2009-Characterization of Biochar From Fast Pyrolysis and Gasification Systems

Characterization of Biocharfrom Fast Pyrolysis andGasification SystemsCatherine E. Brewer,a,b Klaus Schmidt-Rohr,c Justinus A. Satrio,a and Robert C. Brownaa Center for Sustainable Environmental Technologies (CSET), Iowa State University, Ames, IA 50011; [email protected](for correspondence)b Department of Chemical and Biological Engineering, Iowa State University, Ames, IA 50011c Department of Chemistry, Iowa State University, Ames, IA 50011

Published online 5 August 2009 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ep.10378

Thermochemical processing of biomass produces asolid product containing char (mostly carbon) andash. This char can be combusted for heat and power,gasified, activated for adsorption applications, orapplied to soils as a soil amendment and carbonsequestration agent. The most advantageous use of agiven char depends on its physical and chemicalcharacteristics, although the relationship of charproperties to these applications is not well understood.Chars from fast pyrolysis and gasification of switch-grass and corn stover were characterized by proxi-mate analysis, CHNS elemental analysis, Brunauer-Emmet-Teller (BET) surface area, particle density,higher heating value (HHV), scanning electron mi-croscopy, X-ray fluorescence ash content analysis,Fourier transform infrared spectroscopy using aphoto-acoustic detector (FTIR-PAS), and quantitative13C nuclear magnetic resonance spectroscopy (NMR)using direct polarization and magic angle spinning.Chars from the same feedstocks produced under slowpyrolysis conditions, and a commercial hardwoodcharcoal, were also characterized. Switchgrass andcorn stover chars were found to have high ash con-tent (32–55 wt %), much of which was silica. BETsurface areas were low (7–50 m2/g) and HHVsranged from 13 to 21 kJ/kg. The aromaticities fromNMR, ranging between 81 and 94%, appeared toincrease with reaction time. A pronounced decreasein aromatic C��H functionality between slow pyroly-

sis and gasification chars was observed in NMR andFTIR-PAS spectra. NMR estimates of fused aromaticring cluster size showed fast and slow pyrolysis charsto be similar (�7–8 rings per cluster), while higher-temperature gasification char was much more con-densed (�17 rings per cluster). � 2009 American Insti-tute of Chemical Engineers Environ Prog, 28: 386–396, 2009Keywords: switchgrass, corn stover, char quality,

solid-state 13C NMR

INTRODUCTION

Thermochemical processing of biomass hasreceived significant recent attention as a platform foreconomically producing energy and chemicals frombiorenewable resources [1, 2]. Product compositionfrom these processes varies with reaction conditionsand includes noncondensable gases (syn or producergas), condensable vapors/liquids (bio-oil, tar), andsolids (char, ash). In fast pyrolysis systems, dry bio-mass is heated very rapidly (up to 10008C/s) in theabsence of oxygen and the products quickly removedand quenched to maximize production of bio-oils.Traditional charcoal-making typically employs slowpyrolysis conditions: slow heating rates (1–208C/min)in the absence of oxygen, and long char residencetimes (hours to days). Gasification uses higher tem-peratures and some oxygen (less than the stoichio-metric ratio) to produce a noncondensable gas rich inhydrogen and carbon dioxide. Both fast pyrolysis andgasification yield some amount of char, typically 15–20% and 5–10% of the feedstock mass, respectively.� 2009 American Institute of Chemical Engineers

386 October 2009 Environmental Progress & Sustainable Energy (Vol.28, No.3) DOI 10.1002/ep

How to best use this coproduct depends on the localeconomic circumstances and the char properties.Combusting the char to supply process heat is com-mon [3, 4], while a few chars may be suitable forfurther activation to be used in higher-value adsorp-tion applications [5, 6].

Use of coproduct char as biochar, that is, charfrom biomass applied to soil as a soil amendmentand/or a carbon sequestration agent, is anotheroption [2]. Although biochars have been used formillennia in some cultures’ agricultural practices,current interest in biochars stems from the investiga-tion of terra preta soils in the central Amazon. Thesedark, incredibly fertile soils have been shown tocontain man-made charcoal that functions as soil or-ganic matter [7–9]. The link between char propertiesand their efficacy in soils, however, is not wellunderstood, much less how to engineer the processconditions to produce desired biochar properties.This is especially true for chars from gasification andfast pyrolysis; most research in this area has focusedon product yields and char combustion properties[4, 10–12].

The purpose of this research was to provide athorough characterization of chars produced undertypical fast pyrolysis and gasification conditions usinglocally common feedstocks: switchgrass and corn sto-ver. This characterization serves as the initial step inan overall engineered biochar production scheme.The next steps would include soil incubation andcrop growth studies using chars from these processes,the formulation of desired biochar properties basedon soil tests, and finally, the engineered productionof chars with these properties.

A key aspect of determining char quality for bio-char (and other) applications is the ability to quanti-tatively characterize the forms of carbon present, asthe type of carbon is believed to be related to char’sreactivity and recalcitrance in soil [8, 13–17]. Con-cern has been expressed about ‘‘incompletely’’ pyro-lyzed biomass as it may provide too much bio-avail-able carbon to the soil without enough simultaneousnitrogen, resulting in nitrogen immobilization andtherefore, negative short-term effects on plant yield[15]. Previous studies have used proximate analysisto differentiate between ‘‘volatile’’ and ‘‘fixed’’ car-bon [18], X-ray diffraction to measure carbon crystal-linity [3], FTIR spectroscopy to identify char carbonfunctionality [19, 20], and various solid-state 13C nu-clear magnetic resonance spectroscopy (NMR) tech-niques such as cross-polarization/magic angle spin-ning (CP/MAS) to measure carbon functionality andaromaticity of chars [14, 19, 21–23], and other highlyaromatic materials [24, 25]. The difficulty with all ofthese methods is the semiquantitative nature of theinformation they provide. CP NMR, for example,tends to underestimate the nonprotonated fractionof black carbons due to the slow transfer of hydro-gen magnetization to carbons in the middle of largearomatic structures and is sensitive to signal loss byinteraction with unpaired electrons, rendering up to70% of carbon ‘‘invisible’’ [26, 27]. The direct-polar-ization (DP) or Bloch-decay MAS NMR approach is

superior in most respects [26, 28, 29], since it isinherently quantitative and detects most carbon [26].Further, DP/MAS NMR can be combined with dipolardephasing to quantify the fraction of nonprotonatedaromatic C [30]. This study explores the applicationof these quantitative NMR techniques to study thestructure of fast pyrolysis and gasification chars. Theuse of two complementary NMR methods for esti-mating the size of clusters of fused aromatic rings inchars, based on spectral analysis and 1H-13C dipolardistance, is also demonstrated.

EXPERIMENTAL

Char SelectionSeven representative chars were selected for this

study, one from each thermochemical process foreach feedstock and one commercially available woodcharcoal. Switchgrass and corn stover were obtainedlocally (Story County, IA). Before thermochemicalprocessing, feedstocks were ground in a hammer millto pass a [¼]" screen and dried to <10% moisture.Mixed hardwood charcoal was obtained from a com-mercial kiln (Streumph Charcoal Company, Belle,MO). This char had been used in a biochar soil col-umn nutrient leaching study [31] and was considereda good candidate for comparison.

Slow PyrolysisSlow pyrolysis was performed by placing feed-

stock into a paint-can fitted with a nitrogen purge(1 L/min flow rate) and a thermocouple for tempera-ture measurement. The sealed can was placed into amuffle furnace and heated at �158C/min to 5008C.Corn stover (50 g) was held at 5008C for 30 min;switchgrass (125 g) was held at 5008C for 2 h. Thechar was then cooled under nitrogen flow andstored in sealed glass jars. Mass yield of char was33.2% and 41.0% for corn stover and switchgrass,respectively.

Fast PyrolysisFast pyrolysis was performed on a 5 kg/h capacity

bubbling fluidized bed reactor optimized for bio-oilproduction [5]. The sand bed was fluidized with nitro-gen preheated at 5008C. Char was collected using ahigh-throughput cyclone catch and cooled undernitrogen before being stored in resealable plasticbags.

GasificationGasification was performed on a 3 kg/h capacity

bubbling fluidized bed reactor using an air/nitrogenfluidizing gas (0.20 equivalence ratio). For reactor set-up details, see Meehan et al. [32]. The average steadystate temperature was 7608C for switchgrass and7308C for corn stover. Char was again collected bycyclone, cooled under nitrogen, and stored in reseal-able plastic bags.

Environmental Progress & Sustainable Energy (Vol.28, No.3) DOI 10.1002/ep October 2009 387

Physical PropertiesBrunauer-Emmet-Teller (BET) surface area was

measured by nitrogen gas sorption analysis at 77 K(NOVA 4200e, Quantachrome Instruments, BoyntonBeach, FL). Before analysis, samples were vacuumdegassed at 3008C for 4–16 h (conditions typical forcarbons). Degassing time varied based on the timenecessary to reach a stable surface area measurement.Particle density was measured by helium pycnometer(Pentapycnometer, Quantachrome Instruments) usingdegassed samples from BET analysis and long purgetimes (10 min) to prevent errors due to volatile con-tent outgassing.

Char particle structure and surface topographywere analyzed by scanning electron microscopy(SEM) using a Hitachi S-2460N variable pressure scan-ning electron microscope. Samples were mounted oncarbon disks. Variable pressure mode allowed for ex-amination of insulating samples with minimal samplepreparation. A residual atmosphere of 60 Pa (0.5Torr) of helium was adequate to eliminate chargingfrom samples, while allowing reasonably high magni-fications (up to 15003).

Chemical PropertiesMoisture, volatiles, fixed carbon, and ash content

were determined in triplicate by ASTM proximate anal-ysis method for wood charcoals (ASTM D1762-84,reapproved 2007). Fused quartz crucibles were usedand chars were not ground prior to analysis (mostwere already fine powders). Elemental analysis wasperformed by LECO Corporation (St. Joseph, MI) usingTRUSPEC-CHN and TRUSPEC-S analyzers (LECO).Samples (�0.1 g) with larger particles were crushedusing a mortar and pestle before analysis. Oxygencontent was not able to be determined consistentlydue to high inorganic oxygen content (in the ash)decomposing during analysis. Higher heating value(HHV) of chars was determined by oxygen bombcalorimeter (Parr Instrument Company, Moline, IL)according to Parr Sheet No. 240M, 205M, and 207M.

Mineral content was measured by X-ray fluores-cence (XRF) spectrophotometer (PHILIPS PW2404)equipped with a rhodium target X-ray tube and a4 kW generator. Dry char (4 g) was mixed with X-raypellet mix powder (1.5 g) and boric acid (1 g) for2 min in a puck grinder, and then pressed into a pelletunder vacuum to 25 tons pressure for 15 s. Dry feed-stock was also analyzed to verify that char had notbeen contaminated by sand from the fluidized bed.

FTIR-PASSurface functionality was investigated by Fourier

transform infrared (FTIR) spectroscopy using a Digi-lab FTS-7000 FTIR spectrophotometer equipped witha PAC 300 photoacoustic detector (MTEC Photoacous-tics, Ames, IA). The sample chamber was purgedwith helium for several minutes prior to analysis toprevent the interference of water and carbon dioxide.Spectra of dried feedstock and char samples weretaken at 4 cm21 resolution and 1.2 kHz scanningspeed for a total of 64 coadded scans.

Solid-State 13C NMR13C NMR experiments were performed using a

Bruker DSX400 spectrometer at 100 MHz (400 MHz1H frequency). Quantitative 13C Direct Polarization/Magic Angle Spinning (DP/MAS) NMR experimentswere performed using 4-mm sample rotors at a spin-ning frequency of 14 kHz. The 908 13C pulse-lengthwas 4.5 ls. Sufficiently strong 1H decoupling at gB1/2p 5 72 kHz with the two-pulse phase-modulated(TPPM) scheme was applied during an acquisitiontime of 2 ms. Recycle delays (10–40 s) were deter-mined by the Cross Polarization/Spin-Lattice Relaxa-tion Time/Total Sideband Suppression (CP/T1-TOSS)technique to make sure that all carbon sites were>95% relaxed [33]. Delays were confirmed by a seriesof DP experiments with increasing recycle delays. Toobtain quantitative information on the nonprotonatedaromatic carbon fraction, DP/MAS 13C NMR withrecoupled dipolar dephasing was used [30]. The dipo-lar dephasing time was 68 ls. The total time for DP/MAS and DP/MAS with gated decoupling experimentswas typically 23 h per sample.

Qualitative char composition information, in partic-ular alkyl carbon composition, was obtained withgood sensitivity by 13C CP/TOSS NMR experimentswith samples in 7-mm rotors at a spinning speed of7 kHz, a CP time of 1 ms, a 1H 908 pulse-length of4 ls, and a recycle delay of 0.5 s. Four-pulse TOSSwas employed before detection and TPPM decou-pling was applied for optimum resolution.

The size of fused aromatic rings typical of charcoalcan be estimated based on recoupled 1H-13C dipolardephasing [34]. In short, two 1H 1808 pulses per rota-tion period prevent magic angle spinning (MAS) fromaveraging out weak CH dipolar couplings. Composite908x-1808y-908x pulses were used to reduce effects ofimperfect pulse flip angles. To detect nonprotonatedcarbons with good relative efficiency, DP/TOSS wasused at a spinning speed of 7 kHz, in 7-mm rotorsfor the pyrolysis chars. All experiments on the gasifi-cation char had to be performed in 4-mm rotors,where the pronounced 400-MHz radio-frequencyabsorption due to sample conductivity was lesssevere than for the larger amount of material in the7-mm rotor. The 13C 908 and 1808-pulse lengths were4.5 and 9 ls, respectively. The recycle delays werethe same as used for DP/MAS spectra. Instead of thetotal aromatic signal between 107 and 142 ppm, onlythe signal of nonprotonated C (after 40 ls of regulargated decoupling) was considered in the analysis. Forreference, milled-wood lignin [35] (a better-definedsample than the commercial lignin in Ref. 34) wasrun under the same conditions, with a 60-s recycledelay.

RESULTS AND DISCUSSION

Physical PropertiesSEM micrographs of switchgrass and the three

types of switchgrass char are shown in Figure 1.Overall plant structure was visible in all of the chars.Increased porosity from volatiles escaping during


thermochemical degradation can also be seen. Theparticle size decrease observed in the gasification andfast pyrolysis char is believed to be caused by rapiddevolatilization creating very porous (macroporous)and fragmented chars [36]. In general, gasificationchars are fine powders while fast pyrolysis are veryfine powders [3]. Table 1 shows the particle densitiesand BET surface areas of the representative chars.Surface areas were very low (7–50 m2/g) comparedto commercial activated carbons and increased withprocess temperature and char residence time. Particledensity, also known as solid or true density,increased with ash (mineral) content and processtemperature. It has been suggested that particle den-sity can be used to estimate the charring temperature.As temperature and reaction time increase, the degreeof graphitization increases and char’s particle density(typically 1.5–1.7 g/cm3) approaches that of solidgraphite (2.25 g/cm3) [37]. The presence of minerals,which are denser than most forms of carbon, canalso cause higher apparent particle density in high-ash chars.

Chemical PropertiesResults from proximate and elemental analysis of

chars are shown in Table 1. Switchgrass and cornstover chars had high ash contents (32–55 wt %) atthe expense of carbon content. For most charapplications, this high ash content puts switchgrassand corn stover chars at a disadvantage compared tochars from low-ash feedstocks. Char HHVs (Table 1)are similar to those presented by Boateng and are

comparable to coals [3]. Table 2 lists the ash composi-tion of switchgrass, corn stover, and hardwood charsas determined by XRF. (Because of the nature of thesamples and the calibration method, the relative con-centrations of the elements are accurate, but the over-all mineral content in the char is overestimated.)Corn stover and switchgrass ashes predominantlycontain silica, while hardwood ash contains mostlyalkali metals. Biomass combustion research hasshown that feedstocks containing more silica haverelatively high slagging tendencies [38]. Furthermore,contamination by sand or soil during biomass collec-tion enhances this tendency [38]. For this reason,chars from switchgrass and corn stover (collected byfarming equipment) have three inherent challengescompared to traditional charcoals for use as fuels:high overall ash content, high silica content, andcontamination by soil.

Aromaticity from NMRFigure 2 presents quantitative 13C DP/MAS NMR

spectra of switchgrass slow pyrolysis, fast pyrolysis,and gasification chars. The corresponding quantitativespectra of nonprotonated carbons and CH3 groups,obtained after 68 ls of dipolar dephasing [26], arealso shown (thick lines). The spectra are dominatedby the band of the aromatic carbons around 128ppm, the majority of which are not protonated. Smallsignals of C¼¼O and alkyl groups are also detected.These are seen more clearly in the CP/TOSS spectraof Figure 3, which over-represent the signals ofprotonated C and contain no residual spinning

Figure 1. Scanning electron micrographs of switchgrass (a) feedstock, (b) slow pyrolysis char, (c) fast pyrolysischar, and (d) gasification char.


sidebands. In addition to spectra of pyrolysis chars,the spectrum of the switchgrass feedstock is shownfor reference in Figure 3c. Figure 3d shows the CPspectrum of fast pyrolysis char from corn stover. Therelative fractions of eight types of functional groupsobtained from the DP spectra are compiled in Table3. The sum of of the aromatic C��O, nonprotonatedaromatics, and aromatic C��H fractions gives the totalaromaticity. Although the aromaticities of the differentchars are similar (see first column of Table 5), thefraction of protonated aromatic carbons decreasessignificantly from slow pyrolysis to gasification.

From the abundance and estimated composition ofthe functional groups in Table 3, the elemental com-position can be estimated. Since C��OH andC��O��C groups have similar resonance positions,their ratio cannot be determined; a 50:50 ratio ofethers and OH groups was assumed in the analysis.The resulting fractional oxygen contents listed in Ta-ble 3 take into account ether oxygen shared betweenits two bonded carbons. Since ether oxygen atomsare not protonated, the H contribution is correspond-ingly reduced. Table 4 compares these NMR-derivedcomposition values with those from combustion anal-ysis. The good agreement validates the NMR assign-ments. Both elemental analyses consistently showmuch higher oxygen content for the fast pyrolysischar. The detailed NMR analysis (see Table 3) revealshigher fractions of all kinds of oxygen-containingmoieties.

Degree of Aromatic Condensation from NMRIn addition to determining aromaticity, NMR can

also provide an estimate of the degree of aromaticcondensation. Various research groups haveattempted to use NMR to quantify the fraction ofbridgehead carbons (fbridge) as a measure of thedegree of aromatic condensation. Several groups sim-Ta

ble

1.Composition,physicalproperties,

andhigherheatingvalue(H

HV)ofrepresentativech

ars.

Char

Particle

density

(g/cc)

BETsu

rface

area(m

2/g)

Moisture

(wt%)

Volatiles

(wt%)

FixedC

(wt%)

Ash

(wt%)

C(w

t%)

H(w

t%)

N(w

t%)

S(w

t%)

HHV(as

received)

(MJ/kg)

Switch

grass

S.P.

1.76

50.2

0.9

7.1

39.5

52.5

39.4

1.3

0.7

0.002

15.37

Switch

grass

F.P.

1.78

21.6

2.7

16.4

26.4

54.6

38.7

2.5

0.6

0.21

16.34

Switch

grass

gasification

2.06

31.4

2.5

10.3

34.3

53.0

42.8

1.6

0.8

0.17

15.86

Corn

stoverS.

P.1.54

20.9

1.8

11.1

54.7

32.4

62.8

2.9

1.3

0.05

21.60

Corn

stoverF.

P.1.85

7.0

1.0

14.9

34.4

49.7

37.8

2.5

0.8

0.06

13.83

Corn

stovergasification

1.92

23.9

1.9

5.5

38.5

54.0

38.5

1.3

0.7

0.09

15.29

Switch

grass

F.P.

[3]

ND

7.7

3.8

28.4

42.0

25.9

63

3.7

0.8

ND

19.37

HardwoodS.

P.1.60

19.7

2.6

19.7

63.8

13.9

65.3

2.6

0.6

0.05

22.64

Elementalco

mpositionvaluesarereportedonadry

weightbasis;HHVan

dproxim

atean

alysisresultspresentedonawetbasis.

S.P.,slow

pyrolysis;

F.P.,fast

pyrolysis;

ND,notdeterm

ined.

Table 2. Ash composition of switchgrass, corn stover,and hardwood char samples by X-ray fluorescencespectroscopy prepared by the pressed pellet method.

ElementSwitchgrassF.P. char

Corn stoverF.P. char

Hardwoodchar

Al2O3 0.49 2.33 0.60CaO 3.65 3.80 22.37Cl 0.47 0.59 0.03Fe2O3 0.76 1.87 2.36K2O 6.00 4.03 1.35MgO 1.55 2.02 0.48MnO2 0.15 0.13 0.83Na2O 0.07 0.20 0.06P2O5 3.86 1.19 0.20SiO2 43.62 29.98 5.67SO3 0.99 0.28 0.27Other 0.25 0.64 0.51Total 61.86 47.06 34.73

All values are dry weight %. Elements are representedas their respective oxides. F.P., fast pyrolysis.


ply assumed that all carbons resonating around 130ppm are bridgehead C, and that aromatic C��Hresonances are in a narrow band around 108 ppm[39, 40]; the clear dipolar dephasing by >20% of the130-ppm band in the spectra in Figure 2, however,demonstrates the significant presence of C��H signaland thus shows that this assumption is false.

A better approach was suggested by Solum et al.,who focused on the signals of nonprotonated C(selected by dipolar dephasing) and assigned thosebetween 135 and 90 ppm to bridgehead carbons [24,25]. Still, the spectra in Figure 2 show no indicationof a minimum near 135 ppm that would indicate aspectral separation of bridgehead from other nonpro-tonated aromatic carbons; instead, the spectra suggestthat the bridgehead carbon band extends beyond 135ppm and cannot be reliably separated from smallerbands of other nonprotonated aromatic C. Indeed, intwo coals of high aromaticity, the assumptions of So-lum et al. resulted in a higher fraction of alkylatedaromatics than that of total alkyl carbons, which con-firms that some bridgehead carbon signal wasassigned incorrectly [24]. In addition, the use of crosspolarization from 1H is likely to result in an under-representation of the bridgehead carbons far from thenearest 1H and must, therefore, be avoided in thestudy of chars.

We propose here that the degree of aromatic con-densation can be estimated most reliably by combin-ing two complementary approaches: spectral analysisand long-range H��C dipolar dephasing. From quan-titative 13C NMR spectra, one can estimate the fractionof carbons along the edges of the aromatic rings,vedge 5 1 2 vbridge, which decreases with increasingaromatic ring cluster size. The usually dominant spec-tral contributions to the aromatic edge carbons inchars come from aromatic C��H and aromatic C��Omoieties, whose fractions vCH 5 faCH/far and vC��O 5faC��O/far can be determined quite easily from the 13Cspectra. (far, faCH, and faC��O stand for aromatic carbon

fractions: total aromatic carbon, aromatic carbonbonded to hydrogen, and aromatic carbon bonded tooxygen, respectively.) Together, they constitute theminimum aromatic edge fraction,

vedge;min ¼ vCH þ vC�O (1)

Additional contributions can come from alkyl Cand C¼¼O bonded to the aromatic rings. Thus, theupper limit of the edge fraction is provided by

vedge;max ¼ vedge;min þ valkyl þ vC¼O (2)

If the fraction of C¼¼O groups exceeds that ofalkyls, one can show that some of the C¼¼O must bebonded to the aromatic rings (and thus contribute tovedge,min), but this is not relevant with the samplesstudied here. Table 5 lists the values of vedge,min andvedge,max for the chars studied. Given the relativelysmall alkyl (valkyl 5 falkyl/far) and C¼¼O fractions inchars, the range of vedge between vedge,min andvedge,max is quite limited, particularly for the slow py-rolysis sample.

Figure 2. Quantitative 13C NMR spectra, obtained withdirect polarization at 14-kHz MAS, of three charsmade from switchgrass: (a) slow pyrolysis, (b) fastpyrolysis, (c) gasification char. Thin line: spectrum ofall carbons; bold line: corresponding spectrum ofnonprotonated C and CH3, obtained after 68 ls ofdipolar dephasing. ssb, spinning side band.

Figure 3. CP/MAS/TOSS 13C NMR spectra, highlightingthe signals of alkyl residues, of (a) slow pyrolysis charfrom switchgrass, (b) fast pyrolysis char fromswitchgrass, (c) switchgrass feedstock for reference,and (d) fast pyrolysis char from corn stover.


The geometry of condensation, for example, linear(primary catenation) versus clustered (e.g., circularcatenation), can be assessed based on vedge,max [24]:in linearly condensed systems, vedge > 0.5, sovedge,max< 0.5 excludes linear condensation. Moregenerally, based on the equations given by Solumet al., one can show that the number of carbons, nC,in a cluster relates to the edge fraction by

3=ðvedge � 0:5Þ � nC � 6=v2edge � 6=v2edge;max (3)

with the upper limit for linear and the lower for cir-cular condensation [24]. For instance, if vedge,max �0.4, then there must be more than nC 5 37 carbonsin a cluster. Table 5 lists these minimum cluster sizes.

The second approach, long-range 1H-13C dipolardephasing, probes distances of the aromatic carbonsfrom hydrogen at the edge of the condensed ring sys-tem in terms of the strongly distance-dependent 1H-13Cdipolar couplings [34]. The larger the average 1H-13Cdistance, the slower the dephasing of the 13C signal;thus, a slower decay indicates a larger cluster size. Theslower dephasing for the gasification char compared tothe pyrolysis chars (see Figure 4) indicates a larger aro-matic-cluster size in the former, consistent with thespectral analysis. Curves for specific sites in model com-pounds [34], with two two-bond and two three-bondcouplings, provide an approximate length-scale calibra-tion (see dashed and dash-dotted lines in Figure 4).

The dephasing for the slow pyrolysis char nearlycoincides with the three-bond calibration curve, indi-

Table 3. Quantitative NMR spectral analysis of switchgrass and corn stover chars.

MoietiesLocation (ppm)

Carbonyls Aromatics Alkyls

C¼¼O210-183

COO183-165

C��O0.75H0.5

165-145Cnon-pro/C��H

145-90HCO0.75H0.5

90-50CH1.5

50-25CH3

25-6

Switchgrass S. P. 0.8 1.4 6.7 52 35 1.4 1.6 1.3Switchgrass F. P. 3.3 3.2 10.3 49 24 3.6 2.6 4.0Switchgrassgasification

2.0 2.2 6.4 66 15 4.0 2.0 2.0

Corn stover F. P. 3.5 4.2 11.7 42 27 3.3 3.7 4.8Corn stovergasification

2.5 2.8 8.2 61 18 3.0 2.3 1.3

All values are % of total 13C signal. Cnon-pro, nonprotonated aromatic carbon.S.P., slow pyrolysis; F.P., fast pyrolysis.Error margins: 6 1%.

Table 4. Elemental analysis of switchgrass and corn stover chars from NMR and combustion (in parentheses).

Char C (wt %) H (wt %) O 1 N (wt %)

Switchgrass S. P. (39.4 6 0.4) 1.53 (1.31 6 0.01) 4.9 (�6.3)Switchgrass F. P. (38.7 6 0.2) 1.63 (2.49 6 0.03) 10.1 (�5.7)Switchgrass gasification (42.8 6 0.1) 1.18 (1.60 6 0.02) 7.9 (�3.9)Corn stover F. P. (37.8 6 0.6) 1.82 (2.48 6 0.05) 11.3 (�10.5)Corn stover gasification (38.5 6 0.2) 1.08 (1.29 6 0.01) 8.2 (�7.2)

S.P., slow pyrolysis; F.P., fast pyrolysis.

Table 5. Aromaticities, fractions of aromatic edge carbons, and minimum number of carbons per aromaticcluster in switchgrass and corn stover chars.

Char Aromaticity vCH vedge,min vedge,max nC,min Harom/Halk

Switchgrass S. P. 94 0.37 0.44 0.51 >23 C 4.5Switchgrass F. P. 83 0.29 0.41 0.61 >16 C 1.2Switchgrass gasification 86 0.17 0.25 0.40 >37 C 1.4Corn stover F. P. 81 0.33 0.48 0.72 >12 C 1.2Corn stover gasification 87 0.21 0.30 0.44 >31 C 2

Aromaticity is listed as % of total 13C signal. F.P., fast pyrolysis; S.P., slow pyrolysis.


cating that the nonprotonated carbons in this char areon average at a three-bond distance from the nearest1H. Figure 5a shows a typical fused-ring system thatis compatible with the NMR data, both spectroscopicand dipolar-dephasing, for the slow pyrolysis char.Here, the number of sites (filled triangles andsquares) that dephase slower than the calibrationsites is similar to the number of sites (thin-line ellip-ses) that dephase faster. The corresponding modelfor fast pyrolysis char (see Figure 5b), has a smallerC��H and larger C��O fraction, but features an onlyslightly smaller fused-ring system. The slowerdephasing for the gasification char in Figure 4requires a significantly larger fraction of C at a �3-bond distance from the nearest 1H. The structure inFigure 5c contains many more slow-dephasing (filledtriangles or squares) than fast-dephasing (thin-lineellipses) sites.

Comparison with Previous NMR Studies of CharThe structures with significant clustered aromatic

condensation derived here from the dual NMRapproach are very different from the small, linearlycondensed structures proposed by Knicker for cellu-

lose heated under oxic conditions, based on CP/NMR[23, 41]. The difference may arise primarily from dif-ferences in char production conditions, namely thepresence of oxygen and the lower temperature; fol-lowing the temperature-ring size relationship sug-gested above, a less condensed (but not necessarilylinear) structure would be expected. The oxic heatingconditions may have resulted in local ‘‘hot spots,’’where the temperature exceeded that of the furnacesetting as some of the carbon was exothermicallycombusted.

Form of H in CharsBased on the quantitative NMR analysis, we can

estimate the ratio of the fractions of hydrogenattached to aromatic and alkyl carbons, Harom/Halk.The data in Table 5 show that in spite of the higharomaticity of all chars, aromatic H is strongly domi-nant only in the slow pyrolysis char.

Figure 4. Plot of the area of signals of nonprotonatedaromatic carbons resonating between 107 and 142ppm, under long-range 1H-13C dipolar dephasing.Circles: Gasification char from switchgrass. Squares:Slow and fast pyrolysis char, whose data coincidewithin the error margins of 62%. Dash-dottedline: Carbons 11 and 13 of 1, 8-dihydoxy-3-methylanthraquinone, which are three bonds awayfrom the two nearest protons. Dashed line: Carbon 1of 3-methoxy benzamide, which is two bonds awayfrom the two nearest protons. The new referencedata for lignin (triangles) coincide with this line.

Figure 5. Typical aromatic clusters, derived fromNMR, in (a) slow pyrolysis char, (b) fast pyrolysischar, and (c) gasification char from switchgrass.Symbols label the distance of carbons resonatingbetween 107 and 142 ppm from the nearestproton(s). Thin-line ellipse: Two bonds from multipleH. Thick-line ellipse: Two bonds from one H. Opentriangle: Three bonds from multiple H. Filled triangle:Three bonds from one H. Open square: Four bondsfrom multiple H. Filled square: Four bonds from oneH, or more than four bonds from any H.


FTIR-PASFTIR-PAS is a very fast and easy method to gauge

the ‘‘progress’’ of the pyrolysis reaction. Drying is theonly sample preparation step needed to perform thisanalysis, which eliminates the sample handling anddilution difficulties encountered when pelletizingsamples with potassium bromide [42]. The FTIR-PASspectra in Figure 6 show the progression of switch-grass feedstock to gasification char; the spectra forcorn stover (not shown) were very similar to those ofswitchgrass. The spectrum from commercial hard-wood slow pyrolysis is also shown. The mostdramatic change is the O��H stretch peak around3400 cm21, which dominates the feedstock spectrumbut is almost absent in the gasification char spectrum.Assignment of other peaks important for chars,including the aliphatic C��H stretch at 3000–2860cm21, the aromatic C��H stretch around 3060 cm21,and the various aromatic ring modes at 1590 and1515 cm21, can be found in a paper by Sharma et al.on lignin chars [20]. The series of spectra in Figure 6suggests a gradual loss of lignocellulosic functionalgroups, but the NMR spectra of Figure 3 show that allof the chars contain little, if any, polysaccharideresidues. The relatively strong aromatic C��H stretchin slow pyrolysis char matches well with the largearomatic C��H concentration seen in the NMR spec-tra. It was noted that fast pyrolysis char appeared less‘‘reacted’’ than the slow pyrolysis char, but again thisis not supported by NMR; rather, the higher oxygencontent in the fast pyrolysis char makes the distribu-tion of functional groups more similar to those of theswitchgrass feedstock.

Effects of Synthesis Conditions onChar Structure

NMR showed that the aromaticity of slow pyrolysischar is higher than that of fast pyrolysis or gasifica-tion chars. Tentatively, this can be attributed to thelong residence time (2 h) in slow pyrolysis, compared

to that of fast pyrolysis and gasification in a fluidizedbed reactor (<2 s). Nevertheless, slow pyrolysis charexhibited a cluster size only slightly larger than fastpyrolysis char and much smaller than gasificationchar. This suggests that cluster size is controlledmostly by reaction temperature, not duration.

None of the chars contained recognizable frag-ments of feedstock biopolymers, showing that thereaction time was sufficient for complete conversionto char. This preempts concerns about unreacted,bio-available fractions in carbon sequestration andsoil amendment applications. Elemental analysis andNMR showed consistently that fast pyrolysis char con-tained more oxygen in various functional groups, notjust alkyl C��OH as the feedstock. The somewhatenhanced COO concentration in fast compared toslow pyrolysis char may actually result in a slightlybetter cation exchange capacity. In conjunction withthe similar aromatic cluster size, this suggests that fastpyrolysis char should perform similarly as, if notbetter than, slow pyrolysis char in soil amendmentapplications. Gasification char, though similar to slowpyrolysis char in aromaticity, has much larger aro-matic clusters, suggesting significantly differentproperties. The larger cluster size is proven not onlyby the dual NMR approach introduced here, but alsoby the lower ppm value of the nonprotonatedaromatic carbon band, which is characteristic ofbridgehead carbons [22], and by the observed radio-frequency power absorption due to conductivity ofsufficiently large fused ring systems [43].

CONCLUSIONS

1. Chars from fast pyrolysis and gasification arephysically and chemically different from traditionalhardwood charcoals and chars prepared from herba-ceous feedstocks by slow pyrolysis. The types ofcarbon present appear to depend on process temper-ature and, to a lesser extent, reaction time. None ofthe chars contained a detectable amount of onlypartially pyrolyzed biomass.

2. Fast pyrolysis and gasification chars should beincluded in biochar trials. Their wider range of prop-erties and reaction conditions will offer insight onhow to engineer desirable biochars. The structuralfeatures of fast pyrolysis char suggest favorable prop-erties in this application. Coproduction of biocharand bioenergy may prove to be a more cost-effectiveand resource efficient use of biomass crops and cropresidues.

3. Switchgrass and corn stover chars have highsilica ash content and low surface area, and therefore,will present challenges to traditional char applicationssuch as combustion or activation; the best use ofswitchgrass and corn stover chars may be soil appli-cation depending on the economic circumstances andthe local soil properties.

4. Solid-state 13C NMR using techniques presentedhere (DP/MAS, DP/MAS with dipolar dephasing, CP/TOSS, and DP/TOSS with recoupled 1H-13C dipolardephasing) can provide the quantitative informationneeded to reliably track changes in carbon structure

Figure 6. FTIR-PAS spectra of switchgrass, switchgrasschars, and a commercial hardwood char.


over reaction time and temperature, which will bemeaningful to engineered char production and bio-char testing.

Financial support for this research was providedby a National Science Foundation Graduate ResearchFellowship and an ISU Plant Science Institute Gradu-ate Research Fellowship (Brewer), and the ISU Bio-economy Institute. The authors thank the followingfor their assistance on various aspects of the analysisprocess: undergraduates Daniel Assmann and HernanTrevino, CSET colleagues for providing char samplesand process information, CSET staff on HHV, Xiao-wen Fang on NMR, Warren Straszheim on SEM, ScottSchlorholtz on XRF, and John McClelland and RogerJones on FTIR-PAS. They also thank David Laird forhis valuable comments on agronomic applications ofbiochar.

LITERATURE CITED

1. Wright, M.M., & Brown, R.C. (2007). Comparativeeconomics of biorefineries based on the bio-chemical and thermochemical platforms, Biofuels,Bioproducts and Biorefining, 1, 49–56.

2. Laird, D.A. (2008). The charcoal vision: A winwin win scenario for simultaneously producingbioenergy, permanently sequestering carbon,while improving soil and water quality, Agron-omy Journal, 100, 178–181.

3. Boateng, A.A. (2007). Characterization and ther-mal conversion of charcoal derived fromfluidized-bed fast pyrolysis oil production ofswitchgrass, Industrial & Engineering ChemistryResearch, 46, 8857–8862.

4. Boateng, A.A., Mullen, C.A., Goldberg, N., Hicks,K.B., Jung, H.-J.G., & Lamb, J.F.S. (2008). Produc-tion of bio-oil from Alfalfa stems by fluidized-bedfast pyrolysis. Industrial & Engineering ChemistryResearch, 47, 4115–4122.

5. Fan, M., Marshall, W., Daugaard, D., & Brown,R.C. (2004). Steam activation of chars producedfrom oat hulls and corn stover, Bioresource Tech-nology, 93, 103–107.

6. Azargohar, R., & Dalai, A.K. (2008). Steam andKOH activation of biochar: Experimental andmodeling studies, Microporous and MesoporousMaterials, 110, 413–421.

7. Lehmann, J., Gaunt, J., & Rondon, M. (2006). Bio-char sequestration in terrestrial ecosystems—Areview, Mitigation and Adaptation Strategies forGlobal Change, 11, 395–419.

8. Glaser, B., Lehmann, J., & Zech, W. (2002).Ameliorating physical and chemical properties ofhighly weathered soils in the tropics with char-coal—A review, Biology and Fertility of Soils, 35,219–230.

9. Glaser, B., Haumaier, L., Guggenberger, G., &Zech, W. (2001). The ‘Terra Preta’ phenomenon:A model for sustainable agriculture in the humidtropics, Naturwissenschaften, 88, 37–41.

10. Yanik, J., Kornmayer, C., Saglam, M., & Yuksel,M. (2007). Fast pyrolysis of agricultural wastes:Characterization of pyrolysis products, Fuel Proc-essing Technology, 88, 942–947.

11. Lee, K.H., Kang, B.S., Park, Y.K., & Kim, J.S.(2005). Influence of reaction temperature,pretreatment, and a char removal system on theproduction of bio-oil from rice straw by fastpyrolysis, using a fluidized bed, Energy & Fuels,19, 2179–2184.

12. Tsai, W.T., Lee, M.K., & Chang, Y.M. (2007). Fastpyrolysis of rice husk: Product yields and compo-sitions, Bioresource Technology, 98, 22–28.

13. Baldock, J.A., & Smernik, R.J. (2002). Chemicalcomposition and bioavailability of thermallyaltered Pinus resinosa (Red pine) wood, OrganicGeochemistry, 33, 1093–1109.

14. Novak, J.M., Busscher, W.J., Laird, D., Ahmedna,M., Watts, D.W., & Niandou, M.A.S. (2009).Impact of biochar amendment on fertility of aSoutheastern coastal plain soil, Soil Science, 174,105–112.

15. Joseph, S., Peacocke, C., Lehmann, J., & Munroe,P. (2009). Developing a biochar classification andtest methods. In J.Lehmann & S.Joseph (Eds.),Biochar for environmental management scienceand technology, London: Earthscan, 107–126.

16. Preston, C.M., & Schmidt, M.W.I. (2006). Black(pyrogenic) carbon: A synthesis of current knowl-edge and uncertainties with special considerationof boreal regions, Biogeosciences, 3, 397–420.

17. Schmidt, M.W.I., & Noack, A.G. (2000). Black car-bon in soils and sediments: Analysis, distribution,implications and current challenges, Global Bio-geochemical Cycles, 14, 777–793.

18. Antal, M.J., Allen, S.G., Dai, X., Shimizu, B., Tam,M.S., & Gronli, M. (2000). Attainment of the theo-retical yield of carbon from biomass, Industrial &Engineering Chemistry Research, 39, 4024–4031.

19. Cheng, C.-H., Lehmann, J., Thies, J.E., Burton,S.D., & Engelhard, M.H. (2006). Oxidation ofblack carbon by biotic and abiotic processes,Organic Geochemistry, 37, 1477–1488.

20. Sharma, R.K., Wooten, J.B., Baliga, V.L., Lin, X.,Geoffrey Chan, W., & Hajaligol, M.R. (2004).Characterization of chars from pyrolysis of lignin,Fuel, 83, 1469–1482.

21. Czimczik, C.I., Preston, C.M., Schmidt, M.W.I.,Werner, R.A., & Schulze, E.-D. (2002). Effects ofcharring on mass, organic carbon, and stablecarbon isotope composition of wood, OrganicGeochemistry, 33, 1207–1223.

22. Knicker, H., Totsche, K.U., Almendros, G., &Gonzalez-Vila, F.J. (2005). Condensation degreeof burnt peat and plant residues and the reliabil-ity of solid-state VACP MAS 13C NMR spectraobtained from pyrogenic humic material, OrganicGeochemistry, 36, 1359–1377.

23. Knicker, H., Hilscher, A., Gonzalez-Vila, F.J., &Almendros, G. (2008). A new conceptual modelfor the structural properties of char producedduring vegetation fires, Organic Geochemistry,39, 935–939.

24. Solum, M.S., Pugmire, R.J., & Grant, D.M. (1989).Carbon-13 solid-state NMR of Argonne-premiumcoals, Energy & Fuels, 3, 187–193.


25. Kelemen, S.R., Afeworki, M., Gorbaty, M.L.,Sansone, M., Kwiatek, P.J., Walters, C.C., Freund,H., Siskin, M., Bence, A.E., Curry, D.J., Solum, M.,Pugmire, R.J., Vandenbroucke, M., Leblond, M., &Behar, F. (2007). Direct characterization of kero-gen by X-ray and solid-state 13C nuclear mag-netic resonance methods, Energy & Fuels, 21,1548–1561.

26. Smerni, k R.J., & Oades, J.M. (2000). The use ofspin counting for determining quantitation insolid state 13C NMR spectra of natural organicmatter, Part 2: HF-treated soil fractions, Geo-derma, 96, 159–171.

27. Simpson, M.J., & Hatcher, P.G. (2004). Determina-tion of black carbon in natural organic matter bychemical oxidation and solid-state 13C nuclearmagnetic resonance spectroscopy, Organic Geo-chemistry, 35, 923–935.

28. Mao, J.D., Hu, W.G., Schmidt-Rohr, K., Davies, G.,Ghabbour, E.A., & Xing, B. (2000). Quantitativecharacterization of humic substances by solid-statecarbon-13 nuclear magnetic resonance, Soil Sci-ence Society of America Journal, 64, 873–884.

29. Hammes, K., Smernik, R.J., Skjemstad, J.O., &Schmidt, M.W.I. (2008). Characterisation and eval-uation of reference materials for black carbonanalysis using elemental composition, colour,BET surface area and 13C NMR spectroscopy,Applied Geochemistry, 23, 2113–2122.

30. Mao, J.D., & Schmidt-Rohr, K. (2004). Accuratequantification of aromaticity and nonprotonatedaromatic carbon fraction in natural organic matterby 13C solid-state nuclear magnetic resonance,Environmental Science & Technology, 38, 2680–2684.

31. Laird, D., Fleming, P., Wang, B., Horton, R., &Karlen, D.L. (2008). Impact of soil biochar appli-cations on nutrient leaching. 2008 GSA-SSSA-ASA-CSA Joint Meeting. Houston, TX.

32. Meehan, P.M., Jones, S.T., & Brown, R.C. Feed-stock potassium as a gasification promoter, Bio-mass and Bioenergy, Submitted for publication.

33. Mao, J.-D., Hu, W.-G., Schmidt-Rohr, K., Davies,G., Ghabbour, E.A., & Xing, B. (2000). Quantita-tive characterization of humic substances by solid-state carbon-13 nuclear magnetic resonance, SoilScience Society of America Journal, 64, 873–884.

34. Mao, J.D., & Schmidt-Rohr, K. (2003). Recoupledlong-range C-H dipolar dephasing in solid-state

NMR, and its use for spectral selection of fusedaromatic rings, Journal of Magnetic Resonance,162, 217–227.

35. Mao, J., Holtman, K.M., Scott, J.T., Kadla, J.F., &Schmidt-Rohr, K. (2006). Differences between lig-nin in unprocessed wood, milled wood, mutantwood, and extracted lignin detected by 13Csolid-state NMR, Journal of Agricultural and FoodChemistry, 54, 9677–9686.

36. Scala, F., Chirone, R., & Salatino, P. (2006). Com-bustion and attrition of biomass chars in a fluid-ized bed, Energy & Fuels, 20, 91–102.

37. Downie, A., Crosky, A., & Munroe, P. (2009).Physical properties of biochar. In J.Lehmann &S.Joseph (Eds.), Biochar for environmentalmanagement science and technology, London:Earthscan, 13–32.

38. LindstroIm, E., OIhman, M., Backman, R., & Bos-troIm, D. (2008). Influence of sand contaminationon slag formation during combustion of woodderived fuels, Energy & Fuels, 22, 2216–2220.

39. Mann, A.L., Patience, R.L., & Poplett, I.J.F. (1991).Determination of molecular structure of kerogensusing 13C NMR spectroscopy. I. The effects ofvariation in kerogen type, Geochimica et Cosmo-chimica Acta, 55, 2259–2268.

40. Wei, Z., Gao, X., Zhang, D., & Da, J. (2005).Assessment of thermal evolution of kerogen geo-polymers with their structural parameters mea-sured by solid-state 13C NMR spectroscopy,Energy & Fuels, 19, 240–250.

41. Kaal, J., Brodowski, S., Baldock, J.A., Nierop,K.G.J., & Cortizas, A.M. (2008). Characterisationof aged black carbon using pyrolysis-GC/MS,thermally assisted hydrolysis and methylation(THM), direct and cross-polarisation 13C nuclearmagnetic resonance (DP/CP NMR) and the ben-zenepolycarboxylic acid (BPCA) method, OrganicGeochemistry, 39, 1415–1426.

42. McClelland, J.F., Jones, R.W., Luo, S., & Seaver-son, L.M. (1993). A practical guide to FT-IR pho-toacoustic spectroscopy. In P.B.Coleman (Ed.),Practical sampling techniques for infrared analy-sis, Boca Raton, FL: CRC Press, 107–144.

43. Jiang, Y.J., Solum, M.S., Pugmire, R.J., Grant,D.M., Schobert, H.H., & Pappano, P.J. (2002). Anew method for measuring the graphite contentof anthracite coals and soots, Energy & Fuels, 16,1296–1300.


2009-Characterization of Biochar From Fast Pyrolysis and Gasification Systems

Documents

Transcript of 2009-Characterization of Biochar From Fast Pyrolysis and Gasification Systems