Where is research on fossil fuels going in times of climate change? A perspective on chemical enhanced oil recovery

In a society where the global search for renewable and sustainable sources of energy, chemicals and materials has rapidly become a priority, due to the challenges posed by climate change, one may think that there is no place anymore for research regarding extraction and production of fossil fuels. However, there is no denying that coal, oil and gas still constitute the major source of energy worldwide by far, and humanity will be dependent on them still for decades. As a matter of fact, the global demand of fossil fuels is still increasing rather than decreasing. However, fossil resources are consumed at a much higher rate than they are generated. Only focusing on crude oil, the chances of finding new conventional reservoirs are nowadays low, and this conventional sources are quickly depleting, therefore it is necessary to turn more and more the attention to unconventional sources, such as carbonate reservoirs, tar sands, oil shales and deep off-shore reservoirs. Chemical enhanced oil recovery (cEOR) techniques are already well established in improving macroscopic and microscopic oil displacement, increasing oil production after primary and secondary methods are exhausted. However, these are mostly implemented in reservoirs that normally possess favorable conditions. Research can still contribute in making these processes more efficient and sustainable, especially for unconventional oil, for as long as we will need to extract fossil fuels. After supporting the idea that exploitation of fossil resources cannot stop just yet, this perspective article will highlight recent advances in the field of cEOR, mentioning new developments in traditional methods, such as polymer and surfactant flooding, as well as newer methods, such as nanoparticles, preformed particle gels, and smart waterflooding, plus the various possible new combinations (hybrid EOR). This paper will finally try to shortly outline future directions of cEOR research, with emphasis on sustainable methods, such as CO2 flooding combined with carbon capture and storage strategies, and newly available research tools, such as reservoir-on-a-chip and machine learning.

Fulfilling the ever-increasing global demand of energy, chemicals and materials, while reaching carbon neutrality, is one of the most important objectives of this century. The target set in the Paris Agreement-that is to limit global warming to well below 2°C, and possibly 1.5°C-involves obtaining negative CO 2 emissions by 2050. [1] It has been estimated that the amount of emissions that needs to be avoided by that time, to reach such goal, is the astounding total of 800 Gtons of CO 2 . [2] That is why terms such as sustainability, carbon footprint, and circular economy have rapidly become prominent in energy research and all related fields. [3,4] It is immediately obvious that negative emissions would be impossible, if we continue to use carbon-based fuels as source of energy, directly or indirectly, without doing anything about it. In coal, gas or other hydrocarbons, the energy is stored in C-C and C-H chemical bonds, and it can be only harvested conveniently by reaction with molecular oxygen present in the atmosphere, which ultimately results in formation of water and CO 2 when the combustion is complete, which is also hard to achieve. Humankind has been doing this for centuries-since the start of the first industrial revolution-producing an excess of CO 2 in the atmosphere, that now constitutes by far the main greenhouse gas. The impact of anthropogenic climate change is nowadays undeniable. [5] Starting around 1850, the level of atmospheric CO 2 went from the stable 280 ppm of pre-industrial era, to the current 420 ppm. [6] The main contribution to emissions comes from the energy sector and transport (73%), followed by agriculture, forestry and land use (18%), then industrial processes (5%, prevalently cement and chemicals), and waste (3%). [7] The inverse process of converting CO 2 into energy storage molecules, performed in nature via photosynthesis, does proceed at a much slower pace. Therefore, the unbalance between formed and captured CO 2 is destined to remain, if we do not put all our efforts toward an energy transition, to make us independent from fossil fuels, as quickly as possible.
It seems that the best-or maybe even the only-way to transition to a perfectly carbon neutral society, is to immediately stop using carbon-based fuels altogether, in favor of renewable and clean sources. But are renewable energies currently enough to meet the global demand? It should not come as a surprise that the answer is no, but maybe not many have an idea of how far we are from replacing fossil fuels with sustainable and green alternatives. Research in renewable energy have made incredible progress in the last decades, yet, fossil fuels still contributed to more than 84% of the total energy consumption worldwide in 2019 (Fig. 1). [8] In absolute values, the energy provided by fossil fuels went from about 20.000 TWh in 1950 to more than 140,000 TWh in 2019, and it is still increasing. Renewables are catching up very slowly with their contribution growing from 6% of the total in 1965, mostly from hydropower dams, to 11% in 2019. [8] Global investments in fossil energy are still substantial and, according to the latest IEA report, although renewable dominates new power investments, upstream oil and gas investment is still expected to increase by 10% in 2021. [9] Projections tell us that the demand of oil, coal and gas will continue to increase until 2050, rather than decrease. [10] Additionally, it has been estimated in 2012 that to replace one unit of fossil fuel, keeping into account economic and human factors, about ten times more units of non-fossil fuels are necessary. [11] There are also other issues that make difficult replacing completely fossil fuels, even if enough renewable energy would be available. Renewables are mostly stored as electric energy, which is not always a suitable carrier. For example, some heavy transports (trucks, planes, ships) cannot be conveniently powered by batteries, due to their low volumetric efficiency, [12] while fossil fuels are ideal in this respect-which is logical, considering that engines and transportations were designed around them. Moreover, the materials used for batteries are relatively rare, expensive, and need a good waste management and efficient recycling protocols. [13] In this respect, biofuels could represent a better alternative to electric power in the transport sector, provided that waste generation in their production is kept to a minimum, but it will require still time and effort to make them competitive with fossil fuels. Also soil contamination due to biofuel spillage may represent a problem. [14] Conversion of other renewable energies into biofuels, making them more accessible, has been proposed and evaluated. [12] It should be also considered that biofuels come mainly from biomass, therefore indirectly from solar power, on a time scale which may still be too long to be exploited sustainably. It can be pointed out that things can still be greatly improved, as a big amount of biomass waste is produced by agriculture and forestry industry (140 Gtons/year) and it is mostly incinerated or discarded. [15] Nonetheless, biofuels still produce CO 2 emissions, and are more expensive to produce than fossil fuels. It should also be considered that we are already in deficit of bio-resources, as the Earth Overshoot Day testifies, meaning that energy and materials from biomasses cannot cover the full demand, even potentially, in the present conditions.
Hydrogen is often proposed as a suitable clean energy carrier. As its sole combustion product is water, it does not constitute a problem in terms of emission, and it does not present pollution issues due to incomplete or inefficient conversion. It also has sufficiently good volumetric efficiency, as it is a gas and it can be compressed. However, its carbon footprint depends on the way used to produce it. Ideally, it can be made in a completely sustainable way, with zero-net emissions, when produced from completely renewable sources, such as water electrolysis powered by solar energy (so-called green hydrogen); more economically viable processes are those used for example in brown and gray hydrogen (from steam reforming of coal and methane, respectively); finally a more sustainable version is "blue hydrogen", where the associated production of CO 2 is contained by in situ carbon capture and storage (CCS). [16] Biohydrogen can also be made from biomass. [15] Figure 1. Energy consumption by source from 1965 to 2019. [8] Other renewables includes biomass, geothermal, and waste energy. Reproduced under CC-BY licence.
These technologies can make H 2 a promising future fuel, [17] but again this does not necessarily means lower CO 2 emissions.
Decarbonization via CCS and CCU (carbon capture storage and utilization, respectively) technologies constitutes a key to reach negative emissions by 2050. [2,16,18] In this respect, chemical EOR has a big role, as CO 2 EOR contributes significantly to CCU technologies (estimated to reach up to 1.8 Gtons/year of CO 2 by 2050). [16,18,19] This will be discussed in more detail in the next section. CO 2 can be chemically reduced to chemicals, such as urea, polycarbonates and methanol, but not all these processes are economically viable. [16] Liquid fuels can be produced from CO 2 and water, but the catalysts needed are based on noble metals, therefore extremely expensive. [20] Other conventional decarbonization methods include production of concrete for constructions, and production of microalgae. It is difficult to foresee which of these processes will be scaled up, if any, as it all depends on unpredictable economic scenarios. [16] Another important aspect of the dependence on fossil fuels of modern society, is that the majority of the chemical industry is built around them. Fossil fuels are the main source of platform chemicals, used in chemical industry to produce any sort of products and materials, such as pharmaceuticals, detergents, paints, and of course a vast quantity of plastics and other polymers. Only considering plastics, their production surpassed the 350 Mtons in 2015, totaling more than 8000 Mtons ever produced (of which 600 have been recycled and 800 incinerated), and it is still increasing at nearly exponential rate. [21] This requires a significant use of carbon-based molecules, mostly obtained by oil refineries. There is nowadays the possibility to obtain many of the common building blocks from alternative bio-based sources, for example polyethylene can be made from bioethanol, via conversion to bioethylene. [22] However, the quantities that can potentially be produced from renewable feedstocks are still far from meeting the current and future demand of such chemicals. It has been estimated (in 2013) that even if all the bioethanol produced worldwide (used as biofuel for the transport sector) would be converted to bioethylene, this could potentially cover the 25% of the total demand of such chemical. [23] Increasing such amount, would come with the cost of competing with the food production industry, as the substrates used are sugarcane and starch sources.
The problem of sustainability in polymers has been extensively addressed, [24] and it is obviously strictly connected to climate change. Besides the more immediate problem of environmental pollution due to waste plastic accumulation, a large amount of CO 2 emissions comes from polymers and chemicals industries. Today many polymeric materials are designed with recycling and sustainability in mind, [24] but this is not always possible: a classic example is constituted by car tires, for which a cradle-to-cradle recycling is made particularly challenging, due to the difficulty to selectively de-vulcanize the constituting rubber. [25] The same can be said for many composite, elastomeric and thermoset materials.
The use of bio-based sources to produce chemicals and plastics is an active research field. [15] Cellulose, hemicellulose and lignin-based residues are promising substrates for any sort of chemicals and precursors of bioplastics and other biomaterials. As mentioned above, the large amount of waste biomass produced by agriculture and forestry industries, shows that there is a big potential for bio-based chemicals, if technologies feasible to convert this waste into useful products keep to be implemented. [15] However, the current production capacity of bioplastics is about 4 Mtons (as of 2017), less than 1% of the global year plastic production. [21] Despite all the efforts, still the majority of plastic materials and chemicals are produced from fossil fuels, and only a small fraction of these can be actually recycled.
To conclude this introduction, it has to be said that the factor mainly influencing governments policies toward sustainability or not, it is not a technological, but an economical one. Even if potentially 100% of the fossil sources could be replaced by renewable ones, this would cause instability in the global economy, and it would alter the geopolitical landscape in unpredictable ways. The use of waste is limited by transportation and processing problems, and the use of bio-based source needs to also constantly take into account competition with food.
This reinforces the idea that small steps at the time need to be taken, therefore our independence from fossil fuels may still be far to come, and it may last long after 2050.
Can we make oil production more sustainable? Recent advances in cEOR Once suggested that we cannot interrupt extraction and transformation of fossil-based fuels any time soon, we remain with the challenge to make these processes as sustainable as possible, in order to limit greenhouse gas emissions. Here, the focus will be on the production of crude oil, and in particular on chemical enhanced oil recovery (cEOR) methods, therefore thermal and miscible methods will not be discussed. General definitions and description of primary, secondary and enhanced (or tertiary) oil recovery will not be given here. The interested reader is referred to books and reviews on the subject, that gives a complete general overview. [26][27][28][29] The first observation to be made, is that both conventional and unconventional oil resources are limited, fewer new reservoirs are discovered, and their accessibility decreases with time, making more challenging to maintain economically feasible extraction processes. It is generally accepted that all the sources of cheap oil have already been found and exploited, therefore any additional discovery or technology used to increase oil production comes with increasing challenges and costs. [30] For example, an overview of about 1500 EOR projects made in 2010, showed that the majority of projects involved sandstone reservoirs, [31] but the attention is now shifting more toward carbonate reservoirs, more difficult to exploit because of low porosity, severe heterogeneity, and oil-to mixed-wet rock properties. [32,33] Other targets of EOR projects are deep off-shore reservoirs, tar sands and shale oils, characterized by very heavy oils and harsh conditions. All the methods discussed here potentially targets all these types of reservoirs, but the majority of studies have been performed for carbonate ones.
cEOR methods are implemented in many reservoirs all around the world, where primary and secondary recovery methods leave more than 60% of the original oil in place (OOIP) behind. [34] The first cEOR methods have been developed in the 1960s (although preliminary tests in the US and Canada date back to the 1940s), and peaked in the 1980s, following the increase in oil prices, due to the 1970s energy crisis in the Western world. [31] In more recent times, oil prices have normalized and implementation of cEOR projects has slowed down to some extent, but research remains quite active. [28,30,[34][35][36] cEOR involves the injection of chemicals, or mixture thereof, into reservoirs, with the main goal to mobilize oil left behind by primary and secondary recovery methods.
Traditional cEOR methods include polymer flooding, surfactant flooding, alkali flooding, and all the possible combinations of these three, such as surfactant-polymer (SP flooding), and alkali-surfactant-polymer (ASP flooding). [37,38] Other chemical methods involve injection of CO 2 , foams, nanoparticles, and low-salinity water. Again, we can have combinations of all the mentioned methods, as well as their combination with non-chemical EOR (hybrid cEOR). Microbial EOR is in fact a biological process, rather than chemical, and it will be mentioned here mostly as part of some hybrid methods. [39][40][41] In Scheme 1 the chemical methods are summarized, separating traditional ones from more advanced ones.

Polymer flooding
Polymer flooding is the oldest and most used cEOR method, and it has been extensively reviewed. [30,[42][43][44][45][46] It dates back to 1950, when it was shown that increasing the viscosity of the water phase used for water flooding, could improve oil mobility, achieving a better macroscopic sweeping efficiency of the reservoir. Surprisingly, the polymers used have basically not changed in almost 70 years of existence of this technology, the main one being partially hydrolyzed polyacrylamide (HPAM), which is essentially a random copolymer of acrylamide and acrylic acid, followed by polysaccharides (biopolymers) such as Xanthan gum. HPAM is ideal in sandstone reservoirs, which are homogeneous, possess relatively high permeability, low temperature and salinity, and are water-wet. Moreover HPAM adsorption in this type of reservoirs is limited. [30] HPAM is one of the cheapest synthetic polymers, is nontoxic (although the main constituting monomer, acrylamide, is carcinogenic), soluble in water, and it can be easily synthesized with high molecular weight, which gives it excellent viscosifying properties at relatively low concentration. A recent survey conducted on 72 documented polymer flooding project around the world, showed that 92% of them used HPAM. [47] The main disadvantages of HPAM are related to chemical and mechanical degradation, extensively occurring in the most difficult reservoirs, characterized by high temperatures, high salinity, high shear rates due to low permeability. These degradation cause reduction of molecular weight, with subsequent loss of viscosity, and increased rock adsorption and "salting out" phenomena, due to conversion of neutral acrylamide groups to charged acrylic acid, more sensitive to ionic strength and presence of divalent cations (Ca 2+ and Mg 2+ ). [30,43] The latter is especially true in carbonate reservoirs, that contains high concentrations of such cations.
The shift to more difficult and non-conventional reservoirs, has encouraged the research of alternative polymers to HPAM, able to better perform in such harsh conditions. Typical approaches are based on modified versions of HPAM. These are realized adding associative hydrophobic groups, able to enhance rheological properties, and substituting acrylamide and acrylic acid units with other hydrophilic monomers with better hydrolysis and salt resistance, such as sulfonated ones. The interested reader can consult several sources for more details. [30,[42][43][44][45][46] A recent interesting approach consists on using surface active polymers, or polymeric surfactant, as flooding agent. [28,36,48,49] This goes in the direction of combining surfactant-polymer (SP) flooding into one single chemical, with several potential advantages. [36] These can simultaneously increase viscosity of the sweeping fluid, as in regular polymer flooding, but also decrease IFT with the oil phase, causing additional oil mobilization via emulsification, capillary reduction, or wettability alteration, mechanisms typical of surfactant flooding. [30,50] Another non-standard approach can be based on playing with polymer architecture (hyperbranched vs linear), to tune the viscoelastic properties. [51,52] Particularly interesting is also the use of thermoviscosifying polymers for high-temperature reservoirs. [30,[53][54][55] This approach introduces the use of "smart" polymers, able to display the desired properties (in this case an increase in viscosity), only when and where required, at the high temperature of the reservoir.
With respect of polymer flooding, it can be expected that this will continue to be a trend: studying optimized versions of modified HPAM, with "smart" properties, to meet the requirements for harsh reservoirs conditions.

Low-salinity and "smart" waterflooding
In recent times, much attention has been given to low-salinity polymer flooding, [56] and to the related low-salinity water injection (LSWI), sometimes also called smart waterflooding. [57][58][59] The latter is not, strictly speaking, a cEOR method, because it only uses water without added chemicals, however it is included in tertiary methods, as it is performed after regular waterflooding. [58] It could be also argued that low-salinity brines tested for LSWI are prepared by dilution of seawater, but often include the addition or replacement of specific ions (hence the definition of "smart" water), therefore some chemical treatment is indeed performed. Regardless of classifications, LSWI has proven effective in sandstone reservoirs with high salinity connate water, already more than 20 years ago, but it has gained popularity only in more recent times. [60] It is generally accepted that the main mechanism of oil recovery is via wettability alteration, but other factors may be in play, depending on the types of oil and rock formation. [58,59] Moreover, the presence of certain type of ions instead of others, can reduce scaling phenomena and chemicals precipitation. LSWI is obviously very attractive, as a green and cheap EOR method.
LWSI still suffers of the drawback of having low viscosity, therefore the sweeping efficiency remains low, especially with highly viscous oils. A synergy can be created by combining it with polymer flooding. Low-salinity polymer flooding (LSP) has gained attention is recent times. [56,60] This can be considered a hybrid cEOR method, more of which will be discussed later.
LSP presents some advantages over traditional polymer flooding. First of all, due to low salinity, the viscosifying ability of the polymer (HPAM) are enhanced, since the low ionic strength reduces the electrostatic screening between charged units (acrylic acid) present on the polymer chain, resulting in chain expansion and higher hydrodynamic volume. [43] In core flood experiments, LSP has shown considerably lower polymer retention and better long-term injectivity, with respect to high salinity polymer flooding. [56] However, the salinity is increased in the produced water, because of rocks dissolution, which causes viscosity loss during the recovery process. Overall, LSP is a promising and cost-effective improvement of traditional polymer flooding, and it will probably be relevant in future oil recovery research.
Recently, the research group of the author of this perspective has proposed and investigated polystyrene-b-poly (methacrylic acid) amphiphilic block copolymers, [61][62][63] as agents for lowsalinity polymer flooding. [64] This system, due to its structure and self-assembly behavior, can be considered between an associative polymer and a nanoparticle, and it presents interesting rheological and interfacial behavior, also in presence of surfactants. [65] Surfactant flooding Surfactant flooding is another well-established cEOR method, used alone or in combination with polymers and/or alkali, in SP, AS and ASP flooding, since the 1960s. [50] As for polymer flooding, the use of surfactants in cEOR is well established, and it has been reviewed several times. [32,[66][67][68] Generally speaking, the efficiency of surfactants depends on different mechanisms and many factors (IFT reduction, wettability alteration, adsorption, salinity, temperature, viscoelastic properties), therefore the optimization of a surfactant for EOR applications will greatly depend on the reservoir and crude oil characteristics, probably even more than for polymer flooding.
Initially, the most common surfactants used for EOR were petroleum sulfonates, cheap and easy to make, but with time more sophisticated structures have been synthesized, in order to obtain ultralow IFT values and better salinity and temperature tolerance. From this point of view, gemini and/or zwitterionic surfactants possess superior characteristics, therefore these are the most studied for EOR nowadays, for difficult reservoir conditions. [50] Also very popular are biosurfactants and bio-based surfactants. The former kind is produced by microorganisms, directly in situ, therefore these are more connected with microbial EOR, a different method, not treated here. The latter refers to surfactants synthesized from biomass. These can have comparable properties than other synthetic surfactants, possessing the clear advantage of being prepared from renewable and more sustainable sources.
As for polymer flooding, the research is moving toward the direction of making surfactants that would be more effective in harsh reservoir conditions, namely high temperatures and salinity. Intensive research resulted in increasingly complex structural modification to the most simple surfactants, such incorporation of ethoxylated and propoxylated units in ionic surfactants to improve solubility in high salinity conditions, till the development of ultralow IFT systems such as zwitterionic and gemini surfactants. [50] Polymeric surfactants, able to combine beneficial effects of surfactants and polymers, analogous to SP flooding, also show some promise for future developments, although often they still present limitations due to high rock adsorption and low salt tolerance. [36] Nanoparticles Nanoparticles (NP) for cEOR are possibly the most popular recent advancement in the field, and one that shows promise to be further developed in the future. [35,[69][70][71][72][73][74] Nanoparticles can improve oil recovery by effectively altering rock wettability, by decreasing IFT between water and oil phase, and by affecting viscoelasticity of the flooding fluid. Moreover, they can create synergistic effects with polymers and surfactants, preventing their adsorptions onto rock formation, therefore they are often used in combination, in NP-polymer, NP-surfactant, and NPpolymer-surfactant flooding. Also low-salinity waterflooding has been combined with NP flooding with excellent results in carbonate reservoirs. [71] Some NP employed in cEOR, such as silica, are cheap, easy to prepare and relatively environmentally safe. Some new development in NP flooding involves nanohybrids, such as polymers grafted on inorganic NP, [70,75] and allpolymer nanoparticles made by self-assembly of amphiphilic block polyelectrolytes. [64] Preformed gel particles (PPG) These are cross-linked polymer particles, which received great attention recently, especially for hybrid EOR methods (see later). [76][77][78] These are used especially in mature oilfields, where high permeability zones are present due to fracturing, causing channeling, poor sweep efficiency, and high water cut. Unlike solid particles, gel particles are deformable, therefore easier to transport through pores. [78] Sometimes, small amounts of nanoclays are added to improve mechanical strength of the gels. PPG are often made of the acrylamide as the main monomer, therefore they suffer the same issues of HPAM in harsh reservoir conditions, such as degradation and shrinking due to high salinity. As anticipated, PPG are often used in combination with other EOR methods, namely low-salinity water, [79] surfactants, [80] and polymers. [81] Supercritical CO 2 Carbon dioxide, the main component of greenhouse gases, can be employed as flooding and/or foaming agent for cEOR (in surfactants-alternated-gas, or SAG injections. [82] ). This method has had a resurgence in recent times, even though it is known and used since the beginning of EOR, because it "kills two birds with one stone", having the possibility to combine EOR and CCSU. [16] CO 2 -EOR can be categorized as a gas method, or a miscible one (when used in supercritical state), but it can be included in chemical methods, CO 2 being not an inert gas, but a weak acid, able to react with bases present in the reservoir or in the oil and water phases. The main difference between previous and modern ideas about CO 2 -EOR, is in the simultaneous optimization of oil recovered and CO 2 stored. As estimated in 2019, 5% of EOR projects utilizes CO 2 , but it has been estimated that potentially it could be employed in 90% of world reservoir, and it can potentially capture up to 140 Gtons of CO 2, if optimized for capture rather than for oil production. [16] However, it has also been estimated that the process may not be overall energetically favorable. [83] As anticipated, CO 2 is also used in combination with surfactants, for foam EOR. [82,84,85]

Hybrid methods
The necessity to improve oil recovery from difficult reservoirs, where often a single cEOR method show limitations, has increased the research in hybrid methods, where synergies between different recovery mechanisms are exploited. A clear definition of hybrid EOR method does not exists, but it seems to include several combinations of chemical and non-chemical methods.
Typical hybrid methods that have gained recent interests mostly uses combination of low-salinity water with other chemical and non-chemical EOR methods, such as smart waterassisted foam flooding (SWAF), [86] low-salinity polymer flooding, [56] low-salinity ASP, [38] and low-salinity PPG. [79] Lowsalinity water is beneficial for the efficiency of all these cEOR techniques, simply because salts are known to be detrimental for the properties of the chemical agents, as discussed before for the single categories. NP stabilized foams and emulsions are also among the promising new hybrid EOR methods, [87] due to the stabilizing effect of NP on multiphase systems. Surfactant/ PPG is also considered a hybrid method. A comprehensive list of all the hybrid methods proposed for EOR would be too long to compile, but virtually any combination of known EOR methods can be included.

New tools for cEOR research
Next to the rise of new chemical agents, or the improvement of existing ones, several modern tools have disclosed the possibility to make the study of phenomena related to oil recovery more systematic and in depth, giving us a better understanding of the mechanisms. Historically, chemical flooding processes have been investigated experimentally, by characterizing and testing different chemical formulations in core flood, sand packs, or related setups, to try to predict how they would perform in a real EOR process. However, data are mostly empirical, and general observations are difficult to make, as the results obtained in a single study are very dependent on the particular settings used, and on many interdependent variables. Additionally, relevant properties such as viscoelasticity, IFT, wettability, mobility ratio, capillary number, that can be measured or estimated ex situ, can be very difficult if not impossible to estimate in the porous system. Finally, the large heterogeneity of reservoirs makes practically impossible to find a chemical formulations or conditions guaranteed to work out of the laboratory scale.
Major contributions to a deeper understanding of cEOR can come from tools such as microfluidics (reservoir-on-a-chip), mathematical modeling and simulations, and machine learning methods. These give the possibility to give a deeper understanding of the phenomena involved in cEOR, in a reproducible way, and can make results more standardized across research groups all over the world involved in EOR research, allowing to speed up advancements in the field.
Microfluidic devices are designed to study flow of liquids in micro-channels, therefore their direct application in studying cEOR should come quite naturally, although it did not receive much attention until recent times. In these devices, capillary force, surface tension (or better IFT), and viscoelasticity govern the flow mechanisms, as in real EOR processes, and they can be easily monitored, much better that in real rock formations. Indeed, microfluidic devices allow to directly observe and monitor over time what happens during flow at the pore scale, which is not possible with core flood, where only postmortem analysis is possible, to some extent. These characteristics made microfluidics very popular in the last decade for the investigation of EOR (reservoir-on-a-chip concept). [88,89] The geometry of the device can be controlled to a high degree, even to replicate real rock formations, thanks to the mature 3D printing technologies available. The wettability of the channels can also be tuned, by selecting different materials (which can be glasses or plastics), but also by chemically and physically treating them. It is my opinion that microfluidics will be a leading technique to study EOR in the future research.
Another approach involves mathematical modeling of cEOR processes. Models have been developed and implemented to simulate flooding processes in reservoir, some dating back to the early 1970s. [90] Of course, the massive improvement of computational power made these models more and more refined and accurate with time. [30,91,92] The general approach is based on solving numerically equations related to flow of viscoelastic fluids in porous media, such as Navier-Stokes equations or Darcy's laws, with appropriate boundary conditions. The difficulties come from trying to take into account all the phenomena happening during flow, such as chemicals degradation, hydrolysis and adsorption. Another crucial aspect to consider, is an appropriate simulation of reservoir conditions, which need to be included in the model, to improve the reliability of these methods. [91] Basically any cEOR method has been studied via mathematical simulations. As anticipated, the growing computational power of modern processors suggests that mathematical modeling of EOR processes will keep being relevant in future research, to give a better understanding of the processes and possibly helping predict their efficiency and the impact of single variables on the overall recovery. A very recent and still young development in cEOR research involving informatics tools, is given by the use of machine learning, [93,94] also for EOR coupled with CCS. [19,95] This approach also will most likely become more and more relevant in the future, as machine learning is becoming prominent in any research field.
What is the future of cEOR research? Some conclusions and perspective In this perspective article, I tried first to convey the message that, although efforts from the scientific community must be focused on realizing a complete energy transition, we cannot completely neglect research and advancements related to fossil fuels. Our goal should be trying to integrate as much as possible the various possible sources of energy, chemicals, and materials, in order to keep meeting the growing global demand, while keeping at bay CO 2 emissions. All this must be done with sustainability principles in mind, which can be applied also in the oil industry. It is unlikely that by 2050 we will be able to sustain ourselves solely on renewables. In this respect, chemical EOR methods will still be playing a relevant role for many years, therefore it is important that they become more efficient, so that we can obtain more energy and useful products, with lower emissions. Obviously, CCSU strategies are necessary to reduce carbon footprint of fossil fuels. For this reason, CO 2 -EOR coupled with CCSU is certainly an approach that needs attention, and indeed it is foreseeable that this will become a prominent EOR method. CO 2 flooding can be also implemented in foam EOR, when coupled with surfactant (for example in SAG injection). Besides experimental work, LCA and even machine learning methods can represent fundamental tools for the assessment and possible advancement of this particular techniques. Other cEOR methods, especially those involving polymers and surfactants, have reached maturity, but the current challenge is to obtain efficient extraction in difficult reservoirs, such as low permeability carbonate ones, tar sands, or shale oils. Recent research in surfactants and polymer flooding, has shown that the efficiency in difficult reservoirs can still be improved, by designing "smart" systems. The introduction of nanoparticles, or preformed particle gels, opened up several new promising possibilities to overcome the known problems of stability and efficiency of ASP formulations. Particularly promising in recent years seems to be the use of low salinity or "smart" brine as a tertiary recovery, alone or in combination with other techniques. The possibility to improve oil recovery by simply altering the salinity of water is very appealing from an environmental point of view, and it can be cost-effective. Speaking of environmentally friendly methods, probably the use of biopolymers and biosurfactants as alternatives to fully synthetic chemicals should be investigated and exploited more. The progress here is probably hindered by the fact that these molecules can be easily degraded by microorganisms during flooding, which requires the use of biocides in the formulation, making bio-derived products less attractive.
Additionally, the structure of fully synthetic chemicals can be more easily tuned to show the desired properties.
Other aspects related to the impact of cEOR techniques on the environment that should be further investigated, are related to the separation of produced oil and water, and possible recycling or purification of the water phase. These have not been treated here, but it is important to keep them in mind in future research.
Finally, it can be expected that significant advancements in cEOR research can be made thanks to the support of newly developed or improved tools. It is foreseen that microfluidics can provide a unique tool for the study of cEOR processes more systematically, and with better visualization of the phenomena. Also, IT tools such as mathematical simulations and machine learning can become more and more relevant as computational power grows, and more data are gathered by experimentalists in the laboratories or directly from the oilfields.
In conclusion, the challenges posed by global climate change cannot be solved by a single approach to the production of chemicals and energy. New and old sustainable technologies-including the ones related to EOR methods-needs to be further developed, exploited and integrated in a global combined effort.

Conflict of interest
The author declares no conflict of interests.