Biomass fly ash in concrete: SEM, EDX and ESEM analysis

Available online at www.sciencedirect.com

www.fuelfirst.com

Fuel 87 (2008) 372–379

Biomass fly ash in concrete: SEM, EDX and ESEM analysis

Shuangzhen Wang a, Larry Baxter a,*, Fernando Fonseca b

a Department of Chemical Engineering, Brigham Young University, 350 CB, Provo, UT 84602, United Statesb Department of Civil Engineering, Brigham Young University, Provo, UT 84602, United States

Received 29 September 2006; received in revised form 7 April 2007; accepted 3 May 2007Available online 11 June 2007

Abstract

This document summarizes microscopy study of concrete prepared from cement and fly ash (25% fly ash and 75% cement by weight),which covers coal fly ash and biomass fly ash. All the fly ash concrete has the statistical equal strength from one day to one year after mix.Scanning electron microscopy (SEM), Energy dispersive X-ray (EDX) and environmental scanning electron microscopy (ESEM) analysisshow that both coal and biomass fly ash particles undergo significant changes of morphology and chemical compositions in concrete dueto pozzolanic reaction, although biomass fly ash differs substantially from coal fly ash in its fuel resources.� 2007 Elsevier Ltd. All rights reserved.

Keywords: Biomass fly ash; EDX; SEM

1. Introduction

Biomass-coal cofiring enjoys overwhelming advantagesin terms of cost, efficiency, technical risk and implementationtime among renewable energy options for power generationin many regions of the world. However, current Americansociety for testing and materials (ASTM) and American con-crete institute (ACI) standards preclude none coal derived flyash from use in concrete. This investigation studies the poz-zolanic reactivity of biomass (and coal) fly ash in concretethrough a variety of image and chemical analyses.

This document is written in a chemical compositionshorthand common to the concrete industry but foreignto the great majority of chemists and chemical engineers.This shorthand notation is useful because the detailedchemical composition of the reactants and products inany practical concrete application are both cumbersomeand only approximately known. Table 1 summarizes thecorrespondence between traditional chemical specificationsand the shorthand notation used in this document and will

0016-2361/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.fuel.2007.05.024

* Corresponding author. Tel.: +1 801 422 8616; fax: +1 801 422 0151.E-mail addresses: [email protected] (S. Wang), larry_bax-

[email protected] (L. Baxter).

be especially useful for those unfamiliar with the tradi-tional nomenclature used by the concrete community.

The cement hydration reactions contain the hydrationof calcium silicates, tricalcium aluminates and ferritephase. The hydration of calcium silicates (take C3S as anexample) produces C–S–H gel and CH, as follows [1]:

2C3S + 11H!C3S2H8 + 3CH ð1Þ

The Ca/Si ratio in C–S–H gel varies in a wide range, butnormal values lie between 0.8 and 2.1, with an average of1.5 [2]. The other cementitious reaction product, CH, formsa layered (often referred to as laminar) cement paste [2].

Pozzolanic reactions refer to chemical reactions betweenactive SiO2 and the CH, as follows [1]:

CH + S + H!C–S–H ð2Þ

The stoichiometric coefficients of this reaction are notfixed. The complicated reaction network summarizedabove leads to a variety of products collectively calledC–S–H gels, similar to the products from cementitiousreaction illustrated in Eq. (1).

C–S–H gels account for the main strength of concrete,with CH contributing little to concrete build up; for pozzo-lanic reactions in this paper, fly ash combines and

Table 1Abbreviations of cement compositions in civil engineering field

Abbreviations C S A F M K N S C HActual meaning CaO SiO2 Al2O3 Fe2O3 MgO K2O Na2O SO3 CO2 H2O

S. Wang et al. / Fuel 87 (2008) 372–379 373

consumes CH and forms new C–S–H gels, which contrib-ute more to the strength build up of concrete.

2. Experimental procedures

A suite of concrete samples that represent the range ofproperties likely to be encountered during biomass cofiringwith coal provide the baseline data for this document. Thecementitious portion of these samples contains neat cementas control and coal fly ash replacing 25% cement by mass.The fly ash has four resources: (1) coal fly ash, includingClass C and Class F; (2) wood fly ash from pure wood com-bustion; (3) blended biomass fly ash that cover Wood C(80% Class C blended with 20% wood fly ash by mass)

Fig. 1. Scanning electron micrographs (SEMs) of cement and fly ash. (a

and Wood F (80% Class F blended with 20% wood flyash by mass); and (4) cofired fly ash that cover SW1(80% Galatia coal fired with 20% switchgrass by mass)and SW2 (90% Galatia coal fired with 10% switchgrassby mass). The detailed properties of fly ash and fly ash con-crete including mix design, strength, freezing and thawingand rapid chloride permeability appear elsewhere, whichindicate that in related to all the above properties tested,biomass fly ash (except pure wood) has equal or better per-formances than coal fly ash in concrete [3,4].

Image analysis of specimens obtained from fragments ofconcrete samples used primarily for compressive strengthanalyses provide a sample suite for this investigation.Images primarily from unpolished surfaces of these frag-ments allow evidence of crystallization and other chemical

) Cement; (b) Class C; (c) Class F; (d) SW1; (e) SW2 and (f) Wood.

Fig. 2. Laminated CH crystal in Wood mix (658-day). (a) CH crystal and (b) EDX.

Fig. 3. Barely reactive fly ash particle in Class F mix (56-day).

374 S. Wang et al. / Fuel 87 (2008) 372–379

reactions to be developed. Scanning electron microscopy(SEM)-based images come from either Philips XL30 envi-ronmental scanning electron microscopy (ESEM) FEG orJEOL JSM840a, both of which are located in the Micros-copy Lab of Brigham Young University (BYU). ESEM

Fig. 4. The spectrum of reactive fly ash particle in SW1 mix (573-day

concrete slabs are either polished by 0.5 lm alumina parti-cles at BYU Geology Lab or prepared as polished thin sec-tions (24 mm · 46 mm · 30 lm) by Wagner PetrographicCompany in Provo, UT.

3. Results and discussion

3.1. Raw materials

SEM pictures of fly ash and cement appear in Fig. 1.Wood fly ash particles have irregular shapes while otherash particles show approximately spherical shapes.

3.2. Reactive fly ash particles by SEM and energy dispersive

X-ray (EDX) analysis

Fig. 2 shows the crystal of CH, the product of cementhydration and the reactants of pozzolanic reaction, which

). (a) SW1 fly ash particle; (b) EDX point 1 and (c) EDX point 2.

Fig. 5. Reactive class C fly ash particle one year.

Fig. 7. Reactive SW1 fly ash particle one year.

S. Wang et al. / Fuel 87 (2008) 372–379 375

has a laminar morphology in (a) and supported by EDXanalysis in (b) [1,5]. The fly ash particles in the concretemix react slowly with CH to form C–S–H gels, and the pro-cess goes on for years [2,6–8]. Fig. 3 shows the barely reac-tive Class F fly ash particles in a 56-day-old concretesample because the particles’ surfaces are quite smooth,which are only partially coated with some foreigncomponents.

Fig. 4a shows a broken cenosphere SW1 fly ash particlein concrete mix of 573-day. The shell and inner compo-nents come from different resources and have differentcompositions, which are verified by the EDX spectrum inFig. 4b and c.

Chemical composition of spot A

0

20

40

60

80

100

FeL MgK AlK SiK S K CaK Total

Elements

8KV

15KV

Chemical com

0102030405060708090

100

FeL MgK AlK

E

%(m

ol)

%(m

ol)

a

c

Fig. 6. Spot chemical analysis

3.3. Chemical analysis by ESEM

All the reactive fly ash particles samples are polishedthin sections except the raw SW1 and SAW fly ash particlesin Figs. 9 and 15, which are concrete slabs polished atGeology Lab of BYU.

Discussions about acceleration voltage about quantita-tive analysis are listed below. There are several major ele-ments in fly ash concrete, Al, Si, Ca and Fe with theabsorption edge voltage as 1.560, 1.840, 4.037 and7.110 KV, respectively. Those four elements can be only

FeL MgK AlK SiK S K CaK Total

Chemical composition of spot B

0

20

40

60

80

100

Elements

%(m

ol) 8KV

15KV

8KV

15KV

position of spot C

SiK S K CaK Total

lements

b

of Class C fly ash particle.

Chemical analysis of SW1 fly ash particle

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

1 2 3 4 5 6 7 8 9 10

Point

Ca/

Si (

mo

l)

Line A

Line B

Fig. 8. Line A and B scanning of SW1 fly ash particle (one year).

Fig. 10. Reactive SW2 fly ash particle one year.

376 S. Wang et al. / Fuel 87 (2008) 372–379

detected above their absorption edge voltage accordingly.For a good quantitative analysis of one element, the accel-eration voltage should be at least 2–3 times of its absorp-tion edge voltage. For example, for a better quantitativeanalysis, the acceleration voltage should be at least 8.074(4.037 · 2) KV for Ca and 14.220 (7.011 · 2) KV for Fe,respectively. However, higher acceleration voltage leadsto higher excitation volumes for the same density of mate-rials (if assuming the density of C–S–H gel 2.0 g/cm3 andfly ash particle 2.5 g/cm3, the diameter of the excitation vol-ume will be approximately 1.2 lm at 8 KV and 2–2.5 lm at15 KV). Therefore, the accurate quantitative chemicalanalysis of some tiny part of fly ash will be rough becauseof possibly incorporating detected areas beyond the tinyspecifically interested area (say here, if the dimensions areless than 1 lm). In this sense, a dilemma arises betweenaccurately quantifying the chemical composition of thesmall spot and the heavy elements interested, dependingon the objective of the detection. In real situation, for thequantitative chemical analysis, if there is no significantamount of iron in the sample, the acceleration voltage ischosen as 8 KV with scanning distances between the neigh-boring points more than 1 lm, thus avoiding the severelyinterfering effects from near neighborhood.

A reactive Class C fly ash particle was studied in Fig. 5.The quantitative chemical analysis of three individual

a b

Fig. 9. Chemical analysis of raw SW1 fly ash particle. (a

points A, B and C on the reaction ring is tested at 8 KVand 15 KV, respectively; the results were listed in Fig. 6.For this fly ash particle, chemical compositions do notchange much for the three points detected at 8 and15 KV, respectively. This measurement has verified the for-mer discussions of acceleration voltage: 2–3 times of theabsorption edge voltage is enough for accurate quantitativechemical analysis.

The line scanning was applied to SW1 fly ash particle inFig. 7. Line A of 10.71 lm in length was scanned with 5seconds dwelling time on each equally spaced point. LineB was 2.23 lm long and followed the same ideas. FromFig. 7, line A was on the main body of fly ash particleand inert; line B was on the reaction ring of ash particle.Further chemical analysis in Fig. 8 has confirmed this fact:line A has really flat Ca/Si mol ratio in the neighborhoodof 0.02, which is consistent with bulk chemical analysis ofSW1 fly ash by XRF (Ca/Si (mol) = 0.04) [3]; and line Bhas a much higher ratio of Ca/Si (mol) in the range of0.5–1.7, within the typical range of C–S–H gel [1].

Fig. 9 illustrates a raw SW1 fly ash particle with thechemical analysis. It can be seen that the results in Figs.

0.000

0.005

0.010

0.015

0.020

0.025

1 2 3

Position

Ca/

Si (

Mo

l)

) Raw SW1 fly ash and (b) ESEM chemical analysis.

Chemical composition of reactive SW2 ash particle

Ca/Si (mol)

1 2 3 4 5 6 7 8 9 10

Point

0

0.20.4

0.60.8

11.21.41.61.8

Ca/

Si (

mo

l)

Fig. 11. Chemical composition of reactive SW2 fly ash particle.

Fig. 12. Two reactive SW2 fly ash particles with smaller size.

Chemical analysis of SW2 fly ash (particle 1)

Ca/

Si (

mo

l)

1 2 3 4 5 6 7 8 9 10

Point

0

0.2

0.4

0.6

0.8

1

1.2

Ca/Si (mol)

Fig. 13. Chemical analysis of SW2 fly ash with smaller particle size(particle 1).

Chemical analysis of SW2 fly ash (particle 2)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

1 2 3 4 5 6 7 8 9 10

Point

Ca/

Si (

mo

l)

Ca/Si (mol)

Fig. 14. Chemical analysis of SW2 fly ash with smaller particle size(particle 2).

S. Wang et al. / Fuel 87 (2008) 372–379 377

7–9 are very consistent: the raw SW1 fly ash particles havethe Ca/Si mol ratio about 0.02, which is close enough tothat of its bulk chemical composition, 0.04 [3].

A reactive SW2 fly ash particle with approximate diam-eter of 11 lm was illustrated in Figs. 10 and 11. In Fig. 10,points 2–9 are on the main inert body of the particle, whichare confirmed by chemical analysis in Fig. 11 (Ca/Si molratio about 0.05). Likewise, point 1 and 10 are in the reac-tion ring of the particle supported by the chemical analysis(Ca/Si mol ratio of 1.6) in Figs. 10 and 11, respectively.Again, this ash particle (Ca/Si mol ratio of 0.05) is repre-

sentative of the bulk chemical analysis (Ca/Si(mol) = 0.05) of the raw SW2 ones [3].

Two reactive SW2 fly ash particles of smaller size (diam-eters about 5 lm) were illustrated in Figs. 12–14. Accelera-tion voltage was fixed at 8 KV due to the small particle sizeand the small distances between neighboring scannedpoints. For particle 1 in Fig. 12, points 1, 2, 3 and 8, 9,10 are on the reaction ring part and points 4, 5, 6 and 7are on the solid and inner part. Points 2 and 9 with Ca/Si mol ratio of about 0.7 from Fig. 13 imply the particlehas undergone significant pozzolanic reaction. Points 5and 6 of particle 1 in Fig. 12 are located on its inner parti-cle. It is clear from Fig. 13 that points 5 and 6 have Ca/Simol ratio of 0.05. The image in Fig. 12 and the chemicalanalysis in Fig. 13 indicate that they most probably remaininert.

Particle 2 in Figs. 12 and 14 shows the similar situationas particle 1. Points 2 and 8 on the ring have Ca/Si molratio as 0.4–0.9, a strong evidence of pozzolanic reaction;point 5 on the solid inert particle body has Ca/Si (mol)ratio as 0.01, a clear evidence of being inert.

A comparison of SW2 fly ash particles in Figs. 10 and 12indicated that smaller fly ash particles have higher reactionextent than the larger ones, which could be attributed todifference relative surface areas, because the Ca/Si molratio near the centers of the larger (Fig. 10) and smaller(Fig. 12) particles is quite close to each other. Therefore,the particle size has important influences on pozzolanicreactivity.

A raw SW2 fly ash particle with chemical analysis wasillustrated in Fig. 15. Comparison of raw and reactiveSW2 fly ash particles from Figs. 10–14 indicates that rawSW2 fly ash particle do have very low Ca/Si mol ratio lessthan 0.1, which is consistent with its bulk chemical analysisby XRF (Ca/Si (mol) = 0.05) [3].

A reactive iron rich particle was illustrated in Fig. 16.For line A, iron mol percentage was about 5% at 8 KVand 55% at 15 KV, respectively; therefore, for better quan-titative chemical analysis, acceleration voltage was fixed at15 KV. On line A in Fig. 16, points 2–9 are on the mainbody of the iron particle, and the Ca/Si (mol) ratio is about0.05; but point 1 and point 10 near the reaction ring have

0.000

0.005

0.010

0.015

0.020

0.025

0.030

1 2 3

Position

Ca/

Si (

Mo

l)

a b

Fig. 15. Chemical analysis of raw SW2 fly ash particle. (a) Raw SW2 fly ash and (b) ESEM chemical analysis.

Reactive iron ash particle

0

0.5

1

1.5

2

2.5

3

1 2 3 4 5 6 7 8 9 10

Point

Ca/

Si (

mo

l)

Line A at 15KV

Line B at 15KV

Fig. 17. Chemical analysis of reactive iron fly ash particle one year.

Fig. 16. Reactive iron ash particle one year.

378 S. Wang et al. / Fuel 87 (2008) 372–379

Ca/Si (mol) ratio of 0.3, which is a strong evidence of poz-zolanic reaction. Further investigation of line B on the

reaction ring with Ca/Si mol ratio from 1.8 to 2.6 con-firmed the strong pozzolanic reaction (see Fig. 17).

4. Conclusions

By comparisons of morphology, spot chemical analysisof raw fly ash particles and the ones staying in concretefrom 2 months to 1.5 years, the following conclusionscan be drawn:

1. Within one year period curing in concrete, many fly ash(coal and biomass) particles have formed a ring of reac-tion product surrounding them, with inner cores remain-ing inert.

2. Smaller fly ash particles have higher reactivity due totheir higher relative surface areas.

3. Fly ash particles with high iron contents (>50% mol) canhave significant pozzolanic reactivity.

4. Acceleration voltage should be carefully chosen forgood quantitative analysis in concrete: (1) 8 KV workswell if no significant amount of elements with atomicnumbers greater than calcium; otherwise, a higher accel-eration voltage is recommended based on the rule of 2–3times of the absorption edge voltage of the interestedelement.

References

[1] Mindness S, Young JF, Darwin D. Concrete. 2nd ed. Upper SaddleRiver, NJ 07458: Pearson Education; 2003.

[2] Taylor HFW. Cement Chemistry. 2nd ed. 1 Heron, Quay, London,E144 JD: Thomas Telford; 1997.

[3] Wang S, et al. Biomass fly ash in concrete: mixture proportioning andmechanical properties. Fuel, doi:10.1016/j.fuel.2007.05.026.

[4] Wang S, et al. Durability of biomass fly ash in concrete: freezing andthawing and rapid chloride permeability tests. Fuel, doi:10.1016/j.fuel.2007.05.027.

[5] Helmuth Richard. Fly ash in cement and concrete. Skokie, IL60077: Portland Cement Association; 1987.

S. Wang et al. / Fuel 87 (2008) 372–379 379

[6] Hewlett Peter. Lea’s chemistry of cement and concrete. 4th ed. JordanHill, Oxford, OX2 8DP: Butterworth-Heinemann Linacre House;1997.

[7] Menetrier D, Jawed I, Sun TS, Skalny J. ESCA and SEM studies inearly C3S Hydration. Cement Concrete Res 1979;9:473–82.

[8] Ghose A, Pratt PL. Studies of the Hydration Reactions and Micro-structure of Cement-Fly ash Paste. in: The conference ‘‘Effects of FlyAsh Incorporation in Cement and Concrete.’’ Editor, Sydney Dia-mond, 1981, pp. 82–91.