E ec o Hyd ophobic C oss-Linke s in S ong Base Gel-Type Resins
on he Adso p ion Kine ics and Capaci y o Pe luo oalkyl
Subs ances
Flo ian Junge,
⊥
Fiona E. Ruckbeil,
⊥
Regina Gni ss, Raine Haag, Alejand o Lo en e, Fabio Lo enz,
Sunil P. M. Menache y, Aki S. Ruhl,*Alexande Spe lich, Ana Zida , and Ola Wagne *
Ci e This: ACS EST Wa e 2025, 5, 4435−4447
Read Online
ACCESS Me ics & Mo e A icle Recommenda ions *
sı Suppo ing In o ma ion
ABSTRACT: The pe sis ence and wa e mobili y o pe - and poly luo oalkyl
subs ances (PFAS) ha e led au ho i ies wo ldwide o lowe egula o y limi s o
p e en ad e se heal h e ec s. Remo al ia adso p ion on ac i a ed ca bon can
be ine icien due o he unspeci ic su ace in e ac ion, while ion exchange esins
wi h posi i e cha ges and hyd ophobic chains can o e as e kine ics and
imp o ed emo al. In he e, no el ca ionic esins we e syn hesized by c oss-
linking polye hylenimine, ollowed by me hyla ion. To ob ain c oss-linked
pa icles and in oduce hyd ophobic in e ac ing moie ies in one single syn he ic
s ep, alipha ic, luo ous, and silicone-based oligoe he s we e used as c oss-linke s.
These ca ionic adso ben s we e compa ed wi h wo s a e-o - he-a s ong base gel- ype ion exchange esins and g anula ac i a ed
ca bon in iso he m and kine ic s udies. The newly de eloped adso ben s showed signi ican ly as e emo als o all es ed long- and
sho -chain PFAS. The luo ous ca ionic adso ben achie ed equilib ium loadings ha we e compa able o hose o he s a e-o - he-
a adso ben s o all PFAS wi h i e o mo e pe luo ina ed ca bon a oms.
KEYWORDS: PFAS adso p ion, wa e ea men , ion exchange esins, luo ine− luo ine in e ac ions, siloxane, polye hylenimine
■INTRODUCTION
The excessi e, un es ic ed use o luo ina ed o ganic
molecules and luo opolyme s esul ed in a global con ami-
na ion o wa e , soil, and o ganisms wi h pe - and
poly luo oalkyl subs ances (PFAS). The high pe sis ence,
nega i e heal h e ec s, and mobili y in he wa e cycle o
many PFAS ha e led o se e al coun ies, o example, in
No h Ame ica and Eu ope, in oducing s ic limi s on he ng/
L le el in d inking, su ace, and g oundwa e .
1,2
These new
limi alues ep esen a majo challenge o wa e ea men .
The mos commonly used p ocess a p esen is adso p ion on
g anula ac i a ed ca bon (GAC) in ixed-bed il e s.
3,4
In
complex eal wa e s, adso p i e emo al by GAC can be
uneconomical, as compe ing adso p ion o o he wa e
cons i uen s p e en s selec i e emo al, and only low capaci ies
and ea ly b eak h oughs a e achie ed, especially o sho -chain
PFAS.
5,6
In addi ion, long e en ion imes wi h co espond-
ingly la ge sys ems a e equi ed due o slow kine ics. Fas e
kine ics and imp o ed emo al o sho -chain PFAS can be
achie ed h ough he use o ion exchange ma e ials (IX), which
a e inc easingly being used success ully o PFAS emo al on
an indus ial scale.
7
The hyd ophobic alkyl chains o he
polys y ene backbone o he IX i sel suppo he binding o
he hyd ophobic luo oalkyl ail g oups, while he posi i e
cha ge o hese esins binds he head g oups o he
p edominan ly nega i ely cha ged PFAS, mos no ably
pe luo oca boxylic acids (PFCA) and pe luo osul onic acids
(PFSA).
8
The mos highly posi i ely cha ged polyme s a e he
p o ona ed o me hyla ed de i a i es o b anched polye hylene
imine (bPEI), linea polye hylene imine (lPEI), and poly-
( inylamine) (PVAm) because hey possess he highes
con en o amine g oups o any polyme . The syn hesis o
PVAm and lPEI equi es he use o p o ec ing g oups o
p e en imine au ome iza ion o b anching, espec i ely.
9,10
Thus, bPEI emains he only one o he h ee polyme s ha is
accessible in la ge scale h ough a one-s ep ca ionic polyme -
iza ion. Hype b anched s uc u es, like hose ound in bPEI,
a e also known o hei gene ally supe io mul i alen ac i i y
compa ed o linea polyme s,
11
which made hem p omising
candida es o be es ed as PFAS adso ben s.
12−16
Howe e ,
p is ine bPEI is unsui able as an adso ben as i is a wa e -
soluble liquid.
10
Bi- o oligo unc ional elec ophiles a e
commonly used o c oss-link bPEI o ob ain insoluble c oss-
linked PEI (cPEI) pa icles.
17−23
Besides he high densi y o
Recei ed: Janua y 22, 2025
Re ised: May 28, 2025
Accep ed: June 2, 2025
Published: June 13, 2025
A iclepubs.acs.o g/es wa e
© 2025 The Au ho s. Published by
Ame ican Chemical Socie y 4435
h ps://doi.o g/10.1021/acses wa e .5c00094
ACS EST Wa e 2025, 5, 4435−4447
This a icle is licensed unde CC-BY 4.0
Downloaded ia 217.9.40.134 on Oc obe 6, 2025 a 13:07:48 (UTC).
See h ps://pubs.acs.o g/sha ingguidelines o op ions on how o legi ima ely sha e published a icles.
po en ial ca ionic ammonium g oups in cPEI, he bene i o
using cPEI as an adso ben is he possibili y o e icien ly
in oduce di e en hyd ophobic moie ies du ing he c oss-
linking s ep. Po en ial c oss-linke s include di- o polyalde-
hydes,
18,22,24−29
epichlo ohyd in,
23
diisocyana es,
21,30−33
diha-
lides,
17,19,24,34
diac yla es,
35
di-N-hyd oxysuccinimide es e s
36
and di- o iepoxides.
12,20,22,35,37−39
Howe e , i is bene icial
o e ain a high amoun o amine g oups du ing he c oss-
linking o bPEI since only amines can be p o ona ed o
alkyla ed o posi i ely cha ged ammonium g oups subse-
quen ly. Diepoxides a e he e o e he only ype o po en ial
c oss-linke s ha p ese e amines as unc ional g oups wi hou
equi ing he addi ion o any base o educing agen s and
wi hou he o ma ion o hyd olysis-p one linke s. Fu he ,
many diepoxides, especially diglycidyl e he s, a e comme cially
a ailable o can be con enien ly p epa ed om comme cial
p ecu so s. In con as o he indus ially used bu yla ed
ammonium g oups,
40
we chose o c ea e he qua e na y
ammonium g oups on he PEI-based ma e ials h ough
me hyla ion due o s e ic epulsion. No ewo hy, wo ecen
s udies obse ed ha ei he ime hyla ed luo ina ed
41
o
dodecyldime hylammonium-based adso ben s
42
end o ha e
be e PFAS adso p ion han hei ibu yla ed de i a i es.
Highly luo ina ed alkyl chains in e ac wi h each o he in
he o m o speci ic luo ine- luo ine in e ac ions.
43
Hence,
modi ica ion o adso ben s wi h luo oalkyl g oups, p e e ably
in he polyme backbone han in he side chain,
44
is a common
and la gely success ul design p inciple o boos he adso p ion
o PFAS.
45,46
Ne e heless, he e ec i eness o luo ophilic
modi ica ion o PFAS adso ben s is s ill subjec o scien i ic
deba e
47,48
and PFAS emo al echnologies should ideally
a oid in oducing addi ional PFAS- unc ionalized ma e ials
in o he en i onmen . Thus, we de eloped no only a luo ous
low-molecula -weigh pe luo opolye he (PFPE)-cPEI bu
also wo mo e sus ainable, luo ine- ee cPEI adso ben s wi h
alipha ic poly(e hylene glycol) (PEG) and silicone-based
poly(dime hylsiloxane) (PDMS) c oss-linke s o compa ison.
This is, o he bes o ou knowledge, he i s IX wi h
dime hylsiloxane unc ional g oups applied o PFAS emo al.
The e ha e been ew examples o PEI/poly(dialkyl siloxane)
composi es being used o o he applica ions.
49−52
PDMS
sha es some p ope ies wi h luo opolyme s, such as low
su ace ene gies, wa e - epellency, and good he mal s abil-
i y.
53,54
He ein, we wan ed o in es iga e whe he PDMS
p o ides a sui able hyd ophobic g oup o he selec i e
emo al o PFAS. Molecula dynamic calcula ions om Ke e
al.
55
sugges ed hyd ophobic-d i en adso p ion o PFSA on o
ino ganic siloxane su aces (kaolini e), indica ing a po en ial
iabili y o his concep .
As he indus y s a ed eplacing long-chain PFAS wi h
sho e ones, a g owing con amina ion wi h sho -chain
PFAS,
56−59
which a e by de ini ion o he O ganisa ion o
Economic Co-ope a ion and De elopmen (OECD)
60
all
PFCA and PFSA wi h less han six luo ina ed ca bon a oms
a ises. Consequen ly, he e icien emo al o sho -chain PFAS
becomes a challenge o inc easing impo ance. The e o e, we
chose o ocus on ou sho -chain PFAS ( i luo ome hane-
sul onic acid, TFMSA; pe luo open anesul onic acid, PFPeS;
pe luo obu anoic acid, PFBA; and pe luo ohexanoic acid,
PFHxA) as analy es. We in es iga ed he adso p ion iso he ms
and he kine ics o hei adso p ion on o PFPE, PDMS, and
PEG c oss-linked PEI and benchma ked hem agains he
pe o mance o a comme cialized, s a e-o - he-a IX o PFAS
emo al.
■METHODS
Ma e ials. Sodium hyd ide (60 w % dispe sion in mine al
oil), 15-c own-5 (98%), bis(3-(oxi an-2-ylme hoxy)p opyl)-
e mina ed polydime hylsiloxane (PDMS c oss-linke , p oduc
numbe : 480282, Mn= 800 g/mol), sodium bica bona e
(NaHCO3, 99.5%), sodium chlo ide o he model solu ion
(NaCl, 99%), and polye hylene glycol diglycidyl e he
(PEGDE, p oduc numbe : 475696, Mn= 500 g/mol) we e
pu chased om Sigma-Ald ich (Tau ki chen, Ge many).
B anched polye hylenimine (bPEI, ca alogue numbe : 19850,
Mw= 10 000 g/mol) was pu chased om PolySciences
(Hi schbe g an de Be gs asse, Ge many). Dime hyl sul a e
(99%), 18-c own-6 (99%), m-CPBA (70% - 75% in wa e ),
po assium iodide (99%), d y THF (99.5%), sodium chlo ide
o he ion-exchange (99.5%), aqueous ammonia solu ion (25
w%), THF (99.6%), me hanol (>99%), and allyl b omide
(98%) we e pu chased om Ac os O ganics/The mo Fishe
Scien i ic (Schwe e, Ge many). 2,2’-((Oxybis(1,1,2,2- e a-
luo oe hane-2,1-diyl))bis(oxy))bis(2,2-di luo oe han-1-ol)
( luo ina ed e ae hylene glycol, 98%) was pu chased om
abc (Ka ls uhe, Ge many). Po assium ca bona e (K2CO3,
99%) and sodium sul a e (Na2SO4, 99%) we e pu chased om
Ca l Ro h (Ka ls uhe, Ge many). DCM (≈100%) and p opan-
2-ol (Reag. Ph. Eu .) we e pu chased om VWR (Da ms ad ,
Ge many). A gon (Alphagaz, 99.999%) was pu chased om
Ai Liquide (Dusseldo , Ge many). Deu e a ed chlo o o m
(99.8%) was pu chased om Deu e o (Kas ellaun, Ge many).
Sodium i luo ome hanesul ona e (95%) was pu chased om
To on o Resea ch Chemicals (To on o, Canada). All o he
PFAS s anda ds we e pu chased om Camp o Scien i ic
(Be lin, Ge many): PFBA (99.8%), PFHxA (99%), pe luo -
ohep anoic acid (PFHpA, 96%), pe luo ooc anoic acid
(PFOA, 95%), pe luo ononanoic acid (PFNA, 95%),
pe luo odecanoic acid (PFDA, 99.9%), pe luo obu anesul-
onic acid (PFBS, 97%), PFPeS (98%), sodium pe luo ohex-
anesul ona e (PFHxS, 98%), and po assium pe luo ooc ane-
sul ona e (PFOS, 99.8%). All compounds we e used as
ecei ed.
Comme cialized Adso ben s. The bi uminous GAC
Hyd a in 30N
61
was pu chased om Donau Ca bon
(F ank u , Ge many). The wo polys y ene s ong base
anion exchange esins PFA694E
62
and Lewa i TP108
63
we e
pu chased om Pu oli e (Philadelphia, Uni ed S a es) and
Lanxess (Cologne, Ge many).
Syn he ic P ocedu es. Fluo ina ed Te ae hylene Glycol
Diglycidyl E he (PFPE C oss-Linke ). Sodium hyd ide (NaH,
60 w % in mine al oil, 2.93 g, 3.0 equi ) was added o a
Schlenk lask wi h d y e ahyd o u an (THF, 80 mL).
Fluo ina ed e ae hylene glycol (9.99 g, 1.0 equi ) was
added slowly (in ense oam de elopmen !) o he NaH/THF
mix u e unde ice cooling. A e he addi ion o allyl b omide
(6.32 mL, 3.0 equi ), he mix u e was hea ed o e lux o 24 h.
The eac ion was e mina ed by he addi ion o p opan-2-ol (6
mL) and wa e (5 mL) a 0 °C. The o ganic sol en was
e apo a ed, and he esidue was dilu ed wi h wa e (50 mL)
and dichlo ome hane (DCM, 100 mL). The phases we e
mixed ho oughly. The o ganic laye was sepa a ed, washed
wi h wa e (2 ×50 mL), and m-CPBA (pu i y: 70−75%, wi h
emaining wa e and me a-chlo obenzoic acid, 23.2 g, 4.0
equi ) was added. The eac ion mix u e was s i ed o 4 days
ACS ES&T Wa e pubs.acs.o g/es wa e A icle
h ps://doi.o g/10.1021/acses wa e .5c00094
ACS EST Wa e 2025, 5, 4435−4447
4436
a oom empe a u e. The eac ion mix u e was washed wi h
sa u a ed aqueous sodium bica bona e solu ion (4 ×100 mL)
and wa e (2 ×100 mL), ollowed by d ying wi h Na2SO4,
il a ion, and e apo a ion o sol en . The c ude p oduc was
pu i ied by column ch oma og aphy (silica, hexane/e hyl
ace a e 2/1 →0/1). The PFPE c oss-linke (9.83 g, 77%)
was ob ained as a sligh ly yellow liquid. 1H NMR (400 MHz,
CDCl3): δ= 3.97−3.85 (m, 6H), 3.52 (dd, J= 11.5 Hz, 6.0
Hz, 2H), 3.16−3.13 (m, 2H), 2.79 ( , J= 4.5 Hz, 2H), 2.61
(dd, J= 5.0 Hz, 2.6 Hz, 2H) ppm. 13C NMR (101 MHz,
CDCl3): δ= 122.4 ( ), 116.9−112.3 (m), 73.0, 70.0, 49.5, 39.8
ppm. 19F NMR (376 MHz, CDCl3): δ=−77.9 (4F), −88.6
(4F), −88.9 (4F) ppm. ESI-ToF-MS m/z= 545.0416,
C14H14F12O7Na1+: calc. 545.0446. elemen al analysis (%) C
32.59 H 2.73 N 0.14 S 0.00, calc. C 32.20, H 2.70, N 0, S 0.
Gene al P ocedu e o he C oss-Linking o PEI. bPEI was
dissol ed in THF (70 mL), and an equal mass o he espec i e
c oss-linke was added. The solu ion was e luxed o 24 h
(PFPE and PEG c oss-linke ), while he eac ion ime was
ex ended o 42 h in he case o he PDMS c oss-linke because
no solidi ica ion occu ed a e 24 h. The c oss-linked polyme
was sepa a ed and washed wi h me hanol (2 ×80 mL; PDMS-
c oss-linked polyme : 2 ×40 mL) in a cen i uge. The
unqua e nized PEG-c oss-linked polyme was u he washed
wi h DCM (2 ×80 mL). The unqua e nized PFPE- and
PDMS-c oss-linked polyme s could no be success ully
sepa a ed om he DCM ia cen i uga ion o il a ion. The
c oss-linked polyme s we e hen d ied and g ound in liquid
ni ogen.
Unqua e nized PEG-C oss-Linked PEI (uPEG-cPEI). The
eac ion was pe o med acco ding o he gene al p ocedu e
desc ibed abo e wi h bPEI (11.1 g) and PEGDE (9.69 mL).
Unqua e nized PEG-c oss-linked PEI (uPEG-cPEI) (21.1 g)
was ob ained as an elas ic colo less solid.
Unqua e nized PDMS-C oss-Linked PEI (uPDMS-cPEI).
The eac ion was pe o med acco ding o he gene al
p ocedu e desc ibed abo e wi h bPEI (10.9 g) and he
comme cial PDMS c oss-linke (11.0 mL). Unqua e nized
PDMS-c oss-linked PEI (uPDMS-cPEI) (20.7 g) was ob ained
as an elas ic colo less solid.
Unqua e nized PFPE-C oss-Linked PEI (uPFPE-cPEI). The
eac ion was pe o med acco ding o he gene al p ocedu e
abo e wi h bPEI (8.99 g) and PFPE c oss-linke (8.99 g).
Unqua e nized PFPE-c oss-linked PEI (uPFPE-cPEI) (20.7 g)
was ob ained as an elas ic colo less solid.
Me hyla ion o Unqua e nized, C oss-Linked PEIs. The
espec i e unqua e nized, cPEI (10.0 g), and K2CO3(8.00 g,
57.9 mmol, 0.5 equi ega ding he e hylenimine epea ing uni
unde assump ion o 50 w % PEI con en ) we e added o
dime hyl sul a e (100 mL, 1.05 mol, 9.1 equi ) in a lame-d ied
Schlenk lask. The eac ion was s i ed o 3 days a oom
empe a u e be o e i was ca e ully s opped ( e a ded unaway
eac ion possible!) by he po ion wise addi ion o 25%
aqueous ammonia solu ion (100 mL) a 0 °C. The polyme
was il e ed o and washed wi h wa e (100 mL). Al e na i ely,
cen i uga ion migh be used o sepa a ion, bu decan a ion o
wa e is challenging o he non luo ina ed polyme s. The
polyme s we e d ied and suspended oge he wi h K2CO3
(8.00 g) unde a gon in dime hyl sul a e (100 mL). The
eac ion mix u e was allowed o s i o ano he 4 days a oom
empe a u e. The eac ion was quenched and washed wi h
wa e , as desc ibed p e iously. Addi ionally, he polyme s we e
washed wi h THF (2 ×50 mL) and hen s i ed in sa u a ed
b ine (450 mL) o 1 day a oom empe a u e. The polyme s
we e il e ed o , washed wi h wa e (3 ×50 mL), and d ied a
100 °C. G inding in liquid ni ogen ga e PFPE-cPEI (12.9 g)
as yellowish powde , PDMS-cPEI (9.05 g) as whi e powde ,
and PEG-cPEI (12.2 g) as b own coa se pa icles. All samples
con ained a ew elas ic anspa en chunks ha could no be
success ully g inded. The pa icles we e sie ed di ec ly a e
d ying in high acuum as d y powde ( he pa icles adso b
mois u e wi hin minu es) wi h wo analysis sie es (mesh sizes:
710 and 63 μm; pa icles ou side o his ange we e cu o )
om Re sch (Haan, Ge many) be o e op ical mic oscopy,
PFAS adso p ion, BET and SAXS analysis.
Op ical Mic oscopy. Op ical mic oscopy images we e aken
in Milli-Q wa e wi h an Axio Obse e .Z1 om Ca l Zeiss
(Jena, Ge many) equipped wi h an objec i e EC Plan-Neo lua
5x/0.16 M27 (magni ica ion 5 x) o LD Ach oplan 20x/0.40
Ko Ph 2 (magni ica ion 20 x) and an AxioCamMRm3
came a wi h a 0.63x adap e . Images o 1388 ×1040 pixels
we e eco ded. The so wa e o cap u ing he images was he
ZEN 2012 (blue edi ion) e sion 1.1.2.0 om Ca l Zeiss, and
o pa icle size analysis, he Fiji
64
e sion o ImageJ 1.54k was
used. The pa icle con ou s o each pa icle on one (PFPE-
cPEI and PEG-cPEI), h ee (PDMS-cPEI), o i e images
(TP108) (Figu es S18−S21) we e measu ed manually wi h he
polygon ool due o he he e ogeneous backg ound and
di icul ies wi h he au oma ic de ec ion o he pa icles.
Pa icles ha we e as ly ou o he ocus plane, oo small o be
iden i ied clea ly, pa ially ou side he pic u ed a ea, o
o e lapped ex ensi ely wi h o he pa icles we e no analyzed.
In some ins ances, i was ha d o di e en ia e be ween an
agglome a e and a single pa icle due o he complex pa icle
shapes. These we e coun ed as one pa icle. Ei he he Fe e
diame e (Figu es S22a−S25a), he minFe e diame e
(Figu es S22b−S25b) o he p ojec ed a ea diame e d
calcula ed om he a ea o he d awn polygons A educ i ely
assuming sphe ical pa icles [d= 2 * √(A/π)] (Figu es 2b−d
and S22c) we e epo ed in numbe -weigh ed dis ibu ion
plo s.
Adso p ion Expe imen s. Model Solu ion P epa a ion
and Sample P ocessing. All adso p ion expe imen s we e
ca ied ou in simple model solu ions. Fo his pu pose,
ul apu e wa e (Milli-Q) was mixed wi h he bu e and
elec oly es (0.05 mM NaHCO3and 0.02 mM NaCl). PFAS
we e added o he model solu ion om high concen a ed
s ock solu ions (10 mg/L) o achie e he desi ed s a ing
concen a ion o 10 μg/L. The pH o he model solu ion was
adjus ed o a neu al pH alue (7 ±0.2) by adding 0.1 M
NaOH and HCl. All adso p ion expe imen s we e conduc ed
a oom empe a u e (app oxima ely 22 °C).
A e he espec i e con ac ime and p io o analysis, all
samples we e il e ed using p e insed 0.45 μm memb ane il e s
made o egene a ed cellulose ace a e (Mache ey-Nagel,
Du en, Ge many). Samples we e s o ed a 4 °C in 5 mL
polyp opylene ials pu chased om Th. Geye (Renningen,
Ge many).
Sc eening Expe imen . The h ee cPEI adso ben s we e
compa ed wi h Hyd a in 30N GAC and PFA694E IX in an
ini ial sc eening expe imen . Fo his pu pose, 40 mg o
adso ben (d y weigh ) was added o 1 L o he model solu ion
and con amina ed wi h a PFAS mix con aining PFHpA, PFOA,
PFNA, PFDA, PFBS, PFHxS, and PFOS. Samples we e
agi a ed in 1 L glass bo les on a ho izon al shake , which
we e sealed wi h polyp opylene sc ew caps. All ba ches we e
ACS ES&T Wa e pubs.acs.o g/es wa e A icle
h ps://doi.o g/10.1021/acses wa e .5c00094
ACS EST Wa e 2025, 5, 4435−4447
4437
p epa ed in duplica e, and he i s duplica e was agi a ed o
30 min (p e-equilib ium) and he second o 10 days
(equilib ium).
The pe cen age emo al was de e mined by eq 1 a e he
espec i e con ac ime by compa ing he emaining PFAS
concen a ion in he model solu ion (c) wi h he concen a ion
in a e e ence sample wi hou adso ben (c0), which was
o he wise ea ed in he same way.
= ·
i
k
j
j
j
j
j
y
{
z
z
z
z
z
c
c
emo al 100% 1 ( )
( )
0
(1)
Single poin adso p ion coe icien s Kdwe e calcula ed using
eq 2 o allow be e compa abili y wi hin he da a se s o his
s udy and o he s udies.
=Kq
c
( )
( )
d
(2)
He eby, pa ame e qdesc ibes he PFAS loading (solid-phase
concen a ion) on he adso ben a he con ac ime .
Kine ic Expe imen . In o de o in es iga e he di e en
kine ics o he h ee cPEI adso ben s in mo e de ail, a kine ic
expe imen was ca ied ou o one selec ed PFAS (PFHxA),
and he adso p ion a es we e compa ed wi h hose o he
TP108 IX. Fo his model solu ion, 10 μg/L PFHxA was
p epa ed in ou 5 L ba ches, om which 10 mL was aken o
he de e mina ion o he espec i e e e ence concen a ion c0.
Subsequen ly, 200 mg o adso ben (d y weigh ) was soaked in
10 mL ul apu e wa e (Milli-Q) o 12 h and added o he
emaining 4990 mL model solu ion, esul ing in an adso ben
concen a ion o 40 mg/L. P io soaking was conduc ed o
ensu e ha he kine ics o he adso ben s we e no limi ed by
he p e ious d ying. The dilu ion e ec caused by he addi ion
o 10 mL ul apu e wa e was negligible (0.2%).
E e y ba ch was con inuously s i ed wi h glass-coa ed
magne ic s i e s a 1100 pm, and 6 mL samples we e aken
a e 0.25, 0.5, 1, 4, 7, 24, and 48 h. Ba ch olume change
caused by he sampling was negligible (<0.9%), he bo les
we e sealed wi h polyp opylene sc ew caps du ing he
expe imen and only opened o sampling.
Reac ion kine ic a e cons an s k1and k2we e i ed using
he pseudo- i s o de a e exp ession (PFO, eq 3) and pseudo-
second o de a e exp ession (PSO, eq 5) in i s espec i e
linea ized o m (eqs 4 and 6).
=q q( ) (1 e )
e
k
1
(3)
=
i
k
j
j
j
j
j
j
y
{
z
z
z
z
z
z
q q
qk ln ( )
e
e
1
(4)
=
+
q q k
q k
( ) 1
e
2
2
e2
(5)
= +
q q k q
( )
1 1
e2e
2
(6)
The de i a ions o he PFO and PSO equa ions and a
comp ehensi e heo e ical backg ound can be ound in he
publica ions o Re ellame e al.
65
and Ho and McKay.
66
Fo
he PFPE-cPEI and he TP108 IX, su ace di usion coe icien s
Dswe e addi ionally de e mined using he eely a ailable
modeling so wa e FAST (h ps://www. as -so wa e.de/).
67
The homogeneous su ace di usion model (HSDM) was
selec ed as he modeling app oach, and he assump ion was
made ha he ilm di usion is negligible due o he high
s i ing eloci y and ha in apa icle di usion is he a e-
limi ing s ep. The mass anspo in he adso ben pa icle is
he eby desc ibed by Fick’s second law, whe e is he adial
coo dina e o he adso ben (eq 7):
= +
i
k
j
j
j
j
j
y
{
z
z
z
z
z
q
Dq
q
2
s
2
2
(7)
In o de o examine he in luence o neglec ing ilm di usion
on he e alua ion o he kine ics, he ilm di usion coe icien
kLwas also i ed in a second a ian o he HSDM. Fo his,
he assump ion was made ha he kine ics a e domina ed by
ilm di usion wi hin he i s 15 min o con ac ime. The ilm
di usion coe icien kLwas hen de e mined using eq 8 as
desc ibed by Wo ch e al.
68
whe e amis he o al adso ben
su ace a ea ela ed o he adso ben mass (mA) a ailable in he
ba ch olume (VL):
=
i
k
j
j
j
j
j
y
{
z
z
z
z
z
c
c
m
V
k a ln
0
A
L
L m
(8)
Subsequen ly, he di usion coe icien s we e i ed as in he
i s a ian bu wi h an addi ional conside a ion o ilm
di usion. F eundlich iso he m da a om ou s udy was used o
desc ibe he adso p ion equilib ium using he HSDM. A
de ailed heo e ical backg ound o he HSDM is desc ibed by
Wo ch e al.
68
Iso he m Expe imen . Ba ches o 250 mL model solu ion
con amina ed wi h 10 μg/L TFMSA, PFPeS, PFBA, o
PFHxA, espec i ely, we e p epa ed in 500 mL glass bo les
con aining a ying adso ben concen a ions (0, 10, 20, 30, 40,
50, 60, 70, and 80 mg/L). Samples we e agi a ed o 48 h on a
ho izon al shake (equilib ium) and sealed wi h polyp opylene
sc ew caps 48 h was selec ed as a su icien equilib a ion ime
based on ou own p elimina y es s and is also in line wi h
o he s udies
69,70
ha we e ca ied ou in compa able
concen a ion anges and unde compa able condi ions.
The iso he m model pa ame e s n,KF,KLand qmax we e
i ed using he F eundlich (eq 9) and Langmui iso he m (eq
11) in i s espec i e linea ized o m (eqs 10 and 12).
=q c K
n
F
(9)
= +q c n Klog( ) log( ) log( )
F
(10)
=
+
q q K c
K c1
max
L
L
(11)
= +
q c K q q
1 1 1 1
Lmax max
(12)
PFAS Analysis. TFMSA was analyzed by a high-pe o mance
liquid ch oma og aphy coupled wi h andem mass spec om-
e y (HPLC-MS/MS) me hod desc ibed in de ail by Zeeshan
e al.
71
wi h a limi o quan i ica ion (LOQ) o 1 ng/L.
PFHpA, PFOA, PFNA, PFDA, PFBS, PFHxS, and PFOS we e
analyzed ollowing a HPLC-MS/MS me hod desc ibed in
de ail by Zie zschmann e al.,
72
wi h a espec i e LOQ o 1 ng/
L (PFHpA, PFOA, PFBS, and PFHxS) o 2 ng/L (PFNA,
PFDA, and PFOS).
ACS ES&T Wa e pubs.acs.o g/es wa e A icle
h ps://doi.o g/10.1021/acses wa e .5c00094
ACS EST Wa e 2025, 5, 4435−4447
4438
The concen a ions o PFBA, PFHxA, and PFPeS a e he
adso p ion expe imen s we e de e mined using andem mass
spec ome ic (Agilen 1290 in ini y 2 Se ies HPLC coupled
wi h SCIEX QTRAP 6500+ iple quad upole mass
spec ome e ; LC-MS/MS) analysis. The ch oma og aphic
sepa a ion o hese compounds was conduc ed on a Wa e s
BEH C18 (100 mm ×2.1 mm ×1.7 μm) e e se phase
column. A g adien elu ion o 0.1% ace ic acid and pu e
me hanol ( low a e: 0.25 mL/min; injec ion olume 20 μL)
was used o his pu pose. The compounds we e analyzed in
nega i e elec osp ay ioniza ion (ESI) mode by moni o ing he
ollowing ion ansi ions: PFBA (212.9 →169.0), PFHxA
(313.0 →269.0), and linea PFPeS (349.0 →80.0). The
sou ce pa ame e s o LC-MS/MS a e Cu ain gas; 35 psi, Gas
1; 55 psi, Gas 2; 40 psi, Tempe a u e; 450 °C, Ion sp ay
ol age; −4000 V, and En ance po en ial; −10 V. The
compound-speci ic pa ame e s used o LC-MS/MS analysis,
such as he declus e ing po en ial, collision ene gy, and
collision cell exi po en ial, we e aken di ec ly om li e a u e
( o PFBA and PFHxA
73
and o linea PFPeS
74
). The
espec i e LOQ alues we e 50 ng/L (PFBA and PFPeS) and
10 ng/L (PFHxA).
■RESULTS AND DISCUSSION
The p epa a ion o he h ee cPEI adso ben s ollowed h ee
syn he ic s eps: c oss-linking o bPEI (Mw= 10,000 g/mol),
75
me hyla ion wi h dime hyl sul a e, and ion exchange o
chlo ide (Scheme S1). Diglycidyl e he s o he oligoe he s
PFPE (M = 522 g/mol), PEG (Mn= 500 g/mol),
76
and PDMS
(Mn= 800 g/mol)
77
wi h simila molecula weigh s we e
selec ed as c oss-linke s. The c oss-linking wi h he PDMS
c oss-linke was no iceably slowe han wi h he o he wo
c oss-linke s, and i ook almos wice as long o ob ain cPEI
wi h a simila solid ex u e. The success ul syn hesis o he inal
cPEI adso ben s, PEG-cPEI, PDMS-cPEI and PFPE-cPEI
(Cha 1) was con i med by elemen al analysis (EA) (Table
S2), me cu ime ic i a ion o chlo ide ions (Table S1),
in a ed (IR) spec oscopy (Figu es 1a and S5) and X- ay
pho on (XP) su ey spec a (Figu e 1b and Table S3).
The IR spec a o all cPEI adso ben s show he expec ed
s e ching ib a ion bands o e he g oups (∼1100 cm−1) and
wo bands ha a e commonly a ibu ed o qua e na y
ammonium g oups (∼960 and ∼920 cm−1).
24,34,78,79
The IR
spec um o PFPE-cPEI u he shows he b oad, cha ac e is ic
C−F s e ching ib a ion band be ween 1240 and 1100 cm−1.
Th ee ypical bands o dime hyl siloxanes a e de ec ed a
1259, 1016, and 795 cm−1in he IR spec um o PDMS-
cPEI.
80,81
The quan i ica ion o he XP su ey spec a (Table
S3) shows chlo ine o ni ogen a ios o 0.63:1 o 0.87:1,
indica ing a high deg ee o me hyla ion despi e he high cha ge
epulsion o igh ly packed ammonium g oups in PEI.
9
The
cPEI adso ben s con ain 4.4−4.9 mol/kg o chlo ide ions ha
a e accessible o ion exchange (Table 1), co esponding o a
sha e o qua e na y ni ogen g oups o oughly 70−75%
(app oxima ely 6.6 mol/kg chlo ide would be expec ed i all
ni ogen a oms a e qua e na y ammonium g oups wi h
chlo ide coun e ion). The chlo ide con en o TP108 IX is
less han a hi d o ha and also a bi lowe han p e iously
epo ed chlo ide con en s (anion exchange capaci ies) o
simila esins ha we e de e mined by di e en echniques.
40
The c oss-linke con en s o he cPEI adso ben s we e
Cha 1. Chemical S uc u es o PEG-cPEI, PDMS-cPEI,
and PFPE-cPEI
Figu e 1. (a) IR abso p ion spec a o he cPEI adso ben s wi h
ma ked bands o cha ac e is ic ib a ions ( ull spec um in Figu e S5).
(b) XP su ey spec a o he cPEI adso ben s.
Table 1. Con en s o C oss-Linke in he D ied cPEI
Adso ben s Calcula ed om he Ni ogen Con en
De e mined by Elemen al Analysis (EA) o om he
Oxygen, Silicon, o Fluo ine Con en
a
c oss-linke con en chlo ide
con en
sample
acco ding o
ni ogen
(EA) [w %]
acco ding o
oxygen
(XPS)
[w %]
acco ding o
Si o F
(XPS)
[w %]
acco ding o
i a ion
[mol/kg]
TP108 / / / 1.26 ±0.04
PEG-cPEI 40 ±3 41.2 ±0.9 / 4.36 ±0.09
PDMS-cPEI 30 ±2 49.8 ±0.9 50.5 ±1.1 4.94 ±0.11
PFPE-cPEI 34 ±1 46.1 ±0.9 44.6 ±0.7 4.62 ±0.06
a
The chlo ide con en was de e mined by me cu ime ic i a ion.
ACS ES&T Wa e pubs.acs.o g/es wa e A icle
h ps://doi.o g/10.1021/acses wa e .5c00094
ACS EST Wa e 2025, 5, 4435−4447
4439
calcula ed om he dec ease o he ni ogen con en s as
quan i ied by EA and he ise o he oxygen, silicon, o luo ine
con en in XP su ey spec a. EA indica es c oss-linke
con en s o 30−40 w %, and XP su ey spec a indica e
c oss-linke con en s be ween 41 and 51 w %. This p o es ha
he cPEI adso ben s consis o a compa able amoun o ionic as
well as c oss-linking g oups, allowing us o es ablish
co ela ions be ween he PFAS adso p ion and he ype o
in oduced c oss-linke g oup.
The cPEI adso ben s exhibi a wide ange o mo phologies,
as can be seen in elec on mic oscopic images (Figu es S29−
S31). The d y pa icles ha we e e ained be ween wo
analy ical sie es we e collec ed (mesh sizes: 63 and 710 μm).
This s ep was added o homogenize he adso ben pa icles in
his size ange, which was complica ed due o hyg oscopically
induced agglome a ion. Op ical mic oscopic images we e used
o e alua e he pa icle size dis ibu ions o he cPEI
adso ben s in wa e a e sie ing (Figu es 2 and S18−S25).
Thei median p ojec ed a ea diame e s we e 52 μm (PDMS-
cPEI), 44 μm (PFPE-cPEI), and 26 μm (PEG-cPEI), and hei
median Fe e diame e s we e 77, 62 and 37 μm, espec i ely.
While he cPEI adso ben s showed simila pa icle size
dis ibu ions, he TP108 IX consis s o much la ge pa icles
(app oxima ely 700 μm de e mined a e sie ing as desc ibed
abo e). The cPEI adso ben s we e also checked o po osi y
due o p o ound swelling o he cPEI adso ben s in wa e .
Howe e , nei he SEM (Figu es S29−S31, SEM o TP108:
Figu e S32) o B unaue , Emme , Telle (BET) su ace a ea
de e mina ion (PEG-cPEI: (0.072 ±0.002) m2/g, PDMS-
cPEI: (0.113 ±0.002) m2/g, PFPE-cPEI: (0.076 ±0.002) m2/
g) e ealed any ypes o po es. Simila ly, small-angle X- ay
sca e ing (SAXS) (Figu e S1) did no yield a spec um ha
would be expec ed o highly po ous ma e ials. Only PDMS-
cPEI con ains sca e e s in he small nanome e ange (Figu e
S2). Ne e heless, gi en he ex emely small su ace a ea, i is
unlikely ha his sca e ing o igina es om po es. The
cha ac e iza ion o he adso ben s was comple ed wi h a
s udy o hei he mal s abili y using he mog a ime y. The
h ee cPEI adso ben s a e s able up o 200 °C in a ni ogen
a mosphe e, so hey a e he mally less s able han p is ine bPEI
(s a o deg ada ion a ound 290 °C) bu mo e s able han
TP108 IX (s a o deg ada ion a ound 160 °C) (Figu e S3).
Sc eening expe imen s e ealed ha all adso ben s we e able
o emo e he PFAS in es iga ed, albei wi h di e en kine ics
(Figu e 3). While he h ee cPEI adso ben s we e able o
achie e mo e han 70% o he maximum loading o all PFAS
a e jus 30 min, he IX and especially he GAC showed
signi ican ly slowe kine ics. The slow kine ics o GAC can be
a ibu ed o slow di usion along he su ace and wi hin he
po es o he highly po ous adso ben .
82
In he case o he comme cial PFA694E IX, only sligh
di e ences in he loadings o di e en PFAS could be
de e mined a equilib ium due o he high emo al obse ed
o all PFAS conside ed; i p o ed o be he mos e ec i e
adso ben , pa icula ly o he sho -chain compound PFBS.
PFA694E was used solely o his sc eening as a e e ence
because o a change o supplie du ing he cou se o he s udy.
Howe e , ou own s udies indica e ha only mino di e ences
be ween he PFAS-speci ic esins o he a ious manu ac u e s
a e o be expec ed.
83
Unexpec edly, GAC showed only sligh
di e ences in he emo al o PFAS o di e en chain leng hs.
O he s udies ha e epo ed a s onge in luence o he PFAS
chain leng h on he emo abili y by GAC.
7,84
All cPEI adso ben s we e able o emo e long-chain PFAS
be e han PFAS wi h sho e chains (e.g., Kd,PFHpA = 0.01 L/
kg and Kd,PFDA = 8.60 L/kg o he PEG-cPEI). Wi h he same
numbe o pe luo ina ed ca bon a oms, i was also obse ed
ha PFSA could be emo ed be e han PFCA wi h all
adso ben s (e.g., Kd,PFNA = 0.17 L/kg and Kd,PFOS = 23.23 L/kg
o he PEG-cPEI).
Among he cPEI adso ben s, PFPE-cPEI achie ed he
highes emo als. Wi h a compa able anion exchange capaci y
(see Table 1), he di e ences he e can be a ibu ed o he
di e en c oss-linke s. The hyd ophobic in e ac ions be ween
he pe luo ina ed c oss-linke s and he pe luo ina ed ca bon
ail o he PFAS as well as elec os a ic a ac ions be ween he
dep o ona ed acid g oup and he qua e nized ammonium
g oup on he PFPE-cPEI can be iden i ied as go e ning
adso p ion mechanisms.
85
The PDMS-cPEI showed emo als compa able o ha o
PFPE-cPEI a e 30 min, bu he emo als we e dec eased a e
10 days, which was u he in es iga ed in addi ional kine ic
s udies.
The kine ics o cPEI adso ben s and selec ed IX (TP108)
we e in es iga ed in mo e de ail (Figu e 4). The i ing o he
wo eac ion kine ic models, PFO and PSO, showed ha he
PSO p o ided a be e ma hema ical app oxima ion o he
kine ic cu es o all ou adso ben s (compa e oo -mean-
squa e e o (RMSE) in Table 2). Howe e , simpli ied
conclusions abou he adso p ion mechanism and a e
con olling s ep should no be d awn om his model, as
discussed in mo e de ail in se e al publica ions.
68,86,87
The
poo e ma hema ical i o he PFO could also be linked o he
expe imen design: he cPEI adso ben s al eady each hei
equilib ium loading wi hin he i s ew hou s o he
expe imen ; he le -hand side o eq 4 is no longe de ined
when q( ) app oaches qe. A highe sampling esolu ion be o e
eaching equilib ium could he e o e in luence he e alua ion
o he models and a o he PFO model. Ne e heless, he
PSO allows o a quan i ica ion o he adso p ion a e and
Figu e 2. (a) Op ical mic oscopic image o PFPE-cPEI. Numbe -
weigh ed his og ams including cumula i e cu es o he size
dis ibu ion o he p ojec ed a ea diame e s o (b) PEG-cPEI (n=
387), (c) PDMS-cPEI (n= 200), and (d) PFPE-cPEI (n= 353)
pa icles in op ical mic oscopic images.
ACS ES&T Wa e pubs.acs.o g/es wa e A icle
h ps://doi.o g/10.1021/acses wa e .5c00094
ACS EST Wa e 2025, 5, 4435−4447
4440
should be a he unde s ood as an empi ical equa ion in he
desc ibed con ex o his s udy. The kine ic a e cons an s k2o
he cPEI adso ben s we e signi ican ly highe han hose o he
TP108 IX (e.g., app oxima ely 2000 imes highe o he
PFPE-cPEI). In p ac ice, as e kine ics mean sho e con ac
imes and hus smalle equi ed eac o s.
Expe imen ally de e mined di usion coe icien s ha e a
highe in o ma i e alue o adso p ion mechanisms han he
kine ic a e cons an s
86,87
and can se e as inpu pa ame e s o
u he adso p ion p ocess modeling.
68
Howe e , he calcu-
la ion o su ace di usion coe icien s Ds equi es a highe
compu a ional e o and he addi ional knowledge o iso he m
da a.
68
ADso 7 ×10−15 m2/s was de e mined o he PFPE-
cPEI (RMSE = 4.2 ng/mg) and a Dso 3 ×10−14 m2/s (RMSE
= 32.1 ng/mg) o he TP108 IX, assuming negligible ilm
di usion. Wi h simul aneous conside a ion o ilm and su ace
di usion, a Dso 7 ×10−14 m2/s and a kLo 1 ×10−4m2/s
we e de e mined o he PFPE-cPEI (RMSE = 12.6 ng/mg)
and a Dso 9 ×10−14 m2/s and a kLo 8 ×10−5m/s (RMSE =
18.8 ng/mg) o he TP108 IX. The imp o ed model i
indica es ha , a leas in he case o TP108 IX, ilm di usion
does no appea o be comple ely negligible. In bo h a ian s,
howe e , as e o compa able in apa icle mass anspo was
de e mined o TP108 IX compa ed o he PFPE-cPEI.
One o he key di e ences be ween he PSO and he HSDM
is ha he d i ing o ce o he PSO is he di e ence be ween
he (cons an ) mean equilib ium loading and he loading a
ime , whe eas he HSDM accoun s o he concen a ion
g adien wi hin he pa icle.
68
When using he HSDM, he
pa icle size is di ec ly included in he calcula ion.
The esul s he e o e indica e ha he explana ion o he
as adso p ion by he cPEI adso ben s is no apid in apa icle
di usion, bu a he he small pa icle sizes (52 μm (PDMS-
cPEI), 44 μm (PFPE-cPEI), 26 μm (PEG-cPEI) s 700 μm
(TP 108)) and associa ed la ge ou e su ace a eas (which
also a o s as e ilm di usion, compa e eq 8) as well as
di ec ly accessible adso p ion si es.
The kine ic cu e o he PDMS-cPEI showed ha he
PFHxA concen a ion inc eased again a e abou 4 h. This
Figu e 3. Pe cen age emo als o PFSA and PFCA (wi h di e en numbe s o pe luo ina ed ca bon a oms) a e 30 min and 10 days a an
adso ben dose o 40 mg/L (d y weigh ) o he comme cial PFA694E IX and H30N GAC and he newly syn hesized PEG-cPEI, PDMS-cPEI, and
PFPE-cPEI. Re e ence concen a ions we e 6.4 μg/L PFHpA, 8.9 μg/L PFOA, 6.3 μg/L PFNA, 2.8 μg/L PFDA, 6.6 μg/L PFBS, 8.0 μg/L PFHxS,
and 4.9 μg/L PFOS.
Figu e 4. PFHxA loading a 40 mg/L adso ben dose o e 48 h wi h
PSO model o all adso ben s and HSDM o PFPE-cPEI and TP108.
PFHxA e e ence concen a ions we e 10.9 μg/L (TP108), 10.4 μg/L
(PEG-cPEI), 10.0 μg/L (PDMS-cPEI), and 10.5 μg/L (PFPE-cPEI).
All samples we e analyzed in duplica es. Da a poin s a con ac imes
abo e 4 h we e neglec ed o PDMS-cPEI model i ing (neglec ed
da a poin s in g ay).
ACS ES&T Wa e pubs.acs.o g/es wa e A icle
h ps://doi.o g/10.1021/acses wa e .5c00094
ACS EST Wa e 2025, 5, 4435−4447
4441
phenomenon could occu due o e e sible binding o PFHxA
o ins abili y o he PDMS-cPEI adso ben . PDMS is known
o i s g ea chemical s abili y agains many chemicals, bu i is
eadily deg aded by luo ide ions because o he high s abili y
o Si−F bonds.
88
A measu emen o he e e ence sample
adhe ing o DIN EN ISO 10304-1 (D20)
89
uled ou an
inc eased luo ide concen a ion (c luo ide < 0.1 mg/L). A
deg ada i e luo ida ion o he PDMS c oss-linke seems,
he e o e, unlikely as an explana ion o he ele a ed PFHxA
concen a ions. Displacemen o PFHxA a he adso p ion si es
by compe ing subs ances also appea s unlikely, since a model
solu ion wi h only NaHCO3and NaCl in addi ion o PFHxA
was used, and simila displacemen e ec s should be isible o
he o he cPEI adso ben s. Howe e , s i ing he solu ions o
agi a ion on he ho izon al shake migh ha e pul e ized he
po en ially mechanically less s able PDMS-cPEI pa icles. As
he ini ial emo al wi hin he i s hou s looks p omising, i
would be ad isable o u he in es iga e he unde lying
mechanisms o e e sible PFHxA emo al by he PDMS-cPEI.
Due o he e e sible PFHxA emo al obse ed in he
kine ic expe imen a e 4 h (compa e Figu e 4), he PDMS-
cPEI was omi ed om u he iso he m s udies. The PEG-
cPEI showed signi ican ly lowe emo als han he PFPE-cPEI,
which is why no emo al could be obse ed o he
in es iga ed sho -chain PFAS a low adso ben doses. The
p esen a ion and i o iso he m models a e he e o e omi ed,
and only single-poin adso p ion coe icien s a he highes
adso ben dose (80 mg/L) we e discussed o compa ison.
Bo h he F eundlich and Langmui iso he m models can
desc ibe he iso he m cu es o he TP108 IX and he PFPE-
cPEI (compa e Figu e 5). Howe e , he F eundlich model
appea s o gi e a be e i , as can be seen om a di ec
Table 2. Adso p ion Kine ic Ra e Cons an s o PFO and PSO Model wi h Respec i e Coe icien o De e mina ion R2 o
Linea ized Model and RMSE o Model Cu e
a
PFO PSO
adso ben k1[1/h] qe[ng/mg] R2RMSE [ng/mg] k2[mg/(ng h)] qe[ng/mg] R2RMSE [ng/mg]
TP108 0.1835 267.4 0.9886 23.1 0.0003 364.3 0.9680 15.7
PEG-cPEI −0.0017 40.0 0.0014 35.7 0.1407 34.6 0.9965 9.3
PDMS-cPEI 6.6491 227.9 0.9619 27.0 0.0164 276.5 0.9994 19.8
PFPE-cPEI 0.0247 225.5 0.1143 176.0 0.6071 222.3 1.0000 12.6
a
De e mined o PFHxA a 40 mg/L adso ben dose (compa e wi h Figu e 4).
Figu e 5. Iso he m da a a e 48 h equilib a ion wi h espec i e modeled iso he ms acco ding o F eundlich and Langmui o PFPE-cPEI and
TP108 IX. Re e ence concen a ions we e 12.9 μg/L PFBA, 13.5 μg/L PFHxA, 8.4 μg/L TFMSA, and 11.3 μg/L PFPeS. All samples we e analyzed
in duplica es. The s uc u es o he anions o he PFCA and PFSA a e displayed in he op igh co ne s.
Table 3. F eundlich Model Pa ame e s wi h Respec i e Coe icien o De e mina ion R2 o he Linea ized Model and RMSE
o he Model Cu e
adso ben PFAS n KF[ng/mg/(ng/L)n]R2RMSE [ng/mg]
PFPE-cPEI PFBA 0.30 6.2370 0.5093 6.3
PFHxA 0.84 0.5245 0.8263 95.3
TFMSA 1.09 0.0085 0.6704 26.4
PFPeS 0.94 2.7544 0.9689 56.1
TP108 PFBA 0.38 20.1674 0.9831 15.6
PFHxA 0.36 25.8348 0.9902 20.7
TFMSA 0.33 41.6775 0.9671 35.5
PFPeS 0.27 48.5026 0.6703 71.8
ACS ES&T Wa e pubs.acs.o g/es wa e A icle
h ps://doi.o g/10.1021/acses wa e .5c00094
ACS EST Wa e 2025, 5, 4435−4447
4442
compa ison o he RMSE alues in Tables 3 and 4. Bo h
F eundlich and Langmui a e single-solu e iso he m models.
68
When using adso ben s in which ion exchange con ibu es o
he emo al, such as o TP108 IX and cPEI adso ben s wi h
qua e nized ammonium g oups, i is no co ec o conside a
single solu e sys em, as he PFAS compe e wi h he coun e ion
wi h which he ion exchange is loaded (in his case, chlo ide).
The applica ion o he models he e should he e o e be
unde s ood as empi ical, and he pa ame e s de e mined a e
only alid unde he expe imen al condi ions desc ibed, as
discussed in de ail by Haupe e al.
90
The smalle he
F eundlich exponen nis, he mo e conca e he iso he m shape
is, yielding high loadings a low concen a ions. Fa o able
iso he ms can be obse ed o he TP108 IX, which (wi h he
excep ion o PFPeS) shows a highe loading han he PFPE-
cPEI a he highe adso ben doses. The iso he ms o he
PFPE-cPEI in he es ed concen a ion anges e ealed almos
linea shapes wi h F eundlich exponen s close o 1 (excep o
PFBA).
Fo simila F eundlich exponen s, he a ini ies o he
adso ba es o he adso ben can be compa ed acco ding o
he espec i e KF alues. The es ima e o he TP108 IX esul s
in he ollowing sequence: KF,PFPeS >KF,TFMSA >KF,PFHxA >
KF,PFBA. The highe a ini y o TFMSA (compa ed o PFHxA)
indica es ha he unc ional anionic g oup o he molecule has
a s onge in luence on he emo al o sho -chain PFAS by he
TP108 IX han he leng h o he pe luo ina ed ca bon chain.
The emo al he e is p ima ily domina ed by elec os a ic
a ac ion. The esul s a e di e en in he case o PFPE-cPEI,
which yielded e y low emo als o he sho es chain PFAS,
TFMSA, and PFBA. In con as , in he case o he mos
hyd ophobic compound PFPeS, he PFPE-cPEI achie ed
highe loadings han he TP108 IX. Wi h he PEG-cPEI,
emo als o a simila o de o magni ude could no be achie ed
o any sho -chain PFAS. While he Kd alues a 80 mg/L
adso ben o TFMSA we e abou 6 imes lowe , hose o
PFPeS we e e en abou 150 imes lowe han hose o he
PFPE-cPEI (compa e Kd alues in Table S5).
I can be concluded ha he p esence o he luo ous PFPE
c oss-linke enhances he capaci y o mos PFAS in es iga ed
d as ically. The poo e emo al o he sho es -chain
subs ances (PFBA and TFMSA) wi h simul aneously com-
pa able o e en highe emo al o PFAS wi h a leas 5
pe luo ina ed C a oms (compa e Figu es 3 and 5) indica es
ha hyd ophobic in e ac ions play a mo e impo an ole
du ing he adso p ion o PFAS on o he PFPE-cPEI han on o
he TP108 IX.
■CONCLUSIONS
Th ee ypes o c oss-linked PEI esin pa icles (cPEI) we e
syn hesized, cha ac e ized in de ail, and e alua ed ega ding
hei po en ial as adso ben s o anionic PFAS in compa ison
o an indus ial s a e-o - he-a IX o PFAS emo al. The h ee
cPEI adso ben s di e ed in hei oligoe he c oss-linke
segmen (PFPE, PDMS, o PEG) bu showed compa able
anion exchange capaci ies, c oss-linke con en s, su ace a eas,
and mo phologies. All in es iga ed cPEI adso ben s showed
signi ican long-chain PFAS emo al e iciencies wi hin only 30
min. In con as , PFAS we e only poo ly emo ed by
comme cial IX and GAC in he same ime span. The as
adso p ion a e on o he cPEI adso ben s is e lec ed by he
eac ion kine ic cons an s o he adso p ion o one selec ed
PFAS (PFHxA). The kine ics o all adso ben s could be
desc ibed well using he PSO model. Ne e heless, di usion
coe icien s ha we e calcula ed using he HSDM indica ed
ha in apa icle di usion o PFHxA in he s a e-o - he-a IX
was as e han in he PFPE-cPEI. The cPEI adso ben s
p e e ably adso bed PFAS wi h longe chains, while he
comme cial IX ob ained mo e consis en emo als o PFAS
ega dless o hei chain leng hs. This obse a ion was
con i med wi h iso he m s udies o PFHxA, PFBA, TFMSA,
and PFPeS and he adso ben s PFPE-cPEI and he comme cial
IX. I should be emphasized ha he adso p ion iso he ms
p o ide only an empi ical desc ip ion o he adso p ion p ocess
and a e limi ed o he condi ions used in ou es . O e all, ou
esul s indica e ha hyd ophobic c oss-linked PEI is a
p omising ype o adso ben o he indus ial-scale emedia-
ion o PFAS when as kine ics a e equi ed. The PDMS-cPEI
is a p omising luo ine- ee adso ben o PFAS emo al wi h
e y sho con ac imes. Howe e , in o de o suppo la ge-
scale applica ions o PDMS-cPEI, u he esea ch is necessa y
o ensu e i s s abili y du ing longe ope a ion.
■ASSOCIATED CONTENT
*
sı Suppo ing In o ma ion
The Suppo ing In o ma ion is a ailable ee o cha ge a
h ps://pubs.acs.o g/doi/10.1021/acses wa e .5c00094.
Expe imen al desc ip ions o he me cu ime ic i a ion,
he mog a ime ic analysis, ESI mass spec ome y,
NMR, EA, IR, XPS, SEM, SAXS, BET, as well as
calcula ions, mic oscopic images, his og ams, and single-
poin adso p ion coe icien s (PDF)
■AUTHOR INFORMATION
Co esponding Au ho s
Aki S. Ruhl −Technische Uni e si ä Be lin, Wa e T ea men ,
KF4, 10623 Be lin, Ge many; Ge man En i onmen Agency,
Table 4. Langmui Model Pa ame e s wi h Respec i e Coe icien o De e mina ion R2 o Linea ized Model and RMSE o
Model Cu e
adso ben PFAS qmax [ng/mg] KL[L/ng] R2RMSE [ng/mg]
PFPE-cPEI PFBA 121.5 3.57 ×10−04 0.5261 6.3
PFHxA 1486.7 1.36 ×10−04 0.8505 117.4
TFMSA 1746.4 1.10 ×10−05 0.7743 27.9
PFPeS 1853.7 1.35 ×10−03 0.9648 69.9
TP108 PFBA 384.6 3.29 ×10−03 0.9140 67.7366
PFHxA 473.4 2.81 ×10−03 0.9024 89.5204
TFMSA 393.4 1.42 ×10−02 0.9460 63.8401
PFPeS 297.9 1.81 ×10−02 0.3138 135.7011
ACS ES&T Wa e pubs.acs.o g/es wa e A icle
h ps://doi.o g/10.1021/acses wa e .5c00094
ACS EST Wa e 2025, 5, 4435−4447
4443
