FIELD OF THE INVENTION

The present invention relates to the detection of corrosion, and more particularly to detection of corrosion using electrochemical techniques.

BACKGROUND OF THE INVENTION

Corrosion of medical instruments and objects is an everyday phenome non in hospital environments, for example. It originates, for instance, from residues such as body fluids, which provide an ideal growth substrate for various microbes that might start deteriorating the surface in question through corrosion.

An essential hygienic step in the cleaning of reusable medical instru ments is sterilization by high-temperature steam. However, should the corrosion has already initiated, the instrument needs to be cleaned manually before the ster ilization. If not properly cleaned, iron oxides present on the surfaces will contami nate the whole batch of instruments being sterilized.

The present state-of-art detection of incipient corrosion is based on ma chine vision, which has several shortcomings. Firstly, many aging objects develop darker passivation layer spots, which give rise to a high number of false positives. Secondly, as any visual means of detection is only capable of evaluating surfaces in the field of view, corrosion in "hidden places" such as hinges of clamps, or in the inner sides of structures, such as stents and tubes, cannot be evaluated reliably with existing techniques.

In disclosure US 6264824 B1 corrosion of an object is detected with an electrochemical procedure and more specifically using electrochemical noise measurement. In the disclosure, the corrosion is detected based on corrosion cur rent and the corrosion process is thereby ongoing.

BRIEF DESCRIPTION OF THE INVENTION

An object of the present invention is to provide a method and an appa ratus for implementing the method so as to overcome the above problem relating to detecting corrosion on instruments with complex geometries. The object of the invention is achieved by a method and an apparatus which are characterized by what is stated in the independent claims. The preferred embodiments of the inven tion are disclosed in the dependent claims. The invention is based on the idea of utilization of electrochemical tech niques to enable an analysis of all exposed surfaces. The inspected object is im mersed in an electrolyte, and a current response can be used for detecting the cor rosion. Electrochemical techniques enable to obtain an indication of corrosion even if the corrosion is not visible. Further, with electrochemical techniques, very small amounts of corrosion products on a large non-corroded surface can be detected which would be hard to detect with known methods and devices which are based on visual inspection. Even further, the invention enables to detect corroded sur faces and does not require that the corrosion process is ongoing.

The method and apparatus of the invention are able to provide accurate results with minimal false-positive results. The method does not react to harmless surface alterations and alerts the user only when the presence of iron oxide species is detected.

The invention further enables to inspect objects with complex forms without losing the accuracy of the inspection and without harming or destroying the inspected objects as the corrosion current is not flowing. The inspection pro vides reliable results in connection with non-visible internal structures. The inven tion can also be applied to various automated quality control processes.

BRIEF DESCRIPTION OF THE DRAWINGS

In the following, the invention will be described in greater detail by means of preferred embodiments with reference to the accompanying drawings, in which

Figure 1 shows current waveforms obtained with an embodiment of the invention;

Figure 2 shows the first derivative of current as a function of potential; Figure 3 shows the second derivative of current as a function of poten tial;

Figure 4 shows current waveforms obtained with an embodiment of the invention;

Figure 5 shows a simplified structure of an embodiment of the inven- tion;

Figure 6 shows another current waveform;

Figure 7 shows a modification of the current waveform of Figure 6; and Figure 8 shows a chart relating to machine learning. DETAILED DESCRIPTION OF THE INVENTION

Small patches of iron oxides on the surface of medical implements has been found to be detectable by electrochemical techniques, such as by means of linear sweep voltammetry (LSV) and/or cyclic voltammetry (CV) in a relevant po tential window, for instance between -0.35 V and 0.35 V vs Ag/AgCl reference. Lin ear sweep voltammetry (LSV) and cyclic voltammetry (CV) are procedures that are known as such and are commonly used to study the properties of electrolysis pro cesses.

In the present invention, the object to be inspected is immersed in an electrolyte and the presence of corrosion products on the surfaces of the object is detected by electrochemical techniques based on the current originating from the redox reaction of iron. The object to be inspected may comprise one or multiple medical instruments, for example. The electrochemical techniques include linear sweep voltammetry (LSV) and cyclic voltammetry (CV). With LSV or CV, a sweeping voltage is provided in an electrode and a current response is detected. Based on the current response it can be concluded whether corrosion has occurred on the sur face of the inspected object. The corrosion on the surface of the inspected object may be due to the corrosion process that has brought iron oxides to the surface or the iron oxides may be due to contact with other corroded obj ects, for example. The presence of corrosion which is detected with the invention does not require that the corrosion process is ongoing as the detection is based on the redox reaction of the iron in iron oxide and more specifically on the electrical current originating from the redox reaction.

The voltage sweep rate should be selected to be preferably in the range between 5mV/s and 100 mV/s to facilitate identification of the surface reactions, i.e. the presence of iron(oxides). In LSV or CV procedures a sweeping voltage is ap plied between the electrodes and a current response is measured. As the object or multiple of objects to be inspected which acts as one of the electrodes may have a large surface area, the capacitive current is prevailing. If the inspected object has corrosion, i.e. iron(oxides), the current response to the sweeping voltage displays an alteration of the signal in a certain voltage area. The current is due to the redox conversion between two charge states of iron, the corresponding electron transfer through the external circuit and the ionic charge balancing. More specifically, the two states of iron are Fe ²⁺ and Fe ³⁺ , and the detected current is due to the transfer of electrons between these two states. By complete immersion of the inspected objects, such as medical tools, medical instruments, clinical research instruments, clinical instruments, compo nents, spare parts, or any metallic pieces, into a suitable electrolyte, contact be tween the detecting medium and point of corrosion can be ensured as the electro- lyte is in connection with the object no matter what the shape of the object is. A suitable electrolyte is, for example, a synthetic physiological fluid which is also known as IV-fluid. In the method, an electrochemical reaction is obtained using a setup, in which the electrolyte is preferably the IV fluid. Further, the inspected ob jects in the electrolyte are acting as a working electrode and an electrolyte vessel acts as a counter electrode. Further, a reference electrode is provided to complete the required three-electrode setup. The third electrode is preferably an Ag/AgCl reference electrode. Such a reference electrode is suitable as the physiological fluid is used as an electrolyte is sufficiently high in chloride content. Ag/AgCl reference electrode has been found to be a suitable reference electrode. However, other ref- erence electrodes can also be employed without departing from the invention. The use of IV-fluid is also suitable as it is readily available in hospital environments, for example. Further, the properties of the fluid are not harmful to the instruments which are being scanned. Other suitable reference electrodes include calomel elec trodes (SCE) and reversible hydrogen electrodes (RHE, SHE), for example. With dif- ferent electrodes, a different electrolyte may have to be chosen. Further, the poten tial window in which the voltage change is carried out has to be adjusted according to the selected reference electrode in a known manner. Ag/AgCl reference elec trode is preferred as the electrode is known to be stable and durable in the in tended chemical surrounding. The electrolyte used in the invention is preferably a water-based electrolyte having salt (chloride) content. As mentioned, IV-fluid has the properties which are needed for the electrolyte.

It is important, that the used electrolyte does not react with intact sur faces so that the object from which the corrosion is detected is harmed.

As mentioned above, in the LSV or CV scan voltage is changed and cur- rent between the electrodes is measured. As the corroded surface is typically com paratively small with respect to the non-corroded or intact surface area, the ob tained current originating from the electrolysis reaction is small. The current cor responding to the surface alteration being studied is obtained typically in a certain part of the voltage sweep, and the presence of current indicating the corrosion can be detected from the current response. According to an embodiment of the invention, the current indicating corrosion is detected from the first or second derivative of the current. The first derivative of the current represents the change rate of current and the second de rivative of the current represents the rate of the current change. When a voltage sweep is employed, the resulting current changes approximately linearly, i.e. it in creases or decreases without any abrupt changes. This linear change is due to the capacitive current. However, in a certain voltage range, the current has an addi tional component originating from the redox conversion of iron species if these are present on the surface. This additional current component is indicative of the cor rosion of the object. The current peak or an alteration in the current signal is de tectable as such or from the first or the second derivative of the current. Figure 1 shows an example of current as a function of potential both for a clean sample and for a corroded sample. The current peak is hardly visible in the current of the cor roded sample of Figure 1 at 0,25 Volts. Figure 2 shows the derivatives of the cur rents of Figure 1. It can be seen that a peak is more visible and thus more easily detectable in the derivative of the curve than in the original current waveform of Figure 1. Figure 2 also shows the waveform of the derivative of the clean sample of Figure 1.

Figure 3 further shows the second derivatives of the current response of Figure 1. In Figure 3 the second derivatives of both clean sample and corroded sample are shown. As can be seen, a clear and detectable peak is shown in the wave form of the corroded sample.

The detection of current from the redox conversion of iron species can be detected using circuitry in which the current is measured as digital samples or transformed into digital form. The digital representation of the current waveform can be subjected to mathematical calculation or transformations. These calcula tions may include using different filters for determining whether a current peak is present in the current response. Further, in a preferred embodiment, the first or the second derivative of the current is calculated. As the current is as digital sam ples, the first derivative and the second derivative of the current can be calculated using a suitable software algorithm. The value of the second derivative of the cur rent can be used as an indication of whether corrosion is detected. It is, however, clear that the current originating from the redox reaction can be detected using other means which are able to analyse the current signal. The first or the second derivatives of the current can be determined with respect to time or potential. Figure 4 shows voltammograms which were obtained when the method of the invention was tested. The upper plot shows measurements which were car ried out with corroded test pieces and the lower plot shows similar measurements with test pieces in which the surface is free of corrosion. The potential window in the tests is selected to be from -0,3V to 0,3V. It can be seen that upper plot with the corroded test piece shows an increase in the current when the sweeping voltage reaches approximately 0,18V. Further, a negative peak is visible when the voltage speed is negative at voltages -0,05V. When the voltage increases the positive cur rent peak is due to oxidation of the iron and the negative current peak is due to reduction of the iron. As seen from the lower plot, the pieces which were not cor roded did not show a current peak when tested according to the method.

In the method, the potential window used for the potential sweep can be selected to be from 0 V to 0.35V vs Ag/AgCl reference. Once potential is changed between these values, a current peak is obtained in the current response if corro sion is present. As described above, the current peak is positive and it is obtained when the current is increased.

It has been found out, that in the present method the negative peak due to the reduction of the iron can be detected more clearly than the positive peak. According to an embodiment, the detection of the iron oxides is carried out from the current when the voltage is reduced. Further, in an embodiment, the potential window in which the potential is changed is selected to be -0,25V to 0,1V when the reference electrode is an Ag/AgCl reference electrode. As mentioned above, when other reference electrodes are used, the potential window is changed accordingly. The negative peak or the positive peak refer generally to alteration in the current signal. The alteration in the current signal may detected also with other signal pro cessing procedures than by detecting the peaks. Generally, the change of the cur rent signal with respect to a clean sample may indicate the presence of iron oxides.

The potential window and the sweep rate are selected in such a manner, that the objects which are under inspection are not harmed. If, for example, the sweep rate is too high the current response associated with iron oxide alteration may go unnoticed, and if the potential window is too wide, the intact surfaces might get damaged. Thus the potential window and the sweep rate are preferably se lected in such a manner, that the associated response is obtained only from a cor roded surface, i.e. only when iron oxides are present.

Further, the number of the sweeps, that is the increase and decrease of potential, can have a selected maximum amount. Once a positive indication is found, the process is stopped. If a positive or uncertain indication is obtained, the sweeps are repeated for the set number of times at the maximum. Thus according to an embodiment, the measurement cycle, i.e. the voltage sweep, is repeated. The repeated measurement may be used for ensuring the correctness of the results. The apparatus of the invention, which implements the method, com prises means for immersing the object into electrolyte and means for detecting by electrochemical techniques the presence of corrosion on the surfaces of the object based on current originating from redox reaction of iron.

Figure 5 shows an example of an apparatus for implementing the method. According to an embodiment of the invention, the means for immersing the object 35 comprise a stainless steel member to which the objects to be in spected are connected electrically. The stainless steel member is preferably a stain less steel plate or a stainless steel mesh basket 31 in which the object or objects can be placed. The stainless steel member is electrically connected to the measure- ment circuitry 33 and the object 35 forms the working electrode of the system. The object is immersed into an electrolyte which is preferably held in a stainless steel vessel 32. The stainless steel vessel 32 acts as a counter electrode and is also con nected to the measurement circuitry 33. The setup also comprises a reference elec trode 34, which is preferably an Ag/AgCl reference electrode. The means for detecting the presence of corrosion products comprise the measurement circuitry, which provides a pre-programmed LSV or CV scan. The dedicated measurement circuitry displays the functionality required to produce the LSV or CV scan at an accuracy and sensitivity sufficient for the detection pro cess. For example, the accuracy of the voltage control should be better than lOmV, the accuracy of the sweep rate should better than 2mV/s and the sensitivity of the current measurement should be better than ImA. The accuracy of control and the sensitivity of the measurement allow detecting the current response reliably.

In addition to the measurement circuitry providing the scan, an embed ded software algorithm or software installed on an external unit compares the ob- tained results, i.e. the current response, against a given set of criteria. For example, the comparison may compare the first or second derivative of the current in a spe cific potential window against a limit value. If the first or second derivative of the current exceeds the limit value, the measurement software produces an indication that the object has iron oxide on its surface (corrosion). On the other hand, if the second derivative of the current remains below the set limit value, the measure ment circuitry produces an indication that corrosion was not found. This information may be conveyed to the user simply by using suitable light indications and/or text information on a display visible to the user of the ap paratus. Once the scan has ended, the user of the apparatus can remove the objects and proceed with them according to the obtained results.

In one aspect of the disclosure, machine learning is employed in detect ing corrosion. Specifically supervised learning is employed in teaching the machine learning process. The procedure in the machine learning process comprises meas uring and storing the current response for a selected voltage sweep.

The measuring and storing current response for a selected voltage sweep is carried out without the object. Thus only the used basket or similar is em ployed in a sweep. In the next step, the stored current response of the voltage sweep without the object is removed from the stored current response with the object. The stored current response with the object may be an average of multiple current responses with the same object. The above removal of current response may be carried out simply by subtracting the current values for obtaining prepared data.

Further, in the procedure relating to machine learning, a linear baseline is fitted to the prepared data from a certain potential point to a point of the certain minimum potential point. Further, the baseline is subtracted from the prepared data, and the remaining data is smoothed or filtered using a moving average filter, for example.

In the procedure, the slope of the baseline and the maximum point of the smoothed data are determined and stored. The slope and the maximum point are used as the data relating to machine learning. Each measurement is thus clas sified in a two-dimensional manner.

As supervised machine learning is employed, objects having corrosion on the surface and cleans objects are employed in the above procedure to obtain data which is used as a reference classification. Once classification data is gathered with the procedure, the procedure is used to determine whether the object has cor rosion on its surface or the object is free from corrosion. The data obtained with the above procedure from actual samples are mapped with the reference data. Us ing the reference data the data obtained from the sample is classified with a classi fication method, such as a Nearest-Neighbor classification.

In the above procedure, the current samples are obtained using the measurement of the invention, i.e. current is measured using electrochemical tech niques, such as LSV or CV. The above machine learning procedure is further described in connec tion with Figures 6, 7, and 8 which show examples of obtained data in visualized form. Figure 6 shows an example of the prepared data mentioned above. The pre pared data refers to measured current values during the voltage sweep from which the corresponding current values without any object (i.e. empty basket) is re moved. The measured current values represent an average of three consecutive current measurements.

Figure 6 also shows a baseline which is a fitted curve from the zero-po tential point to point of the minimum potential. In the example shown in Figure 6, the minimum potential is selected to be -0,2 Volts and the highest value of voltage is zero. The voltage in the measurement is swept from the negative side to zero to obtain the current values.

Further, Figure 7 shows a smoothed current curve in which the baseline of Figure 6 is subtracted from the prepared data of Figure 6. The slope of the base line is calculated and the maximum point of the smoothed curve is determined to obtain data used in the classification. Figure 8 shows an example of data that is gathered with known reference objects (with corrosion and without corrosion) to gether with data from the objects from which the corrosion is tested. Thus the amount of data used in the chart of Figure 8 increases with every scanning proce dure carried out.

As actual measurement data is placed in the chart, a certain procedure for determining whether the object to which the measurement data relates is clean or has experienced corrosion. The classification to clean or corroded may be car ried out using Nearest-Neighbor classification, for example. In Nearest-Neighbor classification it is checked whether the nearest existing data point in the chart re lates to a clean object or to an object having corrosion. As the amount of data in the chart of Figure 8 increases, the classification of the object gets more accurate.

Thus according to an embodiment, the detecting of the presence of cor rosion comprises determining the current response in a selected potential range. Determining from the current response data used in a machine learning procedure, the machine learning procedure comprising previously gathered data as described above. Further, the embodiment comprises classifying the object to a corroded sample or to a clean sample based on the determined data and from the previously gathered data using machine learning.

The above-described apparatus is an example of a possible structure with which the detection of corrosion can be carried out. Depending on the size of the object the mechanical structure of the apparatus may be different. It is also un derstandable, that as different objects have different shapes and sizes, the object is not necessarily inspected completely with one procedure or that only a part of an object is inspected. When an object is only partially immersed into an electrolyte, only the immersed portion of the object is inspected. However, when placing the object in the electrolyte in different orientations the complete surface of the object may be inspected using repeated procedures.

It will be obvious to a person skilled in the art that, as technology ad vances, the inventive concept can be implemented in various ways. The invention and its embodiments are not limited to the examples described above but may vary within the scope of the claims.