Guoping Hu,Kathryn H.Smith,Yue Wu,Kathryn A.Mumford,Sandra E.Kentish,Geoffrey W.Stevens

Peter Cook Centre for Carbon Capture and Storage Research(PCC),Particulate Fluids Processing Centre(PFPC),Department of Chemical Engineering,The University of Melbourne,Parkville,Victoria 3010,Australia

Keywords:Amino acids Carbon capture Absorption Solvents

A B S T R A C T The emission of large amounts of carbon dioxide is of major concern with regard to increasing the risk of climate change.Carbon capture,utilisation and storage(CCUS)has been proposed as an important pathway for slowing the rate of these emissions.Solvent absorption of CO2 using amino acid solvents has drawn much attention over the last few years due to advantages including their ionic nature,low evaporation rate,low toxicity,high absorption rate and high biodegradation potential,compared to traditional amine solvents.In this review,recent progress on the absorption kinetics of amino acids is summarised,and the engineering potential of using amino acids as carbon capture solvents is discussed.The reaction orders between amino acids and carbon dioxide are typically between 1 and 2.Glycine exhibits a reaction order of 1,whilst,by comparison,lysine,proline and sarcosine have the largest reaction constants with carbon dioxide which is much larger than that of the benchmark solvent monoethanolamine(MEA).Ionic strength,pH and cations such as sodium and potassium have been shown to be important factors influencing the reactivity of amino acids.Corrosivity and reactivity with impurities such as SOx and NOx are not considered to be significant problems for amino acids solvents.The precipitation of CO2 loaded amino acid salts is thought to be a good pathway for increasing CO2 loading capacity and cutting desorption energy costs if well-controlled.It is recommended that more detailed research on amino acid degradation and overall process energy costs is conducted.Overall,amino acid solvents are recognised as promising potential solvents for carbon dioxide capture.

1.Introduction

The increasing concentration of carbon dioxide(CO2)in the atmosphere is a significant driver of climate change[1].Therefore,reducing the emissions of carbon dioxide has focused attention.In the last decades,many methods[1–4]have been proposed to reduce CO2emissions,which include improving energy utilisation efficiency,developing renewable energy technologies,reforestation and carbon capture,storage and utilisation(CCUS).Of these,CCUS is an efficient way to reduce the emissions of carbon dioxide from energy generation plants and industrial sources.However,there are still barriers hindering CCUS deployment to large scale,such as the high costs of carbon capture processes[5].

A range of technologies including solvent absorption,solid adsorption[6],membrane separation[7],cryogenics[8]and others[9]have been investigated to separate carbon dioxide from gas streams.Monoethanolamine (MEA) solvent is widely considered as the benchmark solvent for efficient separation of CO2.However,the degradation and evaporation of the solvent,and high energy requirement for desorption are still significant challenges to its further deployment[10].In addition,the related emission of degradation products may lead to further pollution problems even though the emission control meets the domestic or international exhaust standard.The development of novel solvents is an important driver to make CCUS more viable.Recently,Bernhardsen and Knuutila[11]and Liang et al.[12]summarised the performance of different amine solvents from the perspective of absorption and cyclic capacity.Hu et al.[13]and Borhani et al.[14]also summarised recent work on potassium carbonate solvents from the perspective of reaction kinetics and further barriers to commercialisation.These works concluded that the solvents developed to date have many drawbacks,and more work is needed to develop better solvent technologies.

Amino acid solvents have been used for removing acid gases from gas streams industrially in a process known as the Alkazid Process marketed by BASF[15].Later,the Siemens PostCap?process was developed and operated in a pilot plant in Staudinger,Germany from 2009.This process and amino acids more generally have attracted much attention because of their ionic nature,low vapour pressure,low toxicity and high biodegradation potential(environmentally friendly)[16].They have similar amine groups to MEA,indicating they may have fast reaction kinetics with carbon dioxide,large carbon dioxide absorption capacities and cyclic capacities.Indeed,some amino acids(lysine[17,18],alanine[19])are reported to have higher CO2loading capacities than MEA due to a second functional amine group and the presence of excessive hydroxide ion binding sites.Wang et al.[20]reported a cyclic capacity of 0.62(molar CO2/molar alanine)in a precipitation process,which is much higher than that of MEA solvents(0.20–0.25).The advantages of fast kinetics and high capacity may lead to its wide implementation across fields including separation of CO2from bio-gas,syngas or flue gas(post-combustion conditions).Lower absorption heat is also an advantage of some amino acids such as alanine and proline[19,21].However,because of this,areas of improvement exist,including the requirement of a high regeneration temperature and the occurrence of an unpleasant smell due to amino acid degradation during solvent regeneration[22].Therefore,a review on recent progress using amino acids solvents for CO2absorption is warranted in order to inspire further studies and innovative solutions for capturing CO2.

A few publications[23–25]report the reaction kinetics of different amino acids as solvents or additives for carbon dioxide capture.These amino acids include taurine,methionine,glutamic acid,sarcosine,alanine,proline,6-aminohexanoic acid,arginine,glycine,histidine,1-aminocyclohexane carboxylic acid,pipecolic acid,2-aminoisobutyric acid,2-piperazinecarboxylic acid,asparagine,aspartic acid,leucine,lysine,serine and valine.Many researchers[26–28]have also investigated the absorption performance of different amino acids.The vapour liquid equilibrium(VLE)data[21,29–34]and physical properties[35–43]including density,viscosity,surface tension,dissociation constants,carbon dioxide solubility and diffusivity are also important parameters which have been widely studied.The aim of this review is to focus on the kinetics and industrial barriers for implementing amino acid solvent processes.The physical properties and VLE studies have not been covered in this review.

2.Screening of Amino Acids

Many studies have been published which screens a range of amino acids to understand their absorption rates with different bases for carbon dioxide capture.Park et al.[24]investigated twelve different amines and amino acids as promoters in potassium carbonate solvents using a bubble reactor.Pipecolic acid and sarcosine were found to be the most effective rate promoters(with 3 wt%amino acid additions in 30 wt%potassium carbonate solvent at 70°C).However,the solubility of pipecolic acid (～50 mg·ml-1) is much lower than that of sarcosine(～1500 mg·ml-1),making sarcosine a better choice.Holst et al.[23,52]studied nine different amino acids based on their potassium salts(0.5 mol·L-1)at 25°C using a stirred cell,among which sarcosine showed the fastest absorption rates.Recently,Hu et al.[25]screened ten amino acids as reactants with carbon dioxide using a stopped flow technique.This was followed by testing five of the amino acids as promoters in potassium carbonate solvents at various concentrations at 25°C and 50°C using a wetted wall column(WWC)apparatus.The results from this study also indicated that proline and sarcosine had the fastest absorption rates and additionally,that both pH and ionic strength are important factors impacting these absorption rates.

3.Chemical Reactions

The chemical structures and the acid dissociation constants(pKa)of the amino acids discussed in this review are shown in Table S1.All the amino acids studied for carbon dioxide separation are primary or secondary amino acids.The chemical structure of a typical amino acid can thus be written as R1NHCHR2COOH.The main reactions with CO2may be written as Reactions 1–3.These reactions have been proven using Raman spectroscopy of reacting products between CO2and amino acid solutions,in which the spectral features of carbamate,carbonate,and bicarbonate were identified[17].The reaction scheme also indicates that the maximum carbon dioxide loading is 0.5(or slightly higher due to the direct reaction of carbon dioxide with water and hydroxide ions)in amino acid solvents(AAS),assuming the amino group is the only reactive group.

The overall reaction rate may be written as Eq.(4).

The reaction between water and carbon dioxide has been widely studied[44,45],and is generally treated as negligible as it is much slower than that of carbon dioxide and amino acids or hydroxide ions at basic conditions.

The reaction rate between hydroxide ions and carbon dioxide can be written as Eq.(5).The concentration of hydroxide is typically low(≤0.01 M)because of the limitations of operational pH.

The reaction between AAS and carbon dioxide can be written as Eq.(6),in which n is the reaction order with respect to AAS.

Different reaction mechanisms[46]have been proposed to explain the reaction between amino acids and carbon dioxide.The most widely used mechanisms are the zwitterion mechanism and the termolecular mechanism.

In the zwitterion mechanism(Fig.1)[26,47,48],amino acids react with carbon dioxide to form a zwitterion(Reaction 7),followed by the deprotonation step in which the zwitterion reacts with a base B)(amino acids,water or hydroxide ions).Reaction 8 gives an example of an amino acid behaving as a base.As discussed previously,the maximum CO2loading for amino acids solvents is approximately 0.5.However,amino acids,water and hydroxide ions can all contribute as a base,so that the maximum loading of amino acid solvents may be higher than 0.5.

Fig.1.A reaction scheme between amino acids and carbon dioxide

The reaction of CO2can be written as Eq.(9)[49].

In the case where there are multiple species that can act as a B,the change in-OOCN+HR1CHR2COO-can be described as Eq.(10).

Assuming quasi pseudo state conditions for the concentration of-OOCN+HR1CHR2COO-,Eq.(9)can be converted to Eq.(11)[48].

Manipulation of Eq.(10)becomes Eq.(12).

Substituting Eq.(12)into Eq.(11),the corrected reaction constant of CO2can be written as Eq.(13).

In the termolecular reaction mechanism[50,51],the overall reaction is described as a single step termolecular reaction(Reaction 14,taking OH-as the base species).

In this case,the corrected reaction constant for amino acids can be written as Eq.(15).

4.Reaction Kinetics between Amino Acids and Carbon Dioxide

In the last few years,many researchers have investigated the reaction kinetics between amino acids and CO2systematically using a range of techniques including the stirred cell[23,51,54–59],stopped flow[28,60–65],wetted wall column(WWC)[26,62,66–71],NMR[61]and string-of-disks contactor[27].Vaidya and Kenig[53]provided an excellent review of the various kinetic measurement techniques available.

A brief summary of the results of these studies are provided in Table 1.Glycine,sarcosine and proline are the most frequently reported solvents due to their high absorption kinetics,solubility and availability.There is no discrepancy in the reaction order for glycine solvents over the temperature range and concentrations investigated,with an overall second order reaction between carbon dioxide and glycine reported.Although the reaction order between CO2and other amino acids is found to vary with experimental conditions,most reported values being between 1 and 2 with respect to the amino acid,indicating k-1＞＞0 in Eq.(13)(assuming the zwitterion mechanism is appropriate to explain the results).Furthermore,as shown in Fig.2,the reaction orders of sarcosine,proline and arginine with CO2have an overall positive correlation with the concentration of amino acid.This may be due to ionic strength effects which are further discussed in Section 5.5.

For comparison,the corrected reaction constants(kobs′)of eight amino acids(see Eqs.(13)and(15))are shown in Fig.3 with an amino acid concentration range from 0.1 to 4.0 mol·L-1at 298 K.It can be seen that these reaction constants all increase with increasing amino acid concentration.At constant temperatures and the differencein kinetics between the amino acids is significant(ranging from 2500 to 100000 s-1at 1 mol·L-1).Sarcosine and lysine show the fastest reaction kinetics while taurine and histidine are comparatively slow.It should be noted that discrepancy between different research exists(Table 1,i.e.reaction order,reaction constants,activation energy)possibly due to different assumptions,methodologies and experimental conditions.Therefore,it is important to choose appropriate data for further calculations or modelling work.

Table 1 A summary of reported kinetics research between CO2 and different amino acids

Fig.2.A correlation between reaction orders and the concentration of amino acids(sarcosine,proline and arginine).

Fig.3.Corrected reaction constants of different amino acids at 298 K.

Fig.4.Arrhenius Equation fitting of the reaction kinetics results between glycine and CO2.

Fig.5.The Arrhenius Equation fitting of reaction kinetics between sarcosine and CO2.

In order to better understand the results in the published literature,the rate constants determined at different temperatures were fitted to the Arrhenius Equation and are presented in Figs.4 to 7.The stopped flow usually indicates slower kinetics(Benamor et al.[64]for glycine,Mahmud et al.[60]for sarcosine and Sodiq et al.[28]for proline)as compared to the results obtained by a stirred cell or wetted wall column.This is due to a limitation of the stopped flow technique whereby dilute solutions need to be used to ensure fast mixing.As shown in Fig.3,the reaction rate constants are strongly dependent upon concentration.Ionic strength was also reported to have a significant impact on reaction kinetics[25,65].Therefore,although the stopped flow technique is quite effective as a fast screening method for assessing the kinetics performance of different solvents,it cannot provide data at concentrations representative of industrial conditions.

Fig.6.The Arrhenius Equation fitting of reaction kinetics between proline and CO2.

Fig.7.The reaction kinetics of MEA and other amino acids with CO2.

As shown in Fig.4,the results from Lee et al.[67]and Park et al.[56]gave much slower reaction kinetics results than others even when both were conducted using a stirred cell apparatus(Table 1)over a wide concentration range(1–3.5 mol·L-1).This discrepancy can be attributed to the fact that the studies were conducted using sodium glycinate rather than the potassium salt.Sodium salts have been demonstrated as less reactive solvents for carbon dioxide absorption[58].The results shown by Mahmud et al.[60]present much lower reaction kinetics compared to other investigators.This may be due to the low operational concentration of amino acids of these tests(tested with stopped flow)which has a low ionic strength.The effects of ionic strength will be introduced in Section 5.5.The rest of the kinetic data agrees well with each other.The Arrhenius Equation for glycine reacting with carbon dioxide can be represented over the temperature range of 275 K to 355 K with Eq.(16).

However,as shown in Table 1,there are discrepancies on the reaction orders between proline and CO2and sarcosine and CO2.An Arrhenius Equation is not representative for all these results(Figs.4 and 5)which may be due to a change in reaction mechanism with concentration or temperature.Different kinetic equations may be needed under different operating conditions.Additional data is required for further understanding of this process over a full range of industrial operating conditions.

In 1996,Versteeg et al.[72]summarised a range of research results of the reaction between MEA and CO2and fitted an Arrhenius Equation with a reaction order of one with respect to MEA(Eq.(17)).

The Arrhenius curves of some other amino acids(histidine,taurine,lysine,threonine and arginine)and MEA are shown in Fig.7.It can be seen that the reaction constants between amino acids and carbon dioxide are comparable or even larger than that of MEA regardless of the reaction order difference.As is known,the reaction order between MEA and carbon dioxide is unity at MEA concentrations lower than 5 mol·L-1[13].In industrial pilot plants,the amine solutions(or amino acids solutions)used for carbon dioxide absorption are usually at high concentrations(＞1 mol·L-1),indicating that the reaction rates of amino acids and carbon dioxide are even higher than that of MEA due to the higher reaction order between amino acids and carbon dioxide(＞1).Therefore,proline,sarcosine and lysine are the most promising solvents for industrial application from the perspective of reaction kinetics.

Penny and Ritter[63]proposed a Br?nsted relationship between the reaction constants of amines with carbon dioxide and the amine dissociation constants at 293 K(Eq.(18)).The Br?nsted relationship was later used to interpret the reaction constant of amino acids by Holst et al.[23]and showed good agreement.In this review,the raw data of eight amino acids(1 mol·L-1)at 293 K from ten publications were extracted and fitted to the Br?nsted relationship as shown in Fig.8.The reaction constants of the amino acids showed overall agreement with the Br?nsted relationship and the reaction constants displayed an increasing trend with increased dissociation constants of amino acids.

Fig.8.The Br?nsted relationship between the reaction constants of amino acids with carbon dioxide and the dissociation constants of amino acids(1:Taurine[51],2:Threonine[59],3:Histidine[71],4,5:Glycine[51,55],6,7:Sarcosine[27,54],8:Proline[58],9:Lysine[70],10:Arginine[26]).

5.Discussion on Industrial Barriers

5.1.Reactivity of different amino acid salts

As shown in literature[58,73],research has suggested that amino acids with a potassium salt are more reactive than with sodium.Majchrowicz et al.showed the reactivity of the potassium salt of L-prolinate was 32%higher than that of its sodium salt when reacting with CO2[58].Aronu et al.[73]compared amine amino acids with potassium amino acids via a semi-quantitative study and concluded amine amino acid salts had better performance than that of potassium salts in terms of both kinetics and cyclic capacity,especially with excessive 3(methylamino)propylamine(MAPA)[74]as a promoter.However,as mentioned in the original article,these results were based on a semi-quantitative study because of the uncertainty of the bubble structure.Later,Aronu et al.[75]completed pilot plant tests using MAPA sarcosinate and results showed that this solvent had slower absorption,lower regeneration energy and lower corrosion than those of MEA solvents.However,the authors mixed equinormal amounts of sarcosine with MAPA in this study.It may be more meaningful to add a higher proportion of MAPA for achieving faster absorption.Therefore,it is recommended that more rigorous experiments be conducted to confirm the advantages of amine amino acid salts.

5.2.Corrosivity of amino acids solvents

Previous work[76,77]has shown that glycinate could increase the corrosion rate of solvents.However,both studies conducted to date have been performed using a potentiostat unit(PGSTAT302N,Metrohm,Germany).Carbon steel was used,and the pH conditions were not provided,which is a key parameter for industrial plant operation.Anh et al.[78]also investigated the corrosion rate of glycine and taurine on carbon steel.Results showed that increased amino acids concentration,operation temperature and CO2loading can increase the corrosion rate.Sodium metavanadate(NaVO3)was proposed as a good corrosion inhibitor.In a pilot plant test conducted by Siemens,a stainless steel bar(1.4571 steel)was used for corrosion tests for 1370 h[79].Results showed no local corrosion and negligible surface abrasion.As the material of a absorber and regenerator can be a significant contributor to the capital costs of a capture plant,more detailed tests are required under industrial conditions including pH,pressure,ionic strength and a wider temperature range(especially for regeneration).

5.3.Loss of amino acids

The degradation of amine solvents[80,81]such as MEA[82,83]and AMP[84]has been studied extensively.However,limited investigations have been conducted on amino acids solvents.Erga et al.[22]mentioned an unpleasant smell during degradation of absorbed glycine solvent,the cause of the smells was not quantitatively investigated.In the pilot plant validation of the Siemens capture process using amino acids(a secondary hindered amino acid neutralised with a strong metal hydroxide base),the degradation(thermal and oxidation)of the solvent was estimated to be below 1%over a year,and the amino acid content in the exhaust due to volatility and degradation was shown to be negligible[79].Because of the very limited number of published studies in this area,more systematic experiments on the degradation of different amino acids are recommended.

5.4.Acidity(pH)sensitivity

As shown in the literature,the pH of amino acid solutions increase with increased amino acid concentration[85]and decreased CO2loading[54].Simons[54]also noticed that as CO2loading increases,both pH and absorption rates decreased,while the mechanism behind this was not discussed in detail.In amine solvents,as the CO2loading increases,the absorption driving force decreases,reducing the overall absorption rate.However,in amino acid solvents,the formation of amino acids shifts with changing pH(Fig.9).It is also well-known that only anions react with carbon dioxide[25,65].Therefore,with pH decreasing,both the reactivity of amino acids solution and the driving force decrease,resulting in much lower absorption rates than that at high pH conditions.In conclusion,it is crucial to control the solvent pH to be high enough to ensure the amino acid is in its reactive form.However,the requirement of a higher pH condition for absorption indicates higher energy consumption because of the difficulty of solvent regeneration.Searching for amino acids with low pKavalues may be an interesting direction of research to enable operation with solvents at lower pH conditions.Adding additional promoters such as piperazine may also be a good choice especially for promoting the absorption rate under high loading conditions because of their low pKaand high reactivity[86].

Fig.9.Transformation of amino acid formations with the change of pH.

5.5.The effect of ionic strength

Many researchers have reported the influence of ionic strength on the reaction between amino acids and CO2,affecting both reaction kinetics and reaction equilibria[87].An exponential relationship(Eq.(19))has been used extensively in the literature to represent the influence of ionic strength[88–90],where the value of b is a function of the solution properties and temperature.The value of b was determined to be 0.1 at 283 K and 0.2 at 303 K for CO2in 0–3 mol·L-1of sodium hydroxide solutions[91],while b values can also be negative for some reactions[92].For amino acids,a few b values have been reported such as 0.38 for sarcosine[27],0.44 for glycine[55],0.57 for alanine[93],0.46–0.67 for histidine[65]and 0.90 for threonine[59].It can be concluded that increasing ionic strength plays a positive role for the reaction between amino acids and CO2.Therefore,adding additional salts may be a method of enhancing the absorption kinetics for amino acids solvents.Most research regarding the reaction kinetics between amino acids and CO2gives reaction orders with no dependency on ionic strength;however,Fig.2 clearly demonstrates that higher concentration leads to higher reaction orders.

5.6.The effect of solubility on the operation of amino acid solvents

As discussed in Section 5.1,different salts can influence the reactivity of solvents for the same amino acid.It should be noted that different salts can also influence amino acid solubility.The salts of sodium,lithium and potassium of sarcosine and proline showed an increasing trend of solubility[94].Thus,different salts may be chosen for different purposes when operating an amino acid process.For example,potassium prolinate may be used to avoid precipitation in the absorber,while sodium prolinate may be used to take advantage of forming a third phase.The precipitation(Fig.10)of amino acids(products including carbamate and protonated amino acids)can potentially increase the carbon dioxide absorption kinetics at high loading(due to the increased driving force provided by precipitating carbamate from the aqueous solution)and the CO2capacity[21,95]while also potentially decreasing the regeneration energy.However,it also brings about the challenge of handling a slurry system.The precipitated solids can potentially decrease the overall mass transfer rates because of the decreased contact area due to the presence of particles[33].A spray column may be used instead of a packed bed column[95].Other problems include blockage of pipelines,heat transfer resistance and erosion due to friction.Therefore,the crystallisation kinetics and equilibria as a function of both temperature and CO2loading must be investigated thoroughly to ensure the operating conditions of the plant can be maintained[79].Detailed research is needed for different amino acids with different chemical additions.For desorption,the precipitated solvents must be heated up to re-dissolve the precipitates and regenerate carbon dioxide.This process can occur on the solid/liquid mixture,or the solids can be separated from the system and heated up separately.The second choice can potentially decrease the heat wasted in heating and evaporating water in the solvents,thus decrease the overall regeneration energy requirements and costs.

Fig.10.Chemical absorption reaction and phase change scheme[96].

5.7.The effect of NOx and SOx on amino acids solvents

Because of the presence of SOxand NOxin emissions such as from coal fired power station flue gases,sulphite,sulphate,nitrite and nitrate can be formed in the solvent.This,on one hand,can further decrease the emission of SOxand NOx,on the other hand can increase the potential of equipment corrosion,deactivate some of the solvent and enhance solvents degradation if these products cannot be easily removed.In the pilot plant tests run by Siemens[79],reclamation of the solvents was accomplished by a sulphur compound removal step and a degradation products removal step.The overall loss of solvents was proved to be negligible.However,an additional step to remove SOxand NOxupstream of absorption may be needed for high sulphur and nitrogen containing fossil fuels.

5.8.Outcomes from large scale amino acid solvent plants

Erga et al.[22]investigated 4 mol·L-1of glycine and sarcosine as solvents for carbon dioxide absorption in a pilot scale.The amino acid solvents required a higher regeneration temperature(140°C)than that of MEA solvent.Knuutila et al.[97]compared 3.5 mol·L-1of sarcosine with 30 wt%of MEA and concluded that sarcosine showed much higher absorption kinetics.However,the energy costs were higher than that of MEA,and the cyclic capacity of 3.5 mol·L-1of sarcosine was limited to 0.2 mol CO2per mole amine group.These results also agree with the modelling results by Majchrowicz and Brilman[98]using 4 mol·L-1of proline and Song et al.[99]using 30 wt%of glycine.Song et al.[99]analysed the VLE data of CO2in sodium glycinate and attribute the high energy consumption to the high CO2solubility at high temperatures(120°C)which results in a large reboiler duty.Conversely,in the pilot plant tests run by Siemens[79],a lower energy consumption than that for 30 wt%of MEA was indicated,but the solvents were not disclosed.The technology was designed to scale up in the FINNCAP-Meri Pori CCS Project[100],which was initially aiming at capturing 1.2 million tons of carbon dioxide a year[79].Unfortunately,the project was abandoned in 2010 because of company strategy and the outcome of various studies.Literature reports that the Siemens technology is still under development[101].However,recent information on this research is quite limited[102].

6.Conclusions and Recommendations

Recent progress on the absorption kinetics of amino acids has been summarised and the engineering potential of using amino acids as carbon capture solvents has been discussed.The reaction order between amino acids and carbon dioxide is usually between 1 and 2(1 for glycine).The second order reaction constants between glycine and carbon dioxide is summarised as kGLY(L·mol-1·s-1)=3.77×1013at 275–355 K.Lysine,proline and sarcosine are reported as having the largest reaction constants with carbon dioxide,with all much larger than that of monoethylamine(MEA).Ionic strength,pH and the cation used are shown to be important factors influencing the reactivity of amino acids.Corrosivity of amino acids solvents and deterioration in the presence of flue gas impurities including SOxand NOxare not considered as large concerns for amino acid solvents.Overall,amino acid solvents are recognised as having good potential to be used for carbon dioxide capture.

Recommendations can be classified into three levels.From a chemical perspective,amino acids with low pKavalues are favourable for both better kinetics and larger pH operational range.From a laboratory work perspective,more work is required to be conducted on fast amino acid solvents(lysine,proline and sarcosine).A systematic study is essential to investigate their corrosivity,degradation,regeneration energy costs,operational pH conditions and the influence of NOxand SOxon their corrosivity and regeneration.Adding extra salts may also be an option to further enhance absorption kinetics.In addition,the precipitation of the amino acids may be a good pathway to reduce desorption energy costs if well-controlled.Thus,the crystallisation of amino acids salts may be thoroughly studied to better understand their behaviour at high concentrations.From the level of industrialisation,pilot plant tests on the performance of lysine,proline and sarcosine is recommended to learn their in-situ long term performance such as their corrosivity,degradation,regeneration energy costs.

Acknowledgements

Infrastructure support from the Particulate Fluids Processing Centre(PFPC),a Special Research Centre of the Australian Research Council,the Peter Cook Centre(PCC)for Carbon Capture and Storage(CCS)Research and CO2CRC are gratefully acknowledged.G.Hu also would like to acknowledge Prof.Alan Hatton in Massachusetts Institute of Technology(MIT)for hosting him to finish the research.

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2018.08.003.