Mineral Nutrient Recovery from Pyrolysis Systems
Jatara Wise,a Donald Vietor,a Tony Provin,a Sergio Capareda,b Clyde Munster,b and Akwasi Boatengc
a
Texas A&M University, Soil & Crop Sciences, College Station, TX; [e-mail] (for correspondence)
b
Texas A&M University, Biological And Agricultural Engineering, College Station, TX
c
United States Dept. of Agriculture-Agricultural Research Service (ARS), Eastern Regional Research Center, Wyndmoor, PA.
Published online 27 March 2012 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/ep.11631
other nutrients when slow pyrolysis of sorghum biochar was
Bioenergy plants such as sorghum, bioenergy rice, corn
soil applied [2]. Recovery of <10% of biomass K in biochar
stover, and switchgrass can be thermochemically converted by
derived from slow pyrolysis indicate K and other mineral
pyrolysis to produce bio-oil, synthesis gas from noncondensable
nutrients could be lost through condensed bio-oil and NCG s
gases, and biochar. The biochar fraction can be recycled back
[2]. In contrast, studies of fast pyrolysis of stored hybrid
to the production field to improve soil physical qualities and
poplar, switchgrass, and corn stover feedstocks indicated
nutrient status. Although previous publications have described
feedstock concentrations of P and K increased during feed-
the beneficial effects of pyrolysis biochar on soil physical
stock aging [3].
properties; relatively little has been published on the recovery
Contrasting nutrient recoveries in biochar derived from py-
of mineral nutrients from pyrolysis co-products. This work
rolysis suggest definitive evaluations of mineral nutrient
quantified the recovery of nutrients (P, K, Ca, and Mg) from
recovery in biochar and other pyrolysis coproducts are needed
pyrolysis coproducts from various feedstocks using two distinct
[1,2]. Studies of the fate of minerals during related combustion
reactors. Nutrient mass balances, on a feedstock basis, were
and thermochemical conversion processes indicate losses
calculated for comparison of the two reactors efficiency in the
through airborne aerosols could occur. Buseck and Pósfai [4]
recovery of the nutrients. The results revealed the recovery
studied the effects of biomass burning on greenhouse gases
of nutrients varied by (1) species, (2) reactor design, and (3)
correlated highly with nutrient mass loss in biochar. Computa- and climate change and concluded that between 20% and
50% of soot particles from both the Southern Ocean and Izana
tions also revealed P recoveries of 93% (fixed-bed reactor) and
regions contained significant K [4]. Li et al. [5] reported that
58% and 55% (fluidized-bed reactor) for pyrolyzed sorghum.
potassium salt particles were the most abundant inorganic aer-
The recovery of key mineral nutrients in pyrolysis coproducts
osol constituent in the smoke from biomass burning [5].
(primarily biochar) is directly related to the feasibility of
While a number of investigators have found high recovery
nutrient recycling through biochar. Ó 2012 American Institute of
of plant nutrients within the biochar, the existence of other
Chemical Engineers Environ Prog, 31: 251 255, 2012*
studies indicating significantly lower conservation of plant
nutrients within the biochar dictates further evaluation is
INTRODUCTION
needed. This work aims to quantify the recovery of nutrients
Pyrolysis is defined as the thermochemical decomposition
(P, K, Ca, and Mg) from pyrolysis coproducts from various
of organic matter in a near oxygen-free environment to pro- feedstocks using two distinct reactors types.
duce liquid, gaseous, and solid coproducts. These respective
coproducts are termed bio-oil, synthesis gas (molecular
MATERIALS AND METHODS
hydrogen and carbon monoxide fractions), and biochar. The
pyrolysis of agriculturally grown bioenergy crops is currently
Preparation of Feedstock
being researched as one method for the generation of fluid
transportation fuels. The resulting solid biochar is being con- High-energy sorghum (Sorghum bicolor), switchgrass
sidered for the sequestration of carbon and as a soil amend- (Panicum virgatum), corn stover (Zea mays), and high bio-
mass rice (Oryza sativa) were grown and harvested under
ment. The bio-oil from biomass pyrolysis can be used
field conditions. Sorghum, switchgrass, and corn stover were
directly in home heating applications and large-scale power
plants or catalytically upgraded to yield specialized fuels and grown on the Texas AgriLife Research Farm near College
chemicals. The noncondensable gaseous (NCG) coproduct of Station, TX, and rice was obtained from the Texas AgriLife
biomass pyrolysis comprises mostly H2 (molecular hydro- Research and Extension Center near Beaumont, TX. Harvesting
gen), CO (carbon monoxide), CO2 (carbon dioxide), CH4 occurred in July 2009 for corn stover, September 2009 for sor-
(methane), and C2H6 (ethane). The combination of H2 and ghum and rice, and January 2010 for switchgrass. Biomass was
CO makes up what is commonly known as synthesis gas rough-chopped to a length of 3 cm during harvest, air-dried,
then oven-dried (608C) before grinding to pass a 2-mm screen.
(syngas) or producer gas.
A limited number of studies have focused on the recovery Before pyrolyzing, biomasses were redried at 658C.
of plant essential nutrients in the biochar. High recoveries of
P and K were reported for biochar derived from fluidized-
Pyrolysis Using Fixed-Bed Pyrolyzer
bed, fast pyrolysis of corn cob and stover feedstocks [1]. In
A fixed-bed reactor system was used for slow pyrolysis of
contrast, Schnell et al. [2] observed low recovery of K and
feedstock samples under controlled conditions. The sealed
pyrolysis reactor comprised a 75-cm-long stainless steel pipe
Ó 2012 American Institute of Chemical Engineers *This article is a
(2.5 cm diameter) enclosed within a ThermolyneÒ tube
US Government work and, as such, is in the public domain in the
United States of America. furnace. Compressed nitrogen (N2) entered the reactor pipe
Environmental Progress & Sustainable Energy (Vol.31, No.2) DOI 10.1002/ep July 2012 251
Table 1. Mean percent recovery of total feedstock P, K, Ca, and Mg in pyrolysis co-products from fixed-bed slow pyrolysis
(ashing/solubilization method).*,**
Species %P Std Dev %K Std Dev % Ca Std Dev %Mg Std Dev
Biochar co-product
Corn stover 49.90b 20.50 30.10a 12.80 60.80a 26.70 61.50a 26.10
Sorghum 90.50a 20.30 4.80c 1.80 55.90a 12.60 38.30b 12.20
Rice 52.10b 3.40 18.40b 0.90 40.70b 2.80 45.10b 2.50
Bio-oil co-product
Corn stover 0.10b 0.10 0.06b 0.04 0.26b 0.11 1.70a 1.45
Sorghum 0.20a 0.20 0.07b 0.00 5.63a 5.69 2.77a 1.99
Rice 0.10b 0.10 0.09a 0.05 0.57b 0.37 2.41a 1.56
NCG co-product
Corn stover 0.16b 0.10 0.09a 0.03 2.09b 0.63 0.06b 0.04
Sorghum 1.00a 1.23 0.12a 0.11 5.41a 3.16 2.92a 1.41
Rice 0.11b 0.06 0.08a 0.02 4.34ab 4.17 0.08b 0.09
Co-products combined
Corn stover 50.09b 20.54 30.25a 12.77 63.18a 26.62 63.30a 25.42
Sorghum 93.63a 20.00 5.02c 1.75 68.89a 11.33 44.04b 13.12
Rice 52.22b 3.36 18.57b 0.96 45.63b 5.68 47.61b 3.17
*Lower case a, b, and c indicates statistical differences (P 5 0.05).
**NCG 5 non-condensable gas.
vessel system and associated auxiliary systems for biomass
feeding and injection, char collection, vapor condensation for
bio-oil recovery, and instrumentation for data acquisition and
control [6].
Biochar, Bio-Oil, and Gas-Trap Processing
For the fixed-bed reactor system, six subsamples of
each feedstock were pyrolyzed and biochar in combustion
boats was composited to represent a replication. Similarly,
bio-oil collected from the chilled beaker and acetone
washes of condenser and plumbing surfaces were compos-
ited to represent a replication for each feedstock source.
The acetone in the acetone bio-oil mixture was evaporated
at room temperature before bio-oil analysis. The bio-oil
was ashed in porcelain crucibles at 8008C for 4 h before
dissolution in 4 mL of concentrated HCl and brought to
volume with deionized water. Inductively coupled plasma
optical emission (ICP) spectroscopy was used to analyze
subsamples of each volume. A 200-mg sample of biochar
was ashed and similarly analyzed. Contents of the acid
trap for NCG s were collected, diluted with deionized
water, and analyzed using ICP analysis. All biomass and
fluidized-bed reactor biochar was analyzed for nutrients
Figure 1. Scatter plot matrix for correlation between DP
using a sulfuric acid digest [7], while all bio-oil was ana-
and mean percent recovery of total P (fixed-bed pyrolysis).
lyzed using the aforementioned ashing method and (ICP)
[Color figure can be viewed in the online issue, which is
trace metal analysis.
available at wileyonlinelibrary.com.]
For the fluidized-bed system, each feedstock was
pyrolyzed three times resulting in three replications for each
feedstock. Biochar and bio-oil from each replication was
collected and stored in separate containers. Bio-char was
analyzed for trace metals using both of the aforementioned
and flowed across the sample containing boat. Bio-oil was
ashing (ICP) and sulfuric acid digestion methods, whereas
condensed in a glass beaker seated within an ice-cooled PVC
chamber. NCG was bubbled through two test tube traps con- bio-oil was only analyzed for trace metals using the ashing
(ICP) method.
taining 1 N HCl in an attempt to capture aerosols containing
mineral nutrients.
Statistical Analysis
Pyrolysis Using Bench-Scale Fluidized-Bed Reactor
For both reactor analyses, analysis of variance along with
The corn stover, sorghum, and switchgrass biomasses were Student s t test was used for means comparison amongst
pyrolyzed using a bench-scale fluidized-bed fast pyrolysis feedstocks. In addition, correlation analysis was used to
reactor. The reactor, located at the Eastern Regional research determine the degree of linear relationship between variables
Center in Wyndmoor, PA, comprised a state-of-the-art reactor (JMP, Version 8. SAS Institute, Cary, NC, 1989 2007).
252 July 2012 Environmental Progress & Sustainable Energy (Vol.31, No.2) DOI 10.1002/ep
Table 2. Mean percent recovery of feedstock P, K, Ca, and Mg in pyrolysis co-products from fluidized-bed fast pyrolysis
(sulfuric acid digestion method).*
Species %P Std Dev %K Std Dev % Ca Std Dev %Mg Std Dev
Biochar co-product
Corn stover 65.44a 16.77 53.08a 3.90 63.17a 10.62 57.80a 2.97
Sorghum 56.52a 9.74 53.95a 8.33 57.88a 8.18 35.97b 5.29
Switchgrass 30.05b 6.68 8.88b 1.20 38.49b 5.84 14.32c 1.44
Bio-oil co-product
Corn stover 5.23a 1.96 2.05a 0.65 5.30a 1.30 7.74a 2.04
Sorghum 2.03a 2.00 0.86a 0.63 2.27a 1.91 3.28a 1.69
Switchgrass 5.49a 4.35 2.12a 1.50 6.09a 4.71 7.66a 4.46
Biochar and bio-oil combined
Corns stover 70.78a 14.81 55.13a 3.25 68.47a 17.67 65.54a 7.10
Sorghum 58.55a 8.45 54.81a 8.04 60.15a 4.72 39.26b 3.55
Switchgrass 35.54b 9.57 10.99b 1.72 44.58b 5.66 21.98c 0.97
*Lower case a, b, and c indicates statistical differences (P 5 0.05).
Table 3. Mean percent recovery of feedstock P, K, Ca, and Mg in pyrolysis co-products from fluidized-bed fast pyrolysis
(ashing/solubilization method).*
Species %P Std Dev %K Std Dev % Ca Std Dev %Mg Std Dev
Biochar co-product
Corn stover 63.35a 25.52 56.95a 0.67 61.89a 15.13 53.60a 0.23
Sorghum 53.26a 11.52 52.38a 4.92 59.05a 9.69 32.40b 6.33
Switchgrass 34.32b 8.37 4.02b 0.45 45.14b 7.27 21.67c 1.95
Bio-oil co-product
Corn stover 5.23a 1.96 2.05a 0.65 5.30a 1.31 7.74a 2.04
Sorghum 2.03a 2.00 0.86a 0.63 2.27a 1.91 3.28a 1.69
Switchgrass 5.49a 4.35 2.12a 1.50 6.09a 4.71 7.66a 4.46
Biochar and bio-oil combined
Corns stover 68.58a 3.56 59.03a 4.01 67.19a 6.82 61.34a 1.81
Sorghum 55.30a 9.90 53.24a 4.65 61.32a 8.74 35.68b 6.62
Switchgrass 39.81b 6.86 6.14b 1.47 51.23b 4.26 29.33c 3.18
*Lower case a, b, and c indicates statistical differences (P 5 0.05).
RESULTS AND DISCUSSION The Pearson s correlation coefficient in Figure 1 suggests
a linear relationship between DP and mean percent recovery
of total P. High Pearson s correlation coefficients were also
Fixed-Bed Reactor Study
found for the remaining macronutrients which suggested
The recovery of total feedstock P, K, Ca, and Mg varied
potential linear relationships. Similar results were found for
amongst species (Table 1). The recovery of total P in sor-
the fluidized-bed system.
ghum was statistically higher than corn stover or switchgrass.
For this study, percent mean recovery varied amongst
The recovery of total K in corn stover was statistically higher
species for a given nutrient. However, as mentioned above,
than sorghum or rice. The low recoveries of K in sorghum
low mean percent recovery of K in sorghum has been
are similar to those observed by Schnell et al. [2]. The high
observed in similar studies. In addition, mass balances
ashing temperature has been suggested to result in K volatili-
revealed low recovery of mass. For corn stover, sorghum, and
zation [8]. Husmoen [9] utilized identical methodology
rice stover coproduct mass balances were 29%, 36%, and 37%,
(i.e., ashing/solubilization) for the analyses of sorghum
respectively. The varying results indicate that optimal nutrient
biochar and reported >85% recovery of both P and K. The
recovery is not exclusive to a single species. Hence, a given
recovery of total Ca also varied amongst species. Ca recovery
nutrient s recovery in biochar following pyrolysis is specific to
for corn stover and sorghum was statistically similar and
feedstock and potentially to reactor operating conditions.
higher than rice. Total recovery of Mg also varied amongst
species. For Mg, corn stover was statistically higher than
sorghum and rice, which were statistically similar.
Fluidized-Bed Reactor Study
To demonstrate that total mineral nutrient recovery was
related to the amount of nutrient mass lost during conversion The recoveries of P, K, and Ca in each biochar and bio-oil
from dry feedstock to biochar a parameter, (Dnutrient) was were statistically similar for corn stover and sorghum (Table
calculated. For phosphorus, (DP) was defined as the mass 2). For biochar, corn stover had similar mean percent recov-
loss of phosphorus (wt %) during conversion to biochar from eries of P whereas corn and sorghum had similar recoveries
dry feedstock. For the fixed-bed system, correlation analysis for K. Both corn and sorghum had higher recoveries of Ca
was done between mean percent recovery of total P (total and Mg than switchgrass. Overall, the nutrient recoveries
P of all three coproducts) and DP. were significantly lower than those reported by Mullen et al.
Environmental Progress & Sustainable Energy (Vol.31, No.2) DOI 10.1002/ep July 2012 253
[1], who reported corn biochar P and K recoveries of 102 recycling of P and K through biochar derived from sorghum
and 89%, respectively. and corn stover, either system is suitable. Fixed-bed slow py-
Recoveries of P, Ca, and Mg in biochar (Table 3) were rolysis is potentially more effective for the recovery of P in
similar for each feedstock when compared with results in Ta- sorghum biochar than fluidized-bed fast pyrolysis.
ble 2. However, recovery of K in biochar remained low
amongst all tested species. When methods are compared in
ACKNOWLEDGMENTS
Tables 2 and 3, the combined coproduct recoveries for P, K,
Fellowship support of Jatara Wise was provided from the
Ca, and Mg did not differ vastly for the two analysis meth-
United States Department of Agriculture National Needs Fel-
ods, again further suggesting that the use of either method
lowship in Bioenergy. In addition, partial support was pro-
was not the reason for low recovery of nutrients.
vided the Department of Energy North Central SunGrant Pro-
The recovery of key plant nutrients within the biochar
gram, the Alfred P. Sloan Foundation Minority PhD program,
and lack of these nutrients in the bio-oil of all three suggests
and the Hispanic Leaders in Agriculture and the Environment
that nutrients are being lost through the gas phase or due to
(HLAE) program at Texas A&M University.
  slagging  in the reactor piping. The fluidized-bed fast pyrol-
ysis setup did not allow for elemental analyses of the the Abbreviations
NCG products. The very low recovery of P, K, Ca, and Mg in ICP Inductively coupled plasma optical emission
NCG Noncondensable gases
the switchgrass suggests that recovery of overall nutrients
and a given nutrient is species dependant. When both reac-
tors are compared, fixed-bed slow pyrolysis had the highest LITERATURE CITED
recovery of sorghum P in pyrolysis coproducts; whereas the 1. Mullen, C.A., Boateng, A.A., Goldberg, N.M., Lima, I.M.,
corn stover coproducts from the fluidized-bed fast pyrolysis Laird, D.A., & Hicks, K.B. (2010). Bio-oil and bio-char
system had the highest recovery of P. High recoveries of P in production from corn cobs and stover by fast pyrolysis.
sorghum could have been a result of higher carbohydrate Biomass Bioenergy 34, 67 74.
concentration in the relatively green sorghum, whereas the 2. Schnell, R. (2009). Sustainable management of biogeo-
other feedstocks had reached senescence. Calcium and Mg chemical cycles in soils amended with bio-resources from
recovery in corn stover and sorghum was similar for both livestock, bioenery and urban systems. Ph.D. dissertation.
reactors, while recovery of K for both reactors was low. College Station, TX: Texas A&M University Libraries.
Combined biochar and bio-oil mass balances for this system 3. Agblevor, F.A., Besler, S., & Wiselogel, A.E. (1995). Fast
were 47% for sorghum and 57% for both corn stover and pyrolysis of stored biomass feedstocks. Energy & Fuels
switchgrass. In a study of the combustion of softwood saw- 9, 635 640.
dust by Boman, it was summarized that K losses could be a 4. Buseck, P., & Pósfai, M. (1999). Airborne minerals and
result of K vaporization into KCl, (KCl)2, K2SO4, and KOH related aerosol particles: Effects on climate and the envi-
during the pyrolysis or ashing processes [10]. Other studies ronment. Proceedings of National Academy of Sciences
have found that considerable amounts of K was found in fil- of the United States of America 96, 3372.
tered Beech wood pyrolysis oil, which likely came from the 5. Li, J., Pósfai, M., Hobbs, P., & Buseck, P. (2003). Individ-
vapor phase [11]. Losses due to K vaporization escaping with ual aerosol particles from biomass burning in southern
NCG s could also be a possibly related to the N2 flow rates Africa: 2, Compositions and aging of inorganic particles.
by both reactors. Another reason for K losses could result Journal of Geophysical Research 108, 8484.
from the lack catalyst anchoring sites, such as carboxyl and 6. Boateng, A., Daugaard, D., Goldberg, N., & Hicks, K.
hydroxyl groups, which are important for preventing K loss (2007). Bench-scale fluidized-bed pyrolysis of switch-
during pyrolysis and gasification [12]. There are a wide range grass for bio-oil production. Industrial and Engineering
of additional factors that could also influence the recovery of Chemistry Research 46, 1891 1897.
mineral nutrients. Factors such as reactor O2 content [13] and 7. Parkinson, J., & Allen, S. (1975) A wet oxidation proce-
the cooling effect of water vapors on solid pyrolysis copro- dure suitable for the determination of nitrogen and min-
ducts [14] can negatively affect nutrient recovery. Other fac- eral nutrients in biological material. Communications in
tors such as low Si/K ratio, especially for bioenergy rice, can Soil Science and Plant Analysis 6, 1 11.
cause slagging and result in nutrient losses within the copro- 8. Gaskin, J., Steiner, C., Harris, K., Das, K., & Bibens, B.
ducts [15]. High moisture content of pyrolysis biomass could (2008). Effect of low-temperature pyrolysis conditions on
also adversely affect the recovery of mineral nutrients [16]. biochar for agricultural use. Transactions of the ASABE
Other factors not evaluated in this study that might address 51, 2061 2069.
why recoveries of these key plant nutrients were low include 9. Husmoen, D.H. (2011). Effects of biochar recycling on
(1) physiological maturity of feedstock, (2) overall nutrient switchgrass growth and soil water quality in bioenergy
profile of feedstock, (3) growing conditions of the feedstock, production systems. M.S. thesis, College Station, TX:
and (4) drying, grinding, and overall handling of feedstock Texas Ag&M University Libraries.
before pyrolysis. In addition, evaluation of complete mass 10. Boman, C. (2005). Particulate and gaseous emissions
balances needs further investigations. These factors likely from residential biomass combustion. Umeå, Sweden:
contribute to significant geographically specific biochar nutri- Umeå University.
ent conservation. 11. Hoekstra, E., Hogendoorn, K.J.A., Wang, X., Westerhof,
R.J.M., Kersten, S.R.A., van Swaaij, W.P.M., & Groene-
CONCLUSIONS veld, M.J. (2009). Fast pyrolysis of biomass in a fluidized
The recovery of P, K, Ca, and Mg varied among feed- bed reactor: in situ filtering of the vapors. Industrial &
stocks for both systems. Volatilization losses from either the Engineering Chemistry Research 48, 4744 4756.
pyrolysis process or the high temperature nutrient analysis 12. Gorrini, B.C., Radovic, L.R., & Gordon, A.L. (1990). On
ashing process could reduce recovery of nutrients, including the potassium-catalysed gasification of a Chilean bitumi-
P and K. Low recovery of mineral nutrients could also be a nous coal. Fuel 69, 789 791.
result of low mass recovery for both systems. From this 13. Reis, V.V., Frederick, W.J., Wag, K.J., Iisa, K., & Sinque-
study, it can be concluded that the recovery of feedstock field, S.A. (1995). Effects of temperature and oxygen
nutrients: (1) varies amongst species and (2) reactor design, concentration on potassium and chloride enrichment
and (3) is highly correlated to feedstock nutrient mass loss in during black-liquor combustion. Tappi Journal (USA),
the biochar coproduct. However, if the primary goal is the 841 852.
254 July 2012 Environmental Progress & Sustainable Energy (Vol.31, No.2) DOI 10.1002/ep
14. Minkova, V., Razvigorova, M., Bjornbom, E., Zanzi, R., and ash properties of biochars from various biomass
Budinova, T., & Petrov, N. (2001). Effect of water vapour components of mallee trees. Energy & Fuels 24, 1972
and biomass nature on the yield and quality of the py- 1979.
rolysis products from biomass. Fuel Processing Technol- 16. Abdullah, H., & Wu, H. (2009). Biochar as a fuel: 1.
ogy 70, 53 61. properties and grindability of biochars produced from
15. Abdullah, H., Mediaswanti, K.A., & Wu, H. (2010). Bio- the pyrolysis of mallee wood under slow-heating condi-
char as a Fuel: 2. Significant differences in fuel quality tions. Energy & Fuels 23, 4174 4181.
Environmental Progress & Sustainable Energy (Vol.31, No.2) DOI 10.1002/ep July 2012 255