronic connections to the stage to maneuver the stage.

A major portion of the device of the invention shown in Fig. 1 includes modified dewar

assembly 15 having a vacuum space maintained at about IQ ⁵ Torr containing a cryogenic sensor

and a cryogenic space containing about twenty liters of a cryogenic liquid.

In particular, dewar 15 shown in Fig. 1 is a modified commercial stainless steel liquid

nitrogen dewar having a cryogen-containing portion 16 for receiving and holding liquid

nitrogen 17, and a vacuum space 18 which thermally insultates the cryogen from room

temperamre. The modification consists of removing a portion of the "superinsulation" 20 of

the commercial dewar and replacing the removed portion with an assembly 22 that includes

outer walls or housing 24, which defines the outer boundary of the vacuum space of the dewar

and for maintaining a vacuum. Housing 24 is welded to the original dewar. Housing 24 includes: an annular plate 26, having a circular opening 27 in its center, and located radially outwardly from central opening 27, on top of the annular plate, is circular channel 49; a

transparent and thin substrate or window 28 located below and spaced apart from annular plate

26; metallic bellows 29, connecting structures 30, plastic flange (ULTEM, a nylon composition) 31 and glass slide window support 32, all of which are strucmres defining the distance between annular plate 26 and window 28 (see Fig. 2).

The modification to dewar 15 also includes bracing assembly 40. Bracing assembly 40

is composed of three footers 41 arranged in a triangular configuration relative to each other, grommet 43 having an annular exterior flange 44 and an annular interior flange 46. The bottom

ends of footers 41 are bolted on ring 45 seated in channel 49. Good mechanical contact between ring 45 and channel 49 is achieved by tying nylon bristles 42 around ring 45. This construction also ensures a weak thermal contact between the bracing assembly 40 generally

maintained at 77 °K and the room temperature plate 26 and window 28. Bracing disk or plate

47 located approximately one inch from plate 26 and parallel thereto has three holes 48 for

receiving complementarily threaded footers held in place by nuts. Plate 47 is connected to the grommet by bolts to exterior flange 44.

The annular plate 26 as shown in Fig. 1 is also connected via threaded rods 60 to

horizontal adjustment annular disk 62 with adjustment screws 63 positioned as shown. Vertical

adjustment nuts 64 on rods 60 are positioned on each side of the annular plate and allow fine movement of window 28 with respect to SQUID 72.

The final modification to dewar 15 is cryogen delivery system 50 that includes stainless

steel bellows 52, copper or brass tube 54, and thermally conducting substrate 56. Stainless

steel bellows 52 is sealed to and is in open communication, at a first end, with the cryogen-

containing portion 16 of dewar 15. The second end of stainless steel bellows 52 is located in

the vacuum space 18 and is in open communication with the annular inner space of the grommet

43. This second end of bellows 52 is seated on the top of interior flange 46 of the grommet.

The purpose of this assembly is to cool the cryogenic sensor SQUID 72 to its operating temperature while simultaneously holding it rigid with respect to the dewar and minimizing the

effects of thermal contraction in dewar walls 16.

The first end of copper tube 54 is received in the bottom inner annular space of the grommet 43 and is seated and soldered on the bottom of interior flange 46 of the grommet and

is in open communication with the annular space of the grommet. Tube 54 extends through vacuum space 18, and through opening 27 in annular plate 26. Located in the second end of

tube 54 and fastened thereto with epoxy is one end of thermally conducting substrate 56 which is rod-shaped. The second end of the rod-shaped substrate is fabricated with a blunt end point,

to which is fastened sensor 72 positioned within an adjustable distance and preferably within a few microns of window 28 the adjustment being provided by adjustment rod 60 and nuts 64.

Specifically, the distance between sensor 72 and window 28 may be as great as 2-3 mm or there may be no distance between them when they are touching. This construction allows for maintenance of sensor temperature at 77 °K while allowing for minute separations between the

sensor and a room-temperature sample.

It is pointed out that at room temperamre the stainless steel bellows 52 exerts a force on bracing assembly 40, holding it in good mechanical contact with annular ring 26. When the

cryogenic liquid is introduced into die dewar so that the liquid passes through the stainless steel bellows 52 and the copper mbe 54, dewar 16 will contract but the position of the point will be

unchanged because bellows 52 will stretch . Tube 54 will also contract, but this movement is

countered by the contraction of footers 41, which are in heat exchange contact with the cryogenic liquid through connection to plate 47 and copper grommet 43.

Of course, by constructing the footers of zinc and other metal parts of copper, brass and

stainless steel as described or of other metals, and noting the physical expansion and contraction

properties of such metals, thermal expansion and contraction of the parts can be anticipated such that the sapphire rod and cryogenic sensor do not move at all or move very little upon cooling or warming of the dewar. Correction of any such movement of the sapphire rod in the

vertical direction can be accomplished by hand, by adjusting vertical adjustment nuts 64. Lower bellows 29 allows for such movement.

As indicated above, an improvement of the invention is the microscopic imaging of a

room temperamre object or sample located outside of the dewar with the ability to bring a

cryogenic sensor within microns of the sample. To this end, the apparatus of the invention includes a scanning sample stage 90 that is located outside the dewar which is a considerable simplification as compared to stages used in previous scanning SQUID microscopes. Such

stages required precision engineering for thermal contract and careful design to overcome the

lack of lubrication in cryogenic temperamres. Notwithstanding the temperature advantage, the requirements of the scarining stage used in the present invention are similar, for instance, to the

stage disclosed in U.S. Serial No. 08/061,102. In particular, the mechanism should have a 1

μm positioning accuracy; it should ideally be non-magnetic and non-metallic, and preferably the stage is motorized. In addition to the capability of moving in the x-y directions, stage 90

should also have capability of moving in the vertical direction in order to raise or lower the

sample to the window of the invention and allow for easy insertion of samples. A vertical translation stage similar to the height adjustment mechanism used in a standard optical

microscope can be used. The vertical translation stage can be mounted above or below the x-y

scanning stage allowing distances of between 2 inches to μm lengths between a sample 91 (see

Fig. 2) on the stage and window 28. In fact, the sample can be touching the window. The vertical translation part of stage 90 used in the present invention is commercially available and

was purchased from Edmund Scientific Co. as part no. J3608. Although this is a metallic stage, a one-inch thick plexiglass stand-off or spacer is placed between a sample and stage 90

to prevent undesirable magnetic interference from the metal of the stage to the cryogenic sensor. Also, limited success in removing residual magnetization was achieved by degaussing

the steel components in stage 90 using a bulk magnetic tape eraser.

The stage 90 and stepper motors 92 for driving the x and y axes of stage 90 are shown

in Fig. 5. To automate the scanning process, motors 92 are used to drive the scanning x-y stage.

Unfortunately, undesirable magnetic fields produced by the motors can easily couple into the

SQUID because there is virtually no magnetic shielding between the SQUID and motors 92. Therefore, it is important to mount the motors as far as possible from the SQUID (about 50 cm) and to envelope them in eddy-current magnetic shields such as 1.5 mm-thick aluminum box 94 as shown in phantom in Fig. 5, which provides magnetic shielding above about 1 kHz.

Motors 92 are mechanically coupled to micrometers 95 and single shafts 96 (x-axis) and a spline 97 coupled to and a right-angle coupler 98 (y-axis) through a 10: 1 reduction right angle

gear box 99. This design allows one to easily change the separation between motors and stage

by simply extending shafts 96. Although magnetically noisy, stepper motors and microstepping drives are used because they provide excellent positioning accuracy.

The CPU with controlling software and peripherals for operating the motor, is shown

in the block diagram of Fig. 6.

In greater detail, thermally conducting substrate 56, in a preferred embodiment, is a

sapphire rod, one-inch long with a 0.25 inch diameter. Sapphire has a large thermal conductivity at low temperamres (at 77°K, about 10 W cm 'deg ^"1 ). Substrate 56 supports the cryogenic sensor 70 which, in a preferred embodiment is a SQUID chip 72 (see Figs. 3a and 3b).

SQUID chip 72 consists of a single 200 mm-thick layer of YBa,Cu ₃ O ₇ as known in the art, is deposited on a 500 μm thick, 10mm by 10mm SrTiOj 24° bicrystal substrate using pulsed layer deposition. See, for instance, R. Gross et al. , "Low Noise Grain

Boundary Junction dc SQUID," Appl. Phys. Lett. Vol 57, p. 727 (1990), herein incoφorated

by reference. The SQUID sensor 70, as shown in Fig. 3a, has a square washer shape with an inner hole size of about 20 μm and an outer size of 60 μm. This geometry gives a measured effective magnetic pick up area of approximately 1.33 x 10 ⁹ m ² . Gold contacts 77 are deposited

onto the chip as shown.

Since the SQUID is small it is difficult to handle. The small chip requires a special mounting procedure which is shown in Figs. 4a through 4c. Once a working SQUID is

obtained on a SrTiO ₃ chip, the chip side of the sensor is epoxied to an end of the one-inch long,

0.25 inch diameter sapphire rod using STYCAST 2850 FT epoxy, creating the strucmre shown in Fig. 4a. To ensure an adequate bond the mating surfaces of these structures were also etched. The SrTiO ₃ epoxy and sapphire are then ground away using a diamond-grit polishing

wheel leaving a 800 μm-diameter tip at the end (see Fig. 4b). Additional epoxy may be used

to coat exposed edges. The tip consists of the disk of SrTiO, containing at least one SQUID

and gold contacts. To make the electrical contact to the surface of this chip, three silver

contacts 78 about 200 nm thick are deposited over the edges of the chip and down the side of

the sapphire rod (see Fig. 4c).

As is known in the art, SQUIDS are usually operated in a negative feedback loop or

flux-locked loop.

To couple magnetic flux into the SQUID for maintaining a flux-locked loop, or for applying the read-out flux required for other imaging schemes, a simple three-turn coil 80 was

wrapped around the sapphire rod as shown in Fig. 2. A mutual inductance of approximately

0.24 pH between the SQUID loop and the coil was measured. It is determined that mutual inductance can be increased by fabricating an electronics squid output feedblock coil 80 directly on the SQUID chip using photolithographic printing techniques known in the art.

A field coil 82, for applying a magnetic field to a sample, has a diameter perpendicular to the longitudinal axis of the sapphire rod and is shown in Fig. 1 , and may be wired to electronic measuring or controlling equipment as schematically shown in Fig. 6.

As discussed above, the design of window 28 separates the SQUID chip 72 which is in vacuum space 18, from a sample 91 which is in air located outside of dewar 15. In order to

obtain images of physical properties with spatial resolutions as fine as 50 μm, the separation of SQUID and sample should not be great and should be within 50 μm. To achieve a better

spatial resolution, the SQUID must be smaller and closer to the sample. In order to accomplish this result, the window 28 must be thin and at the same time sufficiently broad to accommodate

the SQUID chip, which must be positioned within a few μm of the window. The minimum

width of the window depends on the width of the blunt end of the rod 56. In addition, window 28 should be stiff so as not to flex substantially under an atmosphere of pressure. That

is, the window will flex less than its thickness under one atmosphere; flexing of the window

will necessarily increase separation between the SQUID and a flat sample. The window must also maintain a vacuum and it must be chemically inert, non-conducting and non-magnetic so that it does not interfere with the SQUID or react with a sample. Additionally, because it is

very likely that the SQUID sample may accidentally contact the window, the window must be

durable and must tolerate repeated contact with the sample. Finally, it has been noted that the window should preferably be transparent to the human eye (or to infrared or UV radiation

which can be viewed with appropriate imaging systems) to help align the window SQUID and sample prior to a scan of the sample.

It has been found that transparent material having a Young's modulus of about 70 GPa

to 670 GPa should be used. Preferably, this is a single-crystal sapphire (A O ₃ ) which is

transparent and has a Young's modulus of about 50 x ltf psi.

Some materials which are suitable for windows include plastic, diamond, metallized films, MgO, SiN, LaA/O ₃ and combinations thereof, as well as other materials. In addition, as discussed above, such materials should also be transparent. Window 28 is built by creating

a window frame that includes drilling a conical hole in a 1.25 mm thick glass microscope slide

32 using a silicone-carbide tool. The diameter of this hole is lmm on the sample side and 3mm in diameter on the SQUID side (see Fig. 2). The window frame may be made from any

suitably stiff material including epoxy, glass, sapphire, diamond, etc.

Thereafter, a 25 μm-thick, 1 cm x 1 cm single crystal of sapphire 30 is epoxied to the sample side of the glass slide 32 to form the lmm diameter, 25 μm-thick window 28 according

to the following procedure. Having obtained this small window, it is waxed and adhered to a

small glass holder. The window is then epoxied to the window frame 32 while it is still waxed

to the holder. Once the epoxy hardens, the glass holder is removed by boiling the whole assembly in water to melt the wax. The total time in water should be kept to a minimum, because the water tends to temporarily soften the epoxy. The remaining side of the glass slide

is epoxied to plastic flange 31 (see Fig. 2) which is then mounted to the dewar assembly through connecting strucmre 30 as shown in Figs. 1 and 2.

When setting up the microscope for imaging, the alignment of the window with respect to the SQUID is critical. Alignment is accomplished by using an optical microscope and a

mirror to look directly through the thin sapphire window at the SQUID. The roughest approach is to simply move the window by adjusting nuts 64 or screws 63 until it touches the SQUID.

With moderate relative humidity, it is possible to observe water condensing on the window when the SQUID is contacting the window. This is helpful for leveling the window with

respect to the SQUID chip since the fog on the window indicates the location of contact between the SQUID and window.

It is also possible to detect thermal contact between the SQUID and the window by

simply observing the SQUID voltage on an oscilloscope when oscillating flux is being applied

to the SQUID. The degradation of the SQUID performance is quite sudden and significant when SQUID contact is made. Once the window is leveled with respect to the SQUID, the

sample and scanning stage must be leveled with respect to the window. Leveling of the sample

with respect to the plane of motion of the stage ensures that the separation between the sample and SQUID does not change during a scan. Leveling of the stage with respect to the window

is necessary for achieving a small separation because of the relatively broad (1 cm) glass slide on which the sapphire window is mounted (see Fig. 2).

To obtain an image of the physical properties of a sample, individual raster scan lines

are acquired by scanning the sample past the SQUID in the x direction while simultaneously recording the x coordinate and the relevant voltages (static field signal, rf-field signal, eddy-

current signal, etc.) from the SQUID read-out electronics. This is repeated for the sequence of y values to construct the whole image.

The position of the stage is determined exclusively from the stepper-motor position. The control program of the computer 100 can read the stepper-motor position directly from the motor controller board which is mounted in computer 100. However, for proper

synchronization of the x coordinate with the SQUID signal during a scan in the x direction, it

is necessary to provide the data acquisition system with a voltage signal which is proportional to the instantaneous x position of the stage. This is accomplished using an external counter

circuit which simply keeps track of the number of motor steps and adds this to, or subtracts this

from, a position counter depending on the motor direction. The output from this counter is converted to a voltage level by using an integrated digital-to-analog converter. Hence, the

counter simply functions as a position-to- voltage transducer.

Both the SQUID output and the position of the stage are read using an analog-to-digital converter and recorded using a personal computer. The personal computer also controls a scanning operation which generates the grid of position coordinates. Once a data set has been

acquired using a control program, it is converted into an image. In its raw form, the image

data consists of a set of "N" line scans (y-values) with the ith line containing a set of IS/i, data points each having an x coordinate and one or more associated voltage values. To provide an

image, this data is first spatially regularized, i.e. , linearly inteφolated onto a rectangular spaced grid. Then an image rendering program is used to assign a color, or a level of gray to

each grid point.

The sample stage 90 can be moved by manually operating drive screws and a very

simple data acquisition program can be used to record the position of the sample stage using potentiometers attached to the x and y drive screws on the microscope while simultaneously recording the SQUID output. A second program to convert the stream of positions and values

into a set of values on a rectangular grid can be used and finally each value in the grid can be assigned a color and an image can be displayed. This procedure, especially scanning the sample manually, is quite time consuming and

monotonous. Accordingly, a control program has been written and is used to operate the stage,

record the SQUID output and then display an image.

In essence two stepper motors are operated using a controller board which is physically

mounted in computer 100 along with a multifunction input-output (IO) board which is primarily

used as an analog-to-digital converter (ADC) for reading the output from the SQUID

electronics. Hence, both motion control and data acquisition are accomplished using a single personal computer.

Since the SQUID is basically a point-like probe, it must be scanned in a raster pattern to form an image. The raster pattern consists of a series of lines in the x direction at different

values of y, stacked together to form an image.

To take an image with the microscope, the computer first positions the sensor 70 to the

"home" positions = x _san - Jc _oveιshoot , y = y,, where x _stan corresponds to the left edge of the image area, y, is y coordinate of the first scan-line and x _OVCT3/1O0 , is the hysteresis length in the JC scanner.

(see Fig. 7). It is necessary to overshoot the left edge of the image area by ^^ prior to

scanning each line to eliminate the backlash in the scanning mechanism. Next, the x position

of the SQUID is increased at constant velocity (to the right) until x = ς _tart at which point the

computer begins recording the x position and the SQUID signal V _oul . If needed, additional

channels of data, up to the limits of the IO-board, can be acquired at the same time. The data

is continuously sampled until x = x _en at which point the data acquisition for this scan-line ends and the single scan-line is written to the data file. At this point the x position of the SQUID is

"rewound" to x = x _start - _overshoot , and the y- position advanced to y-y ₂ in preparation for acquiring the second scan-line. This procedure is repeated until all N scan-lines are acquired and written to disk.

Once all N scan-lines have been written to the data file, they must be converted into an

image. To accomplish this, the individual scan-lines must be spatially projected onto a

rectangular grid. While the raster lines at y,, y ₂ . . . , y _n are evenly spaced in the y direction, the data points in the x direction are not regularly spaced with respect to each other. That is,

while the data points in any given scan-line may be uniformly spaced, they may be shifted with

respect to an adjacent scan-line. Hence the program must line up the x values in all the scan-

lines by defining an evenly spaced grid in the JC direction and then computing V _oul at each of these points by linearly inteφolating the data from the raw scans.

Once this is done, all that is left to do is to convert the resulting uniform grid of values

into an image by assigning colors or shades of gray to each value in the grid. Various commercially available computer programs can be used to do this. One such program is called "Transform," produced by Spyglass, Inc. in Champaign, Illinois.

An important capability of this software is the ability to select an arbitrary "scan"

variable and "raster" variable. While in the description of the scanning procedure above, x was chosen as the "scan" variable and y as the "raster" variable, with this more generalized

approach to selecting the independent variables, it is equally acceptable to scan the y coordinate and treat x as the raster variable. Furthermore, other parameters for the scan coordinate could

be chosen. For instance, the computer can also control the voltage output from a digital-to-

analog converter in the IO-board which can subsequently be used to set the frequency of an rf

source used for, say, driving the sample. Hence, it is sufficient to produce an image containing frequency versus x by choosing the frequency of the rf source as the "scan" variable and the

position of the x motor as the raster variable. Obviously many other combinations are possible

as well.

Since the "scan" and "raster" variables can be chosen independently, it is naturally possible to use the same program with different microscopes and sensors. This is done by simply choosing the scan and raster variables for the appropriate microscope.

Often it is useful to program a series of image acquisitions to occur under program control. For instance, suppose one would like to take an image, change a parameter, take

another image, change a parameter again and continue, thus producing a sequence of images suitable for a movie. To address this, the program has a variety of prograrmning features which

allow the acquisition of images without operator intervention. This is accomplished by organizing the multiple images into a single "set" which consists of an arbitrary number of

"scenes. " Each "scene" can contain an arbitrary number of "frames." And each frame represents a single image, composed of multiple scan-lines of data points (see Fig. 8). The parameters for a particular "scene" determine which variable is used for the "raster" variable

and which for the "scan" variable. Also the 'scene" determines the values for x _sιart , x _end , y ,

y ₂ , ..., y _N , scan speeds, data sampling rates, and most other imaging parameters. So, if one wants to acquire a number of virmally identical images with only, say, a single parameter

change between images, a single "scene" will be used with multiple frames. The "set" is

simply a sequential ordering of any number of different "scenes" to be acquired. With this

strucmre, it is possible to program virmally any combination.

An important part of data acquisition is recording the operating parameters of the system when the data is taken. In order to assure reproducibility, all relevant parameters must be

saved. With this program, this is done by using a "document" type interface for data sets. After a data set is acquired it is held in a temporary buffer area. The "set" can then be save along with all the relevant parameters in the data acquisition program which were in effect

when the data was acquired. At a later time, this data set can be opened and read back into the program thus returning all the parameters in the program including the buffer file to the state it was in when the data was originally taken. Hence, each of these saved data sets represents

a snap-shot of the state of the program after imaging and hence contains a complete record of

the imaging parameters. This saves a great deal of time usually spent writing down parameter

values in a notebook.

To demonstrate the ability of the microscope to image static magnetic fields, a

ferromagnetic sample and a sample carrying a dc current were imaged. In each case the

microscope is operated in a flux-locked loop.

Figure 9(a) shows a photographic image of the fine printing (microprint) around Benjamin Franklin's portrait on a $100 U.S. Federal Reserve Note. Figure 9(b) shows the

corresponding static magnetic field image of the same region. The ink in this sample is

ferromagnetic and hence produces a substantial magnetic signal. To achieve the best spatial

resolution, the sample was scanned in direct contact with the sapphire window. This is possible

because of the flatness of the sample and the hardness of the window material. The magnetic

fields in the image range from -500 nT (black) to 500 nT (white). These field variations are

about 1000 times larger than the noise threshold of the image and so this sample does not

necessarily demonstrate the field sensitivity of the instrument. However, the small feature size

in the sample does provide a good test of the spatial resolution.

The spatial resolution of the instrument can be deduced from Fig. 10 which shows a

vertical slice (from bottom to top) through the magnetic image in Fig. 9(b) along the line

indicated by the triangle. By measuring the full width at half maximum of the shaφest peak

in this and similar slices, a spatial resolution of about 50 μm is deduced. The data indicates that

the separation between the SQUID and the surface of the sample should be less than 50 μm. This is about a factor of 30 times smaller than any other previously published SQUID-based

system which can image room-temperature samples in air.

In an environment where the microscope would be used for nondestructive testing

applications, the quasistatic fields are more likely to be produced by flowing currents.

Figure 11 is a photograph of a small portion of a printed circuit board showing a number of

copper interconnects and solder points. A current of 100 μA is flowing in the wires as

indicated by the arrows.

Figure 12 is a magnetic image of this sample obtained by scanning at a separation of

about 200 μm between the window and the surface of the circuit board. While the spatial

features in this image are not difficult to resolve, the fields produced by the currents, which

range over about 80 nT, are considerably smaller than before. Hence, it was necessary to first

obtain a background image without the sample, then subtract this from the raw magnetic image

of the circuit board.

The image clearly indicates which conductors are carrying a current and which are not.

Also, by application of the right-hand rule, it is possible to determine the direction of current

flow. Furthermore, magnitudes of the currents can also be extracted in principle by modeling

the field produced by a current / in one of the interconnects. Because the data used to generate the image is quantitative, a fitting algorithm could be utilized to find /. Even in the case where

the density of interconnects causes significant overlap of the fields in the image, more advanced

deconvolution algorithms can be used to extract the directions and magnitudes of the currents.

Note that circuit boards like the one shown inevitably contain magnetic contamination

in the form of small particles of steel, resulting from handling or machining. One such particle

110 is seen in the lower right corner of Fig. 12 where it produces a characteristic dipole

signature. However, even if a contaminant 110 does not have a strong magnetic signature, it

may also be detectable when it is located in a conducting pathway where it could degrade circuit

performance. For example, note the distinctive effect of the current flowing across the solder

point 112 near the bottom of Fig. 12. The perturbation of the current by the hole alters the

field as is easily seen in the image. While the hole is rather larger in this case, the clarity of

the image suggests that much smaller nonuniformities in current flow could be detected. This

technique could ultimately be use to detect small voids or particulate contamination in critical

conducting pathways, possibly beneath the surface, which would otherwise not be apparent until

after the circuit failed.

In some cases, it is not possible or practical to directly inject currents into a sample for

imaging. However, by applying an alternating magnetic field to the sample, alternating eddy

currents can be induced in the sample for probing defects.

To further illustrate the capabilities of the microscope, eddy -current images of a variety

of metallic samples, including an aluminum "lap-joint" assembly as used in aircraft fuselage

construction were tested. A lap-joint is where two sheets of aluminum skin are joined together

and riveted to a support strut. The detection of defects in lap-joints is the goal of some of the

existing eddy-current NDE efforts.

A 14 -scale model of an aluminum (resistivity =0.6 μ cm at 77 K) lap-joint was

constructed. In the scale model, an upper and lower sheet of 0.2 mm-thick aluminum are

jointed to a 0.8 mm-thick support strut by means of 1.6 mm-diameter rivets with 2.5 mm-

diameter heads. The rivet heads are made flush with the sample surface as is done in real

aircraft construction. A crack 114 which extends to a radius of about 3 mm from the center

rivet was placed in the bottom layer of the skin so that it is not visible from either the top or

bottom of the assembled sample. The crack was formed by shearing the metal and then

flattening it again before riveting the sample together; this leaves a very tight crack with no

gap.

Figure 13 shows a 49 kHz eddy-current image of this sample. To obtain the best

results, a linear drive wire oriented in the plane of the sample to induce eddy currents, as

shown in Fig. 14, was used. This induces eddy currents which are strongly perturbed by the

geometry of the crack. In addition to the five rivets, the buried crack which extends diagonally

from the center rivet is clearly resolved in this image. Other strucmre is also visible in this

image. For example, the image indicates the presence of the thicker support strut in the region beneath the rivets.

Since many electrical circuits operate at high frequencies, another potentially important

imaging technique involves imaging the rf fields produced by rf currents flowing in a sample.

For example, silicon-based high-frequency circuits are not designed to function at 77K. The

ability to examine room-temperature samples makes it possible to apply the rf imaging capabilities of a scanning SQUID microscope to this class of samples.

To test the microscope's rf-field imaging capabilities on a room-temperature sample, the rf fields in the vicinity of the wire-meander sample shown in Fig. 15 were imaged. A 400 MHZ rf current is driven in the wire using an rf voltage source. Figure 16 shows the rf image

of the 400 MHZ fields produced over the surface of this sample. The brightest regions

correspond to an rf-field magnitude of about 200 nT. The darkest regions, including the lines

corresponding to locations of the wire, are where the z component of the field is zero. This image emphasizes the point that, when configured for imaging rf fields, the microscope is only

sensitive to field magnitude and not field amplitude.

In previous work, it was found that the maximum frequency at which this technique

functioned properly was about 150 MHZ. Beyond this, cavity mode resonances in the SQUID substrate produced artifacts in the images and effectively degraded the spatial resolution.

Clearly, Fig. 16 shows that the bandwidth of the room-temperature microscope is higher for

this type of imaging. The reason for this improvement is that the substrate size is a factor of

4 smaller in the room-temperature system. By using a smaller SQUID substrate, the frequency

at which the lowest cavity mode will oscillate is increased proportionately, along with the

frequency at which the image deteriorates.

While the above images were made using a SQUID, many other types of cryogenic

sensors could be used. The advantage of using other types of cryogenic sensors is that, when

used in the apparams, they will allow the sensitive microscopic imaging of other physical

properties which the SQUID is not sensitive to.

Such cryogenic sensors include: bolometers for imaging microwave, optical, UV, and

infrared radiation; multiple SQUIDS for more rapidly acquiring images; Hall probes for

measuring magnetic fields from samples; simple junction superconducting devices for

measuring microwave and far-infrared radiation; multiple junction superconducting devices for

imaging magnetic fields or microwave and for infrared radiation; Giant Magnetoresistance or

Collosal Magneto resistance devices for imaging magnetic fields; single electron transistor

devices or Coulomb blockade devices for imaging electric fields and charges and dielectronics;

photocathod and photoresistive devices for imaging optical, UV and far infrared radiation,

cryogenic field effect devices (FET's) for imaging electric fields, and 2-D electron gas devices

((2-DEG) for imaging electric fields and magnetic fields.

While the invention has been described with reference to specific drawings and

embodiments, modifications and variations thereof may be made without departing from the

scope of the invention which is defined in the following claims.