The interplay of porosity, wettability, and redox activity as determining factors for lithium-organic electrochemical energy storage using biomolecules.

Abstract Although several recent publications describe cathodes for electrochemical energy storage materials made from regrown biomass in aqueous electrolytes, their transfer to lithium–organic batteries is challenging. To gain a deeper understanding, we investigate the influences on charge storage in model systems based on biomass‐derived, redox‐active compounds and comparable structures. Hybrid materials from these model polymers and porous carbon are compared to determine precisely the causes of exceptional capacity in lithium–organic systems. Besides redox activity, particularly, wettability influences capacity of the composites greatly. Furthermore, in addition to biomass‐derived molecules with catechol functionalities, which are described commonly as redox‐active species in lithium–bio‐organic systems, we further describe guaiacol groups as a promising alternative for the first time and compare the performance of the respective compounds.


Introduction
Guaiacol groups are common in biomolecules, both in small molecules such as vanillin as well as in macromolecules such as lignin. [1] Although they are not redox-active per se, irreversible oxidation in water results in the formation of the o-quinone/o-hydroquinoner edox pair. [2] This reactionw as first applied prominentlyb yM ilczarek and Inganäsi nc athode materials from lignin, which has abundantg uaiacol units, for green electrochemical energy storaged evices. [3] Since then, many endeavorst oi mprove lignin-based cathodes have been undertaken. [4][5][6][7][8][9][10][11][12] Additionally,c athodes based on polymers from biorefined monomers such as guaiacol and syringol, [13] phenolic acids, [14] and vanillin [15] have been investigated in acidic aqueous electrolytes. Aqueouse lectrolytes are importantf or energy storagei nt hese materials as the established process for the formation of quinones requires the presence of water. [2] Recently,alignin-PEDOT[ PEDOT = poly(3,4-ethylenedioxythiophene)] hybrid material wasi nvestigated as an electrode material in al ithium-organic system.H owever,l ignin mainly serves to dope PEDOT in this system. PEDOT constitutes the majority of the electrode, which limits the sustainability of such am aterial. [16] In polymer-based electrodes, redox-active polymers are always mixed with conductive additives such as carbon nano-tubes, [17] porous carbons, [9,10,15,18] and conductive polymers, [3,19] all of which contributes ignificant capacitive energy storage. Both cyclic voltammograms and galvanostaticm easurements of the mixed cathode materials are used to show influences of both distinct redox-activeg roups (high current only at distinct voltages in cyclic voltammetry and plateau-like galvanostatic behavior) and capacitive behavior (rectangular cyclic voltammogram and triangularg alvanostatic curve). Consequently,i ti s debatable whether the devices should be denoteda ss upercapacitors or batteries. [20] In studies on organic electrode materials, conductive additives such as carbon materials are often consideredp assive parts of the electrode composition. [21,22] To describe charge storagei no rganic materials more thoroughly,s ometimes electrodes without organic active materials are designed, and the resultingc apacity is subsequently subtracted from the total capacity.H owever,s uch investigations suffer from the possible systematic error when neglecting the modificationo fc harge storagep roperties of carbona dditives by organic active materials.
The rapidly increasing number of publications on bio-derived active electrode materials requires the discussion of such possible influencesi nm ore detail. Consequently,w ei nvestigate influences of different components in common biopolymer-based materials for electrochemical energy storage. As model systems, we describe hybrid energy storagem aterials made from model biopolymers and carbon additives. In the former,s everall ow-molecular-weightc ompounds, which have active functionalities common to biomolecules, are immobilized on ap olymer backbone. For immobilization, we make use of the formation of Schiff bases, which can be used to graft aldehydes onto amine-containing polymers, most notably,c hitosan [23] andp olyallyamine. [24][25][26][27] Vanillin is investigated because it contains guaiacol groups. Protocatechuic aldehyde Although severalr ecent publications describe cathodesf or electrochemical energy storagem aterials made from regrown biomass in aqueouse lectrolytes, their transfer to lithium-organic batteriesi sc hallenging. To gain ad eeper understanding, we investigate the influences on charges torage in model systems based on biomass-derived, redox-activec ompounds and comparable structures. Hybrid materials from these model polymers and porous carbon are compared to determine pre-cisely the causes of exceptional capacity in lithium-organic systems. Besides redox activity,p articularly,w ettability influences capacity of the composites greatly.F urthermore, in addition to biomass-derivedm olecules with catechol functionalities, which are described commonly as redox-active speciesi nl ithiumbio-organic systems, we further describe guaiacol groups as a promising alternative for the first time andc ompare the performance of the respective compounds. containsacatechol functionality,t hat is, it may be formed upon the demethylation of the guaiacol unit in vanillin. Finally, 3,5-dihydroxzbenzaldehyde is investigated because it is an isomer of protocatechuic aldehyde. However,a st he hydroxyl groups are found in the meta insteado ft he ortho position, is not redox-active.
The processing of active polymericm aterials with conductive additives, for example, by the preparation of as lurry that containsb oth components, spreading on ac urrent collector, and drying,i saconventionalw ay to create organic batteries. For the formation of electrodes, we chose microporous carbons as conductive additives because of the possibility to synthesize them from biowaste, [28] whichmakes them an attractive sustainable conductive additive, and because of their intrinsically high capacitance that allows the formation of hybride lectrochemical energy storaged evices. [29] The contributions of carbon,t he wettability of the electrolyte, and different functional units to the resultingl ithium-ion-based energy storage devices are discussed. Firstly,w ec haracterize the polymers under investigation and then their composites with microporous carbon to assess the influence of different physicalp roperties and functionalities on charge storage. Finally,the electrochemicalp erformance of these composites are investigated and compared.

Synthesis of polymers and their hybrid materials
Schiff bases of polyallylamine (PAAm) and different aldehydes, namely,v anillin (van), protocatechuic aldehyde (which features two phenolic groups in the ortho position;A -o), and 3,5-diyhdroxybenzaldehyde (which features two phenolic groups in the meta position;A -m), were synthesized in ethanolics olutions as described in the Experimental Section and Supporting Information (Scheme1). The addition of aldehydes to PAAm results in an immediate color change as expected for the formation of Schiff bases, followed by precipitation ( Figure S1) caused by the interaction of leftover amine groupsw ith acidic phenolic groups. [24] During subsequent drying, the formation of Schiff bases proceeds, [30] and washing with anhydrous ethanol yields the pure modified polymers (denoted P-van, P-o, and P-m, respectively).
The successful formationo fS chiff bases is confirmed by using IR spectroscopy (Figure 1), as the resulting polymers contain the functional groups from the respective phenyl ring. Broad bands above ñ = 3000 cm À1 ,f rom hydroxyl groups,a re retained both in the spectra of the polymers and aldehydes, which confirms the retention of the hydroxyl groups during synthesis. Twon ew bands appear in the spectra of polymers at approximately ñ = 2840 and 2920 cm À1 ,w hich indicates CÀH stretching vibrations from the polyallylamine backbone. [31] No aldehyde groups are detected in the spectra of the polymers, but new bands at approximately ñ = 1680 cm À1 represent imine functionalitiesi na ll three samples, which confirms both the successful reaction and the successful removal of any unreacteda ldehydes. [32] Intriguingly,t he shift between the imine and aldehyde peak is very pronounced in the case of van/Pvan andA -m/P-mb ut is rather subtle in the case of A-o/P-o, only by af ew cm À1 ,w hicho riginates from the different wavenumber of the aldehyde carbonyl vibrations in A-o, van, and A-m. The substitution rates in the samples (details are explained in the Supporting Information) calculated using the amount of Nd etermined by using elemental analysis is sum-Scheme1.Reactionofp olyallylaminew ith different aldehydes. www.chemsuschem.org 2020 The Authors. Publishedb yWiley-VCH Verlag GmbH &Co. KGaA, Weinheim marized in Ta ble 1. Substitution rates above 80 %c onfirmt he successful graftingo f aldehydes onto PAAm to yield polymers with ah ighd ensity of redox-activeo rc omparable functional groups.N otably,t he apparent substitution rate of P-m, which exceeds 100 %, is ar esult of the limitations of this methodo f analysis, but other methodsu nfortunately also overestimated the substitution rates of Schiff bases on PAAm in the past. [24,26] This overestimation could be ascribed to as mall amount of water presenti nt he samples, which is visible from the results obtainedf rom the polymers by using thermogravimetric analysis (TGA; Figure S2). Notably,severalindications such as the deviation between multiple measurements and the occurrenceo f sulfur in the samples show that theser esults are not precise. The use of thesem ethods is to ensure comparability between the samples, which they do. Similars ubstitution rates allow us to compare differences causedb yt he redox-activea nd -inactive functional groups.
We formed hybrid cathode materials by combining the polymers and carbon as explained in the Supporting Information and label them C/x in which xi st he corresponding polymer. SEM was used to show the morphology of the obtained materials before [C(pristine)]a nd after ball-milling (C and C/P-o;F igure S3). Upon prolonged milling, the big particles of conductive carbon become smaller.T he morphologies of both C/P-o and the neat carbon after ball-milling are similarb ecause of the tight composite between the polymer and the conductive carbon.W eu sed energy-dispersive X-ray mapping to show a homogeneous dispersion of C, O, and N( FigureS4).
Comparison of the electrochemical performanceofr edoxactive and -inactive polymers C/P-o, C/P-m,a nd microporous carbon are compared in Figure 2i nt erms of their electrochemical performance during cyclic voltammetry (CV) and charge-discharge tests. The redox activity of C/P-o (Scheme S2) is clearly detectablei nt he CV from peaks at approximately 3.3 Vv ersus Li + + /Li (Figure 2a). In the low-voltage region, the voltammograms of C/P-o and C/Pmm atch fairly well, but there are no redoxp eaks at approximately 3.3 Vv ersus Li + + /Li observable in the trace of C/P-m,a s expected for redox-inactive species. Similarly,t he absence of redox peaks is characteristic of purely microporous carbon.
The capacity of microporous carbon in galvanostatic charge-discharge experiments is very low.U pon the addition of P-m and even more so upon the addition of P-o (C/P-m and C/P-o,r espectively), the capacity increases significantly.U pon prolonged cycling, Ca nd C/P-m retain their capacity,w hereas C/P-o shows subsequent decay (Figure 2c), which will be discussedl ater.T he higher capacity of C/P-o than C/P-m can be explained easily by the redoxa ctivity of P-o that results in a slightly bell-shaped galvanostatic discharge behavior of C/P-o ( Figure 2b)a nd as ignificantly higherc apacity.T he range of the increased discharge capacity in C/P-o compared to that of C/P-m (nonparallel discharge curvesshown in Figure 2b However,t he significantly higher capacity observed if we compareC /P-m and microporous carbonc annotb ee xplained by redox reactions as P-m is not redox-active. Ap ossible explanation for this may be the increased wettability of the  carbon surfaces by the electrolyte after functionalization with the heteroatom-rich polymer.C onsequently,n ext we discuss the N 2 and water vapor physisorption behavior to assess the hydrophilicity of the prepared hybrid materials and, therefore, their wettability with solvents of high polarity,s uch as electrolytes ( Figure 3).
During ball-milling, Cl oses most of its internal surfacea rea, probablyc aused by ac ombination of high-energyb all-milling and the blocking of the pores by the binder (cf.,p hysisorption behavior of the pristine microporous carbon in Figure S7). The Brunauer-Emmett-Teller (BET) surfacea rea of the hybridm aterials with the model polymers (C/P-o and C/P-m)a re even lower than that of C, possibly because of the relatively low amount of microporous carbon in the hybrid material and because of the additional blocking of the pores by the active polymers. All samples exhibit high external surfacea reas because of the loose packing of particles.
Water vapor physisorption does not follow the same trend as N 2 physisorption. The water vapor uptake in C/P-o and C/Pmiss imilarly high, and the carbon materialadsorbs significantly lessw ater vapor than C/P-o and C/P-m over the entire range of relative pressure. As the elemental composition of P-o and P-m is the same, this behavior mayb ea scribed to abundant hydroxyl groupsi nt he polymers, which results in mixed carbon-polymer materials with as imilar hydrophilicity.H ydrophilicity is significantly higher in the hybrid materials than in the carbon material, and subsequently,a saresult of the high polarityo fw ater and electrolytes, so is the wettability by polar electrolytes. As the electrolyte reaches al arger part of the internal surfaceo ft he microporous carbon, the contribution to energy storagebythe formation of the electric double layer increases. More chargec an be stored on the surface for both C/ P-o and C/P-m compared to carbon, which leads to ah igher capacity. Similar behavior was described previously in aqueous solutions of polypyrrole with and without chitosan, in which the redox-inactive chitosan increased the capacity of polypyrrole significantly. [33] Notably,b all-milling tends to increase the hydrophilicity of carbon materials, although to al ow extent. [34] Comparison of the electrochemical performanceofc atechol and guaiacol groups Catecholg roups in hybrid materials with microporous carbons contributet ot he overall capacity both through the increased wettability of carbon surfacesb yt he electrolytes and redox activity.S till, catechol groups are rather rarei nn ature, notable exceptionsa re dopamine and tannic acid, which have been described previously for electrochemical energy storage. [35][36][37][38][39][40] In contrast, guaiacol groups (which are chemically similar;c f. Scheme1)a re abundanti nn ature,f or example, in low-value biogenic materials such as lignin. To our knowledge,s uch materials have never been used in secondaryl ithium ion batteries without additional redox active polymers to date. Thus, we will focus on such materials by comparing the electrochemical performance of C/P-o and C/P-van ( Figure 4). Initially,t he dischargec urvesd iffer greatly,b oth in terms of overall capacity as well as shape, but become more similarw ith the increase of the number of cycles. Unlike that of C/P-o, the discharge curve of C/P-van shows no belly shape at the beginning. With the increase of the number of cycles,t he shape changes,a nd ultimately, almostm atches that of the C/P-o discharge curve after approximately 50 charge-discharge cycles. The same behavior is apparent in the CV curves( Figure S8). Although no clear redox peaks are observed in the earlyc ycles of C/P-van,s uch peaks at similar potential as seen in the CV of C/P-o appear with continuing cycling, and the CVs of C/P-van and C/P-o start to resemble each other.W ep erformed EIS for both samples ( Figure S6), which shows the good conductivity of the materials,and C/P-vanexhibits slightly slowerkinetics possibly because of the presenceo fm ethoxy groups.
In aqueous systems, [2] demethylation occurs in the first oxidation step, expressed by the high current in the CV curvesa t voltages approximately 0.2 Vh igher than the reversible redox peak that evolvesi nt he next reduction cycle. Aq uinone-hydroquinone redox pair thus can be formed easily also in molecules that initially contain guaiacol groups.A st his mechanism requirest he presence of water,i ti si mpossible in al ithium-ioncontaining setup in water-free organic electrolytes. Consistently,n os uch first irreversible oxidation peak is observableb y using CV in our setup. Still, galvanostatic and CV curves of C/Pvan indicate the formation of the same quinone-hydroquinone redox couple as that in C/P-o with prolonged cycling (Figure 4 and Figure S8) probably because of the slow demethylation that exposes redox-active functionalities. The mechanism for the involved demethylation process is unclear;h owever,w e suggest as imilar mechanism to that found in aqueous systems, with the difference that fluoride anionss erve as nucleophiles instead of water and form fluoromethane in the process. As no significant irreversible oxidation peak is observed in a distinct early CV cycle ( Figure S9 aa nd b), unlike in the case of similar polymers in aqueous systems (Figure S9 c), we conclude that demethylation requires severalc harge-discharge cycles in al ithium half-cell setup. Irreversible oxidation, which usually results in diminishing capacity with prolonged cycling, is common in organic battery materials. [18] Notably,C /P-van reachesc onstant capacity after approximately 10 charge-discharge cycles. After the maximum was reached, the capacity was constantf or 50 charge-discharge cycles,i nc ontrast to the case of C/P-o for whicht he capacity decreases slowly immediately after it reached the maximum. The behavioro fb oth during continuous chargeand dischargecan be explained by a modeli nw hich both materials pass through three phases: sleeping, living, and dead phases as established for lithiumsulfur batteries by Risse et al. [41] In our case, the sleeping phase represents all the guaiacol groups that did not yetundergo initial oxidation and all the hydroquinone groups that are unreachablet ot he electrolyte. The living phase represents all the quinone-hydroquinone redox pairs that undergo redox reactions in that step and is the only phase that contributes to the Faradaic charge storageo ft he material. Finally,t he dead phase represents all the degraded groups that cannot undergo a redox reaction anymore. Furthermore, slights olubility in the electrolyte and the decomposition of the polymer by traces of water may contributet ot he slow capacity fading in the dead phase. During cycling, units can change irreversibly from the sleeping to livingp hase or from the sleeping/living to dead phase.
The different charges torage behavior with continuous cycling may be explained by differences in the sleeping phases betweenb oth materials. The oxidation of hydroquinone groups happens much faster than that of guaiacol groups, which results in the complete transition of C/P-o from the sleeping phase to the living phase within af ew cycles. The subsequentf ading capacity indicates as low transition to the dead phase. In contrast, the constant capacity of C/P-van after the maximum is reached, which may be ascribed to the continuous slow transition of some guaiacol groups from the sleeping to living phase at as imilarr ate as the transition of other active groupsf rom the living phase to the dead phase. Consequently,t his does not necessarily indicatet he highers tability of the material during cycling. Ultimately,the maximum capacity is lower in C/P-van than C/P-o as partial demethylation (formationo fthel iving species) continues after the maximum capacity is reached. Only after approximately 50 cycles do the capacities of C/P-m and C/P-o become similar because of the formationo fq uinone groups in C/P-van and the irreversible oxidation of C/P-o. The final apparent specific capacity of C/P-van is still slightly lower than that of C/P-o because the specific capacity is calculated from the initial mass of the hybrid material, which includes the methyl groups in P-van. The demethylation of guaiacol groups reduces the weight of the polymerm aterial slightly,w hich makes the assumed gravimetric capacity an underestimation.

New promising sustainable cathode material:C /P-o
We next evaluate the electrochemical behavior of the best-performing material C/P-o in detail. If we compare the charge storageb ehavior of C/P-o, C/P-m, and C ( Figure 2b), the influences of carbon, redox activity,a nd changes in wettability may be estimated as discussed in more detail in the Supporting Information.A ccording to the galvanostaticc harge-discharge experiments, after 10 cycles, reversible redoxr eactions contribute to 18.1 %o ft he charge storageo ft he total electrode material. With only 13.0 %d erived from capacitivec harges torage on the surfaceo ft he unmodified carbon material, changes in wettability of the surface have am ajor influence on charge storage( 68.9 %). Abundant phenolicf unctionalities and imines result in significantly enhanced hydrophilicity,w hichf acilitates capacitive charge storage using polar electrolytes. These results point to the importance of the selection of appropriate bindersa nd carbons in the design of electrode materials from renewable resources.
Further information on the electrochemical behavior of C/Poi sg iven in Figure 5. We performed charge-discharget ests at different current densities (Figure 5a)b etween 0.05 and 0.80 Ag À1 (see Experimental Section for details). Within the first three cycles, the capacity increases ands tarts to decrease slowly afterwards,a sd iscussed above. At higherc harge-discharge rates up to 0.80 Ag À1 ,the capacity decreases only moderately and is restored to almosti nitial valuesa fter the current density returnst o0.05 Ag À1 ,which prominently supports diffusion-and surface-controlled charge storage.D ischarge curves at different current densities (Figure 5b)g enerally appear as almosts traightl ines with as light indicationo fabelly shape, which indicates dominant capacitive behavior with some added distinct quinone-hydroquinone redox reactions, respectively.
CV curves of C/P-o at different rates are presented in Figure 5c.A st he integrated area of the CV curve increases slowly over some tens of cycles ( Figure S10), for better comparability the hybrid materialw as cycled firstly for 100 cycles at 25 mV s À1 followed by cycling at 20, 15, 10, 5, and 0.5 mV s À1 for one cycle each. CV curves show clear oxidationa nd reduction peaks at 3.6 and 2.8 V, respectively,w hich match the range of the belly-shaped behavior in the galvanostatic charge-discharge curves. From the performance at different Figure 5. Detailed analysis of the electrochemicalp erformance of C/P-o. The test was performed in alithium half-cell setup with lithium as the counter electrode and 1 m LiPF 6 in EC/DEC (1:1) as the electrolyte. a) Charge-discharge tests at different current densities as indicated. b) Galvanostatic charge-discharging curves (5 th ,15 th ,25 th ,3 5 th ,a nd 45 th cycle from a). c) CV at differentr ates as indicated. d) Diffusion-controlled processes at 5mVs À1 as calculated from the CV at different rates (seeSupporting Information for details). e) Long-term stability test at 0.1 Ag À1 (every third data pointiss hown).
ChemSusChem 2020, 13,1856 -1863 www.chemsuschem.org 2020 The Authors. Publishedb yWiley-VCH Verlag GmbH &Co. KGaA, Weinheim cycling speeds,t he ratio of Faradaic charge storagea nd double layer capacitance was calculated (details can be found in the Supporting Information). [42][43][44][45] Except for the limits between which the CV experimentsw ere performed, the resulting curve for the capacitive contribution to charge storage (Figure 5d,b lack curve) resembles the CV of C/P-m (Figure 2a, blue curve). In both cases, the capacity is approximately 80 % as high as that of C/P-o (84.8 %c f. Figure 2a,r ed curve and 74.8 %c f. Figure5dg reen curve, respectively), which confirms the importance of hydrophilicity on the charge storage and matching the results of galvanostatic experiments (redox reactions in C/P-o are responsible for 18.1 %o ft he charge storage). Clearly,t he addition of P-o to carbon not only introduces the possibility of Faradaic charges torageb ut also facilitates capacitive charges torage. The notable differences between the calculated capacitive contribution to charge storage shown in Figure 5d andt he charge storagei nC /P-m (Figure 2a,b lue curve) at low and high voltages can be ascribed to the limitations of the assumed model, expressedb yalow coefficient of determination (R 2 ;F igure S11) caused by the sudden increase or decrease in current density at the edges of the CVs. Slight differences between the capacitive charge storage in C/P-m and calculated capacitive contribution to C/P-o as denoted above can be ascribed primarily to the aforementioned collapse of the model at the edges of the CVs.
During long-termc ycling at 0.1 Ag À1 (Figure 5e), the capacity decreases slowly as is observed generallyf or organic systems (see discussion above). However,aremarkable retention of 73.5 %i so bserved after 200 cycles (on average 0.13 %d ecay of capacity per cycle). The high Coulombic efficiency of approximately 97 %m akes P-o ap romisingp olymer for electrochemicale nergy storage applications, comparable to other biomass-derived batteries. [46,47]

Conclusions
Polymers made from polyallylaminea nd redox-active and -inactive aldehydes were used as modelc ompounds for biogenic polymers to elucidate the different contributions to electrochemicale nergy storage in future bio-based batteries and lithium ion capacitors. In the systemsu nder investigation, which included carbon, binder,a nd the (in)active polymer,o nly 18.1 %o ft he capacity could be ascribed to distinct redox reactions. All samples showed as ignificantly higher capacity than microporous carbon itself, and the differencew as ascribed to the increased hydrophilicity of the capacitive hybrid material that causes better wettability by the electrolyte and better lithium transportb ehavior.T herefore, we conclude that as ignificant portion of the capacity of electrochemical energy storage devices based on natural polyphenols stems from capacitive charge storageo nh ydrophilic surfaces in addition to redox activity.F urthermore, the electrochemical performance in al ithium-ion-containing settingf or the model polymers that contain redox-active catechol groups was compared to that with naturally abundant guaiacol groups.B oth showedasimilarp erformance after prolonged cycling, with as lower initial increase in capacity for the guaiacol-containing polymers because of their slow demethylation processes. From these findings, ab iomass-derived hybrid materialw as prepared that showed excellent properties for electrochemical energy storage in lithiumorganics ystems with ac apacity of over 100 mA hg À1 at 0.05 Ag À1 and capacityr etention of 73.5 %a fter 200 cycles at 0.1 Ag À1 .
Synthesis of P-o [24] An aqueous solution of PAAm (0.33 g, 0.87 mmol of repeating units) was added to ethanol (10 mL) and stirred for 10 min. After the addition of A-o (0.12 g, 0.87 mmol) dissolved in ethanol (10 mL), the mixture was stirred for 1h.A fterwards the solvent was removed at 60 8Ca tap ressure of 180 mbar,w hich was lowered to 150 mbar shortly after the beginning of the solvent removal. After the obtained polymer was dried at 80 8Cu nder vacuum for 2h,i t was washed with ethanol by centrifugation (4 times with 50 mL ethanol, 4000 rpm, for 5min). The sample was dried in av acuum oven overnight. More information on the synthesis in included in the Supporting Information.
TGA and combustive elemental analysis TGA measurements were performed by using aN ETZSCH TG 209F1 Libra machine under N 2 at ah eating rate of 10 Kmin À1 . Combustive elemental analysis was performed by using av arioMicro CHNS machine.

Morphology characterization
The morphology at the surface was investigated by SEM by using aZ eiss Leo Gemini 1550 microscope, and the atomic distribution of C, O, and Nw as mapped by using energy-dispersive X-ray spectroscopy (EDX, X-Max, Oxford instruments).

Physisorption
Before the physisorption measurements, the samples were degassed under vacuum at 80 8Cf or 16 h. N 2 physisorption measurements were performed by using aQ uantachrome Quadrasorb SI physisorption instrument at 77 K, whereas water vapor physisorption was measured by using aQ uantachrome Autosorb IQ physisorption instrument at 298 K.

Spectroscopy
FTIR spectroscopy was performed by using aNicolet iS 5FTIR spectrometer (ThermoFisher Scientific).