QuantumTags: Three-Layer Authentication Through Self-Assembly Quantum-Dot Inkjet Printing for Anti-Counterfeiting
Table of Contents
1. Abstract
2. The Problem
3. The Status Quo
3.1 — The Anti-Counterfeiting Paradox
3.2 — Supply Chain Management Needs
3.3 — Current Solutions
4. Nanotechnology for Anti-Counterfeiting
4.1 — The Potential of Nanotechnology for Anti-Counterfeiting
4.2 — The Importance of Quantum Mechanics in Anti-Counterfeiting
5. The Scope: Nanotechnology-Based Solutions
5.1 — Optical Q-IDs
5.2 — Electronic Q-IDs and ePUFs
5.3 — Colour-Shifting Tags
5.4 — The Usage of Quantum Dots
6. The Fundamentals: Physically Unclonable Functions (PUFs)
6.1 — An Overview of PUFs and UNOs
6.1.1 — Challenge-Response Pairs (CRPs in PUFs)
6.2 — The Significance of Atomic Randomness
6.3 — Multi-Level Authentication Technique
7. QuantumTags: The Overview
8. QuantumTags: The Breakdown
8.1 — Fabrication
8.2 — Three-Layer Authentication and New Quantum Dot Advancements
8.3 — Deep-Learning Authentication
8.4 — The Advantages
9. Conclusion
10. Bibliography
1. Abstract
Integrity and trust are at the heart of humanity and allow for peaceful nations, bonds and trust between multiple parties. However, when that integrity is played with, many become overprotective, people cannot enjoy the common object and once trustful systems are tampered with, connections between communities are disrupted. On a global scale, counterfeit goods are where this integrity is played with the most. Counterfeit pharmaceuticals cause the deaths of millions in developing nations2, counterfeit batteries pose risks to everyday items bursting at any moment and overall, these goods cost the global market over one trillion dollars. Current solutions can easily be reverse-engineered and as the counterfeit epidemic surges, the world needs a solution to make counterfeit goods impossible. The integration of nanotags and harnessing the randomness and uniqueness of quantum dots allow for unclonable tags on each product. These tags are verified by the end-user through a deep learning algorithm. The tags are unclonable by any quantum computer let alone any attacker, ensuring the security of millions of lives and billions of dollars.
2. The Problem
The global market of counterfeit goods is currently $1.8 trillion and this number is only increasing. During the COVID-19 pandemic, a greater urgency for counterfeiting occurred as the global demand for medical supplies continued to increase (2). Not only does counterfeiting cost the global economy trillions of dollars, but it also results in fake pharmaceutical pills, costing millions of lives (2). To add on, 500 identity frauds happen on a daily basis, showcasing how counterfeit goods are reaching an epidemic level (10). In 2015, 10% of all luxury goods in Europe were counterfeit, and the number continues to increase (10). The process holds a problem in itself where the inability to find counterfeit goods results in a loss of sales, burdening services and leading to a loss of money (17). Many lives are also put in danger in developing countries when in contact with these fake goods. They make ordinary consumer products dangerous as phones and e-cigarettes blow up in the user’s face due to counterfeit batteries and due to the COVID-19 pandemic, this epidemic has grown by 100x over the past two decades (22).
3. The Status Quo
All current solutions can be reverse-engineered. These existing solutions are analyzed, then counterfeiters recreate the original product and mass-produce the fake goods, causing these new products to be indistinguishable from the original (2).
3.1 — The Anti-Counterfeiting Paradox
There is a mind-boggling paradox in anti-counterfeiting whilst creating a tamper-proof system. There are two types of systems. The first is easy to verify but also easy to clone (eg. holograms) and the second is difficult to clone and difficult to verify (eg. quantum dots)(2). Creating a system which is easy to verify but difficult to clone is the key. With nanotechnology, quantum mechanics only become more complex as the system gets smaller (2). To break this paradox, a smartphone is a solution. A smartphone fires a flash to excite the optical emission from the material and extract the fingerprint. This is both sensitive to the nanoscale variations and is easy and convenient to measure with the phone (2). Therefore, new anti-counterfeiting technologies require this convenience however, old solutions lack this convenience and accessibility (section 3.3).
3.2 — Supply Chain Management Needs
To have a secure supply chain, it is vital for the creation of packaging which can identify the current location and real-time monitoring during each step, resulting in the production of the solution, microprocessors (20). The miniaturization of ID tags leads to intelligent control. The current problem with this intelligent control is visibility, meaning everyone in the shipping process can track the location however, companies lack this end-to-end visibility (20). This traceability would showcase where the raw material comes from, all the way to the transformation site and final delivery. However, the problem lies in the cost and complexity.
3.3 — Current Solutions
One of the many solutions susceptible to attacks are passwords. Passwords are everywhere. Every account made on each website and app requires a password. However, they are typically weak and easily have data breaches (2). With passwords, there can be more letters and numbers which make the password more difficult to break. Although the possibilities are in the millions or billions, a quantum computer can break this encryption (19).
RFID (radio frequency identification) tags and integrated circuits (ICs) are other large players in this market. RFID tags use radio waves to communicate information to users (17). The reader uses a device that emits radio waves and receives signals back from the tag, however, a counterfeiter can duplicate these smart tags (17). Like RFID tags, many people rely on ICs to perform sensitive security tasks and handle sensitive information (14). RFID tags have a tracking system which gives much access to financial information, mobiles, transactions and access cards (14). The reason ICs are greatly used is that their secrets are stored in non-volatile memories meaning their computer memory can store information without power (14). However, these are often vulnerable to invasive attacks as the secrets are always in digital form.
Another common and raving solution is using blockchain. Blockchain identifies the provenance (proof of origin) of a product and has a secure tracking system from all supply chain ends. When attached to a good, it can identify the place of the manufacturer and track the location, but cannot tell the current user (17).
QR codes may be seen as the future of anti-counterfeiting but they are easily reverse-engineered and have numerous limitations (2). They are barcodes read by smartphones which encode information such as a phone number or internet access (17).
4. Nanotechnology for Anti-Counterfeiting
Nanotechnology is the study of manipulating materials on an atomic level called the nanoscale and because every material is investigated at the nanoscale, they become very small, changing their properties such as reactivity, conductivity and solubility as a function of the particle size. Keeping these materials on the nanoscale we can harness their properties and configuration to use them for the desired application. An important note is that on the nanoscale, the colombic attraction becomes stronger (attraction between protons and electrons) which plays a large role in the behaviour of different nanomaterials. However, current nanofabrication causes all nanomaterials to have imperfections and impurities, hindering their application in the real world. But, by harnessing these impurities, counterfeiting can be put to an end.
4.1 — The Potential of Nanotechnology in Anti-Counterfeiting
Current solutions can easily be reverse engineered or replicated with minimal effort by counterfeiters. However, optical quantum IDs provide authentication that is impossible to copy because each nanoscale device has roughly 1000 trillion atoms (2). Even with the most powerful scanning probes, it would take 13 billion years to make an identical clone (2). Nanotechnology allows for unclonability with the smallest size, weight and needed power requirements, for example, properties like metal layer thickness, device geometry or even contact interfaces, all play a role (6). Overall, on the nanoscale, because of the disordered nature of binary-ternary interfaces, each device is automatically unique meaning the properties only add to this uniqueness6. For example, consider a reduced variation in the I-V carbon atoms to less than 1 percent by refining the interfaces between the layers. This would change the molecular-beam epitaxy (MBE), resulting in interfacial variations of a quantum well, creating an entirely different device (6). All in all, this showcases how distorting the nanostructure would affect the results. Specifically, fabrication using self-assembly is well-suited for anti-counterfeiting (read more in section 8). Nanotechnology would also enable roll-to-roll manufacturing, decreasing their price (21). This is where prior to embedding a tag into a system, it is rolled onto a substrate and fed from one roller onto the next.
Furthermore, the novelty brought by the nanoscale is that changing multiple atoms or impurities on a nanomaterial can change all of the properties and the overall material (2) and this novelty is brought to the collected fingerprints. For example, graphene is a single layer of carbon atoms. However, all quantum materials, like graphene, have imperfections due to their manufacturing method (2). This production causes impurity in the atoms, which is why nanotechnology is still in its early stages. Many are not able to reproduce the same properties each time they produce a nanomaterial, however, using this as an advantage, it can cause counterfeiting to be impossible. These impurities also affect how a quantum material emits light (2). This solution of the atomic world allows the laws of quantum mechanics to come into play to produce the QuantumTags (section 8) that are simple to make and impossible to copy (3). At its core, these fingerprints have their unique quantum IDs, are uniquely read and allow for secure authentication3. The track and trace within the supply chain is unique since the solution can be turned on and off in real-time in the supply chain (3).
4.2 — The Importance of Quantum Mechanics in Anti-Counterfeiting
With the rapid advance in manufacturing, complex counterfeit components are increasing in processing power. However, to have certification, it needed to have the use of a secret key which acts as an identity and is then stored onto an integrated circuit (IC — identical to a normal circuit but on a smaller piece of semiconducting material). However, all attackers can learn this key since it must exist as a digital chip and if compromised, the attacker can use the original device6. Instead, devices with randomness are the base of solutions which are unclonable (PUFs and UNOs, section 6). The security for all of these solutions relies upon the impossibility of refabrication6. As the size of a system reduces, a limit is reached. This limit is where quantum confinement takes the properties of the system, causing the nanostructure becomes even more crucial to its properties (6). These confined energy levels are very sensitive to the atomic layer with millions of atoms, ensuring that the probability of creating unique devices is extremely high and the underlying structure is unknown unless dismantled atom-by-atom (6).
5. The Scope: Nanotechnology-Based Solutions
With nanotechnology-based solutions, these QuantumTags are bonding the physical and digital worlds (section 8). Much of these nanotech solutions are either based on PUFs (section 6) or on Quantum IDs (Q-IDs). In short, Q-IDs are harnessing randomness from quantum mechanics on the nanoscale. These create unique identities, or quantum fingerprints, causing each Q-ID to be unclonable since they cannot be predicted nor replicated (3). However, even within these two sub-categories, the options are endless as follows.
5.1 — Optical Q-IDs
By adding the nanomaterials to the surfaces of the nanotags, the Q-IDs are formed. Each product would have a unique Q-ID, acting as a fingerprint, which is impossible to replicate as their imperfections are on the atomic scale (3). Optical Q-IDs are greatly accessible and can be paired with a smartphone. Scanning with a smartphone would allow for verification of the product and identification of a counterfeit (3). These are convenient when manufacturing as they use a transparent micro-thin varnish with the chosen quantum material (figure 1), and this can be deposited on all surfaces, allowing for high-speed printing on varying surfaces3. In short, the app will read the Q-ID. The flashlight of the smartphone is triggered when the scanning begins, shining a light on the quantum-ID. The light is then re-emitted by the Q-ID and is captured by the camera of the phone3. The process is then repeated on different flash intensities to measure the emission dynamics, verifying the quantum material (figure 2).
5.2 — Electronic Q-IDs and ePUFs
Quantum IDs are not solely confined to the physical impurities of nanomaterials but can also be electronic, relating more to the law of physics than the imperfections of light emitted. ePUFs (electronic physically unclonable functions) are a 100 percent secure electronic solution (read more about PUFs in section 6). This is needed as the world is connected through IoT, making large-scale cyber attacks an immense threat for a government that continually relies on these networks (3). Existing PUFs are based on classical properties but ePUFs are able to integrate into these electronic systems. Unique arrangements of atoms and imperfections in nanostructures can make these unclonable devices as they are based on the atomic randomness of quantum confinement3. These quantum ePUFs have the highest bit density, and the best integration potential and need the least number of resources, showcasing their high potential in implementation3. They fit into any electronic product and their simple integration is due to their simplicity, small size, cheapness, low power and operation at room temperature (3). Like many PUFs, they are mass-produced using a standard semiconductor: silicon III-V, which is perfectly aligned with all scalable electronic products (3). They currently give the highest level of security with their form factor: this factor is for irregularity in the shape of the object and the ratio of its volume to the normal volume with the same breadth and height (3). This allows for the global application within microprocessors, phone SIMs, IoT devices, etc.
The fabrication and integration of ePUFs rely on four steps. It begins with constructing the unique Q-IDs. This requires unique data for the ID/key and this data would be created through atomic randomness and quantum effects (3). The second step covers the integration into the electronic system. Due to the high bit density, the ePUFs can hold a greater number of bits within the given memory (3). Step three relies on key extraction where an external stimulus causes an electronic measurement to extract the information. This relies on many resolution levels and verification by humans, hardware and software (3). The fourth step applies this to the varying systems, whether it by IoT or computer hardware, these ePUFs can authenticate any electronic system.
ePUFs are popular and useful as this identification of materials at the atomic level to find quantum fingerprints are more unique than DNA (10). Automatically, thin materials exist in just 2-dimensions and so, when shining a light on these 2D materials, the imperfections emit light back which is used as a signal and is only unique to that section of the material (10). However, since these are electronic and not optical, the signal is turned into a number sequence (acting as the digital fingerprint) and small flakes of these can be added to any system (10). This puts the power of authentication back into the hands of the end-user. Each of these ePUFs has countless challenges which make them not susceptible to breaches. This can be shown by the following equation:
(number of challenges) x (time required to measure each challenge) = the circuits resistor-capacitor (RC) constant = longer than the age of the universe.
5.3 — Colour-Shifting Tags
Colour-shifting tags are another solution which relies on optical properties where numerical characters upon overprinted text have multiple colour-shifting options, allowing for strong and secure brand protection (25). These tags are very durable and have little loss in their characteristics when crumpled (25). These films are simple yet have powerful authentication through observed optical changes with a shift of angles. This perceived change in hue and chroma are varied by fine-tuning the thickness of nanomaterials. Typically, very bright and saturated colours occur by controlling an individual layer’s thickness to obtain specific brightness levels and chroma values (25). The film is fabricated with highly durable materials, resulting in an extremely long life but they require large facilities and expensive machinery (25). Although manufacturing is difficult, this also makes the reverse-engineering process impossible.
This technology is based on optical interference characteristics of thin films made by PVD (physical vapour deposition in large-scale production environments)(25). Complex machinery and processing are required to produce the films with nanoscale uniformity and precision. These are made by deposition very thin layers of high purity metals with specific properties onto a substrate (25). There are two types of these chips: metal-dielectric and all-dielectric.
Metal-dielectric consists of three layers: a semi-transparent metal, a dielectric layer and the reflective layer (figure 3). The semi-transparent metal layer is the top layer and is 5 nm to 10 nm thick. This controls the amount of light reflected from the top surface and the amount of light transmitted to the bottom layers (25). The central layer is hundreds of nanometers thick and selectively controls which wavelength of light undergoes constructive or destructive interference (25). This is based on the index of refraction, which is a measure of a ray of light bending when passing a medium, however, it also depends on the thickness of a layer to match the resonance of the given wavelength25. The bottom layer is the reflector made of a high purity metal and is tens of nanometers thick. This reflects the maximum amount of light back through the filter (25). Overall, the metal-dielectric filters are best suited for the reflection security features. (Figure 4) All-dielectric films have alternating layers of high and low refractive index materials which are used to control the wavelength and colour. The process occurs when the light undergoes interference and is then reflected or transmitted through the device (25).
5.4 — The Usage of Quantum Dots
Nanotechnology consists of various smart materials, one of which is known as quantum dots. Quantum dots (QDs) are artificial semiconducting nanoparticles that transport electrons (1). Semiconductors are materials with conductivity between the conductors and if the particles are small enough, the quantum effects come into play creating varying energies at electrons and holes (absence of electrons). Now, when ultraviolet light hits these semiconducting nanoparticles, they emit light of varying colours depending on the wavelength. These QDs are found in solar cells, fluorescent biological labels and even in quantum computation. The colour emitted relates to the wavelength of energy released which all goes back to two main factors: size and shape. Therefore, by controlling the size, the particles can emit or absorb particular wavelengths of light resulting in different colours. Quantum dots (QDs) are finely tunable dots, roughly 10–100 atoms in diameter, which emit specific colours based on the wavelength of light shone on them (1). QDs have the potential to be used as markers and identify physical items in specific locations in time and space1. The interesting thing about quantum dots is that no two quantum dots are identical, meaning it’s difficult to create a QD with the exact properties (2). The basis of using QDs for anti-counterfeiting is that they emit different colours of light depending on the diameter. There are visible differences in the emission colours even if there is just a difference of a few atoms in width (2).
By pairing these QDs with a blockchain ledger (one of which is the QDXTM Blockchain Ledger), it would allow for a tamper-proof authentication system (1). After the QDs are branded for the specific product, they would be registered and serialized (arranged in series) as they are produced. Then, they would be associated with a cryptographic hash identifier in the blockchain (1). With the basics of blockchain, the hash would follow all encrypted demands, making it almost impossible to guess the length of the hash, preventing attackers from counterfeiting (1).
6. The Fundamentals: Physically Unclonable Functions (PUFs)
6.1 — An Overview of PUFs and UNOs
UNOs and PUFs are two of the main nanotags. UNOs are Unique Objects where the security rests upon the impossibility of refabrication. However, there are no restrictions on what the attacker might know regarding the internal structure of the fingerprint6. The applications are typically within bank notes, passports, access cards, etc. However, they need the trusted external measurement each time the fingerprint is needed (6).PUFs on the other hand are Physically Unclonable Functions which use disordered systems to derive different unique responses and do not need digital storage (6). PUFs are generated by applying different challenges to the PUF and the challenge-response pairs (CRPs) authenticate the device (section 6.2)(6).
6.1.1 — Challenge-Response Pairs (CRPs in PUFs)
PUFs are generated by applying different challenges to the PUF and the challenge-response pairs (CRPs) authenticate the device6. These CRPs can be thought of as input-output systems where the challenges are the inputs and the responses are the outputs. The database of the challenge (Cn)-response (Rn) pairs are created by the manufacturer and stored online, allowing for the user to take any possibility of challenges and check for counterfeit goods when authenticating (6). Each CRP is unique, unpredictable and repeatable while another device contains identical CRPs as they are impossible to identically fabricate, even by the manufacturer (6). The CRPs require a database where the CRPs are recorded and used before each communication and once used, the CRP is erased from the database. There are multiple authentication routes to take with CRPs one of which is multiple CRP-based authentication systems; however, this requires a database large enough to meet the security considerations (6). An alternative route is using a single CRP to authenticate a device. The manufacturer can store the certificate that contains the sole response from the PUF with a signature from the manufacturer’s private key6. When the authentication is required, the PUFs response is remeasured and while the signature and manufacturer’s private key are verified, the stored response is extracted and a check is performed to determine if the two values agree (6).
6.1 — An Overview of PUFs and UNOs
PUFs can be categorized into two sections: strong and weak PUFs. The weak PUFs are known as Physically Obfuscated Keys (POKs) which generate keys from small sets of CRPs, and they continue to scale with the size and complexity. Normally, the derived key is a secret through an internal measurement in the embedding hardware, again relying on atomic randomness (6). Strong PUFs on the other hand have complex input-output behaviour. This is where the available set of CRPs are scaled exponentially and the security will rely upon the attacker not being able to determine this behaviour (6). Examples of strong PUFs include SRAM PUFs (state of static random access memories) or arbiter PUFs. Storing digital information onto a device that is resistant to physical attack is difficult and expensive so, the solution relies on extracting the key information from complex physical systems, PUFs. However, restricting access to the PUFs is important as a middleman must be prevented from finding out the response, so, by combining the response with a hash of a program, the PUF can only be accessed with a particular function (11). The process would be a chain as follows: challenge → PUF → response → hash → secret.
The PUFs are individual chips, which are characterized by their responses and are compared along the supply chain to ensure authenticity. Even if an attacker gains some data about the chip, they are guaranteed to generate more errors or variations than the chip2. In fact, with the same challenges and different PUFs, the output is different. These CRPs are treated as the fingerprint of the PUF which is integrated into the device or object (figure 6)(16). An overview of how PUFs work can be broken down into four steps, even though PUFs can be used in various ways13. (1) It begins when the uncontrollable nanoscale process variations make unique patterns. (2) Then, the unique silicon fingerprint is manufactured where each memory cell stores a binary digit. (3) The fingerprint is identified as a strong cryptographic key. (4) The user keys are then wrapped and encrypted with the PUFs key.
PUFs are so diverse that they can use or replace existing systems and make them unclonable, for example, rather than ICs (section 3.3), PUFs derive secrets from complex physical characteristics and store them onto the physical chip rather than storing them onto a digital memory (14). However, PUFs can extract these complex physical characteristics from ICs, and when combined with nanotechnology, PUFs can aim to provide security claims for ICs (16). PUFs are also very popular because they are a simple alternative to generating unique volatile digital keys in very small hardware devices without the need for tamper sensing mechanisms (16). They are also very easy to build, impossible to duplicate and simple to authenticate. In fact, PUFs are just like keys. Silicon PUFs in specific can be used as unclonable keys. They would have a lock which is the database of a CRPs and to open the lock, the key has to show that it knows the response to one or more challenges (11). Applications of PUFs are extensive. Strong PUFs can generate cryptographic keys as they have a larger number of CRPs, learning to multiple rounds of authentication prior to usage (16). This is done by the prover sending authentication requests, allowing the verifier to send randomly selected challenges and authenticating the received responses (16).
6.2 — The Significance of Atomic Randomness
(Figure 8) Atomic randomness can occur in various different ways, one of which is demonstrated by a simple single electron. The device is a diode where the electron can travel in two paths. The way that the electron chooses is completely random and out of any external control. Whichever path is chosen, the voltage measures the electron across the diode and converts this into a random number, 0 or 1. Eventually, these binary digits turn into the CRPs in an ePUF, demonstrating the randomness (6).
Atomic randomness can be thought of as atomic barcodes. For example, when graphene was given the Nobel Prize in 2010 it was explained how imperfections change the microscopic properties of a material and how the atomic scale is the most complex stage we can get to (16). Harnessing a difference imperfection for each new ID would require a database for these billions of IDs. Specifically, graphene and many other nanomaterials are expensive when they are engineered however, graphene with imperfections is straightforward as the current production method creates these impurities by default16. It is very difficult to make two identical sheets of graphene, and by harnessing this we can prevent counterfeiting. Another way to think about atomic randomness is sand. By tipping a wheelbarrow of sand twice, the arrangement will never be the same and one cannot rearrange the sand to make the particles look identical either, it is simply too difficult (16). Just as particles of sand cannot be rearranged into identical formations, when talking about CRPs, they cannot be predicted, cloned or simulated in any way, because just like tipping over a wheelbarrow of sand, they are based on pure randomness (16).
6.3 — Multi-Level Authentication Technique
Using four layers of security allows for multi-level visual authentication. (figure 9). An example of a chip is shown to the left. Referring to the figure above, the top layer showcases the logo, and brand elements and has the serialization (serial number). Moving down, the second layer showcases the holographic security pattern with the third layer showcasing the colour-shifting film and the fourth layer representing the base of the tamper-evident format (26). The secure four layers cannot be captured or duplicated with any comparable quality, they are durable and will have zero degradation over their lifetime and they provide brand protection (26).
Not only can this technique be used for colour-shifting tags but it can also be used for PUFs based on a thin film of Silver (Ag), Zinc Oxide (ZnO) and Polyvinyl Pyrrolidone (PVP) (29). Specifically, the PUFs can harness the low adhesion of the silver to a glass base, and the degradation of the material during a ZnO deposition to then induce structural randomness (29).
7. QuantumTags: The Overview
The ideal anti-counterfeiting technique must be inexpensive, mass-producible, non-destructible, unclonable and convenient during authentication. A current anti-counterfeiting technique uses inkjet technologies which have been duplicated by attackers in the past due to their uniform patterns (31). However, the technique itself was very well thought out as it had low production costs, mass production, unlimited pattern design and varying ink materials. Instead, we can use the same technique to keep all of the same advantages but use quantum dots to add an infinite number of patterns (31). This solution is fabricated through inkjet printing using II-VI semiconductor core-shell QDs (31). The overview is with the usage of distributed poly (methyl methacrylate) (PMMA) nanoparticles, they can be added to a substrate with stochastic or random pinning points at a three-phase contact line of ink droplets. This way, upon the solvent evaporation, the three-phase contact lines can be pinned around the pinning points (31). The QDs in the ink droplets would subsequently be deposited on the pinning points, forming these physically unclonable patterns and by utilizing red, green, and blue (RGB) emission QDs, full-colour pictures are generated, allowing for an AI algorithm to determine counterfeit goods (31). The first layer is authenticated with the naked eye, the second layer is authenticated with the deep learning model which decodes the pattern with various focusing degrees, brightness, rotation angles and amplification factors (31).
8. QuantumTags: The Breakdown
8.1 — Fabrication
(Figure 10) The fabrication process revolves around the II-IV semiconductor (RBG emitting) light, chosen due to its high fluorescence quantum yield and stability31. The sizes of the QDs are red — 6.5 ± 2 nm, green — 10.5 ± 3 nm and blue — 11.0 ± 2.5 nm31. The relationship between the size and emission wavelength with quantum confinement showcases that a smaller size will have a larger band gap and therefore, a shorter fluorescence wavelength. (Figure 10) (a) This is determined by the composition of the QD. In this case, the surface had oleic acid and the Fourier-transform (sinusoidal functions) had absorption peaks at 2923 nm and 2854 nm. This corresponds to the number of C-C and C-H stretching sections of the -CH2- groups present in the acid. The surface groups are extremely important for the preparation and stable quantum dot inks as these stabilize the QDs in non-polar solvents for months on end31. (b) The surface modification of the substrate with a partly dissolved PMMA solution is done through spin coating where the solution is on the center of a thin film and is quickly rotated, dispersing the liquid. (c) The substrate is randomly distributed with pinning points of PMMA nanoparticles on the PMMA film. (d) Then, the inlet printing of the quantum dots begins onto the substrate. (e) The pinning process occurs when the QD droplets create the contact line at random pinning points. (f) The central and original ink is torn into smaller droplets during the evaporation process. (g) The droplets are then dried, creating the security label with patterns, and finally protected with a thin transparent film. After the film is cleared, an oxygen plasma treatment (with new advancements, this can be avoided) is used to remove the dust31.
The inkjet printing is conducted by continuously jetting the QD ink droplets with roughly 130 µm on the substrates where two adjacent droplets are 200 µm apart. If each security label had 1000 droplets, 1600 labels would be created within 5 minutes with a single nozzle printing machine (including drying time) however, this can be fastened with new advancements. Then, with the evaporation of the solvent, the contact line continuously slides and shrinks meaning the PMMA areas and pinning points stretch and pin the contact line, distorting the fluid convex and forming irregular QD ink patterns (31). During the solvent evaporation, the QD concentrations gradually increase and when the maximum point is reached, the depositing of the stochastic PMMA points on the three-phase contact line occurs, forming the unique QD patterns. As the droplet shrinks, a smaller volumed droplet is more susceptible to being disrupted by the pinning points, splitting them into several smaller sub-droplets (31). Plus, due to the smaller space between them, stochastic factors, such as airflow, cause the droplets to merge to split irregularly before drying, adding more factors to make the tag unique.
These labels are used as PUF codes. The security labels are coated with a thin, transparent sticky gel film, used to protect the film against damage. For these QDs, strong fluorescence intensity is needed. The maximum intensity for RGB QDs is 459 nm, 532 nm and 631 nm, respectively (31). (Figure 11) This figure represents the fluorescence spectra. (Figure 12) The graph showcases the time-resolved fluorescence decay measurements of RGB QD inks excited at 375 nm. During fabrication, it was proven that even under the same conditions, it still produced entirely different results.
The storage capacity of the labels can be denoted by ɭm where ɭ represents the encoding capacity and m represents the number of PUF patterns (31). Through a binary bit model, the storage capacity estimation of one red QDs is 4.7 x 10^202 meaning that a security label with 1000 red QD patterns would have a capacity greater than 4.7 x 10^202,000, indicating that it is unclonable, even by the manufacturer. These remarkable capabilities of the PUF will not go to waste, in fact, even after two months of being excited with UV light, there were no obvious decreases in fluorescent brightness (31), showcasing the chemical stability. To prove this unique fabrication, in figure 13, images g-i showcase that all samples are printed under the same conditions but are completely different.
8.2 — Three-Layer Authentication and New Quantum Dot Advancements
The labels have easy detection and rich anti-counterfeiting features. The labels have three security levels: a naked eye-decodable micropattern, a portable microscope-decodable pattern and a high-resolution microscope-decodable nano labyrinthine (complex) texture improving by difficulty as the layers progress (31). Overall, the QD ink is printed on premodified substrates with randomly distributed glass microspheres or pinning points. The nanoscale labyrinthine pattern forms around the pinning points due to the three-phase contact line during the evaporation of a droplet. Although inkjet printing is the current method, with new advancements in quantum dot technology, the QDs form nanopatterns through self-assembly after drying in the QD dispersion. In fact, since the dots self-assemble to form a maze pattern near the pinning points, they do not need a substrate hydrophilic distribution meaning it is unnecessary to use plasma treatment equipment, reducing the overall price (32). These self-assembly QDs still use the commercial II-VI semiconductor core-shell QDs, specifically CdSe, CdS and CdZnS. Just as in the previous method, the microspheres are deposited on the substrate by spin coating and then the ink droplets fall onto the substrate with microspheres on the surface (32). They spread out as the solvent gradually evaporates and the QD patterns are deposited as they dry.
(Figure 15) (a) The substrate is covered with the microspheres by spin coating, (b) the substrate after the randomly distributed pinning points have deposited, (c) the inkjet printing of the configured QD-ink onto the substrate, (d) the microspheres randomly pin and slip, (e) the ink droplets evaporate, (f) the final label is formed (32). The labels showcase how complicated the labyrinth network really is because by looking at different surfaces under the UV excitation, light and dark Newtonian rings appear between the lower surfaces of the microspheres and the upper surface of the substrate.
A method to make the labyrinthine pattern even more unique is the deposition of the QDs at different locations within the same droplet range as each range has different priorities which go based on order where the region least affiliated with the microsphere is in the least priority (32). It is vital to also consider droplet transportation and QD migration. Droplet transportation is where the solvent will slide and contract as it evaporates and the microspheres must act as stapling points. QD migration is where the QDs within the droplets cause a Marangoni Reflux Phenomenon which describes the flow of these dots. This phenomenon occurs due to the fact that QD inks with different concentrations of elements are printed. The QDs are deposited at the edges of the slip path throughout evaporation and after the full evaporation process, the deposited QDs self-assemble to form a circular pattern, outlining the shape of the droplet at each moment of the sliding process (32). This entire sliding and contraction during evaporation are due to the reflux motion and the deposition of QDs on microspheres since a significantly higher number of QDs on the microspheres results in higher photoluminescence (PL) brightness used to design the patterns. Overall, the number of microspheres covered by one ink droplet varies from 20 to 100. This three-level system has the first level of microscopic patterns visible to the naked eye, the second level allows portable observation with smartphones and external lenses and the third level must be seen through an advanced fluorescence microscope.
8.3 — Deep Learning Authentication
The QuantumTag system needs a database to store the images of these QDs. In this case, a deep learning algorithm was created as opposed to a machine learning platform which would be more time-consuming and has a false positive rate of 20% when using quantum dots. The ML classification can only be used for PUFs which can be transformed into private keys with their CRPs, however, it is not efficient for QDs (31). Deep learning on the other hand is a type of machine learning where the network learns from much larger amounts of data.
Creating the database has a few steps before it is fully functioning (32). (1) The first step begins with each QD security label being imaged with an advanced fluorescence microscope from the manufacturer. Randomly shifting and rotating the image results in a larger dataset of images where the model learns from the emission light and dimensions of the QD. (2) Once the images are trained on the AI, they are characterized in a general manner based on the geometry of the label and stored in the database on a deep learning engine, all done by the manufacturer. (3) Eventually, when the consumer receives the product, they use their smartphones to read out the PUF codes by taking pictures of the tag on the app. (4) The pictures are automatically sent to the deep learning model for authentication. (5) Finally, the engine immediately sends the authentication results (real or fake) to the user.
(Figure 16) An example is shown in the image on the left. B-G represents a library of six single dot pattern security labels. H-M showcases six fluorescent images taken from B (the genuine product) with varying brightnesses, sharpness, rotation angle, magnification, etc. N-S displays six fluorescent images which are not in the database or conform to the geometry of the original, showing how these are fake. The model is constantly being improved. Six QD security labels were chosen to establish a database. 500 images of the tag were obtained by shifting, rotating and adding varying magnifications to the tags (13). 72 images were then randomly selected, and each showed identical geometric characteristics. These 500 images were then divided into 80% for learning and 20% for validation, and after 1000 learning cycles, the model would have an accuracy of 97 to 100 percent.
8.4 — QuantumTags: The Advantages
The QD-inkjet technique has many advantages in its usage. The developed printing strategy for label fabrication allows for various pattern designs, making mass production a much lower cost. The QD ink is active in fluorescence, guaranteeing readout signals by one’s smartphone. The QD security label is only visible under UV light excitation, offering the nanotech security on the second layer (31).
9. Conclusion
With QuantumTags on every single product, counterfeiters have no way to duplicate any product as it is impossible to copy the exact dimensions and emission/fluorescence spectra of any quantum dot, and this is all verified through the AI application powered by a deep learning engine with <4% error rate. After microspheres are added onto the substrates through spin coating, the quantum dot inks are randomly added onto the substrate and during evaporation, they slip and contract and self-assemble as they dry. They are added onto products, added to the deep-learning database and the rest is in the hands of the user. With just the scan of the chip and the click of a button, users can prevent the danger brought by these counterfeit goods. The accessibility of the application to those most susceptible to counterfeit attacks allows for the rate of counterfeiting to decrease rather than continue this epidemic. The current scope for application is limited as the technology is proved and unravelled, but with the quantum world and nanoscale characteristics, anti-counterfeiting will reach an entirely new level where counterfeit goods are impossible to create.
10. Bibliography
- https://www.globenewswire.com/news-release/2019/09/25/1920707/0/en/Quantum-Materials-Corp-Announces-Anti-Counterfeit-Focus-with-QDX-Ledger-Blockchain-Platform.html
- https://www.forbes.com/sites/daveywinder/2020/08/18/goodbye-passwords-hello-unbreakable-quantum-ids-containing-1000-trillion-atoms-quantum-base-qid-lancaster-university/?sh=3db9b08a6b53
- https://quantumbase.com
- https://www.youtube.com/watch?v=_JELdbk_iGU
- https://2hgci74su71345nya32zsxjw-wpengine.netdna-ssl.com/wp-content/uploads/2017/12/Infographic_Electronic_Logo.pdf
- https://2hgci74su71345nya32zsxjw-wpengine.netdna-ssl.com/wp-content/uploads/2018/07/1-Nature-Sci-Rep-1.pdf
- https://www.accenture.com/t00010101T000000Z__w__/gb-en/_acnmedia/PDF-62/Accenture-2017CostCybercrime-US-FINAL.pdf#zoom=50
- https://quantumbase.com/sse/problem/
- https://quantumbase.com/sse/solution/
- https://2hgci74su71345nya32zsxjw-wpengine.netdna-ssl.com/wp-content/uploads/2017/12/Nature_QRNG_Dec_2017.pdf
- http://people.csail.mit.edu/rudolph/Teaching/Lectures/Security/Lecture-Security-PUFs-2.pdf
- https://www.analyticsinsight.net/how-quantum-id-secured-tag-implements-supply-chain-traceability/
- https://www.google.com/url?sa=i&url=https%3A%2F%2Fwww.intrinsic-id.com%2Fphysical-unclonable-function%2F&psig=AOvVaw2haNGlwTG3jdvSWA_pj0kQ&ust=1653648736128000&source=images&cd=vfe&ved=0CA0QjhxqFwoTCKjdi97__PcCFQAAAAAdAAAAAB
- https://www.csl.cornell.edu/~suh/papers/dac07.pdf
- https://www.design-reuse.com/articles/50693/securing-the-ic-supply-chain-integrating-puf-based-hardware-security.html
- https://sci-hub.hkvisa.net/10.1109/access.2015.2503432
- https://www.natlawreview.com/article/get-real-preventing-counterfeit-product-blockchain
- https://www.bcg.com/publications/2019/stamping-out-counterfeit-goods-blockchain-internet-of-things-iot
- https://www.azonano.com/article.aspx?ArticleID=1932
- https://davidkigerinfo.wordpress.com/2016/09/18/nanotechnology-the-next-supply-chain-management-revolution/
- https://nanopinion.archiv.zsi.at/sites/default/files/briefing_no.8_nanotechnologies_for_anti-counterfeiting_applications.pdf
- https://spectrum.ieee.org/how-nanotech-can-foil-counterfeiters
- https://www.spacedaily.com/news/nanotech-05zzd.html
- https://www.azonano.com/news.aspx?newsID=39195
- https://www.nanosecurity.ca/wp-content/uploads/2020/07/LumaChrome_Whitepaper_BP.pdf
- https://www.nanosecurity.ca/wp-content/uploads/2021/10/4-layers-of-security-with-Nanotech-LumaChrome-Holographic.pdf
- https://www.nanosecurity.ca
- https://www.nanosecurity.ca/livelogo-tamper-evident-security-labels/
- https://www.mdpi.com/2079-4991/12/8/1279/htm
- https://opg.optica.org/oe/fulltext.cfm?uri=oe-25-11-12710&id=366938
- https://www.nature.com/articles/s41467-019-10406-7
- https://sci-hub.hkvisa.net/10.1021/acsami.1c02390
