High-efficiency cogeneration boiler bagasse-ash geochemistry and mineralogical change effects on the potential reuse in synthetic zeolites, geopolymers, cements, mortars, and concretes.

Sugarcane bagasse ash re-utilisation has been advocated as a silica-rich feed for zeolites, pozzolans in cements and concretes, and geopolymers. However, many papers report variable success with the incorporation of such materials in these products as the ash can be inconsistent in nature. Therefore, understanding what variables affect the ash quality in real mills and understanding the processes to characterise ashes is critical in predicting successful ash waste utilisation. This paper investigated sugarcane bagasse ash from three sugar mills (Northern NSW, Australia) where two are used for the co-generation of electricity. Data shows that the burn temperatures of the bagasse in the high-efficiency co-generation boilers are much higher than those reported at the temperature measuring points. Silica polymorph transitions indicate the high burn temperatures of ≈1550 °C, produces ash dominated α -quartz rather than expected α-cristobilite and amorphous silica; although α-cristobilite, and amorphous silica are present. Furthermore, burn temperatures must be ≤1700 °C, because of the absence of lechatelierite where silica fusing and globulisation dominates. Consequently, silica-mineralogy changes deactivate the bagasse ash by reducing silica solubility, thus making bagasse ash utilisation in synthetic zeolites, geopolymers, or a pozzolanic material in mortars and concretes more difficult. For the ashes investigated, use as a filler material in cements and concrete has the greatest potential. Reported mill boiler temperatures discrepancies and the physical characteristics of the ash, highlight the importance of accurate temperature monitoring at the combustion seat if bagasse ash quality is to be prioritised to ensure a usable final ash product.

Introduction

Brazil, the world’s largest sugarcane cultivator, generates over 2.5 million tonnes of sugarcane bagasse ash (SCBA) per annum (Faria et al., 2012), whilst Australia generates some 40 thousand tonnes (CANEGROWERS, 2013). Sugar cane bagasse (left over cane stalk after crushing to extract cane juice) is a significant waste product of sugar processing. Traditionally bagasse was burnt for steam generation within the sugar mill and the resulting heat and energy used to process the sugarcane juice into raw sugar. The primary purpose of these early boilers was to dispose of the excess bagasse and would typically be burnt inefficiently. However, with uncertain profitability of the sugar cane industry over the last few decades, the cogeneration of electricity to increase revenue has gradually become a commercial reality where two of the three NSW sugar mills export electricity to the grid (CANEGROWERS, 2013). Initially this was done during the operating season, but now extends into the slack season. Consequently, the supply of bagasse is insufficient, and mills are supplementing bagasse with fuels such as wood chip, fuel oil, and coal. This is possible because of more efficient boiler design and operating conditions aimed at maximising cogeneration capacity (CANEGROWERS, 2013).

While burning bagasse reduces waste and provides economic benefit through cogeneration, bagasse ash is still produced as a waste material, which requires disposal. The ash from the boilers has been, and is still often blended with mill mud from juice clarification and disposed of back to the cane farms (CANEGROWERS, 2013). Although, this has been a very simple, easy, and cheap means of disposal while avoiding the regulation of environmental agencies, it is often no longer possible to achieve in many jurisdictions (Ahmaruzzaman, 2010). This is largely due to environmental and health concerns (le Blond et al., 2010) and the possibility of alternative value added solutions. Furthermore, the ash from these operations is blended into mill mud during the crushing season. However, cogeneration has become a year round operation and this disposal method may no longer be appropriate (Ahmaruzzaman, 2010). Consequently, alternatives uses need to be sought, and the use of SCBA as a direct absorbent (Rehab and El Anany, 2013), glass-ceramic material production (Teixeira et al., 2014), geopolymers (Castaldelli et al., 2013), and zeolite production (Moisés et al., 2013) have been made. Furthermore, considerable research has been conducted on SCBA waste utilization in the concrete industry, where most of the work has focused on SCBA as either a cementitious replacement, or as a filler material (Cordeiro et al., 2009a; Cordeiro et al., 2009b; Cordeiro et al., 2008). Similarly, with rice husk ash (RHA) and other organic agricultural wastes, potential biogenic silica sources, advanced materials including SiO2, SiC, Si3N4, elemental Si, Mg2Si, and metal-matrix composite have been made (Soltani et al., 2015; Bahrami et al., 2016). Moreover, in their review Soltani et al. (2015) also note that rice husk and/or the ash, may be used as: a rubber filler; energy extraction (thermochemical as for bagasse cogeneration, and biochemical); absorbents (water purifiers, gasifiers, and petroleum; as per Rehab and El Anany, 2013 for bagasse); pozzolans (concretes and cements; also identified by Cordeiro et al., 2008; Cordeiro et al., 2009a; Cordeiro et al., 2009b for baggasse), as an insulator (both in the steel and refactrory brick industries); for production of electronic and solar grade silicas; and pigments. Hence, from this RHA work SCBA should also provide similar products.

The use of SCBA as a feedstock for synthetic zeolite manufacture is potentially an economic and attractive environmentally alternative to local mining or importation. Sugar cane (Saccharum officinarum), a member of the grass family, accumulates amorphous silica within cells typically at an average rate of 1.509% by mass SiO2 per year (Hodson et al., 2005), which is on the upper end of silica accumulation in plants (Lovering, 1959). Parr et al. (2009a) indicate that silica (phytolith) content varies from 1.3–2.6% depending on the varietal. When this organic matter is burnt it results in an ash that is much higher in silica than other boiler ashes such as coal. Moreover, the plant acts as a natural filter, excluding many trace metals from accumulation, which is problematic for coal fly ashes (Hodson et al., 2005). Hence, it is highly likely that SCBAs have properties desirable for zeolite production without the need to provide extra silica; synthesised zeolites have many industrial, environmental and agricultural applications.

Zeolites are alumino-silicate minerals that have a repeating channel or cage-like structure resulting from the repeating silica tetrahedral and alumina octahedral bonding (Deer et al., 1986). These structures have varying size cages or channels, and have strong ion exchange properties due to the negative charge on the framework from the incomplete tetrahedral structure of alumina. This negative charge is balanced usually by an alkali metal cation loosely held in the cage or channel bonding (Deer et al., 1986). Although zeolites are found naturally and have commercial use as such, zeolite minerals can also be synthesized from aluminium and silicon containing admixtures. The advantage in synthesizing a zeolite comes from the control of cage or channel size, control of crystal size (i.e. higher surface area on smaller crystals), and control of purity (i.e. activity of the zeolite) (Garcıa-Martınez et al., 2002). At higher temperatures (>175 °C) more crystalline and compact zeolitic structures develop and often lead to crystallisation of sodalite and cancrinite (frequently found in bauxite refinery residues as desilication products). Whereas, at more moderate temperatures (≈120 °C − 160 °C) more open zeolites (e.g. analcime) develop, and at lower temperatures (≈80 °C) even more open and less-crystalline structures (e.g. zeolite-A and zeolite-X) develop (Coombs et al., 1959; Mashal et al., 2005; Ocanto et al., 2009; Thuadaij et al., 2012). The consequence of decreasing crystallinity is that Cation Exchange Capacity (CEC) tends to increase, and therefore the industrial and environmental applications are likely to increase. Sugar cane boiler ash, because of the high silica content of the cell walls, provides the ash higher Si:Al ratios that influence the zeolites produced (Lin and Hsi, 1995; Mashal et al., 2005; Thuadaij et al., 2012).

SCBA may also be used for the development of geopolymers, cements, mortars, and concretes, which are alkaline activated, and may be used in products for construction works in acidic environments, decorative wares, and garden pavers (Cordeiro et al., 2009a; Cordeiro et al., 2009b; Cordeiro et al., 2008; Faria et al., 2012). Geopolymers a class of inorganic polymer formed by the reaction between an alkaline solution and an alumino-silicate source or feedstock. The hardened material has an amorphous 3-dimensional structure similar to that of an alumino-silicate glass. Joseph Davidovits first described geopolymerisation, in his numerous papers and patents (e.g., (Davidovits, 1988; Davidovits et al., 1994). Geopolymerisation occurs where an alkali metal hydroxide/silicate solution (the activator) and an alumino-silicate fine binder are blended appropriately (Davidovits et al., 1994). Most commonly the activator is a mixture of water, sodium hydroxide and sodium silicate, but other alkali metal systems or admixtures can be used. Unlike zeolitisation, geopolymerisation requires strong alkali reactions; hence the activator needs to be concentrated. Consequently, concentrated waste alkalis may be utilised as activator (Davidovits et al., 1994). Because the binder needs to be fine grained with a high concentration of amorphous alumino-silicates, binders commonly include class-F coal-fly-ash, ground granulated slag, meta-kaolin, or any suitable fine amorphous alumino-silicate material. In the hardened material the resultant products are a rigid chain or net of geopolymer and a pore solution of excess alkali metal ions and un-reacted silicon hydroxide (Davidovits, 1988).

Many cements, mortars and concretes include supplementary cementitious material (SCM) to improve performance and/or reduce cement contents. A suitable SCM material must be pozzolanic in nature, being either siliceous and/or alumino-siliceous materials that may not be naturally cementitious, but form cementitious compounds by reacting with calcium hydroxide (Ca(OH)2) and water (International, 2014). Pozzolanic materials are categorised if its strength activity index (SAI), determined by comparing the compressive strength of a mortar containing the potential pozzolanic material mortar is ≥75% to a control mortar (International, 2012). In addition, the silica, alumina, and iron oxide content (SiO2 + Al2O3 + Fe2O3) must be ≥ 50% for class-C pozzolans, and ≥ 70% for class-F pozzolans. (International, 2013; International, 2012)

Concrete durability in acidic and other aggressive environments (e.g., sewage systems, acid sulphate soils) can be prone to sulphuric acid attack, resulting in rapid deterioration (Beddoe and Dorner, 2005). Some SCMs have been shown to improve concrete durability in these harsh environments (e.g., (Chang et al., 2005), because of the SCM pozzolanic consumption of Ca(OH)2 and aluminates. Excess Ca(OH)2 and aluminate react with H2SO4 producing low density and expansive gypsum and ettringite, which leads to cracking and spalling. Thus, incorporating SCMs in concrete exposed to aggressive environments not only reduces the demand on cement production, but also reduces economic losses due to repairs, maintenance and reduced service life (Chang et al., 2005). Industrial waste materials are often used as pozzolans including coal fly ash (CFA) (Manz, 1999), granulated blast furnace slag (GBFS) (Australia, 2001), and agricultural wastes such as rice husk ash (RHA) (Barbhuiya et al., 2006) and sugarcane bagasse ash (SCBA) (Cordeiro et al., 2009a; Cordeiro et al., 2009b; Cordeiro et al., 2008).

This paper investigates the composition and mineralogy of 3 Northern New South Wales sugar mill fly ash (2 are cogenerating plants), to characterise these materials for potential applications such as zeolitisation, geopolymerisation and/or inclusion in cements and mortars. Specific factors of bagasse combustion and the effects on the resulting ash are also considered.

Methods

Collection of fly ash from the three NSW mills: Hardwood, Condong and Broadwater sugar mills was undertaken from the start of July 2011 until October 2011. Some 41 different ashes were collected and analysed including ashes produced from using bagasse, wood chip, coal and oil, or a combination as fuel for the boilers. Mill operation parameters such as FD airflow and temperature, steam-flow, oxygen and carbon monoxide concentrations, and convection bank temperature were recorded (Table 1).

Table 1

Sample collection data for the three NSW Mills (Harwood HWD; Congdong CDG; Broadwater BWR), including mills operation parameters.

Plant origin	Sample date	Sample time (24h)	Fuel	FD air flow [t/hr]	Steam flow [t/hr]	Oxygen [%]	Carbon monoxide [ppm]	Convection bank temp. [°C]	FD air temp. [°C]
Harwood	30.09.11	15:00	Bagasse	204	114	6.8	n/a	844	27
	4.10.11	14:30	Bagasse	223	131	4.8	n/a	896	27
	5.10.11	15:00	Bagasse	207	119	5.5	n/a	893	25
	6.10.11	15:30	Bagasse	200	126	4	n/a	908	22
	11.10.11	10:30	Coal & oil	193	89	11.1	n/a	761	27
	12.10.11	14:50	Coal, oil & bagasse	216	127	5.1	n/a	869	29
	14.10.11	16:00	Bagasse	182	117	4.3	n/a	858	27
	26.10.11	15:45	Bagasse	225	122	5.1	n/a	873	27
	27.10.11	16:00	Bagasse	226	124	5.5	n/a	856	27
	2.11.11	15:00	Bagasse	240	120	6.7	n/a	851	33
	4.11.11	10:40	Coal, oil & bagasse	226	128	5	n/a	889	28
	8.11.11	15:45	Bagasse	226	128	4.8	n/a	880	32
	9.11.11	15:00	Bagasse	225	124	4.9	n/a	868	34
	10.11.11	16:00	Bagasse	217	130	3.8	n/a	902	33
	17.11.11	14:00	Bagasse	223	123	4.8	n/a	884	31
Condong	10.10.11	13:30	Wood chip (camphor)	245.4	85.2	2.13	1053	509.3	35.3
	11.10.11	12:30	Bagasse & wood chip	222.2	104.4	3.7	214	506.1	40.2
	13.10.11	12:00	Bagasse & wood chip	214.6	106.12	2.46	392	512.1	39
	15.10.11	15:00	Wood chip	231.3	84.94	4.24	1052	506.1	40.5
	16.10.11	17:11	Wood chip	205.6	70.91	6.15	1053	479.6	40.1
	20.10.11	11:30	Bagasse	231.1	100.34	4.26	266	489.2	40.8
	21.10.11	11:45	Bagasse	209.6	86.9	4.12	194	472	45
	27.10.11	15:20	Bagasse	219.9	97.56	3.64	378	490.5	41.3
	28.10.11	11:00	Bagasse & wood chip	218.1	80.71	4.8	161	472.2	47.1
	1.11.11	14:40	Bagasse	173.1	92.7	5.88	107	486	41.4
	3.11.11	13:00	Bagasse	216	94.85	3.79	114	486.9	41.4
	4.11.11	15:00	Bagasse & wood chip	196.9	92.29	5.53	152	494.1	42.6
	5.11.11	11:30	Bagasse & wood chip	183.8	63.7	5.91	214	461.4	49.2
	6.11.11	11:10	Wood chip (mixed)	177.2	60.22	6.64	376	458	43.9
	7.11.11	12:30	Bagasse & wood chip	227.1	89.99	4.67	185	499.2	39.9
	8.11.11	16:00	Bagasse & wood chip	184.6	81.72	4.11	228	480	41
	10.11.11	08:35	Bagasse	214.2	92.82	5.6	114	492.5	41.7
	11.11.11	15:30	Bagasse & wood chip	224	88.36	5.97	211	489	43
Broadwater	15.09.11	06:00	Bagasse & wood chip	235	115	4	n/a	480	254
	20.09.11	06:00	Bagasse & wood chip	235	115	4	n/a	480	254
	26.09.11	06:00	Bagasse & wood chip	235	115	4	n/a	480	254
	26.10.11	06:00	Bagasse & wood chip	235	115	4	n/a	480	254
	2.11.11	n/a	Bagasse & wood chip	n/a	n/a	n/a	n/a	n/a	n/a
	16.11.11	n/a	Bagasse & wood chip	n/a	n/a	n/a	n/a	n/a	n/a
	30.11.11	n/a	Bagasse & wood chip	n/a	n/a	n/a	n/a	n/a	n/a
	12.12.11	n/a	Bagasse & wood chip	n/a	n/a	n/a	n/a	n/a	n/a

X-ray fluorescence (XRF)

Geochemical analyses were undertaken using X-Ray fluorescence (XRF) method for major elements Fe2O3, SiO2, Al2O3, MnO, TiO2, CaO, MgO, K2O, P2O5, and Na2O and for additional elements CuO, NiO, CoO, Cr2O3, PbO, SO2, ZnO, As2O3, SnO, SrO, ZrO, BaO, V2O5, Cl, and LOI. XRF is a standard method for whole sample analysis and samples were submitted to Southern Cross’s Environmental Analysis Laboratory, for XRF analysis. XRF analysis for major elements was made on lithium borate fused glasses, while trace elements analyses were made using pressed powder disc.

Light microscopy

Light microscopy observations were made on samples from both feedstock of bagasse and wood chip (Camphor Laurel; Cinnamomum camphora) from Harwood and Condong mills. Samples were placed into centrifuge tubes and centrifuged at 3500 rpm for 3 minutes; the supernatant was then decanted. This wash process was repeated with distilled water. Approximately 0.5 mL of agitated sample from both samples was removed with a pipette and placed onto too warm microscope slides. Once the water had evaporated, two drops of benzyl benzoate were placed onto each slide, followed by a cover slip. Edges of the cover slip were sealed with Depex mounting medium (Merck Pty. Ltd.) to prevent the benzyl benzoate from dispersing. Slides were scanned on an Olympus CX40 trinocular compound microscope (Olympus Corp.) at 400x magnification, and digital images of phytoliths were captured.

Scanning electron microscopy (SEM)

SEM is widely used to allow high-resolution visualisation of materials and is well suited to characterising phytolith morphology features such as topography and dimensions. Sub-samples of Harwood, Condong, and Broadwater ash were centrifuged in water 4 consecutive times at 4500 rpm for 3 min each, to remove the bulk of the lightweight organic material (charcoal) and concentrate the heavier, silica-carrying material, which were dried at 70 °C for 24 hours, then prepared for SEM.

Specimens were mounted on standard 12 mm SEM pin mounts using double-sided low-contamination carbon dots. The samples were coated with carbon, using a Balzers SCD 050 sputter-coater with carbon-evaporative attachment and carbon thread, to allow x-ray analysis. Both, imaging and analyses were done on a Leica Stereoscan 440 system with Tungsten filament as the electron source at a working distance of 15 to 25 mm, EHT of 10 to 20 kV with a probe current of 50 pA to 200 μA. The medium to high voltage and low beam current provides very good resolution, while minimising potential surface charging problems of flaring, excessive glow and edge effects, or other unwanted distortions, which can occur on carbon coated specimens at higher voltage and amperage settings.

X-ray diffraction (XRD)

X-Ray diffraction (XRD) provides an assessment of the mineralogy of materials and can provide estimates of crystalline, glassy, and amorphous mineral components. In this study it is particularly useful to investigate the amorphous and crystalline form of Si. XRD analyses were conducted using a Bruker D4 Endeavor advance diffractometer using Fe-filtered Co Kα1 light with a graphite monochromator to provide monochromatic Co Kα1 x-rays (1.7886 Å), using a Sol-X detector at room temperature. Samples were analysed at 23 °C from 5.0 − 80.1° 2θ (20.51-1.39 Å d-spacing) using Step of 0.036° 2θ at a scan rate of 1° 2θ per minute. XRD data were collected using XRD Commander software and analysed using JADE 7.0® to determine peak positions, identify phases using ICCD databases.

Results

XRF analysis shows difference in composition for major and minor elements between ashes mainly depending on the fuel (Table 2, Table 3). Coal, oil & bagasse fuel ashes are similar in composition to bagasse fuel ashes (Table 4), however wood chip additions begins to lower the SiO2 and Al2O3 contents. Coal, oil & bagasse and bagasse ash, present an average concentration of SiO2 >70 wt.% and of Al2O3 >13 wt.% (Table 4), while wood chip ashes have much a lower concentration of SiO2 (≈28 wt.%) and of Al2O3 (≈4 wt.%). Wood chip fuel ashes contain higher percentage of CaO (≈10 wt.%), than bagasse ashes (alone or mixed) (∼1 wt.%). Similarly, this is also true for MgO (≈3.5 wt.%) in wood chip compared to (∼1 wt.%) for other fuels except coal and oil (∼0.5 wt.%; Table 4)

Table 2

X-ray fluorescence analysis on fly ash samples from Harwood, Condong and Broadwater mills using different type of fuel for major elements Fe, SiO2, Al2O3, MnO, TiO2, CaO, MgO, K2O, P2O5, Na2O, and for loss on ignition (LOI); all results are in %.

Plant origin	Fuel	Sample date	Fe2O3	SiO2	Al2O3	MnO	TiO2	CaO	MgO	K2O	P2O5	Na2O	LOI
Harwood	Bagasse	30.09.11	3.10	70.68	13.79	0.12	0.84	0.99	0.95	2.57	0.27	1.24	5.45
	Bagasse	4.10.11	3.11	70.57	13.68	0.15	0.82	0.81	0.87	2.41	0.25	1.28	6.05
	Bagasse	5.10.11	3.48	71.33	13.64	0.11	0.83	0.87	0.91	2.50	0.25	1.34	4.74
	Bagasse	6.10.11	2.83	69.27	12.98	0.14	0.75	1.29	0.95	2.69	0.29	1.20	7.61
	Coal & oil	11.10.11	1.83	61.60	14.89	0.08	0.93	0.80	0.52	1.71	0.18	0.79	16.67
	Coal, oil & bagasse	12.10.11	3.13	69.11	14.13	0.10	0.84	0.83	1.00	2.81	0.30	1.30	6.45
	Bagasse	14.10.11	2.87	74.68	12.82	0.13	0.78	0.79	0.88	2.28	0.23	1.05	3.49
	Bagasse	26.10.11	2.71	70.17	17.38	0.07	0.94	0.59	0.94	2.39	0.22	1.12	3.47
	Bagasse	27.10.11	3.21	73.00	13.72	0.14	0.91	1.01	1.02	2.59	0.21	1.50	2.69
	Bagasse	2.11.11	2.86	71.17	12.50	0.10	0.79	1.20	1.03	2.18	0.22	1.33	6.62
	Coal, oil & bagasse	4.11.11	2.86	72.42	13.04	0.13	0.83	0.92	1.05	2.44	0.23	1.42	4.66
	Bagasse	8.11.11	2.76	74.19	12.42	0.08	0.76	0.76	1.09	2.16	0.25	1.07	4.46
	Bagasse	9.11.11	2.94	74.47	11.84	0.08	0.78	0.89	1.03	2.18	0.26	1.21	4.32
	Bagasse	10.11.11	2.35	75.81	11.38	0.08	0.74	0.73	1.01	2.12	0.24	1.15	4.39
	Bagasse	17.11.11	2.63	73.49	11.44	0.10	0.73	1.07	1.13	2.39	0.26	1.64	5.12
Condong	Wood chip (camphor)	10.10.11	1.14	24.39	3.68	0.40	0.36	10.81	4.26	5.31	0.53	0.31	48.81
	Bagasse & wood chip	11.10.11	2.40	68.76	9.46	0.11	0.95	1.46	0.82	2.23	0.37	0.89	12.55
	Bagasse & wood chip	13.10.11	2.08	51.55	7.43	0.15	0.83	4.30	1.35	2.62	0.40	0.86	28.43
	Wood chip	15.10.11	0.80	14.44	2.78	0.33	0.37	9.44	2.88	2.16	0.44	0.28	66.08
	Wood chip	16.10.11	1.20	16.90	3.52	0.36	0.45	10.01	3.18	4.52	0.40	0.32	59.14
	Bagasse	20.10.11	4.29	66.19	12.01	0.15	1.47	1.66	0.99	2.04	0.43	0.85	9.92
	Bagasse	21.10.11	3.89	65.03	13.24	0.19	1.44	1.42	1.11	2.14	0.48	0.63	10.43
	Bagasse	27.10.11	4.77	65.64	14.55	0.16	1.55	1.25	1.29	1.95	0.37	0.96	7.51
	Bagasse & wood chip	28.10.11	3.13	66.64	9.63	0.13	1.19	2.20	1.18	1.63	0.25	0.86	13.16
	Bagasse	1.11.11	3.55	68.01	13.88	0.08	1.32	1.30	1.25	1.91	0.39	0.99	7.32
	Bagasse	3.11.11	3.09	65.86	12.78	0.14	1.04	1.25	1.49	2.56	0.40	0.85	10.54
	Bagasse & wood chip	4.11.11	2.78	60.36	10.02	0.17	1.02	2.76	1.69	2.41	0.40	0.88	17.51
	Bagasse & wood chip	5.11.11	2.20	70.34	7.69	0.17	0.84	2.79	1.53	2.35	0.34	0.72	11.03
	Wood chip (mixed)	6.11.11	2.01	56.20	5.44	0.34	0.89	10.58	3.46	2.77	0.43	0.49	17.39
	Bagasse & wood chip	7.11.11	2.39	54.62	8.60	0.22	0.81	5.33	2.79	3.29	0.39	0.31	21.25
	Bagasse & wood chip	8.11.11	2.77	67.98	13.11	0.10	1.00	1.45	1.15	2.29	0.32	0.89	8.94
	Bagasse	10.11.11	2.48	67.09	10.17	0.09	0.92	1.41	1.32	2.05	0.45	0.97	13.05
	Bagasse & wood chip	11.11.11	2.22	53.11	7.23	0.14	0.71	3.45	1.33	1.84	0.36	0.83	28.78
Broadwater	Bagasse & wood chip	15.09.11	3.89	66.67	12.10	0.12	1.73	1.00	0.60	1.55	0.46	0.27	11.61
	Bagasse & wood chip	20.09.11	3.80	63.14	14.07	0.12	2.03	1.53	0.52	2.10	0.40	0.77	11.52
	Bagasse & wood chip	26.09.11	4.70	64.95	14.77	0.17	1.90	0.96	0.45	1.68	0.30	0.40	9.72
	Bagasse & wood chip	26.10.11	2.32	70.86	7.97	0.15	0.91	1.68	0.61	1.25	0.25	0.08	13.92
	Bagasse & wood chip	2.11.11	4.17	68.96	12.45	0.10	2.15	0.78	1.00	1.55	0.33	0.08	8.43
	Bagasse & wood chip	16.11.11	2.13	52.48	9.61	0.25	1.04	3.02	1.36	1.86	0.39	0.04	27.82
	Bagasse & wood chip	30.11.11	4.92	64.67	12.41	0.11	2.11	0.93	1.22	1.63	0.33	0.81	10.86
	Bagasse & wood chip	12.12.11	4.96	63.66	12.33	0.12	2.12	0.96	1.25	1.67	0.34	0.81	11.78

Table 3

X-ray fluorescence analysis on fly ash samples from Harwood, Condong and Broadwater mills using different type of fuel for trace Cu, Ni, Co, Cr, Pb, S, Zn, As, Sn, Sr, Zr, Ba, V, Cl; all data are mg/kg.

Plant origin	Fuel	Sample date	CuO	NiO	CoO	Cr2O3	PbO	SO2	ZnO	As2O5	SnO2	SrO	ZrO	BaO	V2O5	Cl
Harwood	Bagasse	30.09.11	0.015	<0.001	0.002	0.005	0.004	0.151	0.011	<0.001	<0.001	0.014	0.029	0.052	0.008	<0.001
	Bagasse	4.10.11	0.015	<0.001	0.001	0.006	0.004	0.011	0.01	<0.001	<0.001	0.012	0.03	0.049	0.009	<0.001
	Bagasse	5.10.11	0.015	<0.001	0.002	0.006	0.005	0.011	0.011	<0.001	<0.001	0.013	0.027	0.047	0.007	<0.001
	Bagasse	6.10.11	0.016	<0.001	<0.001	0.006	0.004	0.023	0.012	<0.001	<0.001	0.015	0.028	0.05	0.001	<0.001
	Coal & oil	11.10.11	0.017	<0.001	0.001	0.003	<0.001	0.017	0.024	<0.001	<0.001	0.018	0.032	0.05	0.016	<0.001
	Coal, oil & bagasse	12.10.11	0.015	<0.001	0.001	0.005	0.004	0.011	0.014	<0.001	<0.001	0.014	0.025	0.047	0.008	<0.001
	Bagasse	14.10.11	0.016	<0.001	0.002	0.005	0.006	0.011	0.012	<0.001	<0.001	0.011	0.028	0.049	0.007	<0.001
	Bagasse	26.10.11	0.033	0.011	0.021	0.018	0.023	0.011	0.021	0.014	0.03	0.022	0.04	0.044	0.014	0.008
	Bagasse	27.10.11	0.034	0.01	0.021	0.019	0.024	0.015	0.02	0.014	0.031	0.025	0.048	0.047	0.009	0.007
	Bagasse	2.11.11	0.035	0.016	0.02	0.02	0.021	0.013	0.018	0.014	0.027	0.025	0.047	0.038	0.011	0.013
	Coal, oil & bagasse	4.11.11	0.034	0.01	0.022	0.02	0.022	0.013	0.018	0.014	0.027	0.024	0.045	0.044	0.011	0.007
	Bagasse	8.11.11	0.035	0.01	0.021	0.019	0.022	0.017	0.02	0.014	0.026	0.021	0.039	0.037	0.01	0.009
	Bagasse	9.11.11	0.036	0.01	0.021	0.02	0.023	0.011	0.021	0.013	0.023	0.021	0.043	0.042	0.009	0.012
	Bagasse	10.11.11	0.033	0.016	0.021	0.019	0.022	0.011	0.02	0.014	0.026	0.021	0.04	0.036	0.01	0.012
	Bagasse	17.11.11	0.038	0.01	0.021	0.021	0.021	0.011	0.018	0.014	0.018	0.025	0.043	0.039	0.008	0.013
Condong	Wood chip (camphor)	10.10.11	<0.001	0.004	<0.001	0.003	0.004	0.103	0.007	<0.001	<0.001	0.08	0.038	0.156	<0.001	0.154
	Bagasse & wood chip	11.10.11	0.015	<0.001	<0.001	0.005	0.002	0.092	0.014	<0.001	<0.001	0.014	0.023	0.041	<0.001	0.06
	Bagasse & wood chip	13.10.11	0.012	<0.001	<0.001	0.004	<0.001	0.158	0.009	<0.001	<0.001	0.027	0.021	0.05	<0.001	0.082
	Wood chip	15.10.11	0.033	0.001	<0.001	0.003	<0.001	0.052	0.001	<0.001	<0.001	0.065	0.03	0.077	<0.001	0.154
	Wood chip	16.10.11	0.037	0.003	<0.001	0.005	<0.001	0.065	0.006	<0.001	<0.001	0.079	0.041	0.1	<0.001	0.298
	Bagasse	20.10.11	0.024	<0.001	0.002	0.008	0.004	0.176	0.016	<0.001	<0.001	0.016	0.026	0.047	0.005	0.039
	Bagasse	21.10.11	0.026	<0.001	0.002	0.01	0.003	0.136	0.019	<0.001	<0.001	0.013	0.023	0.051	0.005	0.023
	Bagasse	27.10.11	0.027	0.014	0.021	0.023	0.021	0.118	0.025	0.015	0.056	0.025	0.044	0.037	0.01	0.057
	Bagasse & wood chip	28.10.11	0.047	0.013	0.02	0.023	0.019	0.072	0.019	0.015	0.046	0.03	0.044	0.034	0.007	0.047
	Bagasse	1.11.11	0.029	0.012	0.021	0.02	0.021	0.142	0.026	0.014	0.045	0.025	0.04	0.035	0.01	0.054
	Bagasse	3.11.11	0.032	0.011	0.021	0.019	0.019	0.166	0.031	0.015	0.04	0.025	0.037	0.04	0.008	0.213
	Bagasse & wood chip	4.11.11	0.038	0.04	0.02	0.029	0.018	0.091	0.025	0.015	0.047	0.036	0.045	0.044	0.006	0.168
	Bagasse & wood chip	5.11.11	0.044	0.013	0.02	0.022	0.02	0.173	0.021	0.015	0.038	0.036	0.049	0.042	0.006	0.108
	Wood chip (mixed)	6.11.11	0.105	0.013	0.019	0.028	0.024	0.085	0.023	0.014	0.076	0.100	0.103	0.099	0.004	0.074
	Bagasse & wood chip	7.11.11	0.107	0.015	0.019	0.021	0.019	0.05	0.018	0.015	0.054	0.058	0.054	0.081	0.004	0.113
	Bagasse & wood chip	8.11.11	0.042	0.011	0.02	0.021	0.021	0.124	0.022	0.014	0.036	0.027	0.042	0.046	0.009	0.128
	Bagasse	10.11.11	0.033	0.01	0.019	0.02	0.018	0.089	0.022	0.014	0.034	0.023	0.033	0.037	0.009	0.057
	Bagasse & wood chip	11.11.11	0.045	0.012	0.018	0.02	0.011	0.071	0.019	0.017	0.039	0.03	0.036	0.034	0.004	0.065
Broadwater	Bagasse & wood chip	15.09.11	0.028	<0.001	<0.001	0.009	0.001	0.09	0.018	<0.001	<0.001	0.015	0.032	0.046	<0.001	0.057
	Bagasse & wood chip	20.09.11	0.032	<0.001	<0.001	0.01	<0.001	0.100	0.02	<0.001	<0.001	0.021	0.031	0.047	0.003	0.059
	Bagasse & wood chip	26.09.11	0.03	<0.001	<0.001	0.009	0.002	0.084	0.012	<0.001	<0.001	0.014	0.035	0.051	<0.001	0.061
	Bagasse & wood chip	26.10.11	0.016	<0.001	0.001	0.005	0.003	0.059	0.009	0.002	<0.001	0.014	0.037	0.036	0.003	0.112
	Bagasse & wood chip	2.11.11	0.017	0.013	0.021	0.024	0.02	0.083	0.021	0.014	0.074	0.021	0.05	0.025	0.007	0.083
	Bagasse & wood chip	16.11.11	0.026	0.029	0.019	0.022	0.011	0.057	0.017	0.017	0.046	0.033	0.039	0.034	0.004	0.337
	Bagasse & wood chip	30.11.11	0.017	0.012	0.021	0.022	0.018	0.089	0.022	0.015	0.07	0.025	0.045	0.032	0.008	0.133
	Bagasse & wood chip	12.12.11	0.017	0.015	0.021	0.026	0.018	0.091	0.023	0.015	0.072	0.024	0.046	0.029	0.007	0.148

Table 4

X-ray fluorescence analysis showing average concentration (%) of major elements of loss on ignition (LOI) depending on the fuel used.

Fuel	Fe2O3	SiO2	Al2O3	MnO	TiO2	CaO	MgO	K2O	P2O5	Na2O	LOI
Bagasse	3.16	70.37	13.01	0.12	0.97	1.07	1.07	2.28	0.30	1.13	6.51
Bagasse & wood chip	3.18	63.05	10.56	0.15	1.33	2.16	1.18	2.00	0.35	0.59	15.46
Wood chip	1.29	27.98	3.86	0.36	0.52	10.21	3.45	3.69	0.45	0.35	47.86
Coal & oil	1.83	61.6	14.89	0.08	0.93	0.8	0.52	1.71	0.18	0.79	16.67
Coal, oil & bagasse	3.00	70.77	13.59	0.12	0.84	0.88	1.03	2.63	0.27	1.36	5.56

Trace metal concentrations are very low in all samples and independent of the fuel used (Table 3). However, loss on ignition (LOI) data show a strong fuel dependency, with wood chip fuel fly ashes contain >46 wt.% of un-burnt contaminants (carbon, water, and other volatiles; Table 4), whereas bagasse, and coal, oil & bagasse fuel ashes contain much lower contents of un-burnt contaminants (average 5 to 7 wt.%; Table 4). Similarly, bagasse and wood chip, and coal and oil fuel ash are similar to each other and contain ≈ 16 wt.% as un-burnt contaminants (Table 4). LOI concentrations have previously been shown to influence strength and sulphate resistance in concretes, where low LOI ash provides better strength and sulfate resistance (Chusilp et al., 2009a). Moreover, wood chip and bagasse-based ashes have significantly higher P2O5 contents (≈0.4 wt.%) compared to ash based around fossil fuels (≈0.2 wt.%; Table 4). This increased P2O5 contents for the bio-fuels would tend to suggest that returning ash to cane lands would be more beneficial than fossil fuel ash, because of a potential increased fertilizing capacity, however the high silica is of concern (le Blond et al., 2010).

Although it is somewhat difficult to compare ash between, and within mills due to differences in operation parameters of the boilers, some trends appear. Bagasse fuel ash samples contains more silica at Harwood mill (69–76 wt.%) than at Condong mill (66–68 wt.%), however, bagasse & wood chip fuel ash samples contained similar silica concentration at both Condong and Broadwater mills (Tables 2 & 4). Furthermore, the silica content of same fuel ash samples at a specific mill also varied over time (Table 2). Consequently, the variability between different mills for the silica content, and variability within an individual mill, is most likely from different varietals being harvested at that time. Cane varietals may vary in silica content (phytoliths) from 1.3–2.6 wt.% (Parr et al., 2009a). Consequently, variability between bagasse ashes, even within the same mill, are to be expected. Hence, because of the implied variability in the ash, reuse options must be sufficiently tolerant of the variability within a particular fuel, however sudden changes in fuel type (e.g., bagasse to wood chip for instance), may require some materials being excluded from a reuse option.

Bagasse fuel

Plant silica phytoliths (silicified plant cells) are most prolific in grasses (Parr et al., 2009a) and are particularly prolific in sugarcane varieties (Parr et al., 2009a). Observations of extracted phytoliths from the bagasse ashes indicate that there is some variation in the temperature that phytoliths are exposed to during cogeneration of electricity (Table 1). For example, phytolith morphology ranged from completely intact phytoliths ranging from 5 μm to around 250 μm in size with little sign of modification to some textural changes, some charring and some mild warping (Fig. 1A–D).

Fig. 1

Light microscopy of select fuel ashes from the 2 NSW sugar mills Harwood, and Congdon. A: Bagasse fuel, Harwood (04.10.11): phytolith showing some signs of charring and warping but still identifiable as grass (sugar cane) derived material. B: Bagasse fuel, Harwood (04.10.11): phytolith showing some signs of charring and warping but still identifiable as grass phytoliths. C: Bagasse fuel, Condong (20.10.11): showing mild charring and warping of the phytolith. D: Bagasse fuel, Condong (20.10.11): showing mild charring and warping of the phytolith. E & F: Wood chip fuel, Condong (15.10.11): phytoliths showing small silica particles <5 μm left and what appears to be a large charred sieve tube right.

Wood chip fuel

Unlike grasses, silica in trees, particularly the woody parts, do not generally form silica phytoliths (cell wall silicification) but rather have silica deposits between the cells and in the parts of the trunk associated with water transport. Nevertheless some tree species are prolific producers of phytoliths (cf., (Parr et al., 2009b). Preliminary observations made on the phytoliths extracted from the wood ash samples indicate that silica is present mainly as small particles <5 μm as well as some very large silica plant sections associated with the water or resin transport mechanism of the plant up to 300 μm in size (Fig. 1E & F). There was visible charring of the larger pieces of silica however the smaller silica pieces appear to be for the most part not charred or warped.

Charring of the samples is consistent with the operating temperatures of the boilers (Table 1), where temperatures of 500–800 °C are recorded. Above 500 °C, amorphous silica begins to transform into cristobalite a more crystalline form, and melting of phytoliths occurs at temperatures above 1500 °C. Consequently, although crystallinity may have increased (Table 1), this may not unduly affect zeolitisation, but may impact geopolymer formation.

Bagasse ash samples from the three mills Broadwater, Harwood, and Congdong all contain grass phytoliths, typical for sugar cane. Dimensions range from small (20 × 30 μm) to large (60 × 190 μm) (Fig. 2, Fig. 3). Inner structures and articulations are well preserved, perimeter articulations show partial melting (Fig. 2A); warping appears to be minimal. The ash presents in all stages of globulisation, and minimal amounts of un-burnt organic material are identifiable. Grass phytoliths of which most cereal crops, sugar cane, and bamboo are members of, are typified (Figs. 2B–G and 3B). In addition, many of the phytoliths show partial melting of the articulations on the perimeter, but intact inner articulations structure (Figs. 2E–G and 3A & B). Moreover, show a fragmented character, indicating disintegration or rupturing, which occurs in many samples (Figs. 2E–G and 3C–E). Also in samples displaying disintegration or rupturing (Fig. 2E–G), moderate to strong warping often occurs (Figs. 2F and 3D&F). Furthermore, globulisation (Figs. 2H and 3D) and partial melting of the phytoliths (Figs. 2A and 3A) appears characteristic. These developments may indicate an accumulating effect of high temperature conditions in the boiler on phytolith morphology.

Fig. 2

SEM of select bagasse fuel ashes from the sugar mills in this study. A: cylindrical phytolith without apparent cellular structure: B: grass phytolith with well-preserved inner structure C: small elongated warped and articulated phytolith D: grass phytolith showing partial melting on the perimeter and a fragmented appearance E: strongly warped and partial melting on articulations of phytolith. F: strongly warped, fragmented phytoliths with partial melting G: partly melting and cracked phytolith with preserved inner articulations: H: globulised phytolith showing complete melt.

Fig. 3

SEM of select bagasse fuel ashes from the sugar mills in this study. A: smooth phytolith; surface appears to be affected by partly melting. B: grass phytoliths showing partial melting (bilobate features); rounded outer articulations and preserved inner articulations. Phytolith also shows advanced warping in long cell (curvature in upper right half of main phytolith). C: large number of different phytoliths showing partial melting, with round articulations and globulisation D: phytolith with intact inner articulations and rounding of perimeter articulations and globulisation, E: long phytoliths with smooth and warped surfaces indicating partial melting, some globulisation is occuring.

Wood chip ash samples (Camphor Laurel; Cinnamomum camphora) under the SEM exhibits strong charging, suggesting larger amounts of un-burnt organic material (Fig. 4). The wood-like structures (biochar), are readily identifiable (Fig. 4A & C) which shows the elongate cell structure and checkerboard appearance (Fig. 4C). Moreover, there is a distinct lack of any identifiable grasses or other phytoliths (Fig. 2, Fig. 3). However, some Si-based minerals of unknown origin are present (Fig. 4D) together with Ca-based minerals (Fig. 4B & D), these may well be soil particles, but may also be of plant origin. Apart from globulisation and partial melting of the ash, an increased formation of Ca-minerals can be observed (Fig. 4B), which together with globulised material (Fig. 4E) are the only sources of Si, however the source of Si-based materials (Fig. 4B & D) are not determinable as being from soil particles, or phytolith material.

Fig. 4

Camphor Wood chip fuels, A: wood-like structures (biochar), B: Ca-minerals and one possible phytolith or Si-based mineral of other origin, C: large amounts of organic biochar material D: a Si-based globulised mineral of undetermined origin, E: large fly ash particles showing slight rounding from moderate melting and strong warping, globulised material upper left.

Bagasse & wood chip fuel

Bagasse and wood chip ash samples appear to contain considerable quantities of large pieces of charcoal (Fig. 5). More importantly, phytoliths are warped (Fig. 5A, B & D) and the articulations are showing a rounding in the perimeter and on the surface from partial melting (Fig. 5B & H). Some sample appears to contain greater numbers of phytoliths than others (Fig. 5C, E & F), some of which are partially melted and warped (Fig. 5B), whereas others are well preserved with articulations (Fig. 5G & H) and surface structures (Fig. 5H) intact; the fly ash particles are of variable sizes and show partial and full globulisation (Fig. 5E & G). The vast majority of the phytoliths show the grass origin of high articulation (Fig. 5B, G & H), but are show in all stages of fragmentation and melting, with articulations well preserved as well as eroded. The high degree of warping in the samples would indicate exposure to temperature, whereas the partial melting suggests much higher temperatures. Partial globulisation of material also provides evidence of high temperatures, whereas some preserved organic material (Fig. 5C) appears more sponge-like rather than wood-like (Fig. 4A & C).

Fig. 5

Bagasse & Wood chip fuel, A: warping in elongated phytoliths, B: Bagasse & Wood chip fuel, Condong (08.11.11): warped phytolith (bilobate); as if in the process of globulisation C: long phytolith with surface structure preserved, D: warping in grass phytolith, E: variety of particles, incl. long phytoliths and globulised ash, F: long phytolith, slightly warped edges G: grass phytolith with well-preserved inner articulations but rounded perimeter H: slightly warped phytolith with rounded articulations from melting.

Coal & oil fuel

The sample coal & oil fuel ash shows significant charging (white glow) under the electron beam, but contains very few identifiable phytoliths; those that present are of elongated smooth appearance of 30–120 μm length and c.a. 10-μm width, with no apparent cellular structures. The phytoliths that are observable typically show some warping (Fig. 6A) or appear as irregular structures with partial melting and/or fully globulised (Fig. 6A).

Fig. 6

Coal, and Oil fuel, and Coal, Oil & Bagasse fuels A: elongated phytolith with warping and beginning globulisation of the fly, B: phytoliths of morphology consistent with grass phytoliths in the mid-bottom section and right side mid-section of the image C: fully globulised particles D: phytolith with partially melted articulations.

Coal, oil & bagasse fuel

Coal, oil & bagasse fuel ash also shows substantial charging (white glow) under the electron beam (Fig. 6A & B), which suggests the presence of un-burnt organic material reacting with the beam. However, irregular structures that cannot be characterised by EDAX as unidentifiable tend to confirm the presence of organic material; carbon is atomically too lightweight to be analysed by EDAX producing x-ray with a too low energy. Phytoliths appear in three different morphologies: triangular cellular (seed pot-like, Fig. 6A), elongated and flat cellular. Many phytoliths show warping and flat cellular phytoliths (Fig. 6B) with a morphology consistent of grass phytoliths (Fig. 2, Fig. 3), which suggests they originate from the bagasse being added to the coal and oil fuel. The diversity of mill fuels (i.e. coal, oil, and bagasse) is reflected in Fig. 6C and D. Moreover the sample coal, oil, and bagasse fuel ash shows many different types of particles with more fully globulised particles (Fig. 6C), which are larger than in the bagasse-only samples (Figs. 2 & 3). In addition, some phytoliths show significant rounding (partial melting) of the articulation (Fig. 6D).

Observations of XRD data (Fig. 7) across all the different fuels show that α-, or low-quartz is dominant. The α-, or low-quartz peaks are sharp, suggesting that quartz silica is mostly crystalline, although some samples contain a broad arch in the data from ≈5.5 to ≈3 Å, this is not pronounced (Fig. 7). This weak arch in the XRD data indicates that the proportion of glassy and semi-amorphous mineral content is relatively small in most XRD analyses and the low background compared to the peak height ratio. However, wood chip fuels (Camphor Laurel; Fig. 7C) shows that the peak counts are relatively low ≈35 counts per second (Fig. 7C) compared to other fuel samples where up to ≈100 counts per second for bagasse fuel (Fig. 7A), bagasse and wood chip (Fig. 7B), ≈180 counts per second mixed wood chip (Fig. 7D), and ≈200 counts per second coal, oil, and bagasse (Fig. 7E). Given the quantities of un-burnt material seen in wood chip fuel ash (Fig. 4A & C), it is most likely that the glassy and semi-amorphous mineral content is un-burnt carbon rather than glassy and semi-amorphous silica.

Fig. 7

X-ray diffraction analysis for different fuel ashes, A: Bagasse ash, B: Bagasse & Wood chip fuel ash, C: Wood chip (Camphor Laurel) fuel ash, D: Wood chip & mixed fuel ash, E: Coal, Oil & Bagasse fuel ash. F: Wood chip fuel (camphor Laurle) ash with the amorphous arch removed. Peaks are: synthetic quartz (ICCD 00-046-1045); d-100 peak for cristobalite ((ICCD 00-046-1045); calcite (ICCD 01-071-3699); kutnohorite (ICCD 00-011-0345); and calcium aluminium sodium silicate (feldspar/feldspathoid ICCD 01-089-1466).

Moreover, wood chip fuel samples show in addition to the silica peaks, identifiable mineral peaks for calcite and kutnohorite minerals (Fig. 7) as well as other soil minerals such as anorthite and albite (calcium aluminium sodium silicate; feldspar/feldspathoid ICCD 01-089-1466), which suggests a transfer of some soil material during cane harvesting. The calcite and kutnohorite are consistent with Ca being in contact with a reducing (high CO) burn (Table 1). Only, low intensity peaks are identified as α-, or low-cristobalite (d = 4.04, the d100 peak for cristobalite; ICCD 00-046-1045). Similar peaks were found in all other fuel types, but not all samples (data not shown), but were found particularly in samples containing bagasse as their fuel (e.g., Fig. 7A–E). The lack of amorphous silica, and or α-, or low-cristobalite, is at odds with the reported abundance of these components in phytoliths (Barbhuiya et al., 2006; Heaney, 1994; Hodson et al., 2005) and suggests that combustion processes are playing a role in controlling mineralogy; these data are also at odds with other bagasse ashes (Barbhuiya et al., 2006; Chusilp et al., 2009b).

Discussion

The different analysis performed on the 41 ash samples revealed high silica content forming phytoliths in the bagasse and bagasse & wood chip fuel ashes, but not in the wood chip fuel ashes (Tables 1 & 2). In addition, the wood chip addition to bagasse (e.g. Broadwater) does not strongly influence the silica content. However, results from the XRD indicate silica is mostly crystalline α-quartz, rather than the expected α-cristobalite, most typical in phytoliths (Barbhuiya et al., 2006; Heaney, 1994; Hodson et al., 2005), whereas others indicate more amorphous and glassy silica forms.

Silica transformations

Data from the XRD analyses, the operating temperatures of each mill and the SEM and light microscopy of the sample ashes, indicate that the silica in the ash is being transformed to different polymorphs and melted. At low pressures there are 3 groups of silica polymorphs (Table 5), each with 2 closely related members; a low-temperature member given a prefix α-, and a high-temperature member with a β-prefix; although some prefer using low- or high-prefix (Heaney, 1994; Hemley et al., 1994). However, of all silica polymorphs, only α-quartz is stable at normal ambient conditions; all other silica polymorphs will transform into α-quartz given sufficient time. Although other silica polymorphs (Table 5) are meta-stable under different conditions, these conditions are mostly at high temperature and/or pressures, although some may occur at low temperatures and pressures under conditions where quartz is stable (Table 5) (Heaney, 1994; Hemley et al., 1994). Generally, at normal atmospheric pressures α-quartz, will at 573 °C, transform into β-quartz, and upon further heating to 870 °C into β-tridymite, and to β-cristobalite at 1470 °C; at 1705 °C β-cristobalite finally melts (Heaney, 1994; Hemley et al., 1994).

Table 5

Low Pressure polymorphs of silica (After http://www.quartzpage.de/atf.html).

Quartz Polymorph	Stable at	Meta-stable at
high- or β-polymorph
β-Quartz	573–870 °C	-
β-Tridymite	870–1470 °C	117–870 °C
β-Cristobalite	117–870 °C	270–1470 °C
low- or α-polymorph
α-Quartz	< 573 °C	-
α-Tridymite	-	< 117 °C
α-Cristobalite	-	< 270 °C

However, β-quartz does not usually transform to β-tridymite unless it is somewhat impure (Heaney, 1994), and often the β-quartz-tridymite transition is skipped and the sequence looks like this:

In addition, pre-existing α-tridymite and α-cristobalite within materials, will convert to their β-forms at 114 °C and 270 °C respectively, whereas pre-existing amorphous silica will convert to β-cristobalite at approximately 1000 °C (Heaney, 1994; Hemley et al., 1994). In RHA production Soltani et al. (2015) note that a transition between amorphous silica can occur between 700 and 900 °C. Moreover, as-long-as temperature changes are relatively slow, the polymorph transitions (Eqs. (1) & (2)) are fully reversible except for a return back to amorphous silica. For example, amorphous silica once taken to β-cristobalite will cascade to β-quartz and then to α-quartz (Eq. (2)).

A much greater polymorph complexity however, is generated when the temperature is increased or decreased rapidly (Heaney, 1994; Hemley et al., 1994). For example very rapid heating of α-quartz shows that it will still undergo a phase transition to β-quartz, but the β-quartz will then “skip” the β-cristobalite transition directly melting at a much lower temperature of 1550 °C (Heaney, 1994; Hemley et al., 1994). Hence, amorphous silica can be rapidly burnt, and become a silica melt and globulizes at ≈1550 °C without transitioning any other phases (Heaney, 1994; Hemley et al., 1994). Consequently, the data seen here suggests, that burn temperatures are below 1550 °C because of limited globalisation (Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6), and a dominance of α-quartz in the XRD patterns (Fig. 7).

Similarly, if a silica melt is cooled quickly, it will avoid mineral phases and its disordered liquid structure will be preserved as a super cooled liquid, or an amorphous silica glass; lechatelierite when found in nature (Heaney, 1994; Hemley et al., 1994). However, there is no well-defined melting point for silica glass as it slowly turns into a very viscous liquid upon heating. Furthermore, if the β-tridymite and β-cristobalite polymorphs are rapidly cooled below their respective transition temperatures, their crystal structure is first preserved until they transform into polymorphs with closely related structures, which is fully reversible even with relatively rapid temperature changes (Heaney, 1994; Hemley et al., 1994).

Congruently, the XRD data shows (Fig. 7C & F) that both calcite and kutnohorite are present in some ashes, particularly those from wood fuels. Dolomite-like minerals (kutnohorite) and calcite are readily decomposed at temperatures around 750 °C to their respective calcia (CaO) and periclase (MgO) end members (He et al., 2013). Consequently, the presence of kutnohorite and/or calcite would not be expected at burn temperatures of ≈1500 °C. However, Blamey et al. (2011a); Blamey et al. (2011b) note that during a cooling process CaO can revert back to calcite, in the presence of water vapour in a high CO2 vapour at temperatures 4–600 °C. Hence, the presence of calcite and/or kutnohorite in the high carbon (LOI) wood fuel ashes is consistent with a relatively prolonged cooling period, and the dominance of the ash by α-quartz in the XRD patterns through heating to ≈1500 °C (Fig. 7)

Physical changes to phytoliths

Temperature has been shown to promote both chemical and physical changes to the structure of silica. In the case of silica phytoliths produced by plants some of these changes are reported to occur under both oxidizing and non-oxidizing conditions at a range of temperatures. For example, oxidizing conditions or direct contact with fire is widely assumed to induce darkening (charring) of phytoliths (Parr et al., 2009b), while under non-oxidizing conditions silica phytoliths can remain relatively clear.

The physical changes at temperature range from the transformation of silica phytoliths from clear to opaque with brown to black coloration and changes of refractive index (Jones and Milne, 1963). High temperatures are known to promote warping of phytoliths (Runge, 1998). The warping, articulation rounding, and partial globulisation to the silica phytoliths at temperature range from the minor warping at temperatures above >500 °C to beading and the formation of globular structures at temperatures >1200 °C (Runge, 1998). Interestingly the current burn temperatures provided by the sugar millers are ∼500 °C, however, we find definite warping, beading, and, in some cases, the formation of globular silica structures indicating that at least localised burn temperatures are far in excess of the temperatures being provided by the mills.

Importantly, there appears to be considerable differences in the results from individual boilers and batches. For example, materials from Harwood mill exhibits advanced warping in long cells and the inception of melting of the bilobate phytoliths, which based on previous research indicates temperatures of >1200 °C (Fig. 2, Fig. 3, Fig. 4). However, most bi-lobate phytoliths from the Condong Mill (data not shown) have no warping or melting, whereas similar bilobate phytoliths from Broadwater Mill show signs of warping but no melting indicating lower temperature exposure than Harwood Mill (Fig. 2, Fig. 3, Fig. 4). Although some phytoliths from Broadwater Mill do show advanced warping of long cells and the formation of globular silica (Fig. 2, Fig. 3, Fig. 4) suggesting higher temperatures.

More importantly, nearly all the visual inspections show globulisation of some materials (Fig. 2, Fig. 3, Fig. 4) suggesting that localized temperatures in the boilers are greatly in excess of those reported in Table 1. It would seem that locally within the boilers temperatures exceed 1550 °C, the direct melt temperature for impure β-quartz, which has had rapid silica heating (Heaney, 1994; Hemley et al., 1994). Evidence for this partial melting can also be seen in XRD patterns where a broad arch exists in many patterns between 6 and 3.5 Å (e.g., Fig. 7). Arching of the XRD pattern like this is a direct response to the development of glassy phase (semi-amorphous and amorphous materials) typically from the rapid cooling of molten materials, or from the precipitation of very fine crystal structures such as un-burnt carbon. Because of the severe discrepancies between the boiler operation data (Table 1) and the observable temperature effects on phytoliths it is evident that the boiler monitoring point is significantly distal to the seat of combustion.

Silica phase transitions observed (Eqs. (1) & (2)) also have substantial impacts on silica solubility (Table 6), which shows that solubility decreases by about an order of magnitude when silica moves from amorphous forms to more crystalline forms (Holleman and Wyberg, 1985; Rykart, 1995). This discrepancy between the solubility of crystalline and amorphous silica is also true across the pH range as well (Holleman and Wyberg, 1985; Rykart, 1995). Hence, as demonstrated in the XRD data, the dominance of α-quartz (Fig. 7) rather than α-cristobalite and amorphous silica, suggests that ash processing into new products might be hampered.

Table 6

Solubility of low pressure polymorphs of silica in water at 25 °C (After http://www.quartzpage.de/atf.html).

Substance	Silica solubility in Water at 25 °C
Macro-crystalline α −Quartz	2.9 mg/L (Holleman and Wyberg, 1985)/6–11 mg/L (Rykart, 1995)
Chalcedony	22–34 mg/L (Rykart, 1995)
α −Cristobalite	6 mg/L (Holleman and Wyberg, 1985)
α −Tridymite	4.5 mg/L (Holleman and Wyberg, 1985)
Stishovite	11 mg/L (Holleman and Wyberg, 1985)
Quartz Glass (amorphous)	39 mg/L (Holleman and Wyberg, 1985)/120 mg/L (Rykart, 1995)

The pozzolanic reaction for cements and concretes (Cordeiro et al., 2009a; Cordeiro et al., 2008) requires silica solubility to provide the additional strengths through the formation of more complex silicates in the paste rather than aluminates (e.g. di-calcium silicate, rather than tri-calcium aluminate). Such changes in cement chemistry components can also have significant impacts on the physical and chemical resistance of the cement paste (Alavéz-Ramírez et al., 2012). Indeed, Arif et al. (2016) note that the pozzolanic activity of SCBA from these high efficiency co-generation boilers is significantly reduced, to the point where they effectively dilute the cement components. Moreover, the presence of high LOI materials, from wood chip fuels also have a detrimental aspect in the sulfate resistance in developed concretes, and developed strength are severely reduced (Chusilp et al., 2009a). Hence winnowing of the high LOI materials would also be required (e.g., flotation).

The ability to solubilise silica from the bagasse ashes investigated also has ramifications on the capacity to produce zeolites (Alavéz-Ramírez et al., 2012), and the types of zeolitic materials produced. Moreover, the low alumina content of the ash suggests that for zeolite production supplementation of the ash with alumina will be required to bring the Al:Si ratio close to 1; a Al:Si ratio of ≈1 allows the production of the 1:1 zeolites of which zeolite-X, −A, and −Y are industrially particularly useful. Several process for increasing silica solubility have been proposed including alkaline fusion with NaOH (Ríos et al., 2012), and ultrafine grinding (Cordeiro et al., 2009b) to provide increased surface area, and additional broken bonds for solubilisation to occur on.

The further processing of a waste through grinding and/or flotation circuits and/or fusion activation invariably adds cost to the resource utilisation. Hence, although such processes Ríos et al. (2009) provides greater yields or performance in the desired applications (zeolitisation, pozzolanic reactivity), it may be more cost effective to accept lower reactivity, lower conversion percentages, and lower potential performance in re-use products (e.g., use as a filler material rather than a pozzolanic replacement material in concrete) (Arif et al., 2016). The readjustment of the bagasse ash from potential pozzolan to filler material means that an on-sale value for the material will require an adjustment down towards the value of sand, unless there are other unrealised uses that can reverse this price adjustment.

Conclusions

Ash from the three NSW sugar mills have been collected and geochemically characterised to determine how the seasonal changes in fuel source affects the suitability of the ash for either zeolitisation, geopolymer manufacture, use as a pozzolanic and/or as a filler material in cements and concretes. Data collected reveals that when wood chip is used as an alternative, or in addition, to bagasse then the silica contents falls to unacceptably low concentrations, and the LOI from un-burnt carbon substantially increases (Table 2). This has important ramifications for using the bagasse ash left over from the burning process when supplementary wood chips are introduced as the ash will not be suitable for geopolymer or zeolites production nor will it be suitable for inclusion in cements or mortars as a filler material without further treatment.

Observational data from the phytoliths present in many of the ashes suggest that burn temperatures are much higher than those derived from the mill-supplied boiler data. Many phytoliths show significant warping of the structures, partial melting and/or complete globulisation of silica materials. Where globulisation is occurring would suggest that burn temperatures are in excess of 1550 °C as dictated by the β-quartz skipping the β-cristobalite phase transition and moving directly to melt. However, the absence of lechatelierite indicates that burn temperatures remain below a maximum of ≈1700 °C. The dominance of α-quartz rather than the expected α-cristobalite in the XRD analysis is testament to the accepted phase transitions associated with heating quartz. Unfortunately these phase transitions observed in the XRD patterns, will affect silica reactivity and subsequent potential of the bagasse ash for zeolitisation, geopolymer formation, as a pozzolanic in concretes or mortars (Arif et al., 2016). Based on the findings the bagasse ashes investigated in this study have more potential use as a fill material in concretes and mortars and are unlikely to command a substantial on-sale value unless the burn temperatures can be controlled more carefully.

An important finding from this work is that the burn temperature and subsequent phase transitions of the silica materials have a substantial impact on the silica reactivity and potential use of the bagasse ash for geopolymers or zeolites or concrete and mortar fillers (Arif et al., 2016). Many past studies (e.g., (Barbhuiya et al., 2006; Heaney, 1994; Hodson et al., 2005), on bagasse ash have been undertaken using carefully controlled temperature burning of the bagasse (e.g., ashing at 550 °C). The reality of as derived from a commercial mill can be very different with material analysis revealing different burn temperature to those measured during the burning process. Hence, for bagasse ash to be large-scale commercially useful, consistency and controllability of the burn temperatures is critical to ensure that cristobalite and amorphous silica are preserved rather than transformation unreactive α-quartz. Importantly, if boiler-monitoring points are to be used for control the burn process and ash by-product quality, it is important that the monitoring points represent the temperature at the combustion seat.

Declarations

Author contribution statement

Malcolm Clark: Conceived and designed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Laure Despland: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Elisabeth Arif, Manuela Anstoetz, Philip Doumit: Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Neal Lake: Analyzed and interpreted the data; Wrote the paper.

Lachlan Yee: Conceived and designed the experiments; Analyzed and interpreted the data; Wrote the paper.

Jeffery Parr: Conceived and designed the experiments; Wrote the paper.

Competing interest statement

The authors declare no conflict of interest.

Funding statement

This work was supported by Sugar Research Australia (SRA; formally SRDC, Sugar Research Development Corporation; grant SRDC SCU03). Financial support was also provided by Australian Biorefining Pty Ltd., (Adam Blunn) as an industry partner to the SRDC SCU03 grant.

Additional information

No additional information is available for this paper