System and Method for Uniform Ion Milling

Abstract

A system and method for the precise and uniform material removal or delayering of a large area of a sample is provided. The size of the milled area is controllable, ranging from sub-millimeter to multi-millimeter scale and the depth resolution is controllable on the nanometer scale. A controlled singularly charged ion beam is scanned across the sample surface in such a manner to normalize the ion density distribution from the sample center toward the periphery to realize uniform delayering.

Claims

1. A system for delayering a sample, comprising: a. a sample stage supporting a sample contained in a vacuum chamber; b. an ion beam source having at least one ion source generating a generally columnar ion beam; c. a deflector diverting said charged ion beam from an initial beam path to raster said ion beam in at least one direction along a surface of said sample over a raster range; and d. a control circuit in electronic communication with said deflector and said ion beam source modifying at least one of: accelerating voltage, ion flux density, spot size, raster rate, raster dimension and dwell time to provide a time-averaged flux of ions per unit area as a function of position and time on said sample to realize substantially planar material removal.

2. The system according to claim 1, wherein said ion beam source is selected from a gaseous type or liquid metal type.

3. The system according to claim 1, wherein said ion beam spot size is in the range of 50 μm to 5 mm.

4. The system according to claim 1 wherein adjustment of said ion beam creates a uniformly milled sample surface independent of sample composition and layer geometry.

5. The system according to claim 1 wherein the ion beam source is arranged to emit the ion beam substantially parallel to the stage and sample.

6. The system according to claim 1, wherein the axis of said ion beam is substantially parallel to and offset from the planar surface of said sample.

7. The system according to claim 5, wherein said deflector directs said ion beam from a path substantially parallel to said planar surface of said sample to a path intersecting said planar surface.

8. The system according to claim 1, wherein said deflector causes said ion beam to move selectively across at least a portion of said sample surface.

9. The system according to claim 8, wherein said raster range of said ion beam is selected from one of: (i) substantially the entire surface of the sample or larger; and (ii) a limited area within the surface of the sample.

10. The system according to claim 1 wherein said sample stage can be adjusted in at least one of the X, Y, Z, and rotational directions.

11. The system according to claim 10, wherein said adjustment in said rotational direction is in the range of 0 and 100 rpm.

12. The system according to claim 1 wherein said sample stage can be temperature regulated between ambient and cryogenic temperatures.

13. The system according to claim 1, further comprising at least one detector for collecting data from said sample, said detector in electronic communication with said feedback circuit, said detector selected from the group consisting of an optical microscope, a CCD camera, a CMOS camera, scanning electron microscope, energy dispersive spectrometer, secondary ion mass spectrometer and an auger probe.

14. The system according to claim 13, wherein said detector provides in situ feedback relating to at least one of the topography of the sample surface and the depth of material removal across the surface of said sample.

15. The system according to claim 14, wherein a scanning electron microscope is configured to vary its accelerating voltage to derive information from the sample surface and sub-structure.

16. The system according to claim 13, further comprising a secondary electron detector for detecting relative beam position and characteristics of said sample surface by impingement of one of said ion beam and an electron beam.

17. The system according to claim 13 further comprising a backscatter electron detector for detecting sample surface and substructure characteristics by impingement of an electron beam operated at varying voltage.

18. The system according to claim 1, wherein said ion beam source further comprises a filament, an ionization chamber, and a focusing lens.

19. The system according to claim 18, wherein said filament is mounted to a cartridge, which is removable from the remainder of the ion source.

20. The system according to claim 1, wherein said milling rate at a given position on said sample is a function of said ion beam dwell time at said position.

21. The system according to claim 1, wherein the ion beam has an incident milling angle range between 0.0 and 15 degrees.

22. The system according to claim 1, wherein the ion beam has an incident milling angle greater than 15 degrees.

23. The system according to claim 1, wherein the rate of change of deflection of said ion beam is variable as it moves across the sample surface.

24. The system according to claim 1, wherein said ion beam has a selectively adjustable energy level between 1 keV to 50 keV, inclusive.

25. The system according to claim 1, wherein said ion beam has a beam current between 1 μA to 1000 μA, inclusive.

26. An isolating device for conveying a flow stream of a gaseous substance between at least two conductive bodies differing in electrostatic potential, said isolating device comprising: a. an insulating tube having an input, and b. a flow restrictor limiting said flow stream to maintain a pressure at said input of said insulating tube to be maintained above a preselected limit to prevent arcing.

27. The isolating device of claim 26, wherein said flow restrictor is an orifice.

28. The isolating device of claim 26, wherein said insulating tube and said flow restrictor are combined as a capillary tube.

29. The isolating device of claim 26, wherein said conductive bodies are components of said ion source.

30. A method of delayering a sample comprising: a. placing a sample on a stage in a vacuum chamber; b. determining the surface topography of a selected layer of said sample; c. determining a thickness, depth profile, and corresponding removal rate for at least a portion of said layer of said sample; d. scanning a controlled ion beam across said surface of said sample to remove at least one of each layer or a selected portion of each layer; e. detecting changes in said surface topography of said sample; and f. adjusting the dwell time of said ion beam as a function of its position on said sample based upon said surface topography, said adjustment causing removal of said material in a substantially planar manner.

31. The method according to claim 30, further comprising selecting a preselected endpoint with respect to at least one of a layer, an interface between two adjacent layers, a portion of a layer, one of a number of layers, a designated thickness, a specific amount of material removed and said sample.

32. The method according to claim 31, wherein said layers of said sample are removed until said preselected endpoint is reached.

33. The method of claim 30, wherein said ion beam comprises ions and neutral particles and said neutral particles are removed prior to impingement of said ion beam on said sample surface.

34. The method according to claim 30, wherein adjusting the dwell time of said ion beam further comprises determining a time-averaged flux of ions per unit area by the formula:
r=a*t.sup.k; wherein: a. variable “t” corresponds to said dwell time; b. variable “k” is a value determined by the relative removal rate of said material with respect to said position on said sample; c. variable “a” is a constant; and d. variable “r” is a radius measured by a distance between a position of said ion beam on said sample and a center of a milling area of said sample.

Description

DESCRIPTION OF THE DRAWINGS

[0041] FIG. 1 is a sectional representation of the resulting delayering produced by a prior art ion beam milling apparatus, showing a concave surface characteristic.

[0042] FIG. 2 is a sectional representation of a delayering result achieved by the present invention.

[0043] FIG. 3 is a schematic representation of the rastering of an ion beam and corresponding ion density distribution along the X axis of a sample surface.

[0044] FIG. 4 is a diagram of the theoretical basis for the algorithm used to control the ion density distribution of the present invention as a function of the milled area radius.

[0045] FIG. 5 is a schematic diagram of the relative position of certain components of the present invention.

[0046] FIG. 6 is a schematic diagram of the present invention.

[0047] FIG. 7 is a schematic diagram of the present invention.

[0048] FIG. 8 is an exploded view of the electron source assembly of the present invention.

[0049] FIG. 9A is a view of the electron source assembly of FIG. 8.

[0050] FIG. 9B is an end view of the electron source assembly of FIG. 9A.

[0051] FIG. 10 is a cross-sectional view of the electron source assembly shown along line 10-10 of FIG. 9B.

[0052] FIG. 11A is an isometric view of the filament assembly and bending tool.

[0053] FIG. 11B is an isometric view of the filament assembly showing use of pins to bend the filament wire.

[0054] FIG. 12 is a partially exploded view of the filament assembly and loading stand.

[0055] FIG. 13A is a top plan view of the filament assembly and loading stand of FIG. 12.

[0056] FIG. 13B is a cross-sectional view of the filament assembly and loading stand shown along line 13B-13B of FIG. 13A.

[0057] FIG. 14 is an exploded view of the ionization assembly of the present invention.

[0058] FIG. 15 is an end plan view of the ionization assembly of FIG. 14.

[0059] FIG. 16 is a cross-sectional view of the ionization assembly shown along line 16-16 of FIG. 15.

[0060] FIG. 17 is an exploded view of the focus assembly and beam steering assembly of the present invention.

[0061] FIG. 18 is an end view of the focus assembly of FIG. 17.

[0062] FIG. 19 is a cross-sectional view of the focus assembly shown along line 19-19 of FIG. 18.

[0063] FIG. 20 is a partially exploded view of the focus assembly of FIG. 18 showing the first electrical input.

[0064] FIG. 21 is a partially exploded view of the focus assembly of FIG. 18 showing the second electrical input.

[0065] FIG. 22 is a partially exploded view of the focus assembly of FIG. 18.

[0066] Like reference numerals refer to like parts throughout the several views of the drawings. The figures are not intended to limit the present invention to the specific embodiment they depict. The drawings are not necessarily to scale.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0067] As shown in the accompanying drawings, the present invention is directed to a system and methods of delayering samples such as IC chips with a rastering ion beam to achieve uniform delayering across the sample surface. As used herein, a layer refers to a uniform thickness of one or more materials laid upon or spread over a surface. The thickness can be equal to one atomic plane up to a few or several microns. The layer may occupy the entire surface or a partial surface. A sample may refer to metals, alloys, semiconductor materials, ceramics, insulators or any other solid materials. A sample may also refer to, but is not limited to semiconductor devices, integrated circuit chips, a layer of metals and dielectrics of any thickness, one or more materials in an area of any size, optical devices, electronic devices, or any combinations thereof. Uniform delayering refers to the removal of one or more layers in a milling process, partially or wholly, wherein the one or more layers or portions thereof may comprise one or more materials; wherein, the one or more layers may be of any desired and uniform thickness.

[0068] Broadly characterized, the present invention provides a system and methods of delayering samples such as integrated circuits and other solid materials on a large area scale. More particularly, embodiments concern an apparatus and method for facilitating the removal of one or more layers from a sample of an integrated circuit with a desired uniform thickness in an area up to approximately 10 mm in diameter, or greater. Embodiments advantageously maintain the planarity of the surface of the sample during delayering of the defined area. Furthermore, a whole delayering process may contain many endpoints at which features, wire patterns, chemical composition, and other interested properties can be analyzed and/or recorded for specification inspection, failure analysis, or 3D reconstructions.

[0069] In a conventional ion beam, ion flux density within the beam typically follows a Gaussian profile; whereby, the center area possesses higher density and the periphery, or tail of the beam, possesses a lower density. When utilizing broad beam ion milling, the center of the sample is therefore milled more rapidly as compared to the periphery, thus precluding the ability to uniformly delayer a semiconductor chip layer.

[0070] Referring to FIG. 1, the delayering result utilizing a conventional ion beam is shown by curve 1 on sample 2. Usually, the milling rate R in the delayering process is determined by the ion density Id per unit milling area A:

R=c*Id/A=c*Id/r*Δr   (1)

where A=r*Δr and r is the radius of a circular milling mark, c is constant.

[0071] The present invention addresses this uneven delayering by creating a relatively small beam with respect to the sample size and scanning or rastering it across the sample surface in a particular and controlled manner to promote uniform milling, shown in FIG. 2, which is independent of both device position and elemental composition.

[0072] Referring to FIG. 2, a milled trench D on Sample 1 is equal to the sum of depths 3, 4, 5 divided by 3.

D=(D1+D2+D3)/3   (2)

Nonplanarity Δd=(|D1−D|+|D2−D|+|D3−D|)/3   (3)

Milling error ΔE=Δd/D   (4)

[0073] The present system 500 and method for delayering samples 1 allows a large area of a sample 1 to be delayered while making the Δd as small as possible, down to a few nanometers, and the ΔE less than 1%, typically 0.1%.

[0074] This technique yields an ion density distribution as shown in FIG. 3. Specifically, an ion beam source 31 emits an ion beam 32, shown as a range in this Figure, onto the surface of sample 1 at an incident angle. The sample 1 rotates along direction 34 during the delayering process. The resulting milled area 33 is therefore circular in shape.

[0075] The speed range of rotation is variable in the range of 0-100 rpm. The direction of rotation can be either clockwise or counterclockwise. The incident beam 32 angle range is 0.0 to 15 degrees. The size of the milling area 33 is adjustable by varying the ion beam 32 scanning parameters, with a typical milled area being in the range of 1 mm to 10 mm in diameter, or greater.

[0076] The ion density distribution along the X-axis may be shown by one of curves 35, 36, 37 or 38. Curve 35 is a uniform density distribution in which the sample center milling rate is high and the milling rate at the periphery is low.

[0077] Milling rate is generally a function of ion dosage or density with a higher density beam producing a higher milling rate. To achieve uniform milling for the delayering purpose, the milling rate across the sample surface needs to be variable to compensate for geometric and elemental composition differences within a chip layer.

[0078] The theoretical density distribution curve 36 in FIG. 3 may not be sufficient to reduce the ion density at the center area. As a result, the milling rate is still higher than that of the periphery. Another possible density distribution is curve 38 in FIG. 3, which may cause too much of a reduction of the beam density on the sample center area; therefore, resulting in a lower milling rate in the center area as compared to the periphery. Curve 37 theoretically has an ion density distribution that produces uniform delayering of the milled area. This curve is unique for given layer characteristics and needs to be determined layer by layer and chip to chip. Because of this factor, it is critical to develop a unique mathematical milling control algorithm such that the corresponding ion density distribution curve 37 is effective in uniformly removing material from layers possessing differing geometries and elemental compositions.

[0079] FIG. 4 shows a diagram of the milling area 41 on the sample surface. The diameter is in the millimeter scale. The goal is to achieve a uniform ion density distribution at every position within the milling area 41.

[0080] To achieve this purpose, a sub-millimeter diameter ion beam is generated and scanned within the desired area and with characteristics to uniformly distribute ion beam density within milling area 41.

[0081] The system thus avoids the drawback of having a Gaussian ion density distribution within the ion beam itself which is a characteristic of broad ion beam technology.

[0082] To achieve the desired function, a sub-millimeter beam is rastered with respect to the sample surface, point by point, with the intensity of the beam being determined by the rate of movement, thus causing the dwell time at each point of the sample surface to be variable. The dwell time At divided per unit milling area A must be the same, or equal to constant C.

Δt/A=C   (5)

or Δt=C*A   (6)

The factor r is a radius 42 within the milling area 41, while Δr is an increment 43 of the radius in FIG. 4 following the equations:

Δt=C*A=C*2πr*Δr

dt=C*2πr*dr

∫dt=C*2πr*∫dr

t=C*2πr.sup.2

r=a*t.sup.1/2 (a is a constant)   (7)

For practical use, it is necessary to extend the equation (7) to the following:

r=a*t.sup.k (a is a constant, 0<k<=1)   (8)

where k is the milling factor. Equation (7) is a special case of equation (8) when k=1/2. When k=1, the ion density distribution is equivalent to curve 35 in FIG. 3. When k is reduced to a value slightly less than 1, the ion density distribution is equivalent to curve 36 in FIG. 3. When k is further reduced, the ion density distribution is equivalent to curve 37 in FIG. 3. A subsequent reduction in k yields an ion density distribution equivalent to curve 38 in FIG. 3. In practice, it is necessary to determine the k value for different material and layer configurations in order to achieve uniform delayering.

[0083] During an ionization and extraction process, some ions can be neutralized by capturing electrons. These neutrally charged particles move purely as a function of their momentum and cannot be steered. Therefore, their impingement upon a sample surface typically results in uneven, uncontrolled milling of the sample surface. To avoid this problem, the present system and method includes a means of specifically steering only charged ion particles. FIG. 5 shows a basic diagram of the present system 500. An initial ion beam 52 is generated and emitted from at least one ion source 31 and is directed to Faraday cup 53. In this condition, the entire beam is directed into the Faraday cup 53 and no milling occurs. In at least one embodiment, the Faraday cup 53 can also be used to measure ion beam current.

[0084] To precisely control the ion beam for uniform delayering, the ion source 31 of the present invention includes Y deflection electrodes which, when voltage is applied, creates an electrostatic field that deflects a portion of the initial ion beam 52 to create an incident ion beam 32 directed off-axis to impinge upon the sample surface 55, as shown in FIG. 5. Because only ions are affected by the electrostatic field, they are directed or steered to create the incident ion beam 32 by Y deflection. Neutral particles are unaffected by the electrostatic field and remain in the initial beam 52 which impacts the Faraday cup 53 rather than the sample. In this manner, only charged particles impact the sample surface 55 for delayering.

[0085] The sample is supported on a sample stage 56 which can be adjusted in height along axis 57 and may be moved in the X-Y directions. The ion source 31 also includes X-X ion beam deflection electrodes which raster, scan or steer the incident ion beam 32 with respect to the sample surface 55 according to Equation (8). Under no deflection conditions, the incident ion beam 32 intersects with the sample rotational axis at a plane above the sample surface 55 and parallel to the axial center line of the ion source 31. The sample surface 55 is perpendicular to the sample rotation axis. The beam scanning direction along the X axis in FIG. 3 must be parallel to the sample surface 55. During delayering, the sample 1 is either fixed or rotatable in the clockwise or counterclockwise direction at an appropriate speed. Rotation normalizes milling across the entire sample surface 55 and compensates for the incident ion beam 32 only being rastered in the X-direction.

[0086] The beam incident angle with respect to the sample surface 55 is determined by the bending signal strength or voltage as it relates to both the electrical potential applied to the Y deflection electrode and the stage height. Utilizing a larger voltage increases the incident angle. Adjusting the sample stage 56 height upward so that it approaches the plane of the initial ion beam 52 without deflection reduces the incident angle. The beam incident angle is normally adjustable in the range of 0 to 15 degrees.

[0087] Turning now to the system 500 overall, as shown in FIGS. 6 and 7, the system 500 includes an ion beam source 31, a sample stage 56, a vacuum chamber 62, a source controller 61 (not shown), a control unit 69 and various detectors. The sample stage 56 is configured and positioned to support a sample 1 thereon, such as but not limited to IC chips, partial IC chips, multiple IC chips, any kind of wafer, or other types of sample material. The sample stage 56 possesses Z-stage height adjustment along the axis 57 as shown in FIG. 5, as well as rotational capabilities and X-Y offset as described previously. Rotation is employed to normalize the effect of the ion milling process on the chip surface. The rotational speed range of the sample stage 56 may be between 0.0 and 100 rpm. The stage 56 may also have tilt capabilities. In some embodiments, the sample stage 56 can be cryogenically cooled to near liquid nitrogen temperature. A thermal heater, temperature sensor, and control electronics allow the sample temperature to be regulated between cryogenic and room temperature.

[0088] The vacuum chamber 62 provides a hollow space in which the ion beam source 31, sample stage 56, sample 1, and various detectors are contained. The vacuum chamber 62 is part of a vacuum system which also includes a primary pump, preferably a turbomolecular type, a secondary or backing pump, valves, drive circuitry and a control system which may be electrical or pneumatic, and is used to create a vacuum or negative pressure condition within the vacuum chamber 62 during use. The vacuum chamber may be mounted to a frame which provides suitable vibration isolation. The vacuum system may include a controller and gauging to output base and operating vacuum levels.

[0089] The sample 1 may be loaded into the system by incorporation of a load-lock assembly possessing both vacuum and inert gas transfer capabilities. With this methodology, the sample can be protected from the environment throughout preparation and subsequent imaging and analysis. This is particularly important for samples which are environmentally sensitive such as, but not limited to, catalysts and lithium ion battery materials.

[0090] The ion beam source 31 is composed of a filament assembly 100, an ionizer assembly 200, and a focus assembly 300, all as shown in greater detail throughout FIGS. 8-21. Specifically, the filament assembly 100, depicted in FIGS. 8-13B, is the primary electron source. It includes mount 110 connected to a filament lid 111 separated by standoffs 112 and springs 113. At least one filament 120 is secured to the mount 110 by a clamp 130 and retainer 140. The filament(s) 120 may be yttrium oxide coated iridium, tungsten, or other suitable materials for producing electrons. Connectors 150 are in electrical communication with the filament 120 and provide electric current to the filament 120 which creates a flow of electrons. Preferably, there is one connector 150 for each filament lead. These filament leads may be bent or formed into shapes by using bending tool 170 in which the filament leads are bent around pins 172, which may be pushed up or down to accommodate wire bending as described more fully with reference to FIG. 11. A stand 180 may be used for loading the filament assembly 100 as more fully described with reference to FIG. 13. A Wehnelt electrode 160 surrounds one end of the filament 120. An electrical bias applied to the Wehnelt electrode 160 draws electrons from the filament 120.

[0091] The filament assembly 100 is physically coupled to the ionizer assembly 200, as shown in FIGS. 14-16. The ionizer assembly 200 includes a Wehnelt support 210 into which the filament assembly 100 is loaded and retained. The Wehnelt support 210 is mounted through a spacer 211 to an adjustment plate 214, which in turn is mounted to an ionizer flange 215 such that the Wehnelt support 210 aligns with an aperture in the ionizer flange 215.

[0092] At least one ion source 220 is included in the ionizer assembly 200 between the terminal end of the Wehnelt electrode 160 and the aperture of the ionizer flange 215. For instance, the ion source 220 may include a G2 electrode 222, also referred to as a G2 aperture, and an E electrode 224, which are separated from one another by an insulator 212. An F electrode 226, or F aperture may also be included among the ion sources 220 and is physically separated from the extractor 224 by an insulator 212.

[0093] The E electrode 224 has two inputs: the receipt of an electrical bias and a capillary gas input 250. The gas input 250 is connected to a capillary tube 252 which is in fluid flow communication with a gas feedthrough 254 for providing the flow of process gas into the ion sources 220. A clamp 255 may secure the gas feedthrough 254 to the ionizer flange 215. The process gas provided by the gas feedthrough 254 may be argon, xenon or any other gases such as inert gases or combinations thereof. The gas flow rate is adjustable in the range of fractions of a standard cubic centimeter per minute (SCCM) to 10's of SCCM and may be preferably around 0.02 SCCM or greater, though other flow rates are also contemplated.

[0094] In a typical gaseous ion source, ionization takes place in a chamber held at a high voltage potential with respect to ground, which creates ions and then provides acceleration of the ion beam to a grounded system, and at a pressure far lower than atmospheric, typically a few Pascal. The supply of gas is usually grounded and at a pressure far above atmospheric. Gas must be moved from a high-pressure state at ground potential to a low-pressure state at high electrical potential. One configuration for gas delivery is a mass-flow controller at ground potential, followed by an insulating tube, usually inside the vacuum. The tube has a large inner diameter so that the pressure drop across the tube is negligible, causing the pressure inside the tube to be only slightly higher than inside the ionization chamber. Most of the pressure drop occurs within the mass-flow controller. Unfortunately, optimal ionization pressure is usually near the minimum of the Paschen curve, so that glow discharges are not easily prevented inside the insulating tube because of the voltage gradient it supports. This generally puts an upper limit on the ionization chamber pressure that can be used, and/or an upper limit on ion beam potential.

[0095] To solve this problem, the present invention uses a capillary tube 252 to deliver the process gas. This capillary tube 252 has a very small inner diameter. In light of the transition from laminar flow prior to the capillary tube 252 to molecular flow inside the capillary tube 252, the pressure profile inside the capillary tube 252 is not linear. Rather, it is nearly flat until close to the end of the tube, with a large pressure drop at the outlet. Consequently, most of the length of the capillary tube 252 is at a pressure much higher than the ionization chamber. With a high pressure inside the capillary tube 252, high voltages can be supported within the ionizer assembly 200 without danger of gaseous discharges inside the tube. This allows essentially arbitrary combinations of voltage, ionization pressure, and flow rate to the ion source 220, resulting in improved performance and reliability.

[0096] Electrical biases are individually applied to each of the G2 electrode 222, the E electrode 224, and the F electrode 226. Electrons emitted from the Wehnelt electrode 160 strike gas atoms in the region surrounding the ion sources 220. An individual collision results in the loss of an electron from a gas atom. Multiple collisions create significant quantities of ions.

[0097] These ions exit the ionizer assembly 200 and are directed into the focus assembly 300, shown in FIGS. 17-21. The focus assembly 300 includes a support plate 310 that physically secures the lens. The open structure of the support plate 310 allows for vacuum pumping of the various ion source 220 regions. A spacer 313 physically and electrically separates an extractor from the ion source 220. An electrical bias is applied to the extractor, which attracts ions emitted from the ion source 220 region. Ions are further controlled by an electrical bias applied to a drift tube 324 and a focus electrode 322.

[0098] The focus assembly 300 also contains a steering assembly 330, which is capable of electrostatically guiding the beam. The steering assembly 330 preferably includes a plurality of individually electrically biased steering electrodes in the form of steering rods 332 which are contained within steering rod insulators 334. Some of the steeling rods 332 are X deflection electrodes while other steering rods 332 are deflection electrodes. The steering rods 332 may be individually electrically biased to allow for the precise steering and scanning of the resulting ion beam with a deflection of approximately +/−15 degrees. A steering rod cover 336 encloses the steering deflector rods 332 and their corresponding insulators 334. The focus assembly 300 also includes electrical inputs 342 and 344 which are positioned along housing 311 of the assembly. An exemplary configuration may include a first electrical input 342 providing 12 kV and up to 20 amps of power. A second electrical input 344 may provide 700V with a current of up to 5 amps. One of these may be used to create the incident ion beam 32 described previously. The other provides the electrical bias for deflecting the resulting ion beam toward the sample and steering the incident ion beam 32 across the sample surface 55, for rastering purposes.

[0099] The ion beam source 31 described above is capable of generating an ion beam having a beam energy in the range of 1 keV-50 keV or greater and may be continuously adjustable with 1 eV accuracy. Higher beam energy can increase the milling rate but result in sample surface damage. Lower energy can produce a smooth sample surface and less damage, but at a reduced milling rate. [0100] 1. The resulting ion beam diameter may be in the range of 50 μm-5 mm and may be adjustable dependent upon the energy level. For example, the beam size may be 50 μm at 10 keV with a current of 0.1-7 μA in some embodiments, and 1.0 mm at 1 keV with a current of 7-15 μA in other embodiments. The ion beam may have a current between 1 μA to 1000 μA, inclusive. Preferably, the current is up to 200 μA. Greater current can increase the milling rate. The current density of the ion beam may be approximately 10 mA/cm.sup.2. The ion source may be operable in a working distance in the range of 10 mm-100 mm with a raster range of up to 12 mm.sup.2 or greater. In at least one embodiment, the raster range may be 10 mm diameter at a 25 mm working distance.

[0101] As can be appreciated from FIGS. 8-21, the modular design of the ion beam source 31 allows for both rapid replacement of the filament 120 and in situ x-y adjustment of the ionizer assembly 200 with respect to the lens.

[0102] The direction of ions emanating from the ion beam source 31 is above and preferably parallel to the sample surface 55, as shown by the initial ion beam 52 in FIG. 5. The incident ion beam angle is adjustable by varying the sample height along the Y axis 57 and the beam Y deflection bending amplitude caused by the steering assembly 300. Using this combination of adjustments, the impingement point of the incident ion beam 32 coincides with the sample rotation axis 34. Utilizing the deflection electrodes of the steering rods 332, the incident ion beam 32 is scanned from edge to edge of the sample surface 55.

[0103] Returning to FIGS. 6 and 7, the system 500 includes a source controller 61 that provides the operative instructions to adjust the ion beam voltage, current, extraction voltage, beam size, gas flow rate, and filament emission. It consists of several high voltage power supplies which control the various assemblies of the ion beam source 31 discussed above, including ion beam energy, focus, ion extraction, as well as the deflection amplifiers for beam rastering in the X direction and deflecting or bending the ion beam in the Y direction onto the sample surface 55, shown in FIG. 5. An adjustable amplitude raster control distributes ion density along the X direction so that uniform delayering of a large area is achieved. The source controller 61 also controls and adjusts the raster rate of the incident ion beam 32 with respect to each location projecting from the sample center to its largest radius.

[0104] The ion beam source 31 can be mechanically moved in the X and Y directions and rotated. Physically changing the X-adjustment without scanning positions the ion beam 32 with respect to the sample rotational axis. Changing the Y movement ensures that the ion source is at the appropriate height with respect to the milling plane. Rotational adjustment without the application of X-X deflection ensures that the ion beam position is parallel to the sample surface when it is affixed to the stage 65. This alignment procedure may be conducted during system set-up and calibration.

[0105] The system 500 may include a variety of detectors that provide information relating to delayering during the milling process. Each of the detectors is in communication with the control unit 69 to provide data to the control unit 69. Detectors provide information relating to chip feature type, feature locations, feature dimensions, and feature chemical composition.

[0106] The system 500 may include a camera 66 and a light source 67 to capture light reflected from the sample surface. Periodically during the milling process, the sample 1 may be illuminated by light source 67. Images are subsequently acquired by camera 66 and sent to the control unit 69. The camera 66 may include various powered objectives for higher or lower resolution imaging. Patterns revealed during delayering may be recognized by image recognition software installed on the control unit 69 for both feature identification and to confirm or adjust the milling process parameters. In addition, the light source 67 may be used to reveal vertical spacing differences between two layers. As milling progresses, interference fringes are developed and correspond to the depth of the milled area. Light reflection and the creation of interference fringes is a very sensitive method for establishing the nonplanarity of the sample surface, even when nonplanarity is on the order of a few nanometers. In an ideal situation, only one interference ring will appear on the image, indicating a very small depth variation between the sample center and its periphery. Increasing quantities of rings correspond to larger depth variations.

[0107] During milling, rings first appearing at the sample center are an indication that the milling rate at the center is greater as compared to the surrounding area. When rings appear toward the periphery of the sample, it is an indication that the center is milling at a lower rate as compared to the periphery. If more rings appear, a large milling rate difference exists between the sample center and periphery. If the width of the single ring is large, the milling rate difference would be small. Reflection images can be analyzed by human eyes or by computer image recognition techniques. Accordingly, the milling factor k can be adjusted manually or automatically.

[0108] The system 500 may also include a secondary ion mass spectrometer or SIMS 64 which can be used to analyze the elemental composition of material sputtered from the milled surface. Sample layers typically possess a combination of light and heavy elements separated by semi- or non-conductive oxide layers, which when sputtered, can be quantified by SIMS. This quantification can be sent to the control unit 69 for the analysis of the chemistry encountered in each layer for the subsequent determination and identification of chip layer components. Chemical composition analysis is achieved at a point, in an area, and by a map.

[0109] As shown in FIG. 7, the system 500 may include an SEM column 75. The SEM column 75 is located directly above the sample stage 56 and generates a beam of electrons that interact with the sample surface 55, yielding secondary and backscatter electrons, as well as x-rays that are subsequently used to quantify depth profile information, provide three-dimensional structural information, and the elemental composition of the sample. A shutter may be positioned adjacent to the SEM objective lens which may be actuated to protect the SEM lens and detectors from the deposition of sputtered particles produced by the delayering process. Acquisition data from different layers by the SEM working in conjunction with the various detectors can be used for the 3-D reconstruction of the sample by creating feature patterns during the delayering process. When a specific feature pattern appears, the delayering process may be adjusted or stopped.

[0110] The system 300 may include a sample stage featuring X and Y translation; whereby, the sample position moves with respect to the impingement point of the electron beam. In cases where the desired resolution of the SEM image restricts the field of view, translating the sample to multiple positions allows imaging of the entire sample surface.

[0111] A secondary electron detector or SED may also be included for imaging electrons emanating from the sample surface excited by electrons from the electron column or ion-induced secondary electrons as created by ions emanating from the ion source 31. A backscattered electron or BSE detector may also be included to detect backscattered electrons from the sample surface. Backscattered electrons are produced by the elastic scattering of a primary electron beam with an atom's nuclei. Varying the accelerating voltage of the incident electron beam causes the electron penetration depth, or interaction volume, to change. Because information is generated by the nucleus of the atom, an indication of its corresponding element becomes known. This is particularly useful when attempting to distinguish material differences between the various layers within the chip.

[0112] The system 500 may also include an energy dispersive spectroscopy or EDS detector 74 which detects x-rays emitted from the sample surface 55 during bombardment by the incident electron beam to characterize the elemental composition of the analyzed volume. Combining BSE and EDS information yields a depth profile of the surface and near sub-surface in the x-y dimensions.

[0113] A wavelength dispersive spectroscopy (WDS) probe, Auger detector, laser profiler, X-ray, and/or other probes may also be used in the system, all of which are not shown in the Figures but inclusion of which would be within the ambit of one skilled in this art.

[0114] The system 500 also includes a control unit 69. The control unit 69 is in communication with all component parts of the system 500 and establishes the pertinent parameters of the system 500. This includes vacuum conditions within the vacuum chamber 62, adjustment of stage movements, and regulation of the output of a mass flow controller to adjust the amount of process gas supplied to the ion source 31. Accordingly, the control unit 69 provides operative parameters to the ion source controller 61 for operating the ion source 31. It also controls the movement/adjustment of the sample stage 56. The control unit 69 includes a user interface for communication with operators of the system. One or more displays may also be provided on which output data from the detectors may be displayed.

[0115] In addition, the control unit 69 receives input from the various detectors and uses this data in the mathematical algorithm to control the ion beam density distribution across the chip surface.

[0116] Various signals can be captured by the detectors point-by-point with the size of the point being roughly the size of either the ion or electron beam. The control unit 69 creates a depth profile map from data generated at the sample center, sample periphery, and at any quantity of points in between. A higher quantity of points translates to a higher resolution map. Depth profile variations with respect to their corresponding position are then input into the algorithm.

[0117] Analysis of the data from the detectors allows for the adjustment of the milling factor k described in equation (8) in order to achieve uniform delayering. For example, when the milling rate at the sample center is higher than on the periphery, the milling factor k must be reduced. If the milling rate at the sample center is lower than on the periphery, the milling factor k must be increased. The amount of adjustment of k described in Equation (8) is variable and is determined by the variation of the milled depth.

[0118] The raster control can be varied to correspond with sample material type, raster amplitude, ion beam energy, ion beam current, and other parameters. Based on this calculation, adjustments to the operative parameters needed to achieve uniform delayering are sent to the relevant system components, such as to the source controller 61 in order to adjust the raster pattern and corresponding positional current density of the ion beam 32. The result is physically altering the ion beam raster pattern in terms of both raster rate and the corresponding current density per point, with the intention of minimizing Δd. Accordingly, the control unit 69 may adjust the milling factor k in near-real time by analyzing the output of the various detectors.

[0119] When the material removal rate for a given layer of a given thickness is identical, the process can be repeated to remove subsequent layers in a stepwise fashion by utilizing identical ion beam control parameters. This method may comprise some quantity of predetermined steps to automatically remove multiple layers from the chip.

[0120] An objective of the system is to precisely control milling rate, combined with determining the process end point. The simplest form of end-point determination is by time; however, for this to be effective, both the layer thickness and milling rate need to be very well understood. The milling time is calculated using a predetermined desired delayering depth because the milling rate is known at specific beam energies for various materials. Once the end point is determined, the control unit 69 stops the process by deenergizing the ion source 31.

[0121] The end point can also be determined by features or specific chemical compositions appearing on the images or analytical data during delayering. If chemical compositions are needed at some specific depth, EDS 74, SIMS 64 or other chemical analysis data is relied upon to identify the end point.

[0122] For system calibration of milling rate over a given area, a multi-layer control sample, with each layer possessing a different elemental composition, should be employed. The layer thicknesses of such sample need to be uniform and on the order of 5-50 nanometers per layer. Thicker layers can be tolerated; however, this increases the overall milling/calibration time.

[0123] Since many modifications, variations and changes in detail can be made to the described preferred embodiments, it is intended that all matters in the foregoing description and shown in the accompanying drawings be interpreted as illustrative and not in a limiting sense. Thus, the scope of the invention should be determined by the appended claims and their legal equivalents.