In this paper we used the Web of Science (WoS) database as a data source and through advanced search (TS= biosensor) we obtained 86,814 papers in the range time (01.01.1974–01.07.2023).

Figure 1 shows the number of publications in the field of biosensors from 1974 to 2022. We have chosen 2022 as the upper limit of the time frame because this is the last full calendar year in the WoS database (2023—not yet over). The graph shows a clear increase in publications over time, indicating a growing interest in biosensors, most likely related to the increase in environmental pollution (Pohrebennyk et al., 2018) and the widening range of biosensor applications [medicine (Grabowska et al., 2014), environmental protection (Klos-Witkowska, 2016), food industry (Dirpan et al., 2023) and the drive to improve existing devices to achieve even better detection quality (Laad and Ghule, 2022)]. It is worth noting the slight decrease in the number of papers published in 2020 (5,875 papers) compared to 2019 (5,902 papers), perhaps due to the impact of the COVID-19 pandemic, which was felt all over the world.

Current research is carried out in two parallel streams: experimental and theoretical.

In the experimental stream, work focuses on improving the quality of detection and the stability of the device (Klos-Witkowska and Kajstura, 2020), while the theoretical stream revolves around the development of mathematical models (Martsenyuk et al., 2021b), often using already existing solutions for this purpose (Martsenyuk et al., 2021a), also recently, the interest of scientists in combining biosensors with artificial intelligence has been noted. An extremely interesting issue is the problem of stability in biosensors, which is a very important problem not only from a scientific point of view, but also from a commercial point of view. The stability of biosensors directly translates into the longevity of use and the ability to operate the device. Therefore, following the latest scientific trends, the issue of biosensor stability was chosen to be analysed in biosensor development based on advanced analytics using CiteSpace 6.2.R4;

(1) Identifying the most important (so called “explosive articles”) that generated the most interest among researchers based on their citability.

We conducted a literature search using the following search strategy using Web of Science Advanced Search: Publication dates were from 01.01.1974 to 01.07.2023 and search criteria were: for biosensor stability (experimental research): [TS=(biosensor)] AND [TS=(stability)] AND [TS=(empirical)] OR [TS=(experimental)] in the Web of Science Core Collection.

We created visual graphs of keyword co-occurrence, keyword clustering and scientific collaboration using CiteSpace 6.2 R4 software. We investigated the current status and trends in the stability of biosensors.

Biosensors are devices that are susceptible to ageing; this phenomenon can be characterised as a decrease in signal over time.

Stability of biosensors is critical for commercial success as biosensors are now used in an increasing number of different applications. Characterisation of stability in terms of shelf life, reusability and the ability to be used continuously is inadequate, although much work has been done in this area. Stability is very important as it is a major factor affecting the operation of the device.

The mechanisms of biosensor ageing are complex and still poorly understood. However, it is known that the loss of stability of a biosensor is the sum of all changes affecting both the biological material used: for example: enzymes (Fernandez-Lopez et al., 2017), antibodies (Oyetayo et al., 2017), as well as the signal mediator (Ricci et al., 2003; Malinauskas et al., 2004) and the binding material of complexes in the matrix (Panjan et al., 2017).

The analysis presented below illustrates the current state of knowledge and trends in biosensor stability research. As shown in Figure 2, the efforts of researchers have focused on chemically modified electrodes, reduced graphene oxide, direct electron transfer, magnetic microspheres, FFT cyclic voltammetry, poly(o-phenylenediamine), LECTIMS and optimization. The analysis presented was carried out on the basis of the most frequently occurring keywords. These were used to create the clusters shown in Figure 2.

The larger the cluster area, the more frequent the occurrence of the keyword. In the analysis carried out, the keywords are also the names of each of the most relevant clusters, those shown in Figure 2.

Thus, it can be seen that the keyword for cluster #1 (#1) is chemically modified electrode, similarly for the others: #0 reduced graphene oxide, #2 direct electron transfer, #6 magnetic microspheres, #8 fft cyclic voltameters, #10 poly(o-phenylenediamine), #11 lectins, #12 optimisation.

The dots in each cluster represent publications; the names of the most prominent can also be seen in the diagram.

The CiteSpace analysis allows us to select the most significant sentences from the publications contained in the clusters.

Thus, cluster: #0 “reduced graphene oxide” was formed on the basis of 495 publications, of which the publications (Bai et al., 2012; Dong et al., 2012; Wei et al., 2012; Xu et al., 2012; Bai et al., 2013; Wang et al., 2013; Nalini et al., 2014; Zhou et al., 2014; Li et al., 2021) are considered to be the most significant, where one can find information on the application of gold nanoparticles and the development of an amperometric biosensor based on gold nanoparticles decorated with graphene nanoparticles for the detection of clenbuterol. Also of interest are electrochemical biosensors based on hemin-modified graphene nanoplatelets for the determination of l-tyrosine levels. Au nanoparticles were used as a stabilizer.

In cluster #1 “chemically modified electrode”; among the 243 publications forming it, the publications (Arkhypova et al., 2003; Gamella et al., 2006; Mu, 2006; Rahman et al., 2006; Wang and Zhang, 2006; Du et al., 2007; Fan et al., 2007; Fanjul-Bolado et al., 2007; Di Fusco et al., 2010; Chalikian, 2016) are highlighted. The main information in this cluster concerns: the optimization of the biosensor design, as well as the influence of experimental variables such as pH, working potential and temperature on the sensor response. The research also focused on improving the analytical characteristics of the fabricated biosensor, studies of direct electron transfer reactions, optimisation of experimental conditions and substrate affinity based on the Michaelis-Menten constant. pH, working potential and temperature are known as stabilizing factors.

Cluster #2 “direct electron transfer” consists of 460 publications. The ten most important publications are Lai et al. (2008), Zhang et al. (2008), Hsu et al. (2009), Kong et al. (2009), Nenkova et al. (2009), Zhang et al. (2009), Zhu et al. (2009), Liu et al. (2010), Lu et al. (2010), and Garay (2015). They contain information on factors influencing electron transfer. These include the results of studies describing the effect of pH, the preparation of new magnetic microspheres coated with chitosan by modifying magnetic carbon-coated iron nanoparticles. The effect of gold nanoparticles on electron transfer was also investigated.

Cluster #6 “magnetic microspheres” consists of 215 papers. The most important of these are papers (Liu et al., 2005; Tan et al., 2005; Shan et al., 2006; Fan et al., 2007; Tong et al., 2007; Liu et al., 2010; Zhen et al., 2011; Heli and Yadegari, 2014; Ltaïef et al., 2017; Villani et al., 2018) devoted to the preparation of new magnetic microspheres, control of sensor potential, detection limits and response speed from the point of view of biosensor stability.

There are 62 publications in cluster #8 “fft cyclic voltameters.” The most important are (Norouzi et al., 2010a; Norouzi et al., 2010b; Norouzi et al., 2010c; Marinov et al., 2010; Bohnenberger and Schmid, 2014; Cheraghi et al., 2017; Narwal et al., 2017; Stepashkin et al., 2018; de Oliveira et al., 2023; Kyomuhimbo et al., 2023). The key information is the application of this method to the study of: pyruvate oxidase, multiwalled carbon nanotubes immobilized on the surface of a glassy carbon electrode by means of a polymer layer of Nafion, the effect of individual components of an enzyme mixture containing gold nanoparticles, acetylcholinesterase, bovine serum albumin and glutaraldehyde on the current output of constructed acetylthiocholine biosensors. The research also included a biosensor for the detection and quantification of organophosphorus pesticides. fft cyclic voltameters allow us to investigate the stability of biosensors.

Cluster #10 “Poly(o-phenylenediamine)” was formed by 130 publications, of which the ten most important are publications (Ahammad et al., 2011; Devi et al., 2013; Zhang et al., 2014; Dervisevic et al., 2015a; Baytak et al., 2015; Dervisevic et al., 2015b; Chen et al., 2015; Tan et al., 2015; Wang et al., 2015), while clusters #11 “Lectins” and #12 “Optimization” were formed by 65 and 163 publications, respectively. The most important for cluster #11 are Bai and Shiu (2014), Gholivand et al. (2014), Oliveira et al. (2014), Ribeiro et al. (2014), Ghanbari et al. (2019), Bravo et al. (2022), Jalalvand (2022), Tvorynska et al. (2022), Zalpour et al. (2022), Ramesh et al. (2023), while for cluster 12 (Hsu et al., 2009; Kong et al., 2009; Liu et al., 2011a; Liu et al., 2011a; Liu et al., 2011b; Pasahan et al., 2011; Yang HC et al., 2011; Yang WY et al., 2011; Makhmudiyarova et al., 2015; Rakhi et al., 2016) with information on operational stability as well as sensitivity, conductivity, detection limits, response speed, detection range, calculation of the Michaelis-Menten constant, response gain.

Diagram 3 (Figure 3) shows an analysis of the most important publications in the cluster that emerged over time. The red circles indicate the most explosive publications, i.e., those that have had the greatest impact on science. These publications were selected based on the highest number of citations of words (extracted from titles, abstracts, descriptors, and identifiers of bibliographic records) found in these publications in the presented time scale.

The time scale in the graph makes it possible to determine when the explosive publication appeared. A detailed analysis is shown in Figure 4.

Thus, in the context of the stability of biosensors, it can be seen that in 2003 there was a publication (Wang et al., 2003, 267 citations) dedicated to the amperometric biosensor of hydrogen peroxide with a sol-gel/chitosan layer as immobilization matrix, which the authors describe as the development of a new hybrid sol-gel/organic composite material based on the cross-linking of the natural polymer chitosan with (3-aoryloxypropyl) dimethoxymethylsilane. This material was used to fabricate biosensors for H₂O₂ amperometry.

A composite membrane was used to immobilize horseradish peroxidase (HRP) on a gold disc electrode. The properties of the sol-gel/chitosan and sol-gel/chitosan-HRP layers were thoroughly characterized by atomic force and Fourier transform infrared microscopy. Using fluorescent tracers, the protein density in the sol-gel/chitosan was calculated to be 3.14 × 10¹² molecules cm⁻². The developed biosensor had a fast response in less than 2 s with a linear range of 5.0 × 10⁻⁹–1.0 × 10⁻⁷ mol L⁻¹ and a detection limit of 2 × 10⁻⁹ mol. L⁻¹. The reaction showed a typical Michaelis-Menten mechanism. The activation energy of the enzymatic reaction was calculated to be 8.22 kJ mol⁻¹. The biosensor retained about 75% of its initial activity after about 60 days of storage in phosphate buffer at 4°C.

This was followed in 2004 by another publication (Wang and Wang, 2004, 415 citations) on a novel hydrogen peroxide sensor based on horseradish peroxidase immobilized on a colloidal Au-modified ITO electrode. The authors described the development of a novel method to fabricate a hydrogen peroxide sensor by immobilizing horseradish peroxidase (HRP) on a colloidal Au-modified conductive ITO glass substrate. The purified glass substrate was first modified with (3-aminopropyl) trimethoxysilane (APTMS) to provide an interface for the deposition of colloidal Au. Next, 15 nm colloidal Au particles were chemisorbed onto the amine groups of APTMS. Finally, HRP was adsorbed onto the surface of colloidal Au. The immobilized HRP showed an excellent electrocatalytic response to hydrogen peroxide reduction. The performance and factors affecting the obtained biosensor were investigated in detail. The obtained biosensor showed a fast amperometric response (within 5 s) to H₂O₂. The detection limit of the biosensor was 8.0 μmol L⁻¹ and the linear range was from 20.0 μmol L⁻¹ to 8.0 mmol L⁻¹. In addition, the obtained biosensor showed high sensitivity, good reproducibility and long-term stability.

This is an article that is particularly important for the development of the issue related to the stability of biosensors. Despite the fact that the topic of biosensor stability is developing rapidly. An analysis conducted with the help of CiteSpace showed that no explosive articles have appeared in the last 5 years, which formed the basis for further scientific issues related to the topic of stability.

The next landmark publication was a paper (Qian and Yang, 2006, 284 citations) describing a composite layer of carbon nanotubes and chitosan for the fabrication of an amperometric hydrogen peroxide biosensor. The paper described the development of a novel amperometric hydrogen peroxide biosensor based on the cross-linking of horseradish peroxidase (HRP) with glutaraldehyde with multi-walled carbon nanotubes/chitosan (MWNTs/chitosan) coated in a composite layer on a glassy carbon electrode. The MWNTs were first dissolved in a chitosan solution. The morphology of the MWNT/chitosan composite layer was then characterised by field emission scanning electron microscopy. The results showed that the MWNTs were well soluble in chitosan and that robust layers could be formed on their surface. HRP was cross-linked with MWNTs/chitosan by glutaraldehyde to prepare a hydrogen peroxide biosensor. The enzyme electrode showed excellent electrocatalytic activity and fast response for H₂O₂ in the absence of mediator. The linear detection range of H₂O₂ (applied potential: −0.2 V) was from 1.67 × 10⁻⁵ to 7.40 × 10⁻⁴ M with a correction factor of 0.998. The biosensor showed good reproducibility and stability in the determination of H₂O₂. There was no interference from ascorbic acid, glucose, citric acid and lactic acid.

The last of the four most explosive scientific publications was a paper (Zhou et al., 2010, 441 citations) on a novel hydrogen peroxide biosensor based on Au-graphene-HRP-chitosan biocomposites. The paper used graphene very effectively to construct an H₂O₂ biosensor. Graphene and horseradish peroxidase (HRP) were co-immobilised in a biocompatible chitosan (CS) polymer, and then a glassy carbon electrode (GCE) was modified with the biocomposite, followed by electrodeposition of Au nanoparticles on the surface to form an Au/Graphene/HRP/CS/GCE layer. Cyclic voltammetry showed that direct electron transfer of HRP was realized, and the biosensor had excellent performance in terms of electrocatalytic reduction towards H₂O₂. The biosensor exhibited high sensitivity and speed. In addition, the biosensor showed good reproducibility and long-term stability.

As can be seen from the most explosive publications, the stability of biosensors is primarily related to the generation of optimal substrates (i.e., receptor layers).

The range of words considered to be the most explosive, i.e., those that appear most frequently in citations, is shown in Figure 5.

CiteSpace 6.2 R.4 software was used to highlight keywords. The tool used performed a keyword analysis, i.e., it extracted the most frequently cited words. The graph shows a time scale indicating the period of the highest citation of a given keyword. It can be seen that although the analysis covered works from 01.01.1974 to 01.07.2023, the graph shows the period from 1991 to 2023. It can be seen that the dominant keyword is “glucose oxidase,” this term has also been cited the longest, as can be seen in the figure. A similar reasoning can be applied to the other words in Figure 5.

Analyzing the entries below, the terms that deserve special attention are those that remain in quotation marks to this day. Among them we can distinguish: “ Graphene oxide,” “high-performance nanocomposite,” “nanocomposites,” “electrochemical sensor,” “platform,” “sensitive detection,” “surface plasmon resonance,” “amplification” and “immunosensor.”

It can be seen that these words refer both to the substances or materials tested, e.g., non-composite, and to the test methods, e.g., “surface plasmon resonance,” but also to the type of biosensors, e.g., “immunosensor,” “electrochemical biosensor,” or to research questions, e.g., “sensitive biosensor,” “platform.”

Based on the most recent citations of keywords (the above terms are related to the stability of biosensors), we can identify scientific trends and directions of development of a given scientific issue. From the point of view of stability of biosensors, it can be seen that research is growing in the direction of electrochemical biosensors and immunosensors in search of materials that are optimal in terms of durability. The experiments mainly use surface plasmon resonance as a leading research method.

From the point of view of scientific development, international cooperation in the area of development of a particular issue is extremely important. Cooperation between countries, institutions and authors is more likely to lead to progress in related research areas.

Figure 6 shows a map of cooperation and networking between countries in the field of biosensor stability. The analysis showed that researchers from 69 countries were working on biosensor stability. The areas marked with circles show the citation rate of representatives of a country (the detailed number of citations of representatives of a country is presented in Table 1). In Figure 6 you can see that the larger the marked area, the higher the citation rate (the graph does not show all countries, but only those whose representatives turned out to be the most active in terms of cited works). Thus, based on the figure below, it can be seen that the most cited works devoted to the stability of biosensors were published by representatives of China (437 citations), but also among the leaders can be distinguished representatives of: Iran (70 citations), United States (57 citations), India (44 citations), Italy (40 citations), South Korea (33 citations), Taiwan (28 citations), Spain (25 citations), France (25 citations), Turkey (24 citations). In 2012, the work of scientists from Poland was also cited.

In the diagram shown in red is the publication of the French representative (as first author) and co-authors (from the United States and Hungary), which generated the widest cooperation in the area of biosensor stability issues. This publication (Theavenot et al., 1999) is extremely important not only in terms of international cooperation, but also in terms of citations and content. The paper is dedicated to the recommendation, classification and definition of electrochemical biosensors. It contains a detailed description of guidelines for reporting biosensor response and calibration characteristics: sensitivity, working and linear concentration ranges, limits of detection and limits of quantification. You will also find information on biosensor selectivity and reliability, response time, reproducibility, stability and device lifetime.

Table 2 presents list of journals in which the most articles related to biosensors have been published. It shows that the highest ranked item in terms of number of citations are the journals: Biosensors and Bioelectronics, Analytical Chemistry, Sensors and Actuators B: Chemical, Analytica Chimica Acta, Talanta and so forward as presented in Table 2.

Comparing the journals in which the most significant articles were published (Figure 4) with the list in which the largest number of articles related to biosensors are found (Table 2), it can be seen that the most cited articles were published in journals: Biosensors and Bioelectronics, Electrochimical Communication, Talanta and Electrochimica Acta. In the table presented (Table 2), the journal Biosensors and Bioelectronics is in the number 1 position, Talanta is in the number 5 position, Electrochimica Acta is in the number 7 position, while Electrochimical Communication is not among the top ten most cited journals.

Thus, it can be seen that there is a kind of correlation between highly cited articles and their impact on the journal’s citability, while the publications with the highest citations do not fully reflect the journal’s cutability.