WO1995027896A1

WO1995027896A1 - Measurement

Info

Publication number: WO1995027896A1
Application number: PCT/GB1995/000801
Authority: WO
Inventors: Bhikhu Ardeshir Unvala
Original assignee: Unvala Limited
Priority date: 1994-04-12
Filing date: 1995-04-07
Publication date: 1995-10-19
Also published as: GB2289338A; GB9407222D0

Abstract

An AC potential drop measuring technique is described, in which frequency (f) is varied continuously over a given range, and the slope of measured AC potential drop over a length of a specimen surface is measured over the frequency range. Over the frequency range, a range of skin depths is obtained, so that a graph of measured voltage against the square root of the frequency reveals the composition of the specimen at various depths beneath the substrate.

Description

MEASUREMENT

The invention relates generally to the technique of AC potential drop measurement, for measuring properties in metals and other conductive surfaces. The invention also relates to apparatus for use in such techniques.

The technique of AC potential drop measurement (ACPD) are well documented, for example in WO 90/01159, and in "Practical aspects of the ACPD technique" published by Matelect Limited of 33 Bedford Gardens, London W8 7EF. Both references are incorporated herein by reference in their entirety. These techniques are known in particular for measuring the depth of surface cracks in metals.

The ACPD technique comprises passing an alternating current through the specimen under test, and measuring the potential drop between two points of contact on the specimen surface, the measured voltage being dependent on the effective resistance of the specimen between the two contacts. By employing alternating current, the current flow can be confined to a very thin "skin depth" adjacent to the surface of the specimen, which leads to increased sensitivity. When a crack or similar defect is present between the contacts, the effective path length for the current is increased, which provides an indication of the depth of the crack. It is known to use different frequencies for ACPD measurements in different metals, in order to obtain optimum sensitivity and noise immunity. The frequency used determines the skin depth (the depth, measured from the specimen surface, to which the majority of the current is confined) , which in turn determines the impedance of the current path and hence the potential drop measurable. Other factors that alter the skin depth, however, are specimen resitivity p and magnetic permeability μ, so that different materials require a different frequency to obtain a given skin depth. The references give the equations relating these factors.

Similar ACPD techniques have also been applied to detect strain in metal surfaces.

According to the invention there is provided a method of determining the structure of a specimen having a conductive surface, characterised by the steps of measuring ACPD values at the surface for a range of frequencies, and determining from the variation of ACPD values over frequency information characterising the composition of the specimen over a corresponding range of depth values beneath said surface.

According to the invention there is further provided an ACPD measuring apparatus having a continuously variable f requency .

The present invention provides a technique in which ACPD measurements are taken for a specimen over a range of frequencies, with the result that information is gained over a range of depths beneath the specimen surface. From the measured ACPD values for a range of frequencies, the specimen composition and/or layer boundary depths can be determined.

In a particular embodiment of the technique, the ACPD values measured over a range of frequencies are plotted on a graph against the square root of the applied frequency. On this graph, points can be identified where there is a discontinuity in the slope of the measured voltage at a certain frequency. This transition represents the boundary between layers beneath the surface of the specimen.

Embodiments of the invention will now be described in more detail, by way of example only, with reference to the accompanying drawings.

Figure 1 is a block schematic diagram of an apparatus suitable for use in accordance with the present invention;

Figure 2 is a graph illustrating schematically the relationship between measured AC voltage, frequency, and the specimen composition in the apparatus of Figure 1;

Figures 3 and 4 show respectively a schematic cross section of a first example specimen, and the resulting graph of measured voltage against frequency; and

Figures 5 and 6 show respectively a schematic cross section of a second example specimen, and the resulting graph of measured voltage against frequency.

In the apparatus of Figure 1, an oscillator 10 is provided, whose frequency f is continuously variable over the range 0 Hz to 500 kHz. Through a current amplifier 12, a current I which alternates with frequency f is caused to flow through a specimen under test 14. By use of a current sensing resistor 16 and a feedback path 18, a constant amplitude of the alternating current I is maintained.

Two voltage sensing electrodes 20 and 22 are provided on the surface of the specimen 14 in the path of the alternating current I. These sense an AC potential drop (ACPD) , which is amplified by a transformer and preamplifier module 24. The ACPD signal is then further amplified at 26 and passed to a phase sensitive detector (PSD) 28. The phase sensitive detector 28 also receives from the oscillator 10 a reference signal indicating the frequency and phase of the applied current I. A variable phase shifter 30 is provided to alter the phase of this reference by a settable amount, in order that the PSD 28 can extract from the ACPD signal a component having a desired phase relative to the current I. In particular, the phase shifter 30 may be controlled by feedback from the PSD 28 in order to maximise the amplitude of the extracted voltage, or for example to extract a component in phase with the current I. Whatever the phase angle selected, the PSD 28 outputs a measured voltage magnitude signal Vm . The magnitude of this signal can be displayed on a digital display unit 32, and is also available as an analogue output signal at 34 and a digital output signal at 36. The digital output may for example comprise an RS 232 serial data interface, which may also indicate the current value of f, I and so forth.

The principle of operation of such an apparatus is known from the references, whose contents are incorporated herein by reference. To summarise, the current I flowing in the surface of the specimen 14 creates a potential drop (voltage) between neighbouring points on the surface of the specimen, which potential drop is detected by the spaced apart contacts 20 and 22. As the frequency f is increased, the current I is increasingly confined to a very shallow surface layer of the specimen, in accordance with the "skin effect", the nominal depth of penetration of the current being referred to as the "skin depth". the skin depth Δ depends upon the resitivity p of the material, and also in particular on the magnetic permeability μ of the material. In the application of such apparatus to crack depth measurement, explained in detail in the references, the confinement of the current to within the skin depth of the surface is exploited, in that cracks in the surface require the current to flow by a longer path, so that the potential drop between the electrodes 20 and 22, for a given spacing between those electrodes, depends upon the crack depth.

Figure 2 illustrates the relationship between the measured voltage Vm and the square root of the frequency f. For a given material and surface geometry (planar surface, cylindrical surface etc.), the graph of measured voltage against square root of the frequency has the form of a straight line of gradient g. As the magnetic permeability of the specimen material increases, however, the gradient g also increases as shown in Figure 2. A material with very low permeability μ, for example aluminium or copper, has a relatively shallow gradient as shown for example at g, . for a material such as iron with a higher permeability μ, a higher gradient such as g is observed. For ease of comparison, all three lines in Figure 2 are shown emanating from a common origin, corresponding to the DC- resistance of the material. Of course different materials will have a different DC resistance.

In conventional applications of the apparatus, for example in crack depth measurement, different permeabilities (hence different skin depths) cause a difference in sensitivity of the technique as applied to specimens of different metals. Therefore, it has been known to provide a selection of operating frequencies f, in order that sensitivity can be maintained. The higher sensitivity when testing high permeability materials is exploited to reduce the operating frequency f, in order to reduce noise effects in particular caused by inductive pick up between the current energizing loop and the voltage sensing loop. One known apparatus provides six frequency settings, for use with a range of materials. In steel, for example, the skin depth Δ at 1 kHz is approximately 0.4 millimetres, while a comparable skin depth of 0.35 millimetres in aluminium requires an operating frequency of 100 kHz.

As mentioned above, in the apparatus of Figure 1, the frequency f of oscillator 10 is continuously variable, or at least variable in many very small steps. In the technique newly described herein, measurements are taken throughout the frequency range from DC to 500 kHz, as a result of which the skin depth can be varied and the ACPD signal measured over a wide range of skin depths. This allows, for example, information to be determined concerning the composition and thickness of various layers of conductive material at and beneath the surface of the specimen 14. Examples of this technique are illustrated in Figures 3 to 6 as follows.

The principle of operation follows from the relevant equations for the ACPD signal, as given in the references. In particular, as mentioned above, the ACPD for a homogenous material or specimen will vary as the square root of the frequency f. If a graph of voltage versus square root frequency is plotted, a straight line will result, whose slope depends on the values of resitivity p and permeability μ. On the other hand, if the material in question is coated with a second material of a different resitivity and permeability, then the graph of voltage against square root of frequency will be generated by the interaction of two lines of different slopes, one corresponding to the substrate the other to the coating.

In Figure 3 a specimen is illustrated in a schematic cross section, having a surface layer 302 of low permeability material (for example copper), deposited on a substrate 300 of higher permeability (iron). Figure 4 shows the resulting graph of measured voltage against square root frequency, over a certain frequency range. Above a certain frequency f_τ, however, the skin depth increases until a significant portion of the current I is in fact flowing in the underlying substrate material (higher μ) . Below frequency f_τ, therefore, the slope of the measured voltage includes a component characteristic of a higher permeability material, and therefore has a higher slope. Therefore, instead of the "copper" slope continuing at lower frequencies, as indicated by the dotted line 402, a discontinuity of slope occurs in the region of the transition frequency f_τ, and the line indicated at 404 in this region has a greater slope. This greater slope combines contributions from both the surface layer and the substrate.

Figure 5 illustrates a second example specimen, having the opposite configuration to that of Figure 3, in that a low permeability substrate 500 is provided with a higher permeability surface layer 502. In this case, the slope of the voltage graph above a certain transition frequency f_τ has a high value, for example as at 600 in Figure 6, while below the transition frequency the slope of the voltage line is less, as shown at 604 in Figure 6. The slope at 604 includes contributions characteristic of both the surface layer and the substrate. A slope at higher frequencies (600) includes only a contribution from the high permeability surface layer, since at such high frequencies the current I is confined entirely within a skin depth which is shallower than the depth of the surface layer.

It should be appreciated that other properties of the materials determine the slope, including in particular the resitivity p. The above description refers only to differences in peremability μ for the sake of simplicity only.

The point on the graph where the two slopes meet is representative of the location of the junction between the coating and the substrate, and hence can be used to obtain a measure of the depth of the coating. It will further be appreciated that current density does not cut off abruptly at the skin depth Δ, but rather the current density decreases exponentially away from the surface, with a characteristic depth known as the skin depth. This means that discontinuities in slope will not be as abrupt as are indicated in the schematic diagrams of Figures 4 and 6, but they will be detectable none the less. A relatively abrupt change in the observed slope will indicate an abrupt change in the material composition, whereas a more gradual change will indicate a more gradual transition in the composition, as would occur for example if significant diffusion of the coating elements has occurred into the substrate material.

For the performance of the technique, the range of frequencies over which the AC frequency f is scanned will depend on the p and μ values of the materials likely to form the substrate/layer system. Analysis of the variation of the resultant voltage against the square root of the frequency can then be performed simply by plotting out the results on a graph, or by using a computer- or microprocessor-based algorithm to identify automatically the position and character of discontinuities. In particular, it would be possible to automate the frequency scanning and voltage measuring process under microprocessor control, either to give a printed graph for visual analysis or to perform automated detection of discontinuities and slopes.

Standard ACPD connection techniques can be used for the apparatus, namely for current delivery contacts and signal pick up contacts. Such techniques are described in the references as will not be detailed further herein, but typically a unitary probe will house two spring-biased current contacts and two spring-biased voltage pick-up contacts housed in a fixed relationship. Alternatively, a method of inducing the alternating current can be used in order to remove the need for the current contacts, and which can also be extended to both inducing the current and using inductive pick up coils to detect the resultant ACPD signal. In this way, a completely non-contacting method of layer depth determination can be obtained, useful for example where a non-conductive coating is provided on the specimen, which it is desired to preserve.

As with the known ACPD techniques, provision can be made to minimise the effect of noise in the measurement system. This is achieved partially through use of the phase sensitive detector 28 supplied with a reference signal from the oscillator 10. Additionally, an unwanted signal known as induced pick-up should also be minimized if undue errors are not to be generated. Techniques for the reduction of induced pick-up are discussed in the references, and will not be detailed further herein. Briefly, however, current focusing, electrical shielding and careful attention to lead positioning are all useful in the elimination of pick-up. Another technique involves measurement of only the in-phase (resistive) portion of the ACPD voltage signal, which is independent of induced pick-u .

Compared with conventional techniques in which frequency is fixed, a further error may arise from the fact that the induced pick-up does not vary with frequency in the same manner as the desired ACPD signal itself. This might impose an overall non-linearity, for example on the graphs of voltage against square root of frequency. A calibration procedure is possible, however, in which the variation in the ACPD signal for a particular geometry of specimen can be predetermined and stored within the measurement apparatus for later recall. This would be achieved using a reference specimen having the desired geometry (for example a simple planar surface) but having a constant bulk composition. Pick-up effects are governed substantially by spaced and signal paths above the specimen surface, and are therefore constant in respective of the composition of the surface, for a given geometry. Therefore, for such a specimen, a calibration curve of ACPD voltage over the entire frequency range can be measured and stored, and later used to normalise the signal from actual multi-layer specimens to remove the pick-up component, either in read time or during post analysis .

In embodiments where the graph of voltage against square root frequency is available in real time, the technique may be useful in the control of layer deposition or surface treatment processes, while these processes are actually being performed. The technique can also be applied in the following applications, although this list is by no means exhaustive. The thickness of surface layers and composition in a case hardened or carburised layer on iron, steel or other alloys may be performed. The thickness and character of a nitrided surface layer of a metal or alloy component or specimen may be determined. The thickness and character of a conductive plated layer on a conductive substrate can be determined, where the plating has been achieved by any common form of plating such as electroplating or electrolesε plating. Of course the techniques described may also be extended to the characterisation of multi-layer systems, where several conductive layers are successively plated (for example in electroplated nickel silver onto steel).

It is also possible to characterise any other conductive substrate and layer or multi-layer system, such as those produced by rolling or explosive welding, where the layers are in electrical contact. Surface changes introduced by corrosion of originally homogenous substrates can be measured, and also surface changes caused by diffusion of other elements into the substrate material, where the material remains conductive. Surfaces and layers may also be characterised that have been altered from their original state by the action of stress, plastic deformation, mechanical treatment, work hardening, heat treatment or physical or chemical diffusion. These changes can be induced in both an originally homogenous substrate, and in systems already consisting of several layers in electrical contact. The technique of the invention is therefore not limited to the embodiments described above, nor to the applications described above. Further modifications and applications will be apparent to those skilled in the art.

Claims

1. A method of determining the structure of a specimen (14) having a conductive surface (302), characterised by the steps of measuring ACPD values at the surface for a range of frequencies, and determining from the variation of ACPD values over frequency information characterising the composition of the specimen (14) over a corresponding range of depth values beneath said surface.

2. A method according to claim 1, characterised in that said determining step includes linearising a relationship between ACPD and frequency.

3. A method according to claim 2, characterised in that said linearising step comprises taking the square root of the applied frequency.

4. A method according to any preceding claim, characterised by the step of determining the location and/or character of a layer boundary, by identifying a discontinuity in the slope of measured ACPD versus frequency.

5. An ACPD measuring apparatus having a continuously variable frequency.

6. An apparatus according to claim 5, characterised in that said frequency is variable continuously or in at least twenty steps over an operating frequency range.

7. An apparatus according to claim 5 or 6, characterised in that said frequency is variable over the range 300 Hz to 100 kHz, continuously or in discrete ranges .

8. An apparatus according to claim 5 or 6, characterised in that said frequency is variable from 30 Hz to 500 kHz, continuously or in discrete ranges.

9. An apparatus according to any of claims 5 to 8 further characterised by means for analysing measured ACPD values to detect discontinuities of slope against frequency.