scieee Science in your language
[en] (orig)

Bacterial-Cellulose-Derived Carbonaceous Electrode Materials for Water Desalination Via Capacitive Method: the Crucial Role of Defect Sites

Author: Belaustegi Ituarte, Yolanda,Panto, Fabiola,Urbina Moreno, Leire,Corcuera Maeso, María Ángeles,Eceiza Mendiguren, María Aranzazu,Palella, Alessandra,Triolo, Claudia,Santangelo, Saveria
Publisher: Elsevier
Year: 2020
DOI: 10.1016/j.desal.2020.114596
Source: https://addi.ehu.eus/bitstream/10810/50126/1/YB.pdf
Con en s lis s a ailable a ScienceDi ec
Desalina ion
jou nal homepage: www.else ie .com/loca e/desal
Bac e ial-cellulose-de i ed ca bonaceous elec ode ma e ials o wa e
desalina ion ia capaci i e me hod: The c ucial ole o de ec si es
Yolanda Belaus egui
a
, Fabiola Pan ò
b
, Lei e U bina
c
, Ma ia Angeles Co cue a
c
, A an xa Eceiza
c
,
Alessand a Palella
b
, Claudia T iolo
d,⁎
, Sa e ia San angelo
d,⁎
a
Tecnalia, Basque Resea ch and Technology Alliance (BRTA), Ma e ials o Ene gy and En i onmen A ea, 48160 De io, Bizkaia, Spain
b
Is i u o di Tecnologie A anza e pe l'Ene gia (ITAE) del Consiglio Nazionale delle Rice che (CNR), 98126 Messina, I aly
c
Depa men o Chemical and En i onmen al Enginee ing, Enginee ing School o Gipuzkoa, ‘Ma e ials+Technologies’ G oup, Uni e si y o he Basque Coun y (UPV/
EHU), 20018 Donos ia-San Sebas ián, Spain
d
Dipa imen o di Ingegne ia Ci ile, dell'Ene gia, dell'Ambien e e dei Ma e iali (DICEAM), Uni e si à “Medi e anea”, 89122 Reggio Calab ia, I aly
ARTICLE INFO
Keywo ds:
Bac e ial cellulose
Raman spec oscopy
La ice de ec s
Capaci i e deioniza ion
ABSTRACT
Elec oso p i e desalina ion is a e y simple and appealing app oach o sa is y he inc easing demand o
d inking wa e . The la ge-scale applica ion o his echnology calls o he de elopmen o easy- o-p oduce,
cheap and highly pe o ming elec ode ma e ials and o he iden i ica ion and ailo ing o hei mos in luen ial
p ope ies, as well. He e, biosyn hesised bac e ial cellulose is used as a ca bon p ecu so o he p oduc ion o
h ee-dimensional nanos uc u es endowed wi h hie a chically po ous a chi ec u e and di e en densi y and
ype o in insic and he e o-a om induced la ice de ec s. The p oduced ma e ials exhibi unp eceden ed desa-
lina ion capaci ies o ca bon-based elec odes. A an ini ial concen a ion o 585 mg L
−1
(10 mmol L
−1
), hey
a e able o emo e om 55 o 79 mg g
−1
o sal ; as he ini ial concen a ion ises o 11.7 g L
−1
(200 mmol L
−1
),
hei sal adso p ion capaci y eaches alues anging be ween 1.03 and 1.35 g g
−1
. The esul s o he ho ough
ma e ial cha ac e isa ion by complemen a y echniques e idence ha he ela i e amoun o oxygena ed su ace
unc ional species enhancing he elec ode we abili y play a c ucial ole a lowe NaCl concen a ions, whe eas
he a ailabili y o ac i e non-sp
2
de ec si es o adso p ion is mainly in luen ial a highe sal concen a ions.
1. In oduc ion
The a ailabili y o d inking wa e is becoming inadequa e o sa is y
he e e g owing demand caused by he inc ease o wo ld popula ion,
enhancemen o en i onmen al pollu ion le el and clima e changes.
Acco ding o he es ima ion by he Uni ed Na ions, mo e han one
billion people cu en ly li e in wa e -s essed egions and his si ua ion
will wo sen in he nex yea s, wi h almos hal he wo ld's popula ion
expec ed o li e in a eas o high wa e -s ess by 2030 [1]. Wa e
sho age is agg a a ed in Coun ies whe e he necessa y in as uc u e
o ake wa e om i e s and aqui e s lack. Clean wa e has been e-
cognised as a basic human igh by he Uni ed Na ions [2]. I s a ail-
abili y and a o dabili y a e among he main social, echnological, and
economical challenges o be aced by he Wo ld in he 21s cen u y.
Since 97% o he Plane 's wa e is seawa e , seawa e desalina ion
po en ially ep esen s he sma solu ion o sca ci y o esh wa e . This
has encou aged he de elopmen o se e al wa e desalina ion ech-
nologies (e.g. e e se osmosis, dis illa ion, elec odialysis) [3]. Among
he cos -e ec i e and ene gy-e icien ones, elec oso p i e desalina ion
looks pa icula ly a ac i e hanks o i s simplici y, easy main enance,
low ene gy consump ion and en i onmen al impac . Also known as
capaci i e de-ioniza ion (CDI), i emo es cha ged ionic species om
wa e ia he elec os a ic in e ac ion wi h he wo elec odes o ming
CDI-cell, o which an ex e nal ol age is applied [4–7]. Low ene gy
equi ed o desalina e one cubic me e o b ackish wa e s (app oxi-
ma ely only one hal compa ed o e e se osmosis [8]) and low ope -
a ing ol age o he cell (≤ 2 V) makes CDI economically compe i i e
and s a egically impo an o a eas wi hou elec ici y g id.
Al hough se e al ac o s (e.g. cell design and ope a ional se ings)
s ongly in luence he CDI pe o mance, physical and s uc u al p op-
e ies o he po ous elec ode ma e ials ep esen a key poin . Besides
high su ace a ea, elec onic conduc i i y, elec ochemical s abili y and
we abili y, also he o al olume, a e age size and connec i i y o he
po es a e impo an , which has made ca bon he ma e ial o mos e-
quen choice o de eloping and ab ica ing po ous elec odes [6,7]. A
wide a ie y o epo s on nanoca bons is a ailable in he li e a u e
h ps://doi.o g/10.1016/j.desal.2020.114596
Recei ed 1 May 2020; Recei ed in e ised o m 19 June 2020; Accep ed 25 June 2020
⁎
Co esponding au ho s.
E-mail add esses: [email p o ec ed] (C. T iolo), [email p o ec ed] (S. San angelo).
Desalina ion 492 (2020) 114596
0011-9164/ © 2020 The Au ho s. Published by Else ie B.V. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/BY-NC-ND/4.0/).
T
[9–11]. Al hough endowed wi h good CDI pe o mance, many o hem
su e om ela i ely complica ed manu ac u ing p ocesses and high
p oduc ion cos . The sea ch o new highly pe o ming ma e ials wi h
low cos and easy syn hesis ou es is c ucial o he de elopmen and
he p ac ical applica ion o he CDI echnology.
Bac e ial cellulose (BC) is he mos abundan lignin- ee biopolyme
in na u e [12,13]. Al hough i exhibi s he same chemical s uc u e as
plan -de i ed cellulose, (C
6
H
10
O
5
)
n
, BC is endowed wi h highe c ys-
allini y, wa e -holding capaci y, mechanical s eng h, and pu i y
[12,14]. I s syn hesis occu s in h ee s ages, namely i) polyme iza ion o
he glucose in o long and unb anched chains (β-1,4-D-glucan) linked
wi h hyd ogen bonding; assembly o ii) se e al p o o ib ils, consis ing
o a ew o pa allel chains, in o 2–4 nm-wide nano ib ils, and iii) o
bundles o nano ib ils in o 20–100 nm-wide nano ibbons, inally gi ing
ise o a h ee-dimensional (3D) ne wo k o in e wo en andomly-o -
ien ed nano ibbons (BC pellicle) [12,15]. Owing o in e ib illa oids
a he a ious scales, he 3D cellulosic a chi ec u e is endowed wi h
hie a chical po osi y.
This ex acellula biopolyme is biosyn hesised h ough he po es o
he memb ane o some bac e ial s ains including hose belonging o
gene a Rhizobium, Gluconace obac e , Ag obac e ium, Sa cina and
Alcaligenes. Among bac e ia able o sec e e bundles o cellulose chains
as pa o hei me abolism, Gluconace obac e is he genus wi h he
highes BC p oduc ion abili y [15]. Classi ied in he amily Ace -
obac e aceae, i con ains se e al s ains, including Gluconace obac e
xylinus he mos s udied among hem [12–18]. Recen ly, Gluconace o-
bac e medellinensis (ID13488), a s ain isola ed om homemade ui
inega in Medellin (Colombia) cen al ma ke , has ga he ed g ea a -
en ion due o i s abili y o p oduce c ys alline BC unde highly acidic
condi ions, as desi able o indus ial p oduc ion om acidic esidues
(e.g. was es gene a ed om cide p oduc ion) [19].
Due o i s in e es ing con o ma ion, biocompa ibili y and biode-
g adabili y, BC has a ac ed a en ion in a wide ange o new appli-
ca ions in he biomedical ield [20–22], such as issue enginee ing [23],
d ug deli e y and wound d essing [24,25], in he ood packaging a ea
[26,27] and also in en i onmen al applica ions [28].
Mo eo e , hanks o i s wide a ailabili y (BC is biosyn hesized on a
scale o se e al billion ons pe yea [29]), in he las decades, i has
been ex ensi ely exploi ed as a cheap and g een ca bon sou ce o he
p oduc ion o a a ie y o ca bon-based ma e ials [30], mos equen ly
u ilised in he ield o ene gy s o age. Se e al wo ks a e a ailable in he
li e a u e dealing wi h BC-de i ed elec ode ma e ials o supe -
capaci o s [31–40] and echa geable ba e ies [41–43]. Howe e , e-
cen ly, he use o BC-de i ed ca bon-based ma e ials has been epo ed
also o di e en in e es ing applica ions, such as capaci i e deioniza-
ion [44], abso p ion o oils and o ganic sol en s [45], elec omagne ic
in e e ence shielding [46].
In his wo k, BC is u ilised as a ca bon p ecu so o he p epa a ion
o elec ode ma e ials o he desalina ion o wa e ia he capaci i e
me hod. The elec ochemical pe o mance o he BC-de i ed nano-
ca bons is e alua ed. The e ec o he en ichmen wi h g aphene is also
in es iga ed. The ou s anding esul s ob ained in e ms elec oso p i e
pe o mance a e co ela ed wi h he mic os uc u al p ope ies o he
elec ode ma e ials, as desc ibed by mic o-Raman spec oscopy, and he
c ucial ole o he ca bon de ec s is emphasised.
2. Expe imen al
2.1. P epa a ion o he p is ine ma e ials
2.1.1. Pu e BC
Gluconace obac e medellinensis s ain ID13488 (CECT 8140) was
used o he biosyn hesis o BC memb anes. Bac e ial cells we e g own,
o 14 days unde s a ic condi ions, in a s anda d Hes in-Sch amm (HS)
medium con aining D-glucose (2% w/ ), pep one (0.5% w/ ), yeas
ex ac (0.5% w/ ), disodium hyd ogen phospha e (0.27% w/ ) and
ci ic acid (0.115% w/ ). The incuba ion empe a u e and pH o he
medium we e 28 °C and 4, espec i ely The ob ained BC memb anes
we e ea ed wi h KOH (2 w %) o 24 h a oom empe a u e (RT) wi h
he dual pu pose o i) emo ing non-cellulosic compounds and ii)
gene a ing sub-mic opo es and de ec s in he ca bonaceous ma ix upon
ca bonisa ion [47]. Finally, hey we e eeze-d ied (Fig. S1a).
2.1.2. G aphene-en iched BC
In o de o in es iga e he e ec o he en ichmen o he elec ode
ma e ial wi h g aphene, a composi e cellulose/ educed g aphene oxide
( GO/BC) memb ane was u he p epa ed. This was accomplished ia
ex-si u ins ead o in-si u me hod because o he endency o GO o
p ecipi a e in he cul u e medium gi ing ise o a non-homogenous
dispe sion wi hin he g owing ma ix. As ske ched in Fig. S2, he
memb anes we e imme sed in a g aphene oxide (GO) suspension o
24 h unde sonica ion. A comme cial suspension o GO in wa e
(4 mg mL
−1
) supplied by G aphenea (a e age pa icle size < 10 μm;
ca bon: 49–56%; oxygen: 41–50%; ni ogen: 0–1%) was used. The ob-
ained memb anes we e d ied in an o en a 50 °C o 24 h be ween
Te lon molds and he esul ing GO/BC “pape ” was inally chemically
educed o GO/BC wi h asco bic acid (30 g L
−1
) in an o en o 2 h a
95 °C (Fig. S1b).
He ea e , he pu e BC memb ane and he composi e GO/BC
“pape ” will be labelled as samples p-BC and p-GBC.
2.2. P epa a ion o he BC-de i ed ca bonaceous ma e ials
In o de o ob ain ca bonaceous ma e ials o he p epa a ion o he
elec odes o he CDI cell, he p is ine samples we e p ocessed as de-
sc ibed in de ail below and summa ised in Table 1.
Following Liu e al. [44], a wo-s ep hea ing p ocess was ca ied ou
on sample p-BC. The p ocess consis ed o i) an oxida i e ea men
(polyme s abilisa ion) in s a ic ai and ii) an annealing (ca bonisa ion)
pe o med a 700 °C unde 5.5 pu i y ni ogen (100 cc min
−1
low
a e). Based on he esul s o he he mo-g a ime ic analysis (see Fig.
S3 in he suppo ing in o ma ion), s abilisa ion was ope a ed a 225 °C.
S abilisa ion- and ca bonisa ion- empe a u es we e eached by
5 °C min
−1
hea ing a e and kep cons an o 2 h. Uncon olled cooling
down o RT ollowed bo h he he mal ea men s. In he ollowing, he
pu e-BC de i ed ac i e ma e ial will be labelled as sample BC(N
2
).
In o de o in es iga e he e ec o he en ichmen o he ca bo-
naceous elec ode ma e ials wi h g aphene, h ee composi e samples
we e p epa ed (Table 1). The i s one, below coded as GBC(N
2
), was
ob ained by ca bonisa ion o he GO-con aining p-GBC (Fig. S2). The
hea p ocess was ope a ed a 700 °C o 2 h unde 100 cc min
−1
N
2
Table 1
Codes o he p is ine and p ocessed samples, wi h de ails abou hei p epa a ion.
Code Ma e ial ype and p epa a ion de ails
p-BC P is ine Pu e BC memb ane by biosyn hesis and eeze-d ying
p-GBC Composi e GO/BC “pape ” by biosyn hesis, imme sion in GO suspension, molding and chemical educ ion
BC(N
2
) P ocessed G aphene- ee BC-de i ed ca bon By s abilisa ion in ai and ca bonisa ion o p-BC in N
2
GBC(N
2
) G aphene-en iched BC-de i ed ca bon By ca bonisa ion o p-GBC in N
2
BCG(N
2
) By s abilisa ion in ai o p-BC, imp egna ion wi h dilu ed GO suspension and ca bonisa ion in N
2
BCG(He) By s abilisa ion in ai o p-BC, imp egna ion wi h dilu ed GO suspension and ca bonisa ion in He
Y. Belaus egui, e al. Desalina ion 492 (2020) 114596
2
low.
Two addi ional samples, he ea e labelled as BCG(N
2
) and
BCG(He), we e p oduced as ske ched in Fig. S2 o he o me . BCG(N
2
)
and BCG(He) we e p epa ed by s abilising p-BC in ai a 225 °C, im-
p egna ing he s abilised ma e ial wi h a dilu ed GO-suspension and,
a e d ying o e nigh , ca bonising he esul ing ma e ial a 700 °C o
2 h unde 100 cc/min N
2
o He low, espec i ely. A comme cial sus-
pension o GO in wa e (G aphenea; 4 mg mL
−1
), dilu ed in (¼ / ) in
dis illed wa e (Me ck), was used o he imp egna ion.
2.3. Physicochemical cha ac e isa ion o p is ine and p ocessed ma e ials
The p is ine and p ocessed ma e ials we e ho oughly cha ac e ised.
The mo phological analysis was pe o med by scanning elec on mi-
c oscopy (SEM) by u ilising a FEI S-FEG XL30 mic oscope equipped
wi h an ene gy-dispe si e X- ay (EDX) spec ome e . A NETZSCH STA
449C ins umen was used o he he mo-g a ime ic analysis (TGA).
Few millig ams o he sample we e loaded in an alumina c ucible. The
analysis was ca ied ou in ai . The 20–800 °C ange was in es iga ed
ising empe a u e a a a e o 10 °C min
−1
. The B unaue -Emme -
Telle (BET) me hod was used o he speci ic su ace a ea de e mina-
ion by using ASAP 2010 Mic ome i ics equipmen . Fu he de ails on
he sample p epa a ion and expe imen al da a p ocessing can be ound
elsewhe e [11]. X- ay di ac ion (XRD) pa e ns we e eco ded on a
Panaly ical X-Pe di ac ome e using a Ni β- il e ed Cu-K
α
adia ion
(λ = 1.5451 Å) sou ce. Da a we e collec ed in he 5–80° 2θ-angle ange
wi h a s ep o 0.05°. Raman sca e ing exci ed by a solid-s a e lase
ope a ing a 532 nm (2.33 eV) was measu ed by using a NT-MDT
NTEGRA - Spec a SPM spec ome e equipped wi h MS3504i 350 mm
monoch oma o and ANDOR Idus CCD. The sca e ed ligh om he
sample, collec ed ia a Mi u oyo 100× objec i e, was dispe sed by a
600 lines mm
−1
g a ing. In o de o p e en local hea ing and an-
nealing e ec s, a lase powe o 250 μW a he sample su ace was used.
The spa ial homogenei y o he samples was u he e alua ed by e-
co ding spec a om se e al andom posi ions on each specimen and,
in o de o ha e a eliable pic u e o he en i e sample, he spec a
collec ed in each specimen we e a e aged. X- ay pho oelec on spec-
oscopy (XPS) analysis was ca ied ou wi h a Physical Elec onics
GMBH PHI 5800-01 spec ome e , equipped wi h a monoch oma ic Al-
K
α
(1486.6 eV) sou ce se a a powe beam o 300 W. Spec a we e
acqui ed using pass ene gies o 11.0 eV and 58.0 eV o elemen al
analysis and de e mina ion o he oxida ion s a es, espec i ely. The
binding ene gy egions o C 1s-K 1s(280–300 eV), N 1s(390–406 eV),
and O 1s(520–540 eV) we e examined, aking he C 1sline (284.5 eV)
as e e ence [48]. XPS da a we e in e p e ed by using he on-line lib a y
o oxida ion s a es implemen ed in he PHI MULTIPAK 6.1 so wa e and
he PHI Handbook o XPS [49]. The elemen al concen a ions we e
es ima ed om he a eas unde he pho oelec on peaks weighed by he
ela i e sensi i i y ac o s. The esul s ob ained a e epo ed in Table 2.
Fo u he de ails on he ins umen a ion and expe imen al da a
p ocessing see e s [11,50].
2.4. P epa a ion o he wo king elec odes based on he BC-de i ed
ca bonaceous ma e ials
Fou wo king elec odes we e p epa ed by using he he mally
p ocessed BC-de i ed ca bonaceous ma e ials. Thei codes a e epo ed
in Table 3. Th ee o hem (BCE-1, BCE-2 and BCE-3) we e based on
g aphene-en iched samples; in he case o BCE-1, a physical mix u e o
samples GBC(N
2
) and BC(N
2
) in a 1:3 mass a io was used as an ac i e
ma e ial. The elec odes we e p epa ed ia he p ocedu e p e iously
desc ibed in de ail [11,50]. B ie ly, he ac i e ma e ial (90 w %) was
mixed wi h poly inylalcohol (10 w %), supplied by Tecnalia, and he
esul ing mix u e was dispe sed in e hanol. The slu y ob ained a e
8 h s i ing was cas on o he (Me sen Ibe ica) g aphi e cu en collec o
and d ied in an o en a 60 °C o e nigh .
2.5. Elec ochemical h ee-cell measu emen s
A con en ional h ee-elec ode sys em, using a compu e -con olled
AUTOLAB PGSTAT302N Me ohm po en ios a /gal anos a , was u i-
lised o ca y ou cyclic ol amme y (CV) measu emen s. The wo king
elec ode was p epa ed as-desc ibed in Sec ion 2.4; g aphi e ac ed as
he coun e elec ode, while a s anda d Ag/AgCl elec ode was used as
he e e ence. CV measu emen s we e ca ied ou a RT in 0.1 mol L
−1
NaCl solu ion a di e en scan a es (5, 10, 20, 30, 40, 50 and
100 mV s
−1
); he po en ial ange om −1.3 o 0.4 V was in es iga ed.
The speci ic capaci ance, C
S
(F g
−1
), was calcula ed om he cu en -
ol age cu es as
=Cs
m
I
V
dV
1,
(1)
whe e I(A), V(V), (V s
−1
) and m(g) deno e he esponse cu en , he
po en ial, he po en ial scan a e and he mass o he elec o-ac i e
ma e ial [11].
Elec ochemical impedance spec oscopy (EIS) analysis was ca ied
ou in he 1·10
−1
-1·10
4
Hz equency ange wi h AUTOLAB
PGSTAT302N using he h ee-compa men cell (ampli ude o he al-
e na ing ol age: 1·10
−1
V a ound he 0 V equilib ium po en ial).
A ea o he wo king elec odes u ilised o CV ad EIS measu emen s
was 2 cm
2
. Fu he de ails can be ound elsewhe e [11,50].
2.6. Elec oso p ion expe imen s
Ba ch mode expe imen s in a con inuously eci cula ing sys em
we e ca ied ou o in es iga e he elec oso p ion beha iou o he BC-
de i ed elec odes. The sys em, p e iously desc ibed [11], included a
symme ic elec oso p i e uni cell, wo ese oi s, a pe is al ic Fishe
Scien i ic, Mini-pump, a Hanna Mic op ocesso Conduc i i y/TDS
me e and a DC Lab Powe supply LABPS1503. A ea o he wo pa allel
elec odes o ming he CDI uni cell was 9 cm
2
; a non-elec ically
conduc i e sepa a o was in e posed be ween hem. 100 mL o NaCl
solu ion, kep a 25 °C, we e u ilised in each expe imen ( low a e:
Table 2
Su ace composi ion o he samples, as in e ed om he XPS analysis.
Code A omic concen a ion/a .% O/C
a io
Rela i e weigh /w %
C N O C N O
BC(N
2
)
a
59.3 0.0 40.7 0.69 52.2 0.0 47.8
GBC(N
2
) 71.5 1.9
b
26.6 0.37 65.5 2.0
b
32.5
BCG(N
2
)
a
59.0 0.0 41.0 0.69 51.9 0.0 48.1
BCG(He)
a
52.9 0.0 47.1 0.89 45.6 0.0 54.2
Mix u e
c
62.4 0.5 37.2 0.60 55.5 0.5 44.8
a
Ne alues a e sub ac ion o K a omic concen a ion.
b
F om GO.
c
Values calcula ed om he da a o BC(N
2
) and GBC(N
2
) by using he mass
3:1 a io.
Table 3
Codes o he wo king elec odes and ac i e ma e ials u ilised o hei p e-
pa a ion.
Code Ac i e ma e ial(s) and ela i e code(s)
BCE-0 BC-de i ed ca bon BC(N
2
)
BCE-1 3:1
a
mix u e o BC-de i ed ca bon BC(N
2
) and g aphene-en iched BC-
de i ed ca bon GBC(N
2
)
BCE-2 G aphene-en iched BC-de i ed ca bon BCG(He)
BCE-3 G aphene-en iched BC-de i ed ca bon BCG(N
2
)
a
The 3:1 mass a io was chosen in o de o ob ain compa able GO con en s
o he h ee g aphene-en iched elec odes (~5 w %).
Y. Belaus egui, e al. Desalina ion 492 (2020) 114596
3
7.7 mL min
−1
). A 1.2 V ol age was applied un il he comple e cell
cha ging and equilib ium concen a ion we e eached [11,51].
The sal adso p ion capaci y, SAC (mg g
−1
), was calcula ed as
=SAC C C
m
( ) ,
0
(2)
whe e C
0
(mg L
−1
) and C
(mg L
−1
) s and o he ini ial and he in-
s an aneous equilib ium-concen a ions, espec i ely; τ(L) and m(g)
indica e he solu ion olume and he mass o he (ac i e) elec ode pai ,
espec i ely. To check he ep oducibili y, se e al expe imen s we e
ca ied ou on each ype o elec ode ma e ial, and he co esponding
SAC was calcula ed a e aging o e he alues ob ained in hese ex-
pe imen s.
3. Resul s and discussion
3.1. Physicochemical p ope ies o he p is ine pu e BC and composi e GBC
samples
Figs. 1 and S4–S5 summa ise he esul s o SEM/EDX, XRD and
mic o-Raman spec oscopy (MRS) analyses on he eeze-d ied BC.
Fig. 1a–b shows ep esen a i e SEM images o he samples; lowe -
magni ica ion images can be ound in he suppo ing in o ma ion (SI,
Fig. S4a–b). The highly-po ous ib illa -ne -like mo phology, peculia
o biosyn hesised cellulose [52], is clea ly isible. Bundles o mic o-
ib ils, densely-packed due o he wa e emo al p ocess [13], can be
obse ed, along wi h nume ous mac opo es (in bo h he image inse s).
The andomly-o ien ed mic o ib ils a e in e connec ed o o m a 3D
cellulosic skele on, ea u ed by hie a chical po osi y. The analysis a
highe magni ica ion ac o s e eals he lack o spa ial uni o mi y in
e ms o leng h and hickness o he mic o ib ils and bundles, as well.
Con e sely, he elemen al mapping ia SEM/EDX (Fig. S4c) e idences a
spa ially homogeneous dis ibu ion o ca bon, oxygen, po assium (de-
i ing om ea men in KOH) a he mic o-scale.
Na i e cellulose I is composed o I
α
and I
β
allomo phs. The wo
phases ha e he same molecula building and O
3
–H–O
5
in a-chain
hyd ogen bonding, bu di e en O
2
–H–O
6
in e -chain hyd ogen
bonding and c ys al s uc u e [17]. Thei ela i e amoun depends on
he sou ce o he cellulose [17,53]. In gene al, plan -de i ed cellulose is
ich in I
β
allomo ph, while I
α
is he p edominan allomo ph in cellulose
de i ed om algae and bac e ia. Th ee di ac ion peaks a e de ec ed in
he XRD pa e n o p-BC a ound 14.5°, 16.6° and 22.6° (Fig. 1c). These
majo e lec ions co espond o (100), (010), (110) planes in he I
α
phase, and (1−10), (110), (200) in he I
β
one. In his case, he di -
ac ion pa e ns ha e been indexed acco ding o Mille indexes o
cellulose I
α
, he p edominan allomo ph in p-BC [53].
The c ys allini y index o he eeze-d ied BC was es ima ed, ia he
empi ical o mula χ= 100·[I
(110)
-I
A
]/I
(110)
, p oposed by Segal e al.
[54,55], whe e I
(110)
deno es he in ensi y o he main di ac ion peak
a 22.6° o I
α
phase and I
A
indica es he in ensi y due o he amo phous
con ibu ion a 18° 2θ-angle. The mean c ys alli e size was calcula ed,
om he ull wid h a hal maximum (FWHM) o he (110) peak,
h ough he Sche e 's equa ion, d
BC
=kλ/βcosθ, whe e k(0.89) is he
shape ac o , λis he wa eleng h o x- ay adia ion, βis he FWHM in
adians and θis he B agg's angle [16]. Despi e he ela i ely low
signal- o-noise a io o he pa e n, he ound χ- and d
BC
- alues
(85 ± 5% and 6.1 ± 0.3 nm, espec i ely) we e compa able wi h
hose epo ed in he li e a u e o he cellulose p oduced om HS
medium [16,56].
The measu e o Raman sca e ing om di e en andom loca ions in
Fig. 1. Resul s o (a–b) SEM, (c) XRD and (d) MRS analyses on sample p-BC.
Y. Belaus egui, e al. Desalina ion 492 (2020) 114596
4
he eeze-d ied BC (Fig. S5) poin ed a he lack o spa ial uni o mi y
wi hin he sample, e lec ing he local changes in he mo phology
(Fig. 1a–b) and in he c ys allini y deg ee [17,35]. Thus, in o de o
ha e a eliable pic u e o he en i e sample, he spec a ela i e o he
di e en loca ions sampled we e a e aged. Fig. 1d shows he esul
ob ained. The spec al p o ile is peculia o cellulose I [17,57]. The
skele al bending modes o CCC, COC, OCC, OCO bonds a e de ec ed in
he lowes equency egion o he spec um (150–550 cm
−1
) [17]. The
in ense s e ching modes o CC and CO bonds domina e he
800–1180 cm
−1
egion [17]. A highe equencies (1180–1430 cm
−1
),
he bending modes o HCC, HCO, HCH and COH bonds con ibu e o he
Raman in ensi y [17]. Abo e 2500 cm
−1
, a e y in ense band cen ed
a ~2900 cm
−1
and a weake and b oade one (a ~3350 cm
−1
) a e
de ec ed. The o me o igina es om he s e ching modes o CH bonds,
while he la e is associa ed o he s e ching modes o OH bonds [17].
No cellulose mode con ibu es o he Raman in ensi y be ween 1500
and 2500 cm
−1
.
The in ensi ies I
380
and I
1096
o he Raman bands a 380 and
1096 cm
−1
we e u ilised o es ima e he c ys allini y o he eeze-d ied
BC ia he empi ical o mula χ′ = [(I
380
/I
1096
)-0.0286]/0.0065, p o-
posed by Aga wal e al. [58]. The so-ob ained χ′- alue (70.4%) was
compa able wi h ha ob ained by he same me hod o BC nanoc ys als
[59] and compa ible wi h he c ys allini y index in e ed om he XRD
pa e n.
The XRD pa e ns o composi e GO/BC, GO/BC o g aphene/BC
samples gene ally exhibi he same di ac ion peaks as pu e cellulose
[18,32,60–62]; a ely he de ec ion o an addi ional peak o igina ing
om he (001) e lec ion o g aphene oxide has been epo ed [63]. In
he p esen case, he compa ison be ween he di ac og ams o samples
p-GBC (no shown) and p-BC (Fig. 1c) e idenced he p esence o no
ex a peak asc ibable o GO. Con e sely, only he inge p in o GO
was clea ly isible in he mic o-Raman spec um o p-GBC (Fig. S6), in
ull ag eemen wi h he esul s epo ed by o he au ho s [32,60].
Di e en om he case o pu e cellulose, he compa ison among spec a
om di e en andom loca ions in he composi e sample e idenced a
g ea deg ee o spa ial homogenei y, in ull ag eemen wi h he li e a-
u e [18,60,62].
3.2. Physicochemical p ope ies o he BC-de i ed ib ous elec ode
ma e ials
Figs. 2 and S7 summa ise he esul s o SEM, MRS, XRD and XPS
analyses on he he mally p ocessed samples. Fig. 2a– shows e-
p esen a i e SEM images o he samples; lowe -magni ica ion images
can be ound in he SI (Fig. S7). As expec ed [36,47], a e ca bonisa-
ion a 700 °C, he 3D a chi ec u e o p is ine ma e ial, esul ing om
andomly-o ien ed in e connec ed cellulosic mic o ib ils, is p ese ed
and he samples exhibi he mo phology ea u ed by hie a chical po -
osi y ypical o BC-de i ed nanocomposi es [33] and o high su ace-
a ea py olysed nanoca bons [31], as well. Mo eo e , he spa ial in-
homogenei y obse ed in pu e cellulose educes o some ex en
(Figs. 2b,e and S7c).
The D- and G-bands, peculia o diso de ed/amo phous g aphi ic
nanoca bons, a e de ec ed in he mic o-Raman spec a o all he sam-
ples (Figs. 2g and S8), which con i ms he ca bonisa ion o cellulose.
The la e , asc ibed o he E
2g
symme y s e ching o all sp
2
bonded
C]C pai s [64], is ega ded as he Raman inge p in o he g aphi ic
o de ing. The o me , o igina ing om he A
1g
symme y in-plane
b ea hing o he a oma ic C ings, is diso de -ac i a ed [64]. Since i s
ela i e ( o he G-band) in ensi y ises wi h inc easing de ia ion om
pe ec hexagonally-o ganised plana ca bon ne wo k [64,65], he D/G
in ensi y a io (I
D
/I
G
) is commonly used o moni o he densi y o he
sp
2
-de ec s in nanoca bons [64,66–68].
Only a sligh inc ease o I
D
/I
G
is obse ed by compa ing he mic o-
Raman spec a o samples GBC(N
2
) and p-GBC (Fig. S8a), which in-
dica es ha he hea ea men a 700 °C on he composi e GO/BC
sample does no p oduce signi ican s uc u al changes in he ca bo-
naceous ma ix. On he con a y, he compa ison among he ou ca -
bonised samples (Fig. 2g) e idences ma ked spec al di e ences. The
“diso de -band” appea s o be he mos a ec ed by he change. I s
wid h inc eases in he o de GBC(N
2
) < BCG(N
2
) < BCG(He) <
BC(N
2
). A p og essi e weakening ( ela i e o he G-band) accompanies
he b oadening. The G-band b oadens oo. I s wid h inc eases o mino
ex en , bu in he same o de as o he D-band. Mo eo e , he band
mo es owa ds highe equencies. The inc ease in i s equency posi-
ion is in he o de GBC(N
2
) < BC(N
2
) < BCG(N
2
) < BCG(He).
Since clus e ing o he sp
2
phase, bond diso de , p esence o sp
2
ings o chains, hyb idisa ion changes o he CeC bonding ac as
compe ing o ces on he shape o he Raman spec a o nanoca bons
[64], he desc ibed spec al changes poin a di e ences in he mic o-
s uc u e o he samples. As o he D-band, Li e al. [69] epo ed
analogous (no explained) spec al di e ences o biomass-de i ed ha d
ca bon mic o-sphe ules ob ained ia ca bonisa ion a di e en em-
pe a u es. In he case o he mic o-sphe ules, he I
D
/I
G
a io inc eased
and he D-band sha pened wi h inc easing ca bonisa ion empe a u e.
The inc ease o ca bonisa ion empe a u e p oduced a simila spec al
e olu ion also in BC-de i ed ca bon ib es [44,46]. Liu e al. [44] as-
c ibed he I
D
/I
G
inc ease o he o ma ion o de ec s. On he con a y,
Dai e al. [46] explained he spec al changes in e ms o an inc ease in
he c ys alli e domains o a educ ion in he quan i y o amo phous
ca bon.
Indeed, only sp
2
ca bon de ec s con ibu e o he D-band in ensi y
[64], whe eas non-sp
2
ones ( ans-poly-ace ylene-like chains, o med a
he zig-zag edges o he de ec i e g aphi ic laye s and Csp
3
phases in-
e connec ing he g aphi ic domains) con ibu e o he Raman in ensi y
a lowe and highe equencies (T-band a 1150–1230 cm
−1
and A-
band a 1450–1530 cm
−1
, espec i ely) [70,71]. Thus, he in ensi ying
o he “sp
2
-diso de band” mo ing om sample BC(N
2
) o sample
GBC(N
2
) is he e unde s ood as he e ec o he inc ease o he sp
2
-de-
ec s wi h espec o he non-sp
2
ones, while he sh inking o he D-and
G-bands is hough o as e lec ing he e olu ion owa ds a na owe
dis ibu ion o bond angles and leng hs, which is consis en wi h wha
p oposed by Dai e al. [46]. Ac ually, only sp
2
-de ec s a e p esen in he
ca bonaceous la ice o GBC(N
2
) (Fig. S9), which appea s o be he mos
(g aphi ically) o de ed sample, despi e hei la ge densi y. Con e sely,
he highes amo phousness deg ee (i.e. he g ea es de ia ion om
pe ec Csp
2
ne wo k), wi h la ge bond (angle and leng h) diso de and
non-sp
2
de ec s, along wi h he sp
2
-ones, pe ains o sample BC(N
2
).
This is a ele an ea u e since, as known [44], he p esence o de ec s
can gene a e mo e accessible su ace a ea and cause an inc ease in he
abili y o accumula e cha ges, which is bene icial o he cha ge ans e
in he adso p ion p ocess.
The b oad di ac ion peaks associa ed o he e lec ion om (002)
and (101) g aphi ic planes a e clea ly isible in he XRD pa e ns o all
he samples (Fig. 2h), con i ming he diso de ed/amo phous na u e o
he in es iga ed nanoca bons [69,72], as well as he e ec i e he mal
educ ion o GO. Ne e heless, in sample GBC(N
2
), he mos in ense
peak is sligh ly na ow and loca ed a highe 2θ-angle alues wi h e-
spec o he emaining samples, which suppo s he conclusions d awn
om he analysis o mic o-Raman spec a (i.e. i exhibi s he highes
deg ee o g aphi ic o de ing). The a e age in e laye spacing was es-
ima ed om he angula posi ion o he (002) di ac ion peak ia he
B agg's law as d
(002)
=λ/2sinθ, whe e λ(0.15451 nm) is he wa e-
leng h o Cu-K
α
adia ion and θis he B agg angle [72]. All he calcu-
la ed d
(002)
- alues (0.35–0.39 nm) exceeded ha peculia o c ys alline
g aphi e (0.34 nm [73]). None heless, he closes alue (0.35 nm)
pe ained o “ he mos o de ed” sample, GBC(N
2
).
As expec ed [47], he XPS analysis e ealed he p esence o
(12.0–13.7 a .%) po assium in samples BC(N
2
), BCG(N
2
) and BCG(He)
(Fig. S10a). I s ela i e amoun was sub ac ed o quan i y he con-
cen a ion o he su ace oxygen unc ional species, belie ed o be
bene icial o we abili y and pseudocapaci ance o he ca bonaceous
Y. Belaus egui, e al. Desalina ion 492 (2020) 114596
5

ma e ial, as well [50,74,75]. The esul s ob ained by analysing he
high- esolu ion XPS spec a o he samples a e epo ed in Table 2.
In ull ag eemen wi h he ou comes o MRS and XRD analyses,
sample GBC(N
2
) exhibi s he lowes oxida ion deg ee (O/C a omic a io:
0.37). The ele an ela i e amoun o oxygena ed su ace unc ional
species in he emaining samples (40.7–47.1 a .%) accoun s o he
obse ed (002) in e laye spacing expansion [72]. I inc eases in he
o de GBC(N
2
) < BC(N
2
) < BCG(N
2
) < BCG(He) (Table 2), ha is in
he same o de as he equency posi ion o he G-band in he mic o-
Raman spec a (Fig. S11). Thus, al hough a con ibu ion om C]C
chains [68] canno be uled ou , he band upshi can be hough o as
he e ec o he elec on ans e om he π-s a es o he oxygen a oms
[67,70,76]. The compa ison wi h he da a ela i e o p e iously s udied
samples suppo s his conclusion.
Finally, he g ea e O/C a omic a io o BCG(He) compa ed o
BCG(N
2
) hin s a a highe capabili y by ni ogen o emo ing he mally
labile oxygena ed species, in line wi h he g ea e mass loss i p oduces
du ing he py olisa ion o ca bon ib es wi h espec o helium [77]. A
possible explana ion o his inding is ha , di e en ly om helium, a
700 °C, ni ogen eac s wi h he oxygena ed unc ional g oups p esen
on he ma e ial su ace wi h e olu ion o oxygen con aining species
(e.g. NO
x
species).
In o de o ge a deepe insigh in o he oxygena ed species p esen
on he sample su ace, he high- esolu ion pho oelec on spec a o C 1s
co e le el we e no malised o he maximum in ensi y o he main peak
and compa ed (Figs. 2h and S10). Samples BCG(He) (Fig. 2h) and
BCG(N
2
) (no shown) exhibi simila p o iles, wi h C=C/C–C bonds in
a oma ic ings (a 284.5 eV), CeO species in hyd oxyls (a 286.1 eV),
C]O species in ca bonyls (a 287.8 eV), O–C=O species in ca boxylic
g oups (a 288.9 eV), and shake-up π–π
⁎
sa elli e (a 289.9 eV)
[33,67,78] con ibu ing o he spec al in ensi y. The C 1sp o ile o
sample BC(N
2
) u he exhibi s a shoulde on he lowe binding ene gy
side o he main peak. In ag eemen wi h he indica ions in e ed om
he analysis o mic o-Raman spec a, his con ibu ion can be asc ibed
o sp-hyb idised C]C species in ans-poly-ace ylene-like chains (a
283.6 eV) [79,80]. The ela i ely high in ensi y o he shake-up π–π
⁎
sa elli e migh be indica i e o he p esence o localised π–elec ons
along he polyme chains [81]. Finally, GBC(N
2
) exhibi s a la ge e-
la i e amoun o CeO species (Fig. S10a) compa ed o he emaining
ma e ials.
3.3. Elec ochemical p ope ies o he BC-de i ed elec ode ma e ials
The p oduced BC-de i ed nanoca bons we e e alua ed as ac i e
ma e ials o CDI elec odes (Table 3) by ca ying ou CV measu emen s
in 0.1 M NaCl solu ions wi h po en ial anging be ween −1.3 and 0.4 V
and 5–100 mV s
−1
sweep a es. Fig. S12 compa es he CV cu es o he
in es iga ed elec odes a ixed po en ial scan a e ( ), whe eas Fig. 3
displays he e olu ion unde gone by he cu es o each elec ode wi h
inc easing . The measu ed cu es con i m he o ma ion o he elec-
ical double-laye (EDL) o all he elec odes. A ixed (100 mV s
−1
),
he CV cu e su ace a ea inc eases in he o de BCE-3 < BCE-
Fig. 2. Resul s o (a– ) SEM, (g) MRS, (h) XRD and (i) XPS analyses on he p ocessed samples. SEM images e e o samples (a, b, d, e) BC(N
2
), and (c, ) BCG(N
2
).
Mic o-Raman spec a a e no malised o he G-band maximum in ensi y o an easie compa ison; s a s ma k spu ious peaks o igina ing om he sample-holde in he
XRD pa e ns; high- esolu ion pho oelec on spec a o C 1sco e le el a e no malised o he main peak maximum in ensi y o an easie compa ison.
Y. Belaus egui, e al. Desalina ion 492 (2020) 114596
6
2 < BCE-1 < BCE-0 (Fig. S12). Fo a gi en elec ode (Fig. 3), he
su ace inc eases wi h inc easing , indica ing he p og essi e educ ion
o ion anspo a ion a e [82].
In o de o mo e deeply unde s and he beha iou o he p esen
elec ode ma e ials, hei elec ochemical p ope ies we e e alua ed
also by EIS analysis, a e y use ul echnique o examine capaci i e
beha iou and in e nal esis ance o he elec ode ma e ials [83], by
eco ding he eal and imagina y pa o he o al impedance as a
unc ion o he equency o he inpu signal. Fig. 4 displays he Nyquis
plo s ob ained om EIS measu emen s pe o med, in aqueous 0.1 M
NaCl solu ion, in he 1·10
−1
-1·10
4
Hz equency ange. Fo a pu e ca-
paci i e-beha iou , he Nyquis plo consis s o a s aigh line pa allel
o he imagina y axis (y-axis) in he low equency egion. Each de-
ia ion om he ideal beha iou causes he slope o he s aigh line o
diminish and app oach 45° in he Nyquis plo . The de ia ion is as-
c ibed o he inc ease o Wa bu g impedance, which is ela ed o he
kine ics o he ions di usion in he solu ion and o he adso p ion o he
ions a he elec ode/elec oly e in e ace [37,84,85].
The g aphene-en ichmen o he ac i e ma e ial seems o a ec i s
elec ochemical esponse. The beha iou o (g aphene- ee) BCE-0 a
lowe equencies (Fig. 4a) is he closes o he ideal one among all
in es iga ed elec odes, since he Nyquis plo consis s o a s aigh line
de ia ing om he imagina y axis only by 14.6°, which indica es a low
Wa bu g impedance. In he g aphene-en iched (GE) ac i e ma e ials,
Wa bu g impedance inc eases, which poin s a a non-uni o m ions
di usion p ocess a lowe equencies, whe e he ions mo emen is
pa ially hinde ed [86]. In ac , a la ge Wa bu g impedance indica es
g ea e a ia ions in ion di usion pa h leng hs and inc eased
Fig. 3. CV cu es o he in es iga ed elec odes.
Fig. 4. Nyquis plo s o he in es iga ed elec odes (inse : mid-high equency egion o he plo s).
Y. Belaus egui, e al. Desalina ion 492 (2020) 114596
7
obs uc ion o he ion mo emen [87]. The inc ease is in he o de BCE-
0 < BCE-1 < BCE-2 < BCE-3. The equi alen se ies esis ance (also
known as he in e nal esis ance, which includes he bulk elec oly e
esis ance and he cha ge ans e esis ance), es ima ed as he in e cep
o he s aigh line wi h he eal axis (x-axis), inc eases in he same
o de , namely BCE-0 (47 Ω) < BCE-1 (60 Ω) < BCE-2
(64 Ω) < BCE-3 (75 Ω).
In he elec ode BCE-1, p epa ed by he use o BC(N
2
) + GBC(N
2
)
physical mix u e, he slope changes a 1.5 Hz (Fig. 4a); a lowe e-
quencies, he Nyquis plo consis s o a s aigh line wi h a slope o
62.5°, whe eas in he middle equency egion ( om 1.5 Hz o 2 kHz), a
semici cle is isible, which indica es he o ma ion o an EDL close o
he elec ode/elec oly e in e ace [88]. This beha iou is indica i e o
he exis ence o wo di e en cha ge ans e mechanisms. I o igina es
om he coexis ence o wo dis inc componen s (namely BC(N
2
) and
GBC(N
2
)) in he hyb id ma e ial, as al eady obse ed in o he physical
mix u es [50].
Fig. 5 shows he speci ic capaci ance (C
S
) o he elec odes based on
he BC-de i ed nanoca bons, calcula ed ia Eq. (1), as a unc ion o .
Among he GE elec ode ma e ials, he wo s pe o mance pe ains o
BCE-3, which exhibi s he lowes capaci y a any a e. Besides, he
dec ease o C
S
wi h he inc ease o om 5 o 100 mV s
−1
, which gi es
a quali a i e measu e o he impac o di usion as he a e-limi ing
p ocess [50], is e y la ge (−90.1%). None heless, a 5 mV s
−1
, i s C
S
alue (131 F g
−1
) exceeds hose measu ed a he same a e o g aphene
hyd ogel and g aphene ae ogel elec odes (35 and 18 F g
−1
, espec-
i ely) in an aqueous 1 M NaCl solu ion [89]; mo eo e , i equals ha
epo ed o comme cial ca bon powde s wi h e y-high su ace a ea a
lowe a e (2 mV s
−1
) [90]. These indings sugges ha he BC-de i ed
componen o he GE-composi e p o ides a bene icial con ibu ion o
he o e all capaci ance. As known [47], upon ca bonisa ion a 700 °C,
he dense 3D ib ila BC skele on con e s in o po ous ca bon wi h
in e connec ed mac opo es (> 50 nm), on whose walls mesopo es
(2–50 nm) and mic opo es (< 2 nm) de elop. The o me se e as
pa hways o e icien ion anspo ac oss he elec ode, he la e a e
bene icial o ion s o age [82].
Compa ed o BCE-3, elec ode BCE-2 exhibi s supe io capaci y and
educed impac o di usion as he a e-limi ing p ocess (C
S
-dec ease
wi h -inc ease is −84.0%). As he p epa a ion o he wo ac i e ma-
e ials di e s only o he ca bonisa ion a mosphe e (He in place o N
2
),
he la ge amoun o oxygena ed unc ional moie ies p esen on
BCG(He) (Table 2)pa ly jus i ies he obse ed imp o emen , since hey
con ibu e o imp o e bo h he we abili y and pseudocapaci ance o
he elec ode [50,74,75]. The speci ic capaci ance o he hyb id
elec ode (BCE-1) is compa able o ha o BCE-2 o ≥ 20 mV s
−1
,
whe eas below his a e i is smalle . In BCE-1, he impac o ion di -
usion as he a e-limi ing p ocess u he lowe s (as inc eases om 5
o 100 mV s
−1
,C
S
diminishes by −75.3%).
BCE-0 ou pe o ms all he GE-elec ode ma e ials, e en in e ms o
C
S
-diminishing wi h he scan a e inc ease (−68.9%). This esul is
opposi e o ha p e iously ob ained o elec ospun ca bon ib es de-
i ed om polyac yloni ile (PAN), which bene i ed om he g aphene-
en ichmen , in spi e o he speci ic su ace a ea con ac ion ha i
caused [11]. The g ea e C
S
- alues o elec ode BCE-0 wi h espec o
he GE-ones sugges ha g aphene-en ichmen o he ac i e ma e ial
( ega dless o he p epa a ion s age a which i is ca ied ou ) ails in
imp o ing he ma e ial elec ochemical pe o mance. A eason o his
di e ence may lie in he expe imen al p ocedu e o he g aphene-en-
ichmen . In he case o elec ospun ca bon ib es, GO was inco po a ed
in o he p is ine solu ion (i.e. be o e he gene a ion o he ib es ia
elec ospinning) [11], which ul ima ely esul ed in an in ima e con ac
be ween GO and he PAN-de i ed ca bonaceous ib es and in a
homogeneous dis ibu ion o i in he ib ous ne wo k. On he con a y,
in he p esen case, GO was added on o he su ace o he al eady syn-
hesised BC, h ough imme sion o imp egna ion me hod (Fig. S2),
which esul ed in a less uni o m dis ibu ion o GO in he bulk o he
GE composi es and a less in ima e con ac be ween i and he BC-de-
i ed ca bonaceous ma ix.
F om he compa ison wi h he li e a u e, i eme ges ha he speci ic
capaci ance o elec ode BCE-0 a 5 mV s
−1
(366 F g
−1
) is g ea e han
ha measu ed by Liu e al. [44], a 2 mV s
−1
in an aqueous 1 M NaCl
solu ion, o an elec ode based on comme cial BC ca bonised a 800 °C
(297 F g
−1
). The be e pe o mance o he p esen BC-de i ed nano-
ca bon migh depend on he BC biosyn hesis condi ions (e.g. bac e ial
s ain, cul u e medium), as well as on he ca bonisa ion empe a u e,
which is c ucial o he po e s uc u e and mic os uc u e o BC-de i ed
ca bon ma e ials [47]. Using aqueous 6 M KOH solu ion as he elec-
oly e, Zhou e al. [91] ob ained a simila ly high C
S
- alue (360 F g
−1
)
o supe capaci o s based on po ous ca bon ib es, p oduced by block
copolyme mic ophase sepa a ion, exhibi ing an in e connec ed ne -
wo k o po es wi h bimodal dis ibu ion, inclusi e o mesopo es
(~10 nm) and sub-mic opo es (~0.5 nm). Ding e al. [47] epo ed an
e en la ge C
S
- alue (430 F g
−1
) o a symme ic supe capaci o using
1 M H
2
SO
4
solu ion as he elec oly e and an ul a-mic opo ous dual-
doped ca bon, possessing highly concen a ed mic opo es (~ 2 nm) and
a conside able amoun o sub-mic opo es (< 1 nm), as he ac i e ma-
e ial. Sub-mic opo es, which signi ican ly enhanced he capaci i e
pe o mance due o po e con inemen e ec [47], esul ed om he
simul aneous ac i a ion/ca bonisa ion o he dense BC p ecu so , wi h
po assium hyd oxide ac ing as he ac i a ing agen .
In sample BC(N
2
), he analysis o he ni ogen adso p ion/deso-
p ion iso he ms e idenced he p esence o sub-mic opo es wi h
size < 0.5 nm, as epo ed by Ding e al. [47]. The speci ic mic o-po e
olume was 0.206 cm
3
g
−1
; he speci ic su ace a ea was 340 m
2
g
−1
.
Al hough compa able wi h hose epo ed o o he BC-de i ed nano-
ca bons [44], he la e alue could be unde es ima ed [92]. In he
composi es only a sligh inc ease o he po e size and a sligh educ ion
o speci ic su ace a ea we e obse ed, in ull ag eemen wi h wha
p e iously obse ed in elec ospun ca bon ib es [11] and wi h he
esul s epo ed in he li e a u e, as well [93].
Ou s anding s o age and desalina ion p ope ies ha e ecen ly been
epo ed o ca bons simul aneously ha ing a hie a chically po ous
s uc u e consis ing o in e connec ed mac o-/meso-/mic o-po es and
high (in insic and he e oa om-induced) de ec densi y [94,95]. Be-
sides, he p edominan in luence o he ca bon de ec s, wi h espec o
he unc ional g oups, on capaci ance has been poin ed ou [96]. As a
ma e o ac , BC(N
2
) possesses he hie a chical po osi y peculia o
BC-de i ed ca bons and, besides, he highes amo phousness deg ee,
wi h a ple ho a o di e en mo phological de ec s (Fig. S9) and he
la ges ac ion o he highly eac i e [70] non-sp
2
de ec s, including
Fig. 5. Speci ic capaci ance o he in es iga ed elec odes as a unc ion o po-
en ial scan a e.
Y. Belaus egui, e al. Desalina ion 492 (2020) 114596
8
Csp
3
species in e connec ing he g aphi ic domains and chains wi h
localised s a es a he edge-si es. The syne gy be ween hese ac i e si es
and he ma e ial po osi y accoun s o i s ema kable elec ochemical
pe o mance [97].
3.4. Co ela ions be ween elec ochemical p ope ies and de ec i eness o
he elec ode ma e ials
In he las decades, inc easing he g aphi isa ion deg ee and de-
c easing he de ec densi y (e.g. by enhancing he ca bonisa ion em-
pe a u e) has ep esen ed he mos common s a egy o imp o e he
elec ochemical pe o mance o ca bon-based ma e ials. Ve y ecen ly,
in coun e - endency, se e al au ho s ha e poin ed ou he c ucial ole
o de ec s in ion s o age and emphasised he bene i s de i ing om
p oducing de ec - ich nanoca bons [92–95,97–100]. Jus o ci e a ew
o cases, Guo e al. [99] ha e epo ed ha ha d ca bon nano ib es wi h
a de ec - iche ex u e exhibi be e a e capabili y and no iceably
longe cycle pe o mance. Yao e al. [97] ha e shown ha de ec s a he
g aphene laye edges in mic opo ous so ca bons p o ide ex a so-
dium-ion s o age si es. Lu e al. [100] ha e a ibu ed o he highe
de ec concen a ion he enhanced Na s o age pe o mance o so
ca bon p epa ed a lowe ca bonisa ion empe a u e. No el models
in ol ing Na-ion s o age a de ec si es, a he han in e cala ion be-
ween g aphene shee s, ha e been u he p oposed o explain he so-
dia ion p ocess in non-g aphi isable ca bons [98].
In he p esen case, he inc ease o he speci ic capaci ance mo ing
om he leas de ec i e elec ode ma e ial (BCE-3) o he de ec - iches
one (BCE-0) p o es ha he ca bon de ec si es play a c ucial ole in he
ion adso p ion p ocess, in line wi h he mos ecen assessmen s on hei
abili y o gene a e mo e accessible su ace a ea, a ou ing he cha ge
accumula ion [44].
I can be en isaged ha , a low scan a es, ime o he ions o he
elec oly e is enough o di use in o he inne po es o he elec odes
[50], whe e hey a e adso bed a he de ec -si es, ega dless o hei
ype. Ac ually, a good co ela ion is ound by plo ing he speci ic ca-
paci ance a 5 mV s
−1
scan a e (Fig. 6a) agains he ela i e in ensi y o
all he Raman “de ec -bands” (I
T
+I
D
+I
A
)/I
G
. This indica es ha all
ypes o de ec s con ibu e o he ion adso p ion and o he o ma ion o
EDL a he lowes po en ial scan a e. The non-linea i y o he ound
ela ionship migh e lec he di e en chemical eac i i y o he a -
ious ype o de ec s [68,70,101,102]. Wi h inc easing scan a e, he
di usion becomes a e-limi ing [103] and he mo ing ions do no ha e
ime enough o accumula e on o he less accessible su ace egions,
which a e p og essi ely excluded [50]. Unde hese condi ions, he ion
adso p ion akes place p e e en ially a he mos eac i e non-sp
2
si es
on he po e su ace, e.g. single acancies and dandling bonds a zig-zag
edges [101,102]. Ac ually, he speci ic capaci ance a 100 mV s
−1
scan
a e (Fig. 6b) linea ly co ela es wi h he in ensi y a io, (I
T
+I
A
)/I
D
, o
he bands o igina ing om non-sp
2
and sp
2
de ec s. The inc ease in he
densi y o he non-sp
2
de ec s wi h espec o sp
2
ones is able o accoun
also o he dec ease in in e nal esis ance om elec ode BCE-3 o BCE-
0 (Fig. S13a), as well as o he a ia ion o C
s
wi h he inc ease o ν. Ion
di usion has ligh e impac as he a e-limi ing p ocess (i.e. less nega-
i e alue o he C
s
- a ia ion) in ma e ials wi h a la ge ela i e amoun
o non-sp
2
de ec s, in line wi h he ou comes o compu a ional s udies
on he eac i i y o ca bon de ec s wi h sodium ions [104].
Rema kably, he da a ela i e o he elec odes based on PAN-de-
i ed ca bon ib es p e iously in es iga ed [11] show he same end
(Fig. S13), e en i , ob iously, hey canno line-up along he cu es o
BC-de i ed nanoca bons owing he la ge di e ences in he o he in-
luen ial ac o s (e.g. speci ic su ace a ea, concen a ion and ype o
su ace unc ional species).
3.5. Elec oso p i e p ope ies o he BC-de i ed elec ode ma e ials
Typically, ma e ials possessing supe io capaci i e pe o mance
exhibi enhanced elec oso p i e p ope ies. Fig. 7a displays he sal
adso p ion capaci y, SAC, o he elec odes BCE-0, BCE-1 and BCE-2, as
calcula ed ia Eq. (2), a a ious NaCl concen a ions. As a gene al
end, as he concen a ion inc eases, he amoun o NaCl emo ed om
he solu ion aises as a esul o he enhanced mass ans e a e o ions
inside he mic opo es and o he educed EDL hickness and o e -
lapping e ec s [11,105,106]. In he p esen case, he inc ease o sal
concen a ion om 585 mg L
−1
(10 mmol L
−1
) o 11.7 g L
−1
(200 mmol L
−1
) causes SAC- alues o inc ease om 60.7 mg g
−1
o
1.35 g g
−1
o elec ode BCE-0, om 55.0 mg g
−1
o 1.03 g g
−1
o
elec ode BCE-1 and om 79.0 mg g
−1
o 1.11 g g
−1
o elec ode BCE-
2. The compa ison wi h he li e a u e (Table 4) clea ly p o es ha hese
ep esen unp eceden ed elec oso p ion capaci ies o ca bon-based
elec odes. A 585 mg L
−1
and 1.2 V, p esen elec ode ma e ials a e
able o emo e a g ea e amoun o sal (55.0–79.0 mg g
−1
) compa ed
o ha emo ed a he same concen a ion and applied ol age by PAN-
de i ed C ib es (13.7–26.8 mg g
−1
), and a 500 mg L
−1
and 2.0 V by
g aphene ae ogel and g aphene hyd ogel (45.9–49.3 mg g
−1
). A he
ini ial concen a ion o 58.5 mg L
−1
(1 mmol L
−1
), he amoun o NaCl
adso bed on elec ode BCE-0 (8.3 mg g
−1
) is nea ly wice ha adso bed
on CNT sponge a compa able concen a ion (4.3 mg g
−1
). The smalle
speci ic su ace a ea o PAN-de i ed C ib es, g aphene ae ogel and CNT
sponge migh be esponsible o hei lowe SAC- alues.
In e es ingly, below 5.85 g L
−1
,SAC inc eases in he o de BCE-
1 < BCE-0 < BCE-2. This s ongly poin s a he c ucial ole o he
elec ode we abili y a lowe solu e concen a ions. Ac ually, a linea
co ela ion (Fig. 7b) is ound be ween SAC- alues a 585 mg L
−1
(10 mmol L
−1
) and he concen a ion o su ace unc ional species
(Table 2), which, as known [11,50,74,75], imp o e he we abili y o
he ca bonaceous elec ode by he elec oly e and hei pseudocapaci-
ance. In he p esence o lowe mass ans e a e o ions and highe EDL
hickness and o e lapping e ec s, he we abili y/accessibili y o he
elec ode su ace and pseudocapaci i e eac ions become ele an ac-
o s o he ion adso p ion.
A 5.85 g L
−1
(100 mmol L
−1
), he elec oso p i e pe o mance o
elec odes BCE-0 and BCE-2 is e e sed, wi h SAC o BCE-0 exceeding
ha o BCE-2. This sugges s ha , a highe sal concen a ions, he
de ec i eness o he ac i e ma e ial becomes he mos in luen ial
p ope y, as i con ols he ela i e amoun o ac i e si es a ailable o
he elec os a ic adso p ion. SAC- alues a 5.85 g L
−1
well co ela e
wi h non-sp
2
o sp
2
de ec densi y a io (Fig. 7c). Despi e he la ge
di e ences in he mani old ac o s in ol ed in he p ocess, he da a
ela i e o he elec odes based on PAN-de i ed ca bon ib es [11],
show he same end (Fig. S14), hus suppo ing hese conclusions.
4. Conclusions
BC is u ilised as a ca bon p ecu so o he p epa a ion o elec ode
ma e ials o he CDI o wa e , which exhibi unp eceden ed desalina-
ion capaci ies o nanoca bons (55–79 mg g
−1
o sal a an ini ial
concen a ion o 585 mg L
−1
).
The ho ough in es iga ion o he physico-chemical and elec o-
chemical p ope ies o he ma e ials p oduced allows d awing he ol-
lowing conclusions:
•The BC-de i ed nanoca bons exhibi a h ee-dimensional a chi-
ec u e ea u ed by hie a chical po osi y, di e en densi y and ype
o la ice de ec s and high con en o su ace oxygen species
(45–54 w %).
•The ca bon ne wo k de ec i eness s ongly a ec s he elec o-che-
mical beha iou o he elec odes, whose speci ic capaci ance in-
c eases wi h he de ec densi y. All ypes o la ice de ec s pa ici-
pa e in ion s o age a lowe scan a es. As he scan a e inc eases, he
di usion becomes a e limi ing and he ions, ha ing no ime en-
ough o mo e, accumula e p e e en ially a he mos eac i e non-
sp
2
si es on he po e su ace.
Y. Belaus egui, e al. Desalina ion 492 (2020) 114596
9