Labeling and Authentication of Electronic Products

The global trend is that the number of counterfeit microcircuits (and not only microcircuits) is increasing, but the effectiveness of detection methods is decreasing. One of the methods used to combat counterfeiting is to label genuine components. And the safest method for labeling is based on physically non-clonable functions (FNF) i.e. such properties of the product that cannot be reproduced due to the natural scatter of the characteristics of parasitic structures, the uncertainty of the results of random processes of production technology.

The real scale of counterfeit electronics is difficult to assess. It is difficult for an electronic equipment manufacturer to identify counterfeit products among the thousands of products used to assemble a system. In some cases, a counterfeit can be introduced several stages earlier in the supply chain and become part of a module or assembly sold by a reputable company. Most manufacturers do not have the resources to trace the actual origin of each part of a product. Sometimes products can actually work, at least for some functions, for a short period of time.

The distribution of the amplitudes of ionization responses and the radiation degradation of parameters with the accumulation of the absorbed dose is one of such FNFs since possesses the necessary properties: the impossibility of reproduction on the one hand, and the homogeneity of the results within one batch of microcircuits or transistors on the other hand. Therefore, it is proposed to use various signs of deterioration of radiation behavior as FNF.

The real scale of counterfeit electronics is difficult to assess. It is difficult for an electronic equipment manufacturer to identify counterfeit products among the thousands of products used to assemble a system. In some cases, a counterfeit can be introduced several stages earlier in the supply chain and become part of a module or assembly sold by a reputable company. Most manufacturers do not have the resources to trace the actual origin of each part of a product. Sometimes products can actually work, at least for some functions, for a short period of time.

Types of counterfeit products

The main types of counterfeit products:

• reused relabeled products;
• cloned products;
• microcircuits with a replaced crystal.

80% of counterfeit products are reused products. They have already been used in gear, equipment, or electronic devices. Waste electronics are collected and sold to developing countries. There, defective devices are disassembled into parts, components are removed from printed circuit boards at very high temperatures, and then sold on the market as new. With the help of special equipment, the dismantled components acquire a commercial appearance – the old inscription is erased, a matte surface is created, a new inscription similar to the former is applied, “solderless leads” are cleaned and created. etc.

The cloned components are manufactured by unauthorized manufacturers who have no legal rights to manufacture the crystal. Cloning itself is the process of copying a design by fraudsters, mainly to reduce the high development costs of a component.

Counterfeiting can be carried out at the chip level, when the counterfeit manufacturer installs a chip in the microcircuit that is worse than the original in terms of characteristics. In addition, a chip with hidden functionality (“hardware trojan”) can be installed that allows you to send secret information to a potential adversary. This can pose a real threat if the chips are used in the military-industrial complex or in military equipment.

Counterfeit detection methods

The main methods for detecting counterfeit components can be divided into physical and electrical. Physical methods include identification, marking control, control of weight and size parameters, X-ray examination, tightness control, scanning acoustic microscopy. Electrical methods include parametric and functional control, as well as electrical thermal training.

Example of double marking

At the stage of incoming control, it is necessary to check the accompanying documentation, to evaluate in detail the appearance and weight and size parameters, to check the uniformity of the batch. Already at the stage of incoming control, the following discrepancies in labeling are revealed (see Fig. 1-2):

• incorrect marking (for example, incorrect date code, spelling errors, translucent old marking);
• poor quality of marking;
• the marking is washed off after exposure to solvents.

Remarking example

In addition, an acoustic microscope is used to check the markings. In the methods of acoustic microscopy, elastic mechanical vibrations are used to ensure the propagation of waves in a solid medium.

Physically non-clonable functions

Despite the strictness of the rules of the technological process for the manufacture of serial microelectronics, the natural spread of parameters, characteristics of parasitic

structures, uncontrolled parameters of the external environment create unique structural elements of the EEE product that cannot be faked. The principle of EEE marking for subsequent identification based on such elements is called the principle of authentication based on FNF. There are several ideas and implementations of the FNF-based authentication process. All of them are united by the fulfillment of several rules:

1. The implementation of an element with FNF should not affect the basic functioning of the product. It is best if such elements “by themselves” exist in any product without additional efforts of designers.

2. The tested characteristics of an element with FNF cannot be faked or reproduced. They should be based on random processes.

3. The results of checking (measuring) the characteristics of elements with FNF should have the property of being reproducible regardless of the laboratory or the manufacturer of the measuring equipment.

4. The results of checking (measuring) the characteristics of elements with FNF should be able to give an unambiguous answer: a genuine product or a counterfeit one. It doesn’t matter based on comparison with a predicted result, reference values, or a comparison with a reference sample.

One of these FNFs is the radiation behavior of an EEE product. The implementation of such a FNF inside microcircuits or powerful discrete devices does not require the creation of special blocks. Radiation behavior for modern technological processes is determined by the degradation of characteristics associated with parasitic structures and associated EEE parameters.

Radiation behavior as a physically non-clonable function

Examples of the application of radiation behavior as characteristics of FNF are given below. The disadvantage of using radiation behavior as FNF for authentication is that it is destructive testing and will have to sacrifice a sample. However, about half of the methods for detecting counterfeit products listed in AS6171 are also destructive (for example, checking for re-labeling), which does not prevent them from being combined and carried out on the same samples.

Let us give examples of the use of radiation behavior as an FNF to identify inhomogeneous samples in a sample.

Powerful MOSFETs (MOSFETs) IRFNG50 from International Rectifier. Two production batches: 0842 Mexico and 1038 USA. From the point of view of functioning, the parameters of the MOPT samples correspond to the norms from the datasheet. The implementation of the FNF consists in measuring the degradation of the threshold voltage for unlocking the transistor depending on the accumulated dose. Dependency graphs are shown in Fig. 1 and 2. The horizontal line on the graph shows the limiting value of the parameter (not less than 2V), which is allowed by the documentation, and its intersection is a criterion for the parametric failure of the sample.

The results of measuring the degradation of the threshold voltage from the absorbed dose in the course of radiation exposure allows us to draw two conclusions: for one batch there are no differences between the samples, for any samples from different batches – the differences are significant and allow them to be unambiguously assigned to the required batch.

LABELING AND VERIFICATION OF THE AUTHENTICITY OF MICROELECTRONICS PRODUCTS BASED ON THE INVENTION OF RADIATION BEHAVIOR

Operational amplifier OP1177ARZ from Analog Devices. Were examined 4 batches of different years of production: 2008, 2010, 2012, 2013. Except for the date of production, externally, the samples look the same in appearance, characteristics and marks on the crystals [14]. In fig. 3 shows examples for two parties 2008 and 2013.

The implementation of the FNF consisted in taking maps of the ionization response of microcircuit crystals. The procedure is performed as follows. First, the parameters of the samples are checked to ensure that they are valid. Then, a part of the plastic case above the crystal is chemically etched away. An open crystal begins to be subjected to the scanning action of a focused (10 μm diameter, 10 ps duration, 1.06 μm wavelength) laser beam simultaneously with the detection of a current surge in the microcircuit power circuits.

Combining the data on the location of the laser shot and the amplitude of the response makes a map of the ionization response, which actually shows the relationship of several physical processes and their characteristics: (a) the local region of the crystal (for which

the effect of focused radiation occurs), (b) the mechanism and features of the collection of the generated excess charge (taking into account the local profile of silicon doping, the geometry and topology of the regions, the re-reflection of laser radiation from the metallization layers, primary recombination, etc.), (c) the movement of this charge from the place of generation to the investigated terminals of the microcircuit (taking into account losses during movement, recombination, recharge of parasitic reactivities). Taking into account the huge number of unpredictable internal factors (including parasitic elements) that affect the final form of the response at the microcircuit pins, the ionization response map can be considered another implementation of a physically uncloned function.

In fig. 4 and 5 show response maps for different batches of OP1177ARZ microcircuits. After the research carried out, there is no doubt that the topology and manufacturing technology of batch 395507 differs from batch 753607, both on the surface and inside the crystal (see Fig. 6.). This was further confirmed by a detailed analysis of the crystal topology in the place of the most significant difference in the response maps.

Thus, the ionization response map is a variant of the FNF implementation in the course of fully automatic scanning, regardless of the functional purpose of the product. Its advantage over layer-by-layer etching and comparison of metallization with the reference is that the ionization response is formed simultaneously by both the metallization map (determines the fraction of laser radiation reaching the silicon regions) and the topology of the silicon regions of the crystal themselves (the analysis of which is usually limited for visual comparison of topology ).

Conclusion

Testing of microelectronic products for identifying signs is one of the components of the strategy to counter counterfeit. But, unfortunately, even the use of all detection methods does not provide a 100% guarantee that the tested product is genuine.

The way out of this situation is to authenticate samples based on physically non-cloning functions.

The article presents examples of the implementation of the authentication of EEE samples based on FNF – radiation behavior of products. As examples, data are given for powerful semiconductor devices, analog microcircuits with a low degree of integration.

Since radiation tests are one of the mandatory types of EEE tests, to verify compliance with the operating conditions of the equipment under conditions, for example, space radiation, the use of the obtained data on the radiation behavior of products for authentication is an additional result of the research.