scieee Open visual document viewer

Selective adsorption of CO2 in TAMOF-1 for the separation of CO2/CH4 gas mixtures

Capelo Avilés, Santiago; Fez Febré, Mabel de; Rodríguez Gómez, Salvador; Cabezas Giménez, Juanjo; Oliveira, Raiana Tomazini de; Gallo Stampino, Irene I.; Galán Mascarós, José Ramón

Abstract

TAMOF-1 is a robust, highly porous metal–organic framework built from Cu2+ centers linked by a L-histidine derivative. Thanks to its high porosity and homochirality, TAMOF-1 has shown interesting molecular recognition properties, being able to resolve racemic mixtures of small organic molecules in gas and liquid phases. Now, we have discovered that TAMOF-1 also offers a competitive performance as solid adsorbent for CO2 physisorption, offering promising CO2 adsorption capacity ( > 3.8 mmol g–1) and CO2/CH4 Ideal Adsorbed Solution Theory (IAST) selectivity ( > 40) at ambient conditions. Moreover, the material exhibits favorable adsorption kinetics under dynamic conditions, demonstrating good stability in high-humidity environments and minimal degradation in strongly acidic media. We have identified the key interactions of CO2 within the TAMOF-1 framework by a combination of structural (neutron diffraction), spectroscopic and theoretical analyses which conclude a dual-site adsorption mechanism with the majority of adsorbed CO2 molecules occupying the empty voids in the TAMOF-1 channels without strong, directional supramolecular interactions. This very weak dominant binding opens the possibility of a low energy regeneration process for convenient CO2 purification. These features identify TAMOF-1 as a viable solid-state adsorbent for the realization of affordable biogas upgrading.

Full text

A icle h ps://doi.o g/10.1038/s41467-025-58426-w Selec i e adso p ion o CO 2 in TAMOF-1 o he sepa a ion o CO 2 /CH 4 gas mix u es San iago Capelo-A ilés 1,2,15 , Mabel de Fez-Feb é 1,2,12,15 , Sal ado R. G. Bales a 3 , Juanjo Cabezas-Giménez 1,2,13 , Raiana Tomazini de Oli ei a 1 ,I eneI.GalloS ampino 1 , An on Vidal-Fe an 4,5 , Jesús González-Cobos 1,14 , Vanesa Lillo 1 , Osca Fabelo 6 , Edua do C. Escude o-Adán 1 , La y R. Fal ello 7 ,JoséB.Pa a 8 , Paolo Rumo i 9 , Gemma Tu nes Palomino 9 , Ca los Palomino Cabello 9 , S e ano Giancola 10 ,SofiaCale o 11 & José Ramón Galán-Masca ós 1,4 TAMOF-1 is a obus , highly po ous me al–o ganic amewo k buil om Cu2+ cen e s linked by a L-his idine de i a i e. Thanks o i s high po osi y and homochi ali y, TAMOF-1 has shown in e es ing molecula ecogni ion p op- e ies, being able o esol e acemic mix u es o small o ganic molecules in gas and liquid phases. Now, we ha e disco e ed ha TAMOF-1 also o e s a com- pe i i e pe o mance as solid adso ben o CO 2 physiso p ion, o e ing p o- mising CO 2 adso p ion capaci y ( > 3.8 mmol g–1)andCO 2 /CH 4 Ideal Adso bed Solu ion Theo y (IAST) selec i i y ( > 40) a ambien condi ions. Mo eo e , he ma e ial exhibi s a o able adso p ion kine ics unde dynamic condi ions, demons a ing good s abili y in high-humidi y en i onmen s and minimal deg ada ion in s ongly acidic media. We ha e iden ified he key in e ac ions o CO 2 wi hin he TAMOF-1 amewo k by a combina ion o s uc u al (neu on di ac ion), spec oscopic and heo e ical analyses which conclude a dual-si e adso p ion mechanism wi h he majo i y o adso bed CO 2 molecules occu- pying he emp y oids in he TAMOF-1 channels wi hou s ong, di ec ional sup amolecula in e ac ions. This e y weak dominan binding opens he possibili y o a low ene gy egene a ion p ocess o con enien CO 2 pu ifica- ion. These ea u es iden i y TAMOF-1 as a iable solid-s a e adso ben o he ealiza ion o a o dable biogas upg ading. The p esence o CO 2 in gas s eams is dele e ious o he en i onmen as emissions, bu also p oblema ic as an undesi able impu i y in indus ial eeds ocks and in me hane-based gas mix u es such as biogas and na u al gas1,2. The o me is a g een eplacemen o he la e , whe e me hane is ob ained om he anae obic e men a ion o biological esidues, as a enewable uel3. I s exploi a ion o e s a unc ional solu ion o he con e sion o biowas e in o usable enewable ene gy while educing g eenhouse gases. Biogas upg ad- ing in o biome hane o be used as a uel (high calo ific alue) which can mee pipeline g ade equi emen s is o g ea en i onmen al, economic, and echnological in e es 4,5. Howe e , he sepa a ion and cap u e o he ca bon dioxide ( ypically a ound 50% in biogas) is an ex emely challenging p ocess as (i) he CH 4 and CO 2 molecules ha e simila kine ic diame e s making he sepa a ion o mix u e di ficul by kine ic/size exclusion; (ii) low-p essu e ope a ions a e p e e ed, close o 1 ba (low d i ing o ce) o a oid high ene gy-consump ion om comp ession wo k; and iii) i needs e y high e ficiency in CO 2 emo al (≥98%)6,7. Recei ed: 2 Sep embe 2024 Accep ed: 17 Ma ch 2025 Check o upda es A ull lis o a filia ions appea s a he end o he pape . e-mail: sgiancola@o ches asci.com;S.Cale o@ ue.nl;j [email protected] Na u e Communica ions | (2025) 16:3243 1 1234567890():,; 1234567890():,; Gas-liquid abso p ion in chemical sol en s (aqueous amine solu ion)8is he mos ma u e echnology o la ge-scale CO 2 cap u e9,10, including o me hane pu ifica ion11. Howe e , his ech- nology has h ee majo d awbacks: high ene gy equi emen o sol- en egene a ion, was e managemen (solid sal s and aces o gaseous compounds due o amine deg ada ion), and a la ge oo p in . This calls o he de elopmen o al e na i e solu ions based on no el and cu ing-edge ma e ials. In pa icula , he sepa a ion o CO 2 /CH 4 mix u es di ec ly in he gas phase would be highly desi able. In his con ex , a p omising p ocess o gas sepa a ion is physiso p ion12,13. Due o he physical na u e o he in e ac ions in ol ed, his p ocess ypically demands lowe ene gy consump ion du ing he egene a ion s ep when compa ed o p ocesses uled by s ong chemical in e ac- ions such as chemiso p ion in solid o liquid adso ben s. Di e en echnologies based on physiso p ion a e a ailable, such as p essu e swing adso p ion (PSA)14,15, acuum swing adso p ion (VSA)16,and he mal swing adso p ion (TSA)17. These echnologies in ol e a leas wo consecu i e s eps. Fi s , CO 2 is selec i ely adso bed by he ac i e componen a high p essu e and/o low empe a u e, and hen he adso ben is egene a ed by CO 2 deso p iona lowe p essu e/ acuum (PSA/VSA) o /and a highe empe a u e (TSA). Se e al ypes o po ous ma e ials ha e been in es iga ed as phy- siso ben s o CO 2 emo al18. Benchma k zeoli es ypically exhibi compe i i e sepa a ion pe o mance, cha ac e ized by high adso p- ion capaci y and equilib ium selec i i y19. Howe e , while ad an a- geous o CO 2 emo al om d y emission o p ocess s eams, his selec i i y is o en comp omised in he p esence o wa e . Wa e molecules p e e en ially coo dina e wi h exposed me al si es, ou - compe ing CO 2 20. Fu he mo e, hese ma e ials necessi a e high egene a ion ene gy19,21,22, which consequen ly inc eases he ope a- ional expendi u e (OPEX) in indus ial plan s s i ing o high CO 2 15,20. Ca bon-based ma e ials o e he ad an age o low cos and easy egene a ion, al hough hey a e limi ed by a lowe adso p ion capaci y and selec i i y. Nowadays, he limi a ions o he adso ben s a e also limi ing he comme cial use o ca bon cap u e (CC) echnologies jus o niche applica ions, whe e he e is no al e na i e. Sui able ma e ials combining high adso p ion capaci y a low/mode a e wo king p es- su es, able o disc imina e CO 2 agains molecules wi h simila dimen- sions (e.g., CH 4 and N 2 ) wi h high selec i i y, wi h low-ene gy egene a ion and as adso p ion/deso p ion kine ics would o e plausible oppo uni ies o ealize CC no only o biogas upg ading bu also in many addi ional fields. Especially i CC becomes economically compe i i e wi h cu en CO 2 emission igh s. Me al-o ganic amewo ks (MOFs) a e c ys alline and po ous ma e ials23,24 p epa ed by he sel -assembly o me al ions o clus e s wi h o ganic ligands o linke s o o m e icula s uc u es25,26.Due o hei excep ional p ope ies27 such as high accessible su ace a eas (up o 6200 m2g–1)28, ex a high po osi ies (up o 90%)29, unabili y in po e dimensions and mo phology30, e sa ile unc ionali y31 as well as he mal and chemical s abili y32, MOFs a e ecognized as p omising ma e ials o selec i e gas sepa a ions po en ially able o o e come in insic limi a ions o common adso ben s33–35. E en hough nowa- days mo e han70000 MOFs ha e been disco e ed, jus a ew o hem ha e shown p omising ea u es in o adso p ion/deso p ion p o ocols o as memb ane componen s o CO 2 cap u e23,36. Some o he mos ema kable candida es include CALF-20, a Zn-based oxala e-b idged amewo k wi h high pe o mance inCO 2 cap u e om flue gas, andan ex ao dina y obus ness unde humid condi ions37;o Mg 2 (dobpdc), a Mg-based e amine- unc ionalized amewo k wi h ex ao dina y cyclabili y and s abili y du ing CO 2 cap u e om flue gas. Howe e , his las ma e ial is cha ac e ized by chemical CO 2 adso p ion and needs a high ope a ing empe a u e (100°C)38.In hecaseo CO 2 /CH 4 sepa a ions o biogas upg ading, he mos ele an example is MUF- 16 (Co(Haip) 2 , Haip = 5-aminoisoph halic acid)39.ThisMOFisable o cap u e ca bon dioxide om hyd oca bons wi h excep ional selec i - i y al hough wi h mode a e CO 2 adso p ion capaci y40.Mo eo e , he low isos e ic hea o adso p ion allows o an easy egene a ion s ep. Ano he in e es ing MOF is Qc-5-Cu-sql, a Cu- quinoline-5-ca boxylic acid sup amolecula ne wo k, also showing excellen CO 2 /CH 4 selec- i i y ia he molecula sie ing mechanism. Howe e , also in his case mode a e CO 2 adso p ion capaci y is epo ed41.Mo eo e ,a ho ough explana ion o he selec i i y o MOFs o adso bCO 2 isusually missing. Whe eas chemiso p ion is easie o e alua e om s uc u al da a, he physiso p ion mechanism emains di ficul o assess. In si u expe i- men s a e e y a e o disc imina e be ween: (i) size-exclusion p inci- ples, (ii) kine ically con olled p ocesses, o (iii) he modynamically con olled p ocesses42. TAMOF-1 is he fi s in a se ies o homochi al MOFs based on na u al amino acid de i a i es, by ans o ma ion o he α-amino uni in o a iazole g oup43,44. TAMOF-1 ([Cu(S-TA) 2 ].xH 2 O, S–HTA = (S)-3- (1H-imidazol-5-yl)-2-(4H-1,2,4- iazol-4-yl)-p opanoic acid) is easily syn hesized a la ge scale om low-cos aw ma e ials jus by eac ion o a coppe (II) sal wi h L-his idine de i a i e (imidazole-5-ylme hyl)- (1,2,4– iazol-4-yl)ace a e (L 1 ) in wa e . This ma e ial has a 3D ne wo k, made om 10 Å wide, helicoidal, in e communica ed channels, deco- a ed wi h mul iple dangling unc ional g oups ca boxyla e, iazole, and imidazole (Fig. 1) exhibi ing a BET-specific su ace a ea o 980 ± 50 m2g–1and a mic opo e olume o 0.38 cm3g–1.TAMOF-1 exhibi s excep ional wa e s abili y, wi hs anding bo h hyd a ion/ dehyd a ion cycles wi hou s uc u al deg ada ion o po osi y loss44. Mo eo e , ac i a ion o gas sepa a ion is achie ed unde mild con- di ions (353–393 K unde a sweep gas flow), p ese ing he MOF c ys allini y—a a e ea o Cu-based MOFs and su passing he equi emen s o ypical MOFs and zeoli es45,46. In p e ious epo s, we disclosed he pe o mance o TAMOF-1 as a s a iona y phase o he ch oma og aphic sepa a ion o acemic mix u es o o ganic subs ances hanks o i s po osi y and homochi ali y47. The la e was u he exploi ed o he kine ic eso- lu ion o chi al subs a es by ca aly ic coupling. Mo eo e , he cha - ac e is ic shape and size o i s channels and po es, as well as he Cu me al cen e , make TAMOF-1 a good candida e o sepa a ing also a wide a ie y o ola ile o ganic compounds such as benzene−cyclo- hexane sys em and xylene isome s (posi ional isome s econnec ion). The sepa a ion o hese molecules has been p o ed in bo h he liquid and gas phases48. Fu he mo e, TAMOF-1 has been epo ed o be capable o sepa a ing C 2 H 2 /C 2 H 4 and C 2 H 2 /CO 2 mix u es49. Fig. 1 | C ys al s uc u e o TAMOF-1. Rep esen a ion o he c ys al s uc u e o TAMOF-1, showing he ne wo k o open 10 Å-wide channels. Colo code: Cu, deep blue; O, ed; N, ligh blue; C, black. Hyd ogen a oms omi ed o cla i y. A icle h ps://doi.o g/10.1038/s41467-025-58426-w Na u e Communica ions | (2025) 16:3243 2 He e, we epo ha TAMOF-1 is a highly p omising physiso ben o esol e CO 2 /CH 4 mix u es and, in a b oade con ex , aiming o cap u e CO 2 . In pa icula , his ma e ial has highly p omising CO 2 adso p ion capaci y and CO 2 /CH 4 selec i i y. Mo eo e , i allows o a low ene gy (ambien empe a u e) egene a ion. When compa ed wi h o he a ailable adso ben s, including o he MOFs, he TAMOF-1 pe - o mance appea s e y p omising also in e ms o ope a ion cos s. We an icipa e ha his ma e ial may help o b idge he gap be ween e ec i e adso p ion and a o dable egene a ion. No ewo hy, he adso p ion/deso p ion pe o mance emains du able o e ime, e en in he p esence o wa e , wi h only a small educ ion obse ed when exposed o highly concen a ed H₂S. By combining specificexpe i- men al (neu on di ac ion and IR spec oscopy) and heo e ical ools (Mon e Ca lo and molecula dynamics), we ha e ho oughly eluci- da ed he molecula mechanism o he physiso p ion and anspo p ope ies o gas molecules wi hin he TAMOF-1 ne wo k a he o igin o i s p omising pe o mance. Resul s and discussion Gas adso p ion Single gas (CO 2 ,CH 4 ,andN 2 ) adso p ion iso he ms o TAMOF-1 powde up o 10 ba s we e measu ed in he 293–353K empe a u e ange (Fig. 2a). As a unc ion o empe a u e, all iso he ms main ain hei cha ac e is ic shapes, wi h gas up ake inc easing as he em- pe a u e dec eases, ypical o a physiso p ion p ocess. Type I iso- he ms a e ob ained o CO 2 , wi hou eaching a pla eau (sa u a ion) in he in es iga ed p essu e ange. Fo example, a 298 K, CO 2 adso p ion capaci ies o 3.8 and 6.5 mmol g–1a e ob ained a 1 and 10 ba s, espec i ely. Linea iso he ms a e ob ained o bo h CH 4 and N 2 ,wi h adso p ion capaci y ollowing he o de CO 2 >>CH 4 >N 2 , indica ing selec i e p e e en ial up ake o ca bon dioxide. F om he di e en adso p ion models a ailable (See SI), he dual- si e Langmui -F eundlich model shows he bes fi ing o he CO 2 adso p ion iso he ms(Supplemen a yFigs. 7–11). See, o ins ance, he compa ison wi h he single-si e Langmui -F eundlich model in he logq s. logPplo (Supplemen a y Fig 12). This indica es ha a leas wo di e en adso p ion si es a e esponsible o dominan CO 2 up ake. Supplemen a y Tables 16, 17 show he iso he m pa ame e s and eg ession coe ficien s. The CO 2 /CH 4 selec i i y calcula ed by he ideal adso p ion solu- ion heo y a e epo ed in Fig. 2c and in he Supplemen a y Table 18. A 293 K and 1ba ,45.9, 40.9, and 38.30 selec i i y alues a e ob ained o espec i ely 30:70, 50:50, and 70:30 CO 2 /CH 4 gas mix u es (in he ypical biogas composi ion ange). In e es ingly, da a indica e ha TAMOF-1 is capable o sepa a ing CO 2 /CH 4 mix u es o e a wide empe a u e ange. Fo example, a 1 ba and 353K,TAMOF-1 hasa CO 2 so p ion capaci y and 50:50 CO 2 /CH 4 ideal selec i i y o 1.8 mmol g–1 and 18, espec i ely. Ano he impo an ea u e is he comple e adso p ion/deso p ion e e sibili y ob ained o all he gases (Fig. 2b) wi h no hys e esis obse ed. This beha io aligns wi h he physical na u e o he gas adso p ion phenomena wi hin he TAMOF-1 ne wo k and unde sco es he s uc u al igidi y o his ma e ial. The isos e ic en halpy o adso p ion, ΔH ads was es ima ed ia an indi ec app oach om he CO 2 adso p ion iso he ms a he di e en empe a u es using he Clausius-Clapey on equa ion ( o a de ailed desc ip ion o he calcula ions, see SI). Supplemen a y Fig. 13 shows he e olu ion o –ΔH ads wi h CO 2 up ake. A high –ΔH ads is ound a low p essu e (ze o co e age, –ΔH0 ads=0 ), and hen i apidly dec eases as CO 2 up ake inc eases, app oaching he bulk-phase sublima ion hea o CO 2 ,26–27 kJ mol–1, a an adso p ion capaci y highe han 2mmolg –150. This ene gy ange poin s o a physical gas adso p ion beha io and deno es a he e ogeneous adso p ion p ocess occu ing a mul iple adso p ion si es wi h di e en su ace ene gies51. Sahoo e al.42 conduc ed a p elimina y compa ison o mixed gas phase CO 2 so p ion capaci y and sepa a ion selec i i y o a ious MOFs using IAST (Fig. 2d). While many s udies epo only single gas so p ion capaci y (which is ypically highe han he mixed gas capa- ci y), he e iew ocused on hose ha p o ided mixed gas CO 2 da a. Among he MOFs examined, MUF-16 and Qc-5-Cu-sql-βdemons a ed bo hhighCO 2 so p ion capaci y (1.8 and 1.6 mmol g–1, espec i ely) and high CO 2 /CH 4 sepa a ion selec i i y (6690 and 3300, espec- i ely). In e es ingly, TAMOF-1 ( his s udy) exhibi ed an e en highe mixed CO 2 so p ion capaci y (~ 3.1 mmol g–1)wi haccep ablesepa a- ion selec i i y (> 40). Gi en he lack o es ablished pe o mance benchma ks o indus ial-scale biogas sepa a ion, i is p ema u e o defini i ely assess he po en ial o hese MOFs. We ha e also o men- ion ha IAST pa ame e s a e calcula ed unde idealized equilib ium condi ions and do no ep esen he ac ual sepa a ion pe o mance o adso ben s in dynamic condi ions such as hose ob ained om b eak h ough cu es. Ne e heless, IAST analysis is usually epo ed and used o compa e he gas sepa a ion pe o mance o adso ben s aluable insigh s in o he ela i e adso p ion capabili ies o TAMOF-1 wi h espec o o he adso ben s. B eak h ough measu emen s In Fig. 2e, we epo me hane/ca bon dioxide b eak h ough cu es h ough a TAMOF-1 powde bed a 1 ba and 298 K a di e en CO 2 / CH 4 a ios. Gas sepa a ion pa ame e s ob ained om he b eak- h ough cu es a e epo ed in Supplemen a y Table 19 (See SI o de ails). Pu e me hane (≥99.9%) elu es fi s and speedily om he bed. Regula S-shaped cu es we e ins ead obse ed o CO 2 . E ec i e CO 2 / CH 4 sepa a ion is achie ed in all cases as CO 2 is delayed. This confi ms slowe CO 2 di usion h ough he TAMOF-1 bed. By educing CO 2 mola flow, he ela ed cu es shi owa ds highe b eak h ough imes (B ).B (CO 2 ) inc eases om 16.4 o18.7 min g–1when heCO 2 /CH 4 a io is educed om 50:50 o 25:75. In con as , B (CH 4 ) is almos cons an , independen o he CH 4 ac ion. The CH 4 concen a ion o e shoo s i s equilib ium alue (C/C 0 > 1) be o e e u ning o equilib ium (C/ C 0 = 1). This oll-up e ec indica es p e e en ial adso p ion o CO 2 , which displaces some o he ini ially adso bed CH 4 . The p ecise shape o his o e shoo depends on he me hane flow a e, concen a ion, and adso ben p ope ies20. Adso p ion capaci ies o bo h ca bon dioxide and me hane we e also measu ed. Ni ogen is a gas ha can be p esen in combina ion wi h CO 2 and CH 4 in many gas s eams, hus we also e alua ed CO 2 /CH 4 /N 2 sepa a- ions h ough TAMOF-1 beds. Fo his pu pose, fixed-bed column expe imen s we e pe o med by a ying he CH 4 /N 2 a io a cons an CO 2 concen a ion (50%) (Fig. 2 ). Ca bon dioxide b eak h ough cu es pe ec ly o e lap inde- penden ly o he CH 4 /N 2 a io. This indica es ha CO 2 di usion h ough TAMOF-1 is no a ec ed by he ype o o he gas compo- nen s (CH 4 o N 2 ) in he inle gas mix u e. These gases do no a ec he adso p ion capaci y, which depends di ec ly on he CO 2 con- cen a ion, in ag eemen wi h iso he ms measu es (Fig. 2a). As a consequence, CO 2 adso p ion capaci ies (Supplemen a y Table 19) a e simila (q b = 1.7± 0.06 mmol g–1,q s = 2.5 mmol g–1,CO 2 = 50%). A simila CO 2 /CH 4 selec i i y was also obse ed o a bina y CO 2 /CH 4 (S s =6) o a CO 2 /CH 4 /N 2 e na y (S s = 4) inle gas mix u e wi h he same CO 2 concen a ion (i.e., 50%). I is impo an o men ion ha N 2 , when p esen , exi s as he fi s componen and almos immedi- a ely om he column wi h elu ion ime almos coinciden wi h he dead ime. B eak h ough expe imen s a a iable p essu e condi ions we e also collec ed (Supplemen a y Table 19). By p essu izing he bed, B shi s o highe imes wi h me hane emaining always he fi s gas o elu e. Inc easing bed p essu e om 1.2 o 6 ba a a cons an flow a e, B (CO 2 ) inc eases om 34 o 122 min g–1, and B (CH 4 ) om < 4.7 o 32 min g–1.CO 2 adso p ion capaci ies inc ease om 0.42 o 1.24 mmol g–1, bu CH 4 adso p ion also inc eases om < 0.06 o 0.33 mmol g–1. This esul s in a dec ease in he CO 2 /CH 4 selec i i y as A icle h ps://doi.o g/10.1038/s41467-025-58426-w Na u e Communica ions | (2025) 16:3243 3 a unc ion o inc easing p essu e, om > 7 o < 4 be ween 1.2 and 6 ba . Highe empe a u es accele a e he gas elu ion, and he CO 2 concen a ion p ofiles sensibly shi owa d lowe B and lowe adso p ion capaci y (Supplemen a y Table 19). This is in ag eemen wi h he weake adso ben -adso ba e in e ac ions a highe empe a u es, he as e molecula di usion and he exo he mic cha ac e o he adso p ion p ocess. A negligible empe a u e e ec was ins ead obse ed o CH 4 , ha speedily di uses in o he bed and almos immedia ely exi s om he bed. Sepa a ion e ec i eness dec eases by inc easing he empe a u e, and a empe a u es simila o he ac i a ion one (393 K), CO 2 and CH 4 cu es nea ly o e lap. Good A icle h ps://doi.o g/10.1038/s41467-025-58426-w Na u e Communica ions | (2025) 16:3243 4 sepa a ion capabili y (S b >0.9)was oundup o353K,indica ing applicabili y in a wide empe a u e ange. Regene a ion and s abili y assessmen o TAMOF-1 The egene a ion s ep o a TAMOF-1 bed a e sa u a ion appea s o be qui e simple, a oiding he need o he mal hea ing. TAMOF-1 can be comple ely egene a ed a a cons an empe a u e jus by passing h ough a sweep gas (N 2 ). The co esponding CH 4 and CO 2 deso p ion cu es a e sa u a ion by an N 2 sweep gas s eam we e collec ed o di e en flow a es, empe a u es, and p essu es (Fig. 3) CH 4 deso bs as e han CO 2 , as expec ed. The high CO 2 /CH 4 selec i i y in deso p ion allows o he eco e y o a high pu i y CO 2 ac ion, which can be eused as a commodi y. Mo eo e , he egen- e a ion imes a e ela i ely sho , when compa ed wi h he B imes. Fig. 2 | Adso p ion and deso p ion p ope ies o TAMOF-1. a Adso p ion/deso- p ion iso he ms o CO 2 ,CH 4 and N 2 in a TAMOF-1 powde sample a 298 K up o 10 ba . bAdso p ion iso he ms o CO 2 wi h a empe a u e ange o 293–353 K up o 10 ba . cIAST selec i i y o di e en CO 2 /CH 4 gas mix u es wi h a ying a ios a 25 °C up o 1 ba . dCO 2 so p ioncapaci y e susCO 2 /CH 4 IASTselec i i y o bina y 50:50 / mix u es a 1ba and 293 K (IAST selec i i y alues> 20), adap ed om e . 42.eB eak h ough cu es o CO 2 (solid line) and CH 4 (dashed line) o bina y mix u es wi h di e en CO 2 /CH 4 a ios: 25:75(blue), 50:50( ed), 75:25(g een). CO 2 b eak h ough cu es in di e en inle gas mix u es: 50:50 CO 2 /CH 4 ( ed), 50:50 CO 2 /N 2 (cyan), CO 2 /CH 4 /N 2 =50/ 2 5/25 (magen a). Measu es in (eand )we e pe o med wi h 0.7 g o ac i a ed TAMOF-1 powde . TAMOF-1 was ac i a ed unde acuum (10−1mba ) in bo h column op and bo om sides a 393K o 15h. gCO 2 / CH 4 (50:50) b eak h ough cu es o ac i a ed TAMOF-1 (7.7 g) a 72% ela i e humidi y (RH). TAMOF-1 was ac i a ed unde acuum (10-1 mba ) in bo h column op and bo om sides a 393 K o 15 h. hCO₂/CH₄(50:50) b eak h ough cu es o ac i a ed TAMOF-1 (0.7 g) unde acid gas condi ions (0.9% H₂S, 2.5% H₂O). TAMOF- 1 was ac i a ed unde N 2 flow (170 mil min–1g TAMOF –1) a 393 K o 15 h. All b eak- h ough expe imen s we e pe o med a 298 K and 1.05ba . See Supplemen a y Tables 19–21 o de ailed fixed-bed adso p ion pa ame e s. Fig. 3 | Gas deso p ion dynamics o TAMOF-1. TAMOF-1 deso p ion cu es o CO 2 ( ull line) and CH 4 (dashed line) in a N 2 sweep gas flow o egene a e he TAMOF-1 a e CO 2 -CH 4 b eak h ough expe imen s o (a) di e en concen a ions o equi- molecula CO 2 -CH 4 mix u es, (b) di e en p essu es, (c)di e en flow a es and (d) di e en empe a u es. In all cases, he p ocesses we e pe o med a cons an p essu e, empe a u e, and o e all flow a e du ing bo h, he b eak h ough and deso p ion s eps. TAMOF-1 (0.7 g) was ac i a ed unde N 2 flow (170 mL min–1 g TAMOF –1) a 393 K o 15 h. A icle h ps://doi.o g/10.1038/s41467-025-58426-w Na u e Communica ions | (2025) 16:3243 5 The egene a ion ime can be accele a ed by inc easing he ni ogen flow a e, o educing he p essu e, o inc easing he empe a u e. All hese da a highligh he po en ial o TAMOF-1 as an adso ben o p essu e/ acuum/ he mal swing adso p ion p ocesses. The p esence o humidi y is inhe en in all aw biogas mix u es, ypically eaching sa u a ion le els unde wa e 20.F om anenginee ing pe spec i e, and pa icula ly when employing adso ben s unde acuum condi ions, pa ial emo al o wa e om he inle s eam becomes c ucial. This necessi y a ises om he po en ial o wa e condensa ion wi hin he column, leading o obs uc ions ha can diminish acuum p essu e du ing he deso p ion phase52. Howe e , i is impo an o acknowledge ha a ce ain deg ee o wa e apo will in a iably pe sis in he biogas mix u e en e ing he adso ben col- umn. P e ious in es iga ions ha e demons a ed he s uc u al obus ness o TAMOF-1 unde humid condi ions, e en in liquid wa e , wi h he ma e ial exhibi ing comple e eac i a ion. He e, we e alua e he compe i i e adso p ion beha io o CO 2 and H 2 O, as well as he adso ben cyclic s abili y unde dynamic adso p ion condi ions. As depic ed in Fig. 2g, e en in he p esence o high humidi y (~ 72% RH), TAMOF-1 e ains i s CO 2 adso p ion capaci y (2.3mmolg−1)wi hjus a 15% educ ion compa ed o he d y inle mix u e. The CO 2 /CH 4 selec i i y is also main ained wi hin he ange o 4–6. Subsequen o he adso p ion s ep, egene a ion was conduc ed unde acuum a a empe a u e o 80 °C. The sepa a ion pe o mance o TAMOF-1 emained consis en o e mul iple adso p ion-deso p ion cycles, as demons a ed by he o e lapping CO 2 and CH 4 b eak h ough cu es (Supplemen a y Table 20). This s abili y indica es negligible deg ada- ion o he ma e iala e successi e cycles andnea -comple e eco e y o wo king capaci y unde he ope a ing condi ions in es iga ed. Hyd ogen sulfide (H 2 S), a common impu i y ound in aw biogas (~ 0–100 ppm1), p esen s a significan challenge o biogas upg ading. H 2 S is a highly co osi e acid gas ha can comp omise he chemical s abili y o ma e ials used in sepa a ion p ocesses. Fo ins ance, H 2 S can dis up hecoo dina ion bonds be ween o ganicligands and me al cen e s in he MOFs, leading o s uc u al deg ada ion53.Toassess he sepa a ion beha io o MOFs in he p esence o hyd ogen sulfide is he e o e c ucial o hei e ec i e applica ion in biogas upg ading. Analysis o TAMOF-1 in humid (2.5 mol% H 2 O) and acidic gas (0.9 mol% H 2 S) condi ions (Fig. 2h) demons a es i s abili y o sepa a e CO 2 /CH 4 mix u es, e en unde hese condi ions. Howe e , a educ ion o adso p ion capaci ies (16%) is obse ed be ween he fi s and he second cycle, while selec i i y emains ela i ely cons an (~ 3–5). We obse ed a educ ion in he compe i i e adso p ion be ween CO 2 and CH 4 , wi h negligible co-adso p ion o CH 4 .Weassign hisdi e ence o he e y high e en ion ime ound o H 2 S(noH 2 S was de ec ed in he ou le s eam no e en a e he CO 2 equilib ium egime), which compe es o adso p ion si es. Impo an ly, his analysis was pe - o med unde accele a ed condi ions, wi h e y high H 2 S concen a- ion (9000 ppm), almos wo o de o magni ude highe han wha is usually ound in biogases. These da a, he e o e, demons a e a ela- i ely high obus ness o he ma e ial also in s ongly acid condi ions. Howe e ,H 2 S emo alis ecommended p io obiogasupg adingwi h TAMOF-1 in o de o inc ease ma e ial du abili y. S uc u e de e mina ion wi h adso bed CO 2 We a emp ed o localize he p e e en ial c ys allog aphic posi ion o he CO 2 molecules when adso bed in he TAMOF-1 amewo k by X- ay di ac ion analysis. A single c ys al was dehyd a ed, exposed o a CO 2 s eam, and hen cooled down o 100 K. The analysis o he XRD s uc u al da a e ealed esidual densi ies in he channels ha we e compa ible wi h he p esence o CO 2 in he c ys al ma ix. Howe e , he poo selec i i y unde X- ay di ac ion condi ions be ween oxygen and ca bon, in addi ion o he possible p esence o esidual wa e , p ecluded he success ul loca ion o CO 2 molecules in he channels. Fo his eason, we u ned o neu on di ac ion54, whe e ca bon and oxygen a oms can be p ope ly iden ified hanks o hei dis inc neu- on sca e ing leng hs. The neu on di ac ion s uc u e analysis o an ac i a ed TAMOF- 1 single c ys al confi ms he c ys al s uc u e ob ained by XRD (Fig. 4) wi h iangula emp y channels unning along he [111] di ec ion, which is a 3- old axis due o i s cubic symme y (Supplemen a y Tables 1–7). A e exposu e o CO 2 , and cooling down o 100 K, hese channels a e filled wi h CO 2 molecules exhibi ing mode a e diso de (Fig. 4), indi- ca ing ha weak in e ac ions wi h he TAMOF-1 ne wo k a e dominan (Supplemen a y Tables 8–15). Adjacen o his channel, h ee coppe a omssel -assemblein wo o ien a ions ha a e o a edapp oxima ely 45° wi h espec o each o he e e y 13.5 Å along he [111] axis, o ming equila e al iangles o app oxima ely 6.7 Å. This a angemen c ea es a na owe channel whe e CO 2 molecules align hemsel es wi h he oxygen’s lone pai s loca ed be ween he Lewis si es o wo o he coppe a oms in he iangle. These CO 2 molecules a e he only ones ha exhibi a clea , di ec ional in e ac ion wi h any o he a ailable TAMOF-1 si es. We associa e his c ys allog aphic posi ion wi h he highe isos e ic en halpy o adso p ion ound a low co e age. The es o he CO 2 molecules a e loca ed along he channels o he s uc u e wi h appa en ly andom o ien a ions. The in e molecula in e ac ion be ween CO 2 molecules loca ed in he channel is consis en wi h London dispe sion o ces in ag eemen ins ead wi h he lowes isos e ic en halpy o adso p ion app oaching he bulk-phase sublima ion hea o CO 2 . The oxygen a oms in he CO 2 molecules possess lone pai s ha a e di ec ed owa d he mo e posi- i ely cha ged ca bon a oms, which a e si ua ed a dis ances anging be ween 2.6 Å and 3.1Å. Simila ly, he ca boxyla e g oups o he TAMOFin e ac in hesamewaywi h heCO 2 molecules,as heoxygen a oms in he ca boxyla e g oups exhibi lone pai s ha a e di ec ed owa ds heca bona omsin heCO 2 molecule, whicha eposi ioneda adis anceo 3.1Å. I is impo an o no e ha hese neu on di ac ion da a we e ob ained a 100 K. Room empe a u e da a collec ion esul ed in a diso de ed model, indica ing ha oom empe a u e is enough o b eak he dominan in e ac ions be ween CO 2 and TAMOF-1. Spec oscopic analysis To u he in es iga e he in e ac ions be ween CO 2 and he TAMOF-1 amewo k a oom empe a u e, we collec ed IR spec oscopy da a. The IR spec a o CO 2 when adso bed on TAMOF-1 (Supplemen a y Fig. 4) show a band a 2335 cm–1, which, acco ding wi h p e ious epo s55,56, can be asc ibed o physiso bed CO 2 . The less in ense band a 2324 cm–1has also been obse ed be o e and assigned o a combina ion band. These esul s sugges ha he e a e no si e-specifics ong in e ac ions and CO 2 only physically in e ac s wi h he amewo k. To confi m he alidi y o hese esul s, we also in es iga ed CO adso p ion o compa ison. Supplemen a y Fig. 4b shows he C-O s e ching egion o ca bon monoxide adso bed on TAMOF-1. In e - ac ion o he me al-o ganic amewo k wi h adso bed CO esul ed in he o ma ion o an IR abso p ion band a 2139 cm–1,which,acco ding o li e a u e, is assigned o (non-localized) physiso bed ca bon monoxide57, indica ing ha he me al-o ganic amewo k TAMOF-1 does no p esen accessible coo dina i ely unsa u a ed me al cen e s. Compu a ional analysis and adso p ion mechanism We calcula ed he single-componen adso p ion iso he ms o CO 2 , CH 4 ,andN 2 molecules in TAMOF-1 in he 263–333 K, and 1–107Pa ange. The good ag eemen be ween simula ion and expe imen s alida es ou models (Fig. 5). Excess adso p ion, ollowing he p o ocol o Mye s and Monson58, allows compa ison wi h he expe imen . Some ene gy pa ame e s ela ed o ca bon cap u e we e also calcula ed om he single-componen iso he ms o CO 2 and N 2 . We es ima e ha TAMOF-1 exhibi s low pa asi ic ene gy, a olume ic wo king capaci y o 23.46 kg m–3,andafinal CO 2 mola pu i y o he mix u e o 0.904, A icle h ps://doi.o g/10.1038/s41467-025-58426-w Na u e Communica ions | (2025) 16:3243 6 which a e excellen ea u es o ca bon cap u e. The calcula ed adso p ion iso he ms we e fi ed o a dual-si e ( o CO 2 ) and single si e ( o CH 4 and N 2 ) Langmui -F eundlich model using he RUPTURA so wa e59 (solid lines in Fig. 5). Subsequen ly, we p oceeded o com- pu e he bina y iso he ms and de e mine he adso p ion selec i i y o CO 2 :CH 4 and 14 CO 2 :84N 2 mix u es ( ep esen ing he composi ion o d y pos -combus ion flue gas). The esul ing adso p ion selec i i ies a e depic ed in Fig. 5d. In all cases, he adso p ion o CO 2 su passes ha o CH 4 o N 2 , p ima ily due o he s onge in e ac ions be ween CO 2 molecules and he adso p ion su aces. Simula ion shows ha a low alues o p essu e (1–500 Pa) he molecules o ca bon dioxide a e loca ed in he na ow channels o med by he iad o Cu a oms. The e is a dominan in e ac ion be ween CO 2 and wo Cu a oms (as shown in Fig. 4anddiscussedin he SI, neu on di ac ion analysis sec ion), esul ing in high alues o he hea o adso p ion, 50 kJ mol–1, in ag eemen wi h expe imen al esul s (Qs  Uhg DE +RT whe e Uhg DE is he mean in e ac ion ene gy be ween he CO 2 and he TAMOF-1 a e y low p essu e). Howe e , as he CO 2 p essu e inc eases, he CO 2 -CO 2 in e ac ions become s ong enough o displace some o he adso bed CO 2 molecules om hese ene ge ically a o able binding si es. This can be seen om adial dis ibu ion unc ions be ween CO 2 -CO 2 and CO 2 -Cu a oms, displayed in Fig. 6. Figu e 6a shows ha he in e ac ion ene gy be ween he CO 2 molecule and he adso p ion su ace emains s able a a ound Fig. 4 | C ys al s uc u e om neu on da a showing he occupa ion o CO 2 molecules in o he channels in TAMOF-1. a De ail o he CO 2 posi ions wi hin he Cu 3 iangles in hela ges oidspace o TAMOF-1. The dis ances O1C-Cu1(zxy) and O2C-Cu1 a e 2.7726(1) and 2.7068(1) Å, espec i ely. The iew di ec ion is along [111]. Cu1, Cu1(zxy) and Cu1(yzx) a e ela ed by a c ys allog aphic h ee- old axis, whichalso ela es heCO 2 moleculea C1C o woo he s,which a eno shown.C1C/ O1C/O2C and i s wo congene s ha e si e occupancy ac o s o 1/3 as a esul o diso de abou he symme y axis. Colo code: Cu, deep blue; O, ed; N, ligh blue; C, g ay; H, whi e. bView along [111] o he TAMOF-1 open amewo k (g een s ick ep esen a ion) and he posi ions occupied by he gues CO 2 molecules ( ed an de Waals’ adii ep esen a ion) in he TAMOF-1 channels, wi h he wo componen s shown oge he (le ) and indi idually ( igh ). View along [100] (c)and iewalong [111] (d), o pa s o he complex oid s uc u e ha accommoda es he gues CO 2 in TAMOF-1. The oids a e ep esen ed as ligh ed a eas o ex e io su aces and as shadowed a eas o in e io su aces o he oid bounda ies. A icle h ps://doi.o g/10.1038/s41467-025-58426-w Na u e Communica ions | (2025) 16:3243 7 –45 kJ mol–1up o 3 kPa and p og essi ely dec eases wi h p essu e o –23 kJ mol–1.Al hough his endisquali a i ely equi alen o CH 4 and N 2 , he ange o ene gies and p essu es di e s, and he e ec on adso p ion is neglec ed. Fo example, o CH 4 , he in e ac ion ene gy a ies om –25 kJ mol–1(low p essu e) o –12 kJ mol–1(high p essu e). To conclude, he adso p ion selec i i y o CO 2 o e CH 4 o N 2 a in e media e p essu es (Fig. 5d) is mainly explained by wo causes: he elec os a ic in e ac ions be ween CO 2 and he squa e plana Cu si es, and he confinemen o CO 2 inside he TAMOF-1 pocke -cages. The na ow po es o med by he iad o Cu a oms con ibu e o dec ease his selec i i y a high p essu es. The sel -di usi i y coe ficien s (D)o CO 2 and CH 4 we e calcu- la ed using MD simula ions a a ious empe a u es and adso p ion loadings (Fig. 7). CO 2 consis en ly exhibi s lowe di usi i y han CH 4 . A low loading (1 o 4 CO 2 molecules pe uni cell), he molecules o CO 2 a e confined in adso p ion pocke s o med by he iad Cu a oms, esul ing in low di usi i y (i.e., D S CO2 ~10 –10 ms –1). As p essu e inc eases, he molecules o CO 2 di use om he pocke adso p ion si e o he main channel. This leads o a maximum di usi i y a ~ 3 mol kg–1o adso p ion loading o all empe a u es (e.g., a 333 K, he sel -di usi i y is app oxima ely D S CO2 ~10 –9m·s–1). Beyond his loading( h eemoleculespe Cu iad), hemoleculeso CO 2 a e mainly adso bed in he (chi al) channel, p e en ing anspo and causing a p og essi ely dec easing o di usi i y. In con as o CO 2 , o CH 4 molecules, sel -di usi i y eaches i s peak a infini e dilu ion o all empe a u es (see Fig. 7b). The di usion o CH 4 dec eases wi h he inc easeonadso p ion.Asanexample, o 333K and o 1moleculepe uni cell, he sel -di usi i y coe ficien is D S CH4 ~6×10 –9ms –1. B eak h ough cu es o CO 2 and CH 4 we e compu a ionally p edic ed o alida e he expe imen s in a column bed leng h o 5.8 cm and a gas p essu e o 1.05–1.3 ba a 298 K (Fig. 8). The fi ed Langmui -F eundlich pa ame e s we e used o p edic his adso p- ion dynamic beha io . Helium was used as a ca ie gas o ensu e high gas eloci y. Th ee dis inc gas composi ions we e in es iga ed: 95% He and 5% o a CO₂/CH₄mix u e wi h a ios o 1.5:3.5, 2.5:2.5, and 3.5:1.5 ( / ). Ini ially, he column was ully sa u a ed wi h helium. The oid ac ion o he adso ben bed was 0.3. Simula ion da a a e in ag eemen wi h he expe imen al esul s (Fig. 8). The simula ed elu ion gas beha io emains quali a i ely and quan i a i ely con- sis en wi h expe imen al da a ac oss he ange o gas mix u e composi ions in es iga ed (3.75%, 2.5%, and 1.25% / CH 4 ). Al hough a di ec compa ison wi h da a epo ed in Fig. 2is no accu a e, he apid elu ion o CH 4 obse ed in his case may be a ibu ed o i s highe di usi i y a highe gas eloci y (0.015 m s−1) and o i s low gas pa ial p essu e (Fig. 7b). In conclusion, TAMOF-1, a s able, obus , homochi al me al- o ganic amewo k (BET specificsu acea ea=980±50m 2g–1), has shown selec i e adso p ion o CO 2 om CO 2 /CH 4 gass eams, o e ing a p omising pe o mance. Al hough compa ing b eak h ough da a ac oss s udies is o en challenging due o inconsis encies in expe i- men al de ails, he pe o mance pa ame e s exhibi ed by TAMOF-1 appea compe i i e o CO 2 /CH 4 sepa a ions when compa ed wi h s a e-o - he-a (Fig. 2d and Supplemen a y Table 22). As demon- s a ed by s uc u al and heo e ical da a, CO 2 up ake is domina ed by a a ie y o weak in e ac ions. A low co e age, he p e e ed binding si es a e he squa e plana Cu2+ cen e s. Once hese posi ions a e ull, he TAMOF-1 open channels accommoda e addi ional CO 2 molecules wi h weake in e ac ions wi h he dangling unc ional g oups (ca - boxyla e, iazole, imidazole) in he amewo k. Adso p ion iso he ms confi m heCO 2 p e e en ialup akewi h espec oCH 4 o N 2 in a la ge p essu e ange, esul ing in good sepa a ion pa ame e s o CO 2 /CH 4 gas mix u es. CO 2 /CH 4 mix u es can be e ec i ely sepa a ed wi h high selec i i y in a wide CO 2 /CH 4 concen a ion (1–75%), p essu e (1–6 ba ), and empe a u e ange (293–353 K). Elu ion beha io o CO 2 h ough he TAMOF-1 bed depends jus on he CO 2 mola flow and i is no a ec ed by he p esence o he o he gases in es iga ed (CH 4 and/ o N 2 ). TAMOF-1 e ains sepa a ion pe o mance, e en unde high humid condi ions (~72% RH) showing no deg ada ion and high cycling s a- bili y. Sepa a ion capabili y is obse ed also in ex emely acidic con- di ions (9000 ppm H 2 S) al hough a small pe o mance deg ada ion is obse ed. These da a demons a e TAMOF-1 obus ness, also in ha sh acidic en i onmen s al hough H 2 S emo al is p e e ed o inc ease ma e ial du abili y. Fig. 5 | Compu a ional modeling o gas-TAMOF-1 he modynamic p ope ies. a–cCalcula ed excess adso p ion iso he m (emp y ci cles) and expe imen al iso- he ms (solid poin s, see legend o mo e de ails) a 263.15, 303.15, 313, 323, and 333 K. E o s a e es ima ed om he s anda d de ia ion in he adso p ion p o- cesses. d,Selec i i ysi j=qi=qj,whe eiis CO 2 ,andjis CH 4 o N 2 . A icle h ps://doi.o g/10.1038/s41467-025-58426-w Na u e Communica ions | (2025) 16:3243 8 These p ope ies a e c ucial o indus ial-scale CO 2 cap u e applica ions. TheweakCO 2 –TAMOF-1 in e ac ions acili a e acile, low- ene gy egene a ion o he adso ben bed. These a ibu es, coupled wi h high selec i i y, ende TAMOF-1’spe o manceinbiogas upg ading supe io o o he epo ed ma e ials42. Consequen ly, TAMOF-1 appea s well-sui ed o in eg a ion in o p essu e/ acuum/ he mal swing adso p ion p ocesses. A con enien egene a ion p o- ocol, adap able o specific ime and ene gy cons ain s h ough ope a ionalpa ame e uning, is eadily achie able. Cu en e o s a e ocused on scaling and pelle iza ion o TAMOF-1, ad ancing he de elopmen o a CO 2 cap u e and pu ifica ion echnology based on his p omising ma e ial. Me hods Ma e ials All eagen s we e o comme cial g ade and used wi hou u he pu - ifica ion: L-his idine (≥98%, I is Bio ech GmbH), hionyl chlo ide (SOCl 2 ,≥99%, Sigma Ald ich), hyd azine monohyd a e (NH 2 NH 2 ·H 2 O, eagen g ade, 98%, Sigma Ald ich), sodium ca bona e anhyd ous (Na 2 CO 3 , ACS eagen , ≥99.5%, Sigma Ald ich). All sol en s we e o comme cial g ade and used wi hou u he pu ifica ion: HPLC-g ade e hanol, isop opanol, hexanes, ace oni ile and e -bu yl me hyl e he (VWR, Chem-Lab and Sigma Ald ich), N,N-dime hyl o mamide (pep- ide g ade, ≥99.9%, I is Bio ech GMBH) and die hyl e he (≥99%, VWR). Fig. 6 | Compu a ional modeling o CO 2 -TAMOF-1 in e ac ions. a His og am o he CO 2 -TAMOF-1 in e ac ion ene gy a se e al alues o p essu e. bRadial dis ibu ion unc ion o (in e molecula ) pai s, and a oms. Fig. 7 | Gas di usion p ope ies modeling. Sel -di usi i y coe ficien (D) o CO 2 (a)andCH 4 (b) a a ious empe a u es and adso p ion loadings. Solid lines a e fi ed o he da a as guiding e e ences o assis he eade (u.c.= uni cell). Fig. 8 | Expe imen al and compu a ional b eak h ough analysis. Calcula ed (lines) and expe imen al (squa e) b eak h oughcu es o CO 2 andCH 4 a 1 ba and 298.15 K o h ee composi ion a ios: 25 CO 2 :75CH 4 (a),50 CO 2 :50CH 4 (b), and 75 CO 2 :25CH 4 (c). Expe imen s pe o med wi h helium as he ca ie gas (95% / ). Expe imen s pe o med using 7.7 g o ac i a ed TAMOF-1. TAMOF-1 was ac i a ed unde acuum (10−1mba ) in bo h column op and bo om sides a 393 K o 15 h. A icle h ps://doi.o g/10.1038/s41467-025-58426-w Na u e Communica ions | (2025) 16:3243 9