7/27/2019 Colunas Para Teste de Carbon Granular NORIT
1/12
GRANULARACTIVATED
CARBONEVALUATION
7/27/2019 Colunas Para Teste de Carbon Granular NORIT
2/12
Introduction.
NORIT Americas Inc. produces over 150 different varieties
of activated carbon. This diverse product line provides many
options for carbon adsorption. Pilot Column testing can aid
in the process of choosing the best carbon for your purifica-
tion needs.
7/27/2019 Colunas Para Teste de Carbon Granular NORIT
3/12
In liquid phas e a pplica tions, a ctivated ca rbon is used
to remove impurities from potable water, waste-
water, fine chemicals, foods, and pharmaceuticals.
This tec hnica l pa per add ress es how to evalua te
granular activated carbons (GAC) for liquid phase
applications.
The first eva lua tion ste p is to run simple isotherms
(se en below) to dete rmine fea sibility. P ow de red
ca rbon isotherms c an b e used to determine if the
trea tment go a l ca n be met. The proc edure for run-
ning isotherms is available by requesting brochure
number NA 00-3 from NORITAmericas Inc.
Isotherm testing consists of dosing a measured
quantity of powdered carbon into the target solution.Impurities ad so rb o nto the ca rbons s urfac e a nd
are allow ed to rea ch e q uilibrium. While this tes t
provides an indication of how well an impurity may
be removed using activated carbon, it cannot give
definitive s ca le-up da ta for a g ra nula r ca rbon o per-
ation due to a couple of factors:
In a granula r column, dyna mic a ds orption oc curs
because the impurity concentration changes as the
mas s trans fer zone moves through the bed . This
cond ition d oes not exist in a n isotherm test proc e-
dure. Reg eneration effects on multiple cycles ca n-
not be studied s ince the iso therm tes t procedure
uses powdered carbon (only) in a once-through
adsorption process.
Bec aus e of these factors, pilot c olumn tests s hould
be made using the most promising carbons as
indica ted b y the iso therms. This w ill give a mo re
accurate comparison of the carbons, and provide
some scale-up information useful for commercial
system designs.
This text d es cribe s the proce dure for running a
G AC pilot c olumn test a nd eva luating the d a ta.
Data ma nipula tion tec hniq ues a re provided which
allow conversion from test conditions to widely
differing flow rates, effluent concentrations, and
influent co ncentra tion levels. This proced ure ha s
been s ucce ss fully used numerous times to de sign
full scale systems and estimate operating costs.
Isotherm series
7/27/2019 Colunas Para Teste de Carbon Granular NORIT
4/12
GAC Pilot Columns.
A pilot column test system should include at least four
columns operating in series in order to determine the rate of
movement and size of the mass transfer zone through the
carbon bed, (see Figure 1). Pilot column systems can be pur-
chased or leased from NORIT Americas Inc. or assembled
using basic components. They can be constructed from
glass, plastic, reinforced fiberglass, or metal pipe as dictated
by the corrosiveness of the process fluid. The columns should
be designed to allow loading GAC from the top.
Figure 1
PILOT COLUMN SYSTEM
7/27/2019 Colunas Para Teste de Carbon Granular NORIT
5/12
P ilot c olumns a re typica lly ope rated dow nflow.
This is usua lly do ne to s imula te a proposed c om-
mercial de sign. A do wnflow d es ign allow s la rge
G AC vess els to be gra vity fed. It also ensures tha t
the carbon bed remains packed and stable during
the service cycle, resulting in maximum contactbetween the carbon and the feed stream. However,
ca re s hould be ta ken to e nsure the c olumns remain
filled with liquid at all times and are not allowed to
dra in during ope ration. This is often acc omplishe d
by fitting the la st co lumn with a ba ck pressure valve.
In some situations an upflow design is preferable.
One advantage is where suspended solids in the
feed create high press ure drop. It is s ugges ted
that s uspended so lids be removed from the feedstrea m prior to entering a G AC c olumn. How ever,
if this will not be possible in the commercial design,
then the effect of suspend ed s olids should b e
inc lude d in the pilot run. In upflow ope rat ion,
most o f the suspended solids wo rk their wa y up
through the G AC bed without a significa nt increas e
in press ure d rop.
It is recomme nded that the pilot c olumns be ba ck-
washed prior to the start of any test and when
exc essive press ure drop deve lops . This would
involve the w a ter or column-trea ted proc ess stream
to be pumped through the bo ttom of the c olumn
a t a flow rate sufficient to expand the bed. Enough
freeb oa rd sho uld be left to permit 30-50% be d
expans ion during ba ckwa shing. Figure 2 show s
bed expansion for several sizes of Granular DARCO
products in wa ter. The exact ba ckwa sh flow rate
will depend upon the visc osity of the wa ter and the
mea n particle diameter of the GAC. The expans ion
ca n a lso be d etermined visua lly in a glass column.
The ca rbon bed sho uld b e a t lea st 24 inches (61
cm) deep and 1.5 inches (about 4 cm) in diameter.
A smaller column is not recommended since the
wa ll effect beco mes significa nt. NORITAmerica s
Inc. typica lly us es 4 inch (10 cm) dia meter c olumns
in its pilot studies.
The bed c an be supported by g la ss wo ol, wire mesh
sc reen, or a w edg ewire sc reen. A typica l sc reen
size is 60 mesh (U.S. Standard Series) or a 0.008
slot opening.
Figure 2
BED EXPANSION CURVE FOR GRANULAR
DARCO ACTIVATED CARBONS
7/27/2019 Colunas Para Teste de Carbon Granular NORIT
6/12
Selecting Carbon Mesh Sizes.
In the design of a full scale column, GAC mesh
size is very importa nt. NORITAmerica s Inc. offers
DARCO and NORIT brand granular a nd extruded
a ctivate d ca rbons in over a d ozen sizes . Initia lly,
finer mesh carbons show more adsorptive capacitythan coarser carbons because of a significantly
fas ter rate of ads orption. How ever, given time to
reach equilibrium, the coarser carbons will have
the sa me total ca pac ity as the finer ca rbons . The
increas ed a ds orption rate of the s ma ller mesh s ize
ca rbons, how ever, is a t the expense of pressure
drop during c olumn opera tion. The press ure drop
for a 12 x 40 mesh ca rbon is tw ice that of a 12 x
20 carbon, while the pressure drop of a 20 x 40
ca rbon is o ver 3 times tha t of the 12 x 20. Thebes t mes h size for any s pecific a pplica tion w ould
depend upon the dimensions of the ca rbon column
(vesse l), the inlet s tream flow rate &visc os ity, and
the sus pended solids conc entra tion of the inlet
strea m. Figure 3 sho ws the pres sure drop curves
for some mes h sizes of G ranula r DARCO.
Loading the Column.
The c a rbon to be used in the c olumn should first
be thoroughly wetted . This is done to remove air
tra pped in the ca rbon pores tha t wo uld otherwise
prevent the feed liquor from contacting the entire
ca rbon surfac e. In sma ll lab oratory columns, the
carbon is typically slurried with water in a container
and a llowed to soa k. So aking should las t at leas t
30 minutes to a n hour to a llow mos t of the trapped
a ir to diffuse out of the sma lles t pores. In la rge
comme rcial vess els, the c arbon is often left to soa k
overnight prior to ba ckwa shing a nd b eing place d in
service.
Once the ca rbon has soa ked, the fines must be
remove d. C a rbon fines a re pres ent in all gra nula r
a ctiva ted ca rbons. They are genera ted not only in
the ma nufa cturing proces s, but a lso through s hip-
ping a nd handling. If not remove d during loa ding,
fines can cause excessive pressure drop across the
ca rbon column and s horten the c olumns service
life. When loa ding sm a ll co lumns, the c a rbon fines
may be removed b y dec anting, ad ding fresh w ater
and allowing the carbon to settle, and decanting
aga in. This should be repea ted until the wa ter is c lea r,
usua lly 3 or 4 times . In la rge c omme rcia l vesse ls,
the carbon fines are removed by backwashing.
Feed Delivery and Flow Rate Control.
The rec omme nded rang e of conta ct flow is 0.1-3.0
be d vo lumes per hour (B V/hr), depe nding upo n the
deg ree of purifica tion de sired, the type a nd c on-
centration of impurity, the nature of the process
liquor, and pressure drop.
G enerally high levels of impurity loa ding o n the
carbon, high inlet impurity concentrations, andhigh visc os ities w ill req uire a low er inlet flow rate.
The c a rbon w ill perform mo re efficiently a t low flow
rates (longer co ntac t times), but a t the expense o f
the amo unt of liq uid tha t ca n be proces sed through
a column in a given time.
A feed pump s uitab le for ac curate a nd c ontinuous
flow is required since flow rate and total volume of
liquor are the most important controllable variables
in deve loping design da ta . B efore proces s liq uor is
delivered to the ca rbon, suspend ed ma tter should
be removed, preferably by the sa me method
planned for the pla nt sys tem.
Figure 3
PRESSURE DROP CURVES FOR
GRANULAR ACTIVATED CARBONS
7/27/2019 Colunas Para Teste de Carbon Granular NORIT
7/12
Column Operation.
To d evelop reas ona bly good da ta for sca le-up to
full pla nt de sign, it is importa nt to have the p ilot
columns operating as near as possible to the
a nticipa ted plant opera ting cond itions . The mos t
critica l fa ctors, flow ra te a nd feed impurity c oncen-tration, must be constant for the entire test run.
The tempe rature of the full-sca le influent a nd
effluent streams sho uld a lso b e simula ted . This
ma y req uire prehea ting the feed a nd hea t tape &
insulation on the G AC pilot c olumns.
Se veral sa mples o f the feed s trea m should b e ta ken
over the duration of the test to spot any drift in
impurity co nce ntration. Reg ula r sa mples should
be pulled from the botto m of eac h column. Thesamples should be filtered prior to examination if
the acc urac y of the ana lysis could b e a ffected b y
sus pend ed s olids . When the effluent qua lity from
the la st column in the s eries is unacc eptab le, the
test is s topped.
Data Collection.
The d a ta ca n be co mpiled simila r to Figure 4, which
show s the results o f a pilot co lumn study on a n
a ctua l refinery wa ste wa ter. Effluent impurity con-
centrations for all columns a re plotted a ga inst
elapsed time in Figure 5, generating breakthrough
curves. Figure 5 is a plot show ing w hen eac h col-
umn brea kthrough curve exc eed ed the effluent limit
of 30% COD rema ining . The line through the da ta
points in Figure 6 is the B ed Depth S ervice Time
(B DS T) curve. If the flow rate through the G AC is
cons tant throughout the test a nd there a re no
changes in impurity concentration or composition,
then the service time at breakthrough should be a
linear function of carbon bed depth.
Figure 5
BREAKTHROUGH CURVES
FOR REFINERY WASTE
Figure 4COLUMN EXHAUSTION DATA
COLUMN1 COLUMN 2 COLUMN 3 COLUMN 4TIME % COD % COD % COD % CODHRS. REMAINING REMAINING REMAINING REMAINING
25 20 17 13 1250 25 18 14 1175 33 21 16 12
100 45 22 15 13125 62 26 15 11150 67 27 17 13175 76 31 19 12200 80 33 19 13225 85 40 18 14250 88 44 24 16275 91 45 27 18300 96 48 28 20325 98 55 32 24350 97 59 34 24375 100 64 35 27400 99 68 37 26425 100 70 41 29450 100 75 45 31475 97 74 46 34500 100 78 49 34525 100 76 48 35550 98 78 49 37
575 100 76 50 34600 100 79 51 36
7/27/2019 Colunas Para Teste de Carbon Granular NORIT
8/12
Data Evaluation.
The b ed de pth s ervice time (BDS T) eq ua tion is
ba sed on work by B ohart and Ada ms,1 Klotz,2
a nd Dole and Klotz.3 It ha s the follow ing form:
t = ax + b
where t Service time a t breakthrough, hours .
x Bed depth, feet.
a S lope = (1990No /Co)V, hrs ./ft.
b Ord ina te inte rc ept =
16.018/(KC o) ln(Co /C b 1), hrs.
No Carbon efficiency, lb impurity/ft3 carbon.
Co Feed impurity concentration, ppm.
V Linear flowrate , gpm/ft2.
K Adsorp tion ra te constant , ft3
of liq uidtreated per lb of impurity fed to the
system per hr., required for the adsorp-
tion wave front to move through a
bed depth eq uivalent to the c ritica l
bed d epth.
Cef Impurity concentration in effluent at
breakthrough, ppm.
The s lope (a ) of the be d d epth service time line in
Figure 6 represents the time required to exhaust
one foot of the c a rbon b ed (under the tes t co ndi-
tions use d), w hich c an a lso be d efined a s the time
required for the adsorption wave front to move
through one foot of the ca rbon bed . The rec iproca l
of the slope is the ra te a t which the ca rbon bed is
spent. An estima te of the ra te at w hich ca rbon is
used to continuously produce acceptable product
liquid (or the rate at which the carbon must be
a dd ed o r regenerated ) ca n be ma de b y multiplying
this rec iproc a l va lue of the slope b y the bulk density
of the carbon.
The BDS Teq uation c an a lso be w ritten in terms
of bed d epth a s a funct ion of time. This a llow s
the abscissa intercept to be calculated as follows:
Xo = -b/a
Xo is the abscissa intercept and represents the
critica l bed de pth. This is de fined a s the minimum
bed depth required to produce a satisfactory effluenta t time zero. The ordinate interce pt (b) represe nts
the rate of adsorption, or the time required for the
a ds orption wa ve front to pa ss through one c ritica l
bed d epth.
Breakthrough points must b e s pecified for eac h
BDS Teq uation that is de veloped. Obviously, for a
given bed depth, the s ervice time a nd the s lope of
the B DST curve w ill decreas e a s the effluent qua lity
requirements increase.
7/27/2019 Colunas Para Teste de Carbon Granular NORIT
9/12
Changing Process Conditions.
A new B DST eq uation must b e developed if the
linear flow ra te through the bed c hange s. P lotting
a new BDSTcurve will not require additional col-
umn tests. Rather, the new BDS Teq uation can be
ob ta ined simply b y m ultiplying the o riginal s lopevalue by the ra tio of the two rates. The new eq uation
for a cha nge in flow rate is d eveloped a s follow s:
t = V1/ V2ax + b
w here V1 Original flow rate.
V2 New flow rate.
Our experience indicates that the results obtained
by this simplified method w ill be in clos e a greementwith results obta ined b y tes ts pe rformed under
a ctua l flow -ra te c onditions (se e Figure 7). It sho uld
a lso be noted that c hang ing flow rates have little
effect on the ordinate intercept (rate of adsorption).
A new BDSTequation must also be developed if
the c onc entration of the impurities in the feed
cha nges . Aga in, a goo d approxima tion ca n be
obtained by calculation instead of running additional
tes ts. The new s lope is found by m ultiplying the
original slope by the ratio of the impurity concen-
tration of bo th feeds . To find the new o rdinate
intercept, multiply the value of the original intercept
by the same concentration ratio and by the natural
log of the impurity co ncentration. The new eq uation
will have the following form:
t = C1/C2 ax + (C1/C2) ln(C2/Cef 1) b
ln(C1/Cef 1)
w here C1 Original influent c onc entration, p pm.
C2 New influent concentration, ppm.
Goo d a greement is obtained b etween ca lculated
and ac tual values.
Figure 6
BED DEPTH SERVICE TIME RELATIONSHIP
Figure 7
EFFECT OF FLOW RATE CHANGE
ON THE BDST CURVE
7/27/2019 Colunas Para Teste de Carbon Granular NORIT
10/12
Example Problem.
The da ta displayed in Figure 4 wa s p lotte d in Figure
5 to c rea te b reakthrough c urves for the pilot s ystem.
In our example, 30% COD remaining was selected
a s the effluent-limiting c onc entration. Thus, w e ca n
rea d from these curves the breakthrough times a tthe various bed depths:
Bed ServiceColumn Depth, ft. time, hr.
1 3.3 70
2 6.6 160
3 9.9 340
4 13.2 500
These a re plotted in the form of the BDS T eq uation
des cribe d a bo ve. The line through the break-through points generates the eq uation which is
plotted in Figure 6:
t = 44.5x 100
This c a n be a djuste d to reflec t cha nges in operating
co nditions. For exa mple, if the flow rate is d oubled,
the slope of the line is halved and the intercept
rema ins the sa me. The new eq uation bec omes :
t = 22.3x 100
This line, a long w ith a n a ctua l B DS Tline de rived by
physically running the system at twice the flow rate,
is a lso s how n in Figure 7. New eq uations reflec ting
cha nges in influent concentrations c a n also b e
derived by the method o utlined a bove. Examples
of these c hange s a re s hown in Figure 8.
7/27/2019 Colunas Para Teste de Carbon Granular NORIT
11/12
System Sizing.
For a fixed bed adsorber, the necessary carbon
bed depth for the desired s ervice life ca n be rea d
direc tly off of the BDS T curve (Figure 6) or ca n be
ca lcula ted using the BDST eq uation. This is a ll
that is neces sa ry to d etermine sys tem size unlessca rbon regeneration is us ed.
Activa ted c arbon c a n be reac tivated therma lly or
in some ca ses c hemica lly regenerated . In most
applications, thermal reactivation is used to restore
most of the virgin GAC a ds orptive c apa city.
However, c hemica l regeneration with ca ustic, a cid,
or solvents ma y be much more cos t-effective
depending on the nature of the impurities on the
ca rbon. If a chemical regeneration scheme is tobe implemented, the regeneration solution should
be run counterflow to the direction of the normal
se rvice flow. Reg a rdless of the regenerant flow
direction, some o f the GAC a ds orptive ca pa city is
los t with ea ch regenera tion/rea ctivation cyc le. If
use of regenerated or rea ctivated G AC is a n option,
it is very important to tes t the e ffect in the pilot
columns. Many systems whose des igns are ba sed
on virgin GAC da ta are undersized when ope rated
with reactivated carbon.
For additional assistance with GAC system design,
plea se conta ct NORITAmerica s Inc.
References.1 Boha rt, G. S ., and Ada ms, E. Q.,Journal of
American Chemical Society, 42; 523-544, 1920.2 Klotz , I. M., Chemical Reviews, 39; 241-268, 1946.3 Dole, M., and Klotz, I. M., Industrial and
Engineering Chemistry, 38; 1289-1297, 1946.
Figure 8
BDST RELATIONSHIP FOR DIFFERING
INFLUENT CONCENTRATIONS
7/27/2019 Colunas Para Teste de Carbon Granular NORIT
12/12
2001 NORIT NA00-4 2000
www.NORIT.com
NORIT Nederland BV
Amers foort, The Nethe rla nds
Telephone : 31334648911
Telefa x: 31334617429
NORIT (U.K.) Ltd.
Glasgow, Scotland
Telephone : 441416418841Telefa x: 441416410742
NORIT (France) S.a.r.l.
Paris, France
Telephone : 33145910808
Telefa x: 33148673603
NORIT Italia S.p.A
Ravenna, Italy
Telephone : 39544451514
Telefa x: 39544451283
NORIT Deutschland G.m.b.H.
Dsseldorf, G ermany
Telephone : 49211906020
Telefa x: 49211161115
N.V. NORIT Belgium S.A.
Brussels, Belgium
Telephone : 3226750645Telefa x: 3226751119
NORIT (Japan) Co. Ltd.
Mina to-Ku, Tokyo, J a pa n
Telephone : 81352952850
Telefa x: 81352952860
NORIT Singapore Pte. Ltd.
S inga pore, S inga pore
Telephone : 657353066
Telefa x: 657353166
NORIT Americas Inc.
3200 West University Avenue
Marsha ll, TX 75670
800-641-9245
Telephone: 903-923-1000
Telefa x: 903-923-1003
e-mail: info@norit-america s.c om