1 Characterization of Ash in Algae and Other Materials by Determination of Wet Acid

2 Indigestible Ash and Microscopic Examination

3
4 Keshun Liu
5
6
7 Grain Chemistry and Utilization Laboratory
8 National Small Grains and Potato Germplasm Research Unit
9 United States Department of Agriculture, Agricultural Research Service (USDA-ARS)
10 1691 S. 2700 West
11 Aberdeen, ID 83210, USA
12 Tel.: (208) 397-4162
13 Email: Keshun.Liu@ars.usda.gov
14
15
16

© 2016. This manuscript version is made available under the Elsevier user license
http://www.elsevier.com/open-access/userlicense/1.0/
17 Abstract
18 Algae are known for high ash content. It is important to properly characterize ash for value
19 added utilization of algae as food, feed, and feedstock for biofuels. In this study, 12 algae of
20 different sources were measured for proximate composition and mineral profile. Results showed
21 that the relative difference between ash by dry ashing and total minerals by wet digestion
22 increased with ash content. A major cause was soon identified: when using a common procedure
23 of strong attacks for sample digestion before mineral analysis, incomplete digestion existed for
24 most algae samples due to the presence of siliceous materials. It was proposed that algae consist
25 of wet acid indigestible ash (WAIA) and wet acid digestible ash, whereas WAIA is siliceous.
26 Methods to measure WAIA content in the 12 algae, along with oat grain, oat forage, defatted
27 soymeal and fine sand, were then developed based on digestion with nitric acid or sulfuric acid-
28 hydrogen peroxide. For the 12 algae, ash ranged 1.9 to 37.4% dry matter while WAIA by nitric
29 acid digestion varied 0.1% to 25.6%. High correlation between WAIA and ash contents indicates
30 WAIA as an important contributor for algae ash. For identifying what constituted the siliceous
31 materials, all samples in three matrixes (original, ash by dry ashing, and WAIA) were
32 microscopically examined. Because acid digestion had ability to concentrate siliceous materials
33 and maintain their original shape and size, WAIA was the best matrix for microscopic
34 examination. Micrographs of WAIA show three types of siliceous materials in algae: non-
35 diatom cellular structures, diatom cell walls, and sandy particles. It was concluded that high ash
36 content of algae resulted partly from contamination of diatoms and/or sandy particles of geologic
37 origin and that WAIA should be an important quality parameter for algae. Subsequently, several
38 measures are proposed to produce algae with low ash content.
39
40 Key words: Algae ash, wet acid indigestible ash, minerals, dry ashing, wet ashing, acid
41 digestion, contamination

42

43
44
45 Abbreviations
46

2
47 AIA: acid insoluble ash
48 ANOVA: Analysis of variance
49 AOCS: American Oil Chemists‟ Society
50 CHO: carbohydrate
51 DM: dry matter
52 HF: hydrofluoric acid
53 HR: hour
54 ICP-OES: inductively coupled plasma-optical emission spectrometer
55 PBR: photobioreactor
56 PTFE: polytetra-fluoro-ethylene
57 TM: total minerals
58 WADA: wet acid digestible ash
59 WAIA: wet acid indigestible ash
60 WAIAn: wet acid indigestible ash by nitric acid digestion
61 WAIAs: wet acid indigestible ash by sulfuric acid digestion
62

3
63 1. Introduction
64 Algae have been used as animal feed (including aquafeed), human food, fertilizer, and a source
65 for extracting high value molecules [1-3]. In recent years, there has been increased interest in
66 using microalgae as an alternative resource for biofuel production [4]. Key advantages of
67 microalgae include prolific growth rates, the ability to grow on lands that are marginal for other
68 agricultural purposes, and the ability to clean up water resources with excess nutrients [4].
69
70 When evaluating algae for end uses, it is important to assess chemical properties of the algal
71 biomass [5-7]. In addition to proteins, lipids, and carbohydrates, ash content is a very important
72 quality parameter since ash contains inorganic material and is noncombustible. In general, algal
73 biomass is known for high ash content [5]. For example, algae grown in certain locations
74 contained ash as high as 70% dry matter (dm) [8]. This phenomenon not only reduces the
75 amount of ash-free matter, the valuable portion of algae biomass, but also causes a concern for
76 heavy metal toxicity and diminishes inclusion levels for algae uses as feed, food and fertilizer
77 [9]. The inorganic matter from algae also poses major operational problems in biomass
78 combustion systems for energy conversion. These include slagging, fouling, high temperature
79 corrosion, and bed agglomeration. All of these can disrupt normal combustion and fluidization
80 behaviors and, if severe enough, can lead to expensive system shut-downs [10]. For reducing
81 ash in algae, physical pretreatments, such as particle size reduction, sieving and centrifugation
82 were tried [11] but the effect was questionable due to time, water and energy consumed.
83
84 To predict and mitigate ash-related problems effectively, it is crucial to develop a reliable and
85 easily adaptable methodology to properly characterize the ash forming components in algae
86 biomass. The most common method to characterize algae ash has been instrumental analysis for
87 mineral composition [5, 12-15]. Realizing the limitation of mineral analysis alone, Lane et al.
88 [16] characterized algae ash by first subjecting algae samples to sequential chemical extractions
89 with selected solvents, analyzing mineral composition of each fraction, and inferring the mode of
90 occurrence of the main inorganic elements from their distribution patterns among the fractions.
91 This method of combining wet fractionation with mineral profiling had been used in coal and
92 other biofuel materials [10, 17, 18]. Hampel [8] attempted to characterize algal biomass ash by

4
 93 measuring its two main components: silica and heavy metals but found that the method to
94 measure silica was time consuming, expensive, difficult, and prone to errors due to interferences.
95
96 Although studies on algae biomass involving measurement of ash content and/or analysis of
97 mineral composition of whole algae or its fractions can provide information on occurrence of
98 inorganic elements in algae and how this varied among different species of algae and different
99 methods of production and harvesting, none of them can adequately address several important
100 questions regarding the algae ash component. These include: 1) why algae have higher ash
101 contents, some extremely so, compared to other biological materials, 2) why algae have a large
102 variation in ash content, 3) do ash components in algae differ from other biological materials, 4)
103 what is the chemical nature of algae ash, and 5) how to prevent ash built-up during algae
104 production?
105
106 This study was designed to specifically address these questions. It began with collection of 12
107 algae samples from different sources, followed by proximate analysis and mineral composition
108 profiling. It was soon found that mineral analysis based on the nitric acid digestion could not
109 adequately characterize the ash components in algae samples. Careful examination of the method
110 for mineral analysis and review of algae chemistry led to a key reason for inadequacy of the
111 mineral analysis: unlike most other biological materials, algae (particularly diatoms) contain
112 substantial amounts of silica or silicates. Because silicates are non-combustible, they remain as
113 ash upon combustion. Silicate is also extremely resistant to acid digestion. Therefore, even with
114 the use of a very common procedure of strong attacks for sample digestion [19, 20], a substantial
115 portion of ash in algae samples could not be digested before instrumental analysis of minerals.
116
117 Although the problem of incomplete digestion for mineral analysis could be addressed by
118 replacing the strong attack procedure with a total decomposition procedure that involves an
119 additional step of adding hydrofluoric acid (HF) to dissolve silica from algae samples, it is hard
120 to practice in a typical analytical lab due to inherent drawbacks of using HF as a reagent [19,21].
121 In order to develop a reliable and easy-to-adapt methodology to quantitatively characterize the
122 ash component of algae, an alternative yet efficient approach was proposed in this study:
123 Following a strong attack procedure, the insoluble siliceous particulates in the digestion tubes,

5
124 instead of being discarded, were measured gravimetrically upon filtration or centrifugation at the
125 end of the wet acid digestion. By doing so, the new concept that algae ash consists of wet acid
126 indigestible ash (WAIA) and wet acid digestible ash (WADA) was introduced. The objectives of
127 the study were subsequently expanded to include: 1) developing methods to measure the content
128 of WAIA, 2) determining the relationships among total ash, total minerals (TM), WAIA and
129 WADA contents, and 3) identifying chemical nature of the ash component by microscopic
130 examination of all samples in three forms: original matrix (as is), ash by dry ashing, and WAIA.

131 2. Materials and Methods

132 2.1. Materials

133 Twelve algae samples were collected, in a duplicate set, from different sources and producers
134 across the USA. They were named according to the order of receiving and thus the names of
135 producers/providers were omitted to avoid any possible negative impact on their products based
136 on the results of this study. Among the 12 algae, only Algae 4 was grown in a photobioreactor
137 (PBR) system, the rest were all grown in open pond systems. For a few algae samples,
138 information on strains was known, but for many others such information was lacking.
139 Information on production details, harvesting methods, and postharvest handling/processing was
140 also lacking. However, this lack of information did not compromise achieving objectives within
141 this study. Upon receiving, all samples had already been dried by the providers, except for Algae
142 11, which was collected from a nearby waste water treatment facility, centrifuged at 3000 x g for
143 10 min to remove extra water, and dried in an forced air oven at 60ºC overnight.
144
145 Non-algae materials included hulled oat grain (cultivar „Ajay‟), oat forage (cultivar „CDC
146 Dancer‟), defatted soy meal, and sand. Oat grain and forage were grown at the University of
147 Idaho Experimental Station farm in Aberdeen, Idaho. Forage was harvested at a pre-flowering
148 stage and dried in the forced air oven at 60ºC overnight. Defatted soy meal was obtained from
149 the Archer Daniel Midland Co. (DeKalb, IL). Sand was collected from the shore of American
150 Falls Reservoir, Idaho.
151
152 2.2. Sand preparation

6
153 The sand sample was dried in a forced air oven at 60ºC overnight, and then 300 g was sieved
154 with a 200-mesh U.S. standard sieve and a pan, fitted into a sieve shaker (DuraTap, Model
155 DT168, Advantech Mfg. Co., New Berlin, WS), with shaking and tapping for 15 min. The sand
156 of larger particle size on the top of the sieve was discarded. The fine sand from the pan (about
157 40g) was washed with water several times before drying in a forced air oven at 140ºC for 1 hr. It
158 was then stored in a sealed plastic bag.
159
160 2.3. Analysis for proximate composition
161 Dried algae samples were analyzed for moisture, protein, oil, and ash contents. The four non-
162 algae materials were measured only for moisture and ash content. Moisture content was
163 measured by drying in an forced air oven at 105ºC for 3 hr [22], while the oil content was
164 analyzed by an AOCS Official Procedure [23], using a fat analyzer (Model XT 10, Ankom
165 Technology, Macedon, NY, USA) with hexane as the extracting solvent. The total nitrogen
166 content was measured by a combustion method [22], using a protein analyzer (Model FT528,
167 Leco Corp. St. Joseph, MI, USA). In calculating protein content, a conversion factor of 4.78 for
168 nitrogen-to- protein conversion was used [24]. Ash content was measured, using a furnace
169 heated to 600ºC overnight (about 16 hr). Carbohydrate (CHO) was calculated as the difference
170 between 100 and the sum of protein, oil, and ash contents in % dry matter basis. Here, the CHO
171 is defined as a heterogeneous mixture of all those components that were not measurable by the
172 methods for moisture, ash, protein and oil. They include cellulose, lignin and hemicelluloses, as
173 well as starch, pectin, oligosaccharides, sugars, organic acids and pigments.
174
175 2.4. Acid digestion and mineral analysis
176 The algae samples were also measured for contents of individual mineral elements, using a
177 Perkin-Elmer Optima 3200 ICP-OES (inductively coupled plasma-optical emission
178 spectrometer) to quantify constituents in an aqueous solution following nitric acid digestion of
179 the samples. For the nitric acid digestion, the procedure of Anderson et al. [20] was used with
180 modification. Briefly, each algal sample was weighed (125-200 mg) and put into a heavy duty
181 15 ml glass centrifuge tube. Under a fume hood, 3 ml of 30% nitric acid was added into each
182 tube. Tubes were vortexed and left in the fume hood overnight at a room temperature. The next
183 morning, tubes were placed into a block heater and maintained at 70ºC for 1 hr before slowly

7
184 increasing the temperature to 120ºC. Tubes were maintained at 120ºC for 6 hr or until the
185 solution in each tube turned totally clear. Upon cooling, 4 ml of deionized water was added, and
186 each tube was vortexed. After centrifugation at 4000 x g for 10 minutes, the supernatant was
187 saved for instrumental analysis.
188
189 2.5. Measurement of WAIA content by nitric acid digestion
190 WAIA contents in all algae and non-algae samples were measured by a procedure based on nitric
191 acid digestion. The procedure for WAIA measurement was identical to one for nitric acid
192 digestion for mineral analysis described in the previous paragraph except for the following two
193 additional steps. (1) At the beginning of the procedure, each heavy duty 15 mL centrifuge tube
194 was dried in a forced air oven at 140ºC for 1 hr, cooled in a desiccator, and weighed to the
195 nearest 0.1 mg before a sample was added, including a tube as a reagent blank. (2) At the end of
196 wet acid digestion, the content in the tube was centrifuged and any precipitate in the tube was
197 saved instead of being discarded. It was then washed with 8 mL of deionized water twice before
198 being dried in the forced air oven at 140ºC for 1 hr. After cooling in a desiccator, tubes with
199 precipitate as well as reagent blank tube were weighed. The content of wet acid indigestible ash
200 by nitric acid digestion (WAIAn) in a given sample was calculated as follow:
201
202 Wet acid indigestible ash (WAIA) content (% dm) = ((Weight of tube with WAIA – Weight of
203 sample tube before digestion) – (Weight of blank tube after digestion – Weight of blank tube
204 before digestion))/Weight of sample adjusted to dry matter x 100 (1)
205
206 2.6. Measurement of WAIA content by sulfuric acid digestion
207 All algae and non-algae samples were also measured for WAIA content by another procedure
208 based on sulfuric acid digestion. The procedure consisted of the following steps. Samples were
209 weighed (125-200 mg each) and put into heavy duty 15 mL glass centrifuge tubes, which had
210 been pre-conditioned and tared. Four ml of 98% sulfuric acid was then added into each tube
211 under a fume hood. The tubes were placed in a heater block and heated to 120 ºC and
212 maintained at this temperature throughout digestion. After heating for about 30 min, the tubes
213 were vortexed to break up any clumps. To avoid liquid splashing, a low vortex speed was
214 selected and acid resistant gloves were used. After another 30 min heating each tube was

8
215 vortexed again before 30% hydrogen peroxide was added with great caution drop by drop to
216 avoid possible overflow of the digestion liquid from the tube when the sulfuric acid bubbled
217 upon hydrogen peroxide addition. Addition of hydrogen peroxide was repeated one more time.
218 Samples were then continuously heated overnight.
219
220 In the next morning, 30% hydrogen peroxide was added to each tube, drop by drop, until the
221 sample solution in each tube became clear. For some tubes, even though the solution had
222 become clear, some black particles were present. If this happened, heating was continued and
223 hydrogen peroxide was occasionally added drop by drop and swirled until the black particles
224 disappeared.
225
226 At this stage, tubes were removed from the block heater. After cooling for 30 min, 4 ml of
227 deionized water was added and each tube was vortexed. Upon centrifugation at 4000 x g for 10
228 min, the supernatant was decanted. The precipitate (if any) in the tube was washed with 8 mL of
229 deionized water twice before it was dried in the forced air oven at 140ºC for 1 hr. Upon cooling
230 in a desiccator, each tube with precipitate was weighed. The content of wet acid indigestible ash
231 by sulfuric acid digestion (WAIAs) in a given sample was calculated by the same equation used
232 for nitric acid digestion.
233
234 2.7. Calculation of WADA content
235 In this study, the new concept states that algae ash can be divided into two components: WAIA
236 and wet acid digestible ash (WADA), based on their digestibility following a wet acid digestion
237 procedure using strong attacks. Therefore, WADA content could be easily calculated by
238 subtracting the WAIA content measured in the above procedures from the total ash content
239 measured by the ignition method, when all were expressed as % dm.
240
241 2.8. Microscopic examination
242 Samples of algae and non-algae in three matrices (original matrix, ash obtained by dry ashing,
243 and wet acid indigestible ash) were ground with a mortar and pestle, if necessary, and then mixed
244 with deionized water and examined for microstructures with a microscope (Zeiss, Germany)

9
245 fitted with a digital camera, under normal light without filters. Magnification was fixed at 400X
246 (40X objective lens and 10X ocular lens).
247
248 2.9. Data analysis and statistical treatments
249 Replicate data on proximate composition, mineral composition as well as WAIA contents were
250 treated with a JMP software, version 12 (JMP, a Business Unit of SAS Institute Inc., Cary, NC,
251 USA). Analysis of variance (ANOVA) was conducted in order to determine difference among
252 algae samples. ANOVA was also carried out to determine the effect of washing frequency under
253 each of two procedures for WAIA measurement and compare difference between the two
254 procedures. Tukey‟s honestly significant difference test or student test was then used for pair-
255 wise comparisons of means if a factor had a significant effect at p < 0.05.

256 3. Results and Discussion

257 3.1. Differences in proximate composition among 12 algae samples

258 There were significant variations in proximate compositions among 12 algae samples (Table 1).
259 On dm basis, protein content varied from 9.2% to 52.5%; oil content from 0.5% to 31.9%; ash
260 content from 1.9% and to 37.4%; and CHO content from 29.7% to 77.2%. Algae 4, the only
261 sample that was produced in a PBR system, had the lowest ash content (1.9%). These variations
262 were attributed to differences in species, growing methods and conditions, and harvesting
263 methods. The results were consistent with findings of previous reports ([5, 15]. Laurens et al
264 [24] showed that results of compositional analysis varied even with analytical methods used for
265 the same algae samples due to different measurement chemistries. This makes comparison
266 among studies difficult
267
268 In addition, diatoms are known for high ash content [25]. The diatom sample was found to have
269 32.9% ash, which was consistent to the finding of Whyte [26]. For many other biological
270 materials, such as grains, forage, and herbaceous biomass, ash content ranged from 1 to 15%,
271 with most being less than 5% [27]. Therefore, among biological materials, algae not only have a
272 greatest variation but also highest ash content in some samples. This unique and, in most cases,
273 undesirable feature, highlights the importance of further investigation into the ash component of
274 algae.

10
 Table 1. Proximate composition of 12 algae samples

Algae Moisture Protein Oil Ash Carbohydrate

1 5.6 d 48.9 b 1.4 f 6.1 i 43.5 e
2 5.9 d 45.7 c 1.7 f 7.0 h 45.6 d
3 5.5 d 25.0 e 1.6 f 37.4 a 36.0 g
4 4.4 ef 9.2 l 11.6 c 1.9 j 77.2 a
5 (Chlorella) 5.4 de 21.9 g 3.8 e 7.9 g 66.3 b
6 (Diatoms) 7.1 bc 15.3 k 22.1 b 32.9 c 29.7 h
7 4.2 f 20.4 h 0.7 f 36.6 b 42.4 ef
8 7.9 b 39.1 d 0.7 f 14.1 e 46.2 d
9 1.9 g 18.2 j 31.9 a 7.2 h 42.7 e
10 (Spiralina) 5.7 d 52.5 a 0.5 f 6.4 hi 40.6 f
11 6.2 cd 24.5 f 1.7 f 30.3 d 43.5 e
12 11.2 a 19.2 i 8.5 d 12.4 f 59.8 c

Average 5.9 28.3 7.2 16.7 47.8

Mean of replicate results. All expressed as % dry matter basis except for moisture.
Column means bearing different letters differed significantly at P<0.05.
275
276
277 3.2. Mineral composition of 12 algae
278 The ICP-OES instrument was able to detect 21 minerals (Al, As, Ba, Ca, Cd, Co, Cr, Cu, Fe, K,
279 Mg, Mn, Mo, Na, Ni, P, Pb, S, Si, V, and Zn) from digested algae samples. Six of them,
280 including As, Cd, Co, Mo, Pb, and Si, had contents near to or less than their detection limits;
281 their data are not shown. Results indicate that mineral concentrations varied significantly among
282 the 12 algae samples. Macro minerals found in algae included Ca, K, Mg, Na, P, S. Fe, and Al
283 (Table 2), with means as 20.0, 7.4, 3.9, 11.0, 8.4, 6.6, 2.7, and 1.7 mg/g dm, respectively. Ba,
284 Co, Cr, Cu, Mn, Ni, V and Zn were among trace minerals detected, with means of 48, 3, 24, 30,
285 123, 12, 5, and 282 µg/g dm, respectively (Table 3). It is known that some algae genera have
286 the ability to accumulate various inorganic substances within or around the cell, with calcite
287 and/or aragonite, the two most common, naturally occurring, crystal forms of calcium carbonate
288 (CaCO3), as predominant mineral deposits of algae [28]. The process is known as calcification.
289 The observation that among all the minerals measured Ca had the highest average content (20.0
290 mg/g dm, Table 2) was consistent with the previous finding of algae calcification [16, 28].
291
292

11
Table 2. Macro mineral contents (mg/g dry matter) in 12 algae

Algae Ca K Mg Na P S Fe Al

1 3.4 e 8.0 d 2.5 cd 3.2 d 10.0 cd 6.2 c 0.7 d 0.7 ef

2 2.5 e 5.4 ef 3.3 c 5.6 cd 15.5 a 7.1 b 1.2 d 0.2 f
3 46.0 b 5.3 ef 2.5 cd 2.6 d 6.2 ef 6.4 c 8.6 a 2.7 cd
4 2.0 e 5.0 ef 1.0 e 0.7 d 1.2 h 3.6 e 0.1 d 0.0 f
5 (Chlorella) 12 5e 5.7 e 0.6 e 23.0 b 4.0 gh 7.0 b 3.1 bc 1.7 de
6 (Diatoms) 17.0 d 10.5 c 2.2 d 65.0 a 8.9 d 6.4 c 1.0 d 0.7 ef
7 66.0 a 6.2 de 6.2 b 12.0 c 2.6 gh 8.8 a 9.0 a 6.0 a
8 26.0 c 5.3 ef 8.6 a 1.2 d 13.5 ab 8.0 ab 2.4 c 0.7 ef
9 1.9 e 13.5 b 5.7 b 11.0 c 8.9 d 7.0 b 0.4 d 0.0 f
10 (Spirulina) 2.2 e 16.5 a 2.8 cd 5.7 cd 9.6 d 8.3 a 0.4 d 0.1 f
11 29.5 c 4.5 ef 5.3 b 2.0 d 8.1 de 5.8 c 2.7 bc 3.3 bc
12 42.5 b 3.6 f 6.2 b 0.7 d 12.0 bc 4.8 d 3.5 b 3.8 b

Average 20.0 7.4 3.9 11.0 8.4 6.6 2.7 1.7

Mean of two replicate results. Column means bearing different letters differed significantly at P < 0.05.

Table 3. Micro mineral contents (µg/g dry matter) and total mineral content (% dry matter) in 12 algae

Algae Ba Co Cr Cu Mn Ni V Zn Total minerals*

1 14 c 1b 2c 12 c 120 c 3d 1d 25 d 3.5 e
2 64 b 1b 3c 12 c 150 bc 3d 1d 35 d 4.1 e
3 64 b 3b 140 a 9c c 170 b 58 a 7 bc 145 cd 8.1 b
4 6c 8 ab 2c 7c 12 e 3d 1d 12 d 1.4 f
5 (Chlorella) 5c 1b 18 c 26 bc 36 de 9 cd 8b 35 d 4.6 e
6 (Diatoms) 25 c 1b 6c 17 bc 39 de 5d 1d 70 d 11.2 a
7 105 a 15 a 72 b 13 c 160 bc 30 b 20 a 76 d 11.7 a
8 75 b 2b 11 c 165 a 270 a 15 c 4 cd 850 b 6.7 bc
9 30 c 2b 2c 30 bc 71 d 3d 1d 48 d 4.8 de
10 (Spirulina) 5c 1b 9c 6c 34 de 3d 1d 1550 a 4.7 e
11 77 b 1b 8c 19 bc 255 a 6d 7 bc 380 c 6.2 cd
12 107 a 2b 10 c 43 b 165 b 5d 6 bc 160 cd 7.8 b

Average 48 3 24 30 123 12 5 282 6.2

Mean of two replicate results. *Total mineral content was calculated from data of Table 2 and this table.
Column means bearing different letters differed significantly at P < 0.05.

12
293 Mineral composition in various algae has been extensively profiled [5, 12-15]. However, due to
294 the differences in algae species, growing conditions, and harvesting methods and the difference
295 in analytical procedures among studies, it is often difficult to compare the published results.
296 There is lack of systematic research on developing a reliable analytical method for accurate
297 measurement of mineral composition of algae samples.
298
299 In this study, both dry ashing and wet ashing procedures were used. Dry ashing uses a high
300 temperature muffle furnace to vaporize water and burn organic substances. Most minerals are
301 converted to oxides, sulfates, phosphates, chlorides, carbonates, and/or silicates. Anything left
302 after ignition is weighed and amounts to the total ash content [29]. Wet ashing is a procedure to
303 mineralize and dissolve solid samples, by using acids, oxidizing agents or their combinations. It
304 breaks down and removes the organic matrix so that individual minerals are left in an aqueous
305 solution and measured up in a subsequent analysis using spectroscopy techniques [19].
306
307 When contents of individual minerals were added up, TM content could be obtained for each
308 algae sample (Table 3). This value correlated positively with the ash content (Table 1), having
309 an r=0.875, but TM content was always significantly less than total ash content (Fig. 1A). This
310 difference is expected, since ash by dry ashing consists of mineral salts containing oxygen [29]
311 and the oxygen content of ash accounts for much of its weight. In contrast, the TM content
312 obtained by wet ashing and ICP analysis does not include the oxygen content. Furthermore, at
313 low ash content, the difference between the ash content and TM content was about 25% relative
314 to the total ash content. As the ash content increased, this relative difference increased up to
315 80%. The mean value for the relative difference among 12 algae was 50.4%. There was a highly
316 positive correlation for the relative difference (between ash and TM contents) and the ash
317 content, with r = 0.920 (Fig. 1B). Therefore, with increasing ash content in algae, more minerals
318 and/or elements associated with them were not measurable by the wet ashing method. The
319 results indicate that mineral analysis based on the nitric acid digestion could not adequately
320 characterize the ash component of algae.
321

13
322
323
324 Fig. 1. Correlations between the total mineral content obtained by a strong attack wet acid
325 digestion and the total ash content obtained by dry ashing (A) and between the relative difference
326 (between ash content and the total mineral content) and the total ash content (B).
327
328 Incomplete digestion by the nitric acid digestion procedure can be another factor for the
329 noticeable difference between ash and TM contents. For wet acid digestion (wet ashing) of solid
330 samples before mineral analysis, six reagents are typically used: nitric, sulfuric and perchloric
331 acids, hydrogen peroxide, hydrochloric and hydrofluoric acids. The first four contribute
332 principally to the destruction of organic matter while the remaining two ensure the dissolution of
333 inorganic compounds [19]. Although different choices of these reagents have led to many
334 procedures reported in the literature, they can be categorized into four types: total
335 decomposition, strong attacks, moderate attacks, and partial digestions [21]. For routine
336 analysis, a procedure of strong attacks is commonly used. This type of procedure works well for
337 purely organic samples (such as fat) with respect to mineralization and dissolution. However, for
338 samples of mixed composition (organic and inorganic), the procedure of strong attacks cannot
339 ensure the dissolution of silicate compounds and consequently of all elements associated with
340 them. This problem is encountered mainly in analysis of plant materials, since some plant media
341 contain variable concentrations of silica [19, 21].
342
343 Silicon (Si) is the second most abundant element in the earth‟s crust (27% by weight), after
344 oxygen (55% by weight). Silicon does not exist in free form but occurs as silicates or as silica.
345 A silicate is a compound containing an anionic silicon compound. The great majority of the
346 silicates have a basic chemical unit of the tetrahedron shaped anionic group (SiO4) with a
347 negative four charge (-4). Silica is sometimes considered a special silicate. Chemically it is

14
348 known as silicon dioxide, SiO2, with no negative charge [25]. Silica is most commonly found in
349 nature as quartz (a crystalline form). It is also present in various living organisms as biogenic
350 silica (a non-crystalline form). Among the algae, diatoms are unique in that they are enclosed
351 within cell walls made of silica called a frustule, and thus require silicon to grow. The biogenic
352 silica is synthesized intracellularly by the polymerization of silicic acid monomers. This material
353 is then transported to the cell exterior and added to the wall [25]. The presence of small but
354 significant amounts of silica in cell walls of non-diatom algae is also reported [16, 30]. In this
355 study, a procedure featuring nitric acid digestion was used. Because it is one of the commonly
356 used strong attack procedures, incomplete dissolution of silica is expected. This explains the
357 observation that with increasing ash content in algae samples, less minerals were measured by
358 the wet ashing method (Fig. 1B).
359
360 For complete decomposition and dissolution of minerals associated with silicates, a total attack
361 procedure using HF as one of mixed acids is recommended for measuring mineral composition
362 in silicate-containing samples [19, 21]. HF is the most effective mineral acid for breaking up the
363 strong Si-O bond. The disintegration of the silicate lattice leads to the liberation of silicon and
364 other metals bound within the silicate structure. Yet, Si, once liberated, reacts with HF to form
365 SiF4, which volatilizes upon heating during digestion. Therefore, the total attack procedure does
366 not allow for Si measurement, even though it helps dissolute silicates [19,31]. Furthermore,
367 there are two key factors that make the use of HF more difficult to realize in wet digestions.
368 First, in classical wet digestion procedures, glass tubes are often used. Unfortunately, glass
369 material itself is decomposed by HF. Digestion procedures with HF as a reagent must then be
370 performed in polytetra-fluoro-ethylene (PTFE) vessels, which are expensive and oftentimes
371 unavailable in sizes needed. Second, the attack by HF is usually concluded by evaporation to
372 dryness. Although HF reagent does not destroy PTFE vessels, deterioration of the PTFE inner
373 surface during evaporation as a result of sample overheating is prone to occur, since the opacity
374 of PTFE does not facilitate a possible visualization or control of this problem [19].
375
376 Because of the drawbacks of adding HF in an acid digestion system, it is not surprising that
377 among many studies on measurements of inorganic elements in various algae samples [5, 12-15]
378 only one employed a total decomposition procedure [13]. Among the rest, Tuzen et al. [14]

15
379 used a mixture of HNO3 + H2O2 in a microwave digestion system; Fuentes et al. [5] dissolved
380 their algae biomass in a mixture of HNO3 and HCl; and Adams et al. [15] digested their algae
381 samples with HNO3 in a closed vessel (note that this study also used HNO3 digestion but in an
382 open glass test tube). The wet acid digestion procedures reported in the last three studies all
383 belonged to strong attacks. Hou and Yan [12], on the other hand, used instrumental neutron
384 activation analysis, which bypassed wet acid digestion of samples, but they also failed to detect
385 Si.

386 For measuring Si content in in geologic materials, a traditional sample preparation involves
387 fusion with an alkaline flux of a sodium or lithium salt [31], but for biological materials, dry
388 ashing plus lower temperature fusion are generally performed. For example, Reay and Bennett
389 [32] described a sample preparation method to determine total Si content in biological material
390 by ashing and fusing in nickel crucibles with potassium hydroxide containing potassium
391 tetraborate/nitrate. For measuring biogenic silica content, a wet chemical extraction technique
392 consisting of 24 hr time-course digestion and backwards extrapolation to time zero of observed
393 results is commonly used [33]. Yet, all fusion methods introduce high salt concentrations in the
394 measurement solutions and post a high risk of contamination through the reagents. Therefore,
395 these procedures are very time-consuming and prone to errors [8].

396 3.3. Method development for determining wet acid indigestible ash
397 For mineral analysis, a wet acid digestion procedure of strong attacks is easier to use and better
398 adapted to routine analysis than a total attack procedure containing HF. However, upon
399 completion of wet digestion by a strong attack procedure, oftentimes there are some particulates
400 in the digest solution due to incomplete digestion. These particulates are silicates in nature,
401 which have been resistant to the wet acid digestion. To avoid the risk of plugging a sample
402 delivery device of detection equipment, most digestion procedures further require that, before
403 analyzing the solution for mineral contents, the particulates be removed either by filtration or by
404 centrifugation or allowing the sample to settle overnight [20, 34]. Indeed, upon careful
405 examination of the nitric acid digestion procedure used for mineral analysis in this study, it was
406 found that a substantial portion of ash in the high ash algae samples could not be digested. The
407 indigestible ash existed as particulates in the digestion solution. They had to be removed before
408 ICP-OES detection.

16
409 In this study, in order to develop a reliable and easy-to-adapt methodology to quantitatively
410 characterize the ash component of algae, the insoluble particulates in the digestion tubes, instead
411 of being discarded, were measured gravimetrically upon filtration or centrifugation at the end of
412 the wet acid digestion. They are termed as wet acid indigestible ash, abbreviated as WAIA. As
413 an added benefit, by measuring WAIA content in algae samples, our dilemma of having
414 incomplete mineral decomposition and dissolution by using an easy-to-adapt procedure of strong
415 attacks or having complete mineral decomposition and dissolution by using a hard to adapt
416 procedures of total acid decomposition or alkaline fusion is effectively resolved.
417
418 By the definition, the procedure for measuring WAIA is basically a wet ashing method of strong
419 attacks, used routinely for preparing solid samples for mineral analysis. In this study, for method
420 development, two procedures were described for measuring WAIA, based on two commonly
421 used wet acid digestion methods. One was nitric acid digestion, which was also used for mineral
422 analysis by ICP-OES in the same study. The other was sulfuric acid digestion in the presence of
423 H2O2. Compared to the procedure of acid digestion for mineral analysis, there were two extra
424 functional steps needed for the procedure of WAIA measurement. One was that at the beginning
425 of the procedure, empty test tubes were pre-conditioned and weighed. The other was that at the
426 end of acid digestion, the precipitate (that is, indigestible ash), if there was any, was recovered by
427 either centrifugation or filtration with twice water washing, and then dried and weighed. Its
428 content was then calculated based on the original sample weight and expressed as % dry matter
429 basis or as is basis. Obviously, a combined procedure can easily be developed for dual purposes:
430 mineral analysis and measurement of WAIA.
431
432 Again, for method development purpose, a separate experiment was carried out to compare
433 filtration and centrifugation in recovering WAIA in the digestion tube. When using filtration, a
434 glass microfiber filter (Whatman 934-AH) had to be used since strong acids would burn a paper
435 filter. Furthermore, after the first filtration, the insoluble particles contained some residue acid
436 that was not easily washed away by simply adding water and filtering again. This occurrence was
437 particularly problematic with sulfuric acid. In contrast, centrifugation and washing with water
438 was easier and more effective. Therefore, centrifugation is preferable over filtration for

17
439 recovering indigestible ash, and was used throughout this study, while data by filtration was
440 dismissed due to unreliability.
441
442 In another experiment, when centrifugation was used for precipitate recovery, the effect of the
443 frequency of washing the precipitate with deionized water was also assessed. As expected, the
444 procedure without washing gave the highest WAIA value (Table 4). This observation was
445 particularly evident when sulfuric acid digestion was used. The reason was that sulfuric acid had
446 much higher boiling point than nitric acid and it remained in the digestion tube. If no washing
447 was performed, it tended to stick to precipitate particulates and thus contributed the mass of the
448 precipitate. This explains the observation that at 0 washing the average content of WAIAs of 5
449 selected algae samples was as high as 172.19% while that of WAIAn was only 17.07%. As the
450 washing step increased from 0 to 1, there was a significant decrease in values of WAIAn and
451 WAIAs, with WAIAs having a sharper decrease from 172.19% to 28.92%. An increase of
452 washing frequency from 1 to 2 led to another significant decrease in both WAIAn and WAIAs,
453 but less dramatic. For washing frequency beyond 2, the decrease in WAIAn was insignificant.
454 Although a decrease in WAIAs was still evident, it was insignificant due to higher standard
455 deviation. Since increasing washing steps not only required more time but also posed a risk of
456 losing some insoluble ash when decanting supernatants, a procedure with water washing twice
457 was adapted for both methods (nitric acid digestion and sulfuric acid digestion) to measure
458 WAIA values. As shown in Table 4, upon two washing cycles, the value of WAIAs was still
459 systematically higher than that of WAIAn, with average value of 5 algae samples being 15.8%
460 vs. 14.1% dm (Table 4).
461
462
463
464
465
466
467
468
469
470
471
472
473

18
Table 4. Effect of washing frequency with centrifuge on contents of WAIAn and WAIAs in 6 selected algae

WAIAn WAIAs
No. of wash 0 1 2 3 0 1 2 3
Algae 2 2.1 1.2 0.9 1.0 67.9 8.7 1.7 1.3
Algae 3 28.5 26.0 25.6 25.4 127.2 38.7 26.0 25.0
Algae 6 (Diatoms) 21.5 16.6 16.1 15.9 400.8 41.6 19.0 16.3
Algae 7 21.1 18.1 17.6 17.6 82.1 33.5 20.5 20.9
Algae 8 4.7 3.3 2.9 2.9 102.0 10.8 3.6 3.1
Algae 11 24.5 22.0 21.3 21.4 253.1 40.5 24.1 22.2

14.5 14.1 14.0 28.9 15.8 14.8

WAIA content (average) 17.1 a b c c 172.2 a b c c
Standard deviation (average) 0.2 0.2 0.1 0.1 22.1 4.3 1.2 0.4

Mean of duplicate results, WAIA content was expressed in % dry matter.

WAIA: wet acid indigestible ash. WAIAn: WAIA by nitric acid method, WAIAs: WAIA by sulfuric acid method.
Under WAIAn or WAIAs, averaged WAIA values with different superscript letters differed significantly at
P<0.05.

WAIA contents of 12 algae samples, measured by both procedures of nitric acid and sulfuric acid digestion,
are given in Table 5. For easy comparison, ash content of these samples is also given, all on dm basis.
There was a significant variation in WAIA contents among the 12 algal samples. WAIA by nitric acid
digestion (WAIAn) varied from 0.1% to 25.6%, dm, with a mean value of 7.3%, while WAIA by sulfuric
acid digestion (WAIAs) from 0 to 26.1%, with a mean value of 8.1%. For the same 12 samples, as noted
before, the ash content by dry ashing ranged from 1.9 to 37.4% dm, with a mean value of 16.7%.

Besides the 12 algae samples, four non-algae samples: hulled oat grain, defatted soy meal, oat forage, and
sand, were also measured for contents of ash and WAIA. The first three samples represented grain,
processed grain, and crop residue, respectively. Their ash content varied from 2.6 to 11.3% dm. Silica is
expected to be low for these crop materials. Indeed, WAIAn values were found be close to 0% for soy meal,
0.3% for oat grain and 2.3% for oat forage. Sand is a naturally occurring granular material and composed of
finely divided rock and mineral particles. Its composition varies with local rock sources and conditions, but
the most common constituent of sand in inland continental settings is silica, usually in the form of quartz.
The second most common component of sand is calcium carbonate [35]. Upon sieving and water washing,
the

19
 Table 5 Contents of ash, wet acid indigestible ash by nitric acid digestion (WAIAn) or sulfuric acid digestion (WAIAs), wet acid digestable ash by nitric acid
digestion (WADAn), and related paramters in 12 algae and 4 non-algae samples, and total minerals content and related parameters in 12 algae samples.

WAIAn WADAn TM+ (TM+WAIAn) Ash-(TM (Ash-TM-

Sample name Ash WAIAs WAIAn WADAn /Ash /Ash TM TM/Ash WAIAn /Ash +WAIAn) WAIAn)/Ash
% %
Biological, aglae dm % dm % dm % dm % % dm % % dm % % dm %
6.1 1.0 ± 0.1 3.5
1 g 1.2 ± 1.1 c fg 5.1 15.6 84.4 e 56.7 4.4 72.3 1.7 27.7
7.0 0.9 ± 0.0 4.1
2 f 1.7 ± 0.6 c g 6.1 12.9 87.1 e 58.7 5.0 71.6 2.0 28.4
37.4 26.0 ± 0.5 25.6 ± 0.2 8.1
3 a a a 11.8 68.4 31.6 b 21.5 33.7 89.9 3.8 10.1
0.1 ± 0.1 1.4
4 1.9 i 0.0 ± 0.4 c h 1.8 2.6 97.4 f 71.9 1.4 74.5 0.5 25.5
7.9 0.6 ± 0.2 4.6
5 (Chlorella) f 0.5 ± 0.2 c gh 7.4 6.9 93.1 e 58.3 5.2 65.3 2.8 34.7
32.9 19.0 ± 0.9 16.1 ± 0.1 11.2
6 (Diatoms) b b d 16.8 49.0 51.0 a 33.9 27.3 82.9 5.6 17.1
36.6 20.5 ± 2.9 17.6 ± 0.2 11.7
7 a a c 19.0 48.1 51.9 a 32.0 29.3 80.1 7.3 19.9
14.1 2.9 ± 0.1 6.7
8 d 3.6 ± 0.5 c e 11.2 20.3 79.7 bc 47.5 9.5 67.8 4.5 32.2
7.2 0.0 ± 0.1 0.1 ± 0.3 4.8
9 f d h 7.1 1.4 98.6 de 67.6 4.9 69.0 2.2 31.0
6.4 0.0 ± 0.2 0.2 ± 0.1 4.7
10 (Spirulina) g d h 6.3 2.3 97.7 e 73.4 4.9 75.8 1.6 24.2
30.3 24.1 ± 1.0 21.3 ± 0.1 6.2
11 c a b 9.0 70.4 29.6 cd 20.4 27.5 90.8 2.8 9.2
12.4 7.8
12 e 0.9 ± 0.3 c 1.4 ± 0.1 f 11.0 11.3 88.7 b 62.3 9.2 73.5 3.3 26.8

Average values 16.7 8.1 ± 0.7 7.3 ± 0.1 9.4 25.8 74.2 6.2 50.3 13.5 76.1 3.2 23.9
Biological, non-algae
Oat grain 2.6 0.7 ± 0.2 0.3 ± 0.0 2.3 12.2 87.8
Oat forage 11.3 2.7 ± 0.4 2.3 ± 0.0 9.0 20.3 79.7
Defatted soy
meal 6.8 0.0 ± 0.0 0.1 ± 0.0 6.7 1.6 98.4

Average values 6.9 1.1 ± 0.2 0.9 ± 0.0 6.0 11.4 88.6
Geologic
Sand 95.5 85.4 ± 0.4 78.5 ± 0.2 17.0 82.2 17.8

Means of duplicate results. Column means with different letters differed significantly at p <0.05.
TM: total minerals. % dm: % dry matter. WADA was calculated by subtracting WAIA values from ash content values.
474
475 fine sand sample of the present study had 95.5% ash. Yet, its WAIA value ranged from 78.5% by
476 nitric acid digestion and 85.4% by sulfuric acid digestion. The measured amounts presumably
477 were silica (quartz) that was resistant to strong acid digestion. The lost portion of ash upon acid
478 digestion presumably resulted from dissolution of calcium carbonate and other non-silica
479 minerals, which went into the supernatant. Thus, the analytical procedure for WAIA was
480 applicable not only to algae but also to geologic and other biological materials.
481
482 Not surprisingly, correlation analysis showed that WAIAn and WAIAs were highly correlated
483 with each other, with r=0.994 for 12 algae samples (Fig. 2A) and r=0.999 for 16 combined

20
484 samples of algae and non-algae (Fig. 2B). Yet, as noted in Table 4, values of WAIAs were
485 slightly but almost systematically higher than those of WAIAn, and WAIAs values also had
486 higher variation than those of WAIAn. Although additional washing upon sulfuric acid digestion
487 could probably bring the values of WAIAs closer to WAIAn, it was time-consuming, and could
488 lead to even higher variation than WAIAn. In addition, the nitric acid method uses only one
489 reagent while the sulfuric acid method requires a combination of both sulfuric acid and hydrogen
490 peroxide. The former is relatively less labor intensive than the latter. Considering the two
491 aspects, the procedure using nitric acid digestion is preferred over that of sulfuric acid digestion
492 in measuring algae samples for WAIA content.
493
494

495
496
497 Fig. 2. Correlations between wet acid indigestible ash by sulfuric acid digestion (WAIAs) and
498 wet acid indigestible ash by nitric acid digestion (WAIAn) for 12 algae samples (A) and 16
499 algae and non-algae samples (B).
500
501 3.4. Relationships among WAIA, total minerals and total ash contents
502 There was a very high correlation between WAIA and total ash content, by either digestion
503 method, with r = 0.943 for WAIAn vs. ash content and r = 0.946 for WAIAs vs. ash content
504 (chart not shown). Therefore, algae with higher ash content had higher WAIA content. More
505 specifically, for algae samples with ash content < 14.1%, the content of WAIAn varied from 0.1
506 to 2.9% and the content of WAIAs from 0 to 3.6%. Yet, for algae samples having ash content >
507 30.3%, the content of WAIAn varied from 16.1 to 25.6% and the content of WAIAs from 19.0 to
508 26.0%. These observations indicate that: 1) although the two acid digestion methods described

21
509 in this study both belonged to strong attacks procedures [19], they could not digest siliceous
510 materials in algae, and 2) WAIA was a significant contributor of the ash component in algae
511 samples and its amount increased with increasing ash content. Therefore, WAIA content can be
512 an important quality parameter for algae samples.

513 There are intriguing relationships among WAIA, total minerals (TM) and ash contents. Here the
514 ash content was obtained by dry ashing method, while the WAIA and TM contents were
515 obtained by the wet ashing method (after nitric acid digestion, the supernatant was used for
516 mineral analysis while the precipitate was weighed as WAIA content). As shown in Table 5, on
517 average, WAIA represented 25.8% of total ash, while TM represented about 50.3% of total ash.
518 Together, WAIA & TM represented 76.1% of total ash. So, 23.9% of total ash was not
519 measurable by the wet ashing method for measuring both TM and WAIA. In other word, the dry
520 ashing method provided an average of 23.9% more ash value than the wet ashing method even
521 though the latter recovered both TM and WAIA. Similar to the reason for the difference
522 between ash and TM explained before, the reason for this is that the oxygen content contributes
523 significantly to the weight of the ash measured by dry ashing. Although the indigestible fraction
524 (WAIA) included oxygen in the form of slicates or silica, the TM measured by the nitric acid
525 digestion and ICP-OES analysis did not include oxygen and some other elements (such as C
526 from carbonates and Cl from chlorides) that were released from non-silicate minerals.
527 Furthermore, there was a highly negative correlation between WAIAn/Ash and TM/Ash, with r =
528 0.980 (Table 5). Therefore, as the contribution of WAIA toward ash increased (as in algae with
529 high ash content), the contribution of TM toward ash decreased and vice versa.

530 Hampel [8] reported great variation in contents of ash, biogenic silica, and total silica among
531 algae grown on five different locations and harvested at different times, with the ash content
532 ranging 31-81% dm. For analyzing biogenic and total silica contents, the number of harvested
533 samples had to be significantly reduced due to high cost and excessive times required [32].
534 Nevertheless, Hampel [8] found that 1) the content of biogenic Si varied from 0 to 27.6%, with
535 most less than 3%, dm, 2) the content of total silica varied from 3.6 to 67.6%, and 3) the total
536 silica content was generally correlated with the amount of ash measured. In the present study,
537 WAIA, which was equivalent to the total silica content, also correlated well with total ash

22
538 content in 12 samples collected. Yet, the method for WAIA is much easier to adapt than the
539 methods for measuring total silica.

540 3.5. Wet acid digestible ash

541 In the present study, the new concept is proposed that ash in biological (including algae biomass)
542 and geologic material can be divided into WAIA and wet acid digestible ash (WADA), based on
543 its digestibility following a wet acid digestion procedure of strong attacks. By this concept, the
544 content of WADA for all 16 samples was calculated based on the difference between ash content
545 obtained by dry ashing and WAIA content when all were expressed as % dm. Since nitric acid
546 digestion is preferred over sulfuric acid digestion as just discussed, only WAIAn values were
547 used for calculating WADA. Results show that among 12 algae samples, WADA content varied
548 from 1.8 to 19.0% dm, with a mean of 9.4 % dm. This type of ash was a major component of
549 total ash for most algae samples, since it represented 29.6 to 98.6% of total ash with a mean of
550 74.2% for 12 algae samples. In contrast, WAIA was a minor component for the algae samples
551 having ash content less than 14.1% dm, accounting for less than 20.3% of total ash. However,
552 for the 4 algae samples that had ash content over 30%, WAIA contribution was higher,
553 accounting for 48.1-70.4% of total ash. It should be pointed out that the TM content measured
554 by the nitric acid digestion was only a major portion of WADA. The difference between WADA
555 and TM represents another portion of WADA. It equals to the difference between total ash
556 content and sum of TM and WAIA (Table 5, the second column from the right). In other words,
557 this portion of WADA could not be measured by both TM and WAIA analysis. On an average, it
558 represents about 23.9% of total ash (Table 5, the last column from the right). As just pointed out
559 earlier, chemically, this portion consisted of oxygen and some other elements that were released
560 from non-silicate materials from algae samples during nitric acid digestion, but were not
561 measured by the ICP-OES analysis.
562
563 3.6. Microscopic examination
564 To further investigate what constituted WAIA and thus what contributed to the total ash content,
565 WAIA of all 16 samples, resulting from both nitric acid digestion and sulfuric acid digestion,
566 were examined under a microscope for their microstructure. Since WAIAs gave similar
567 microstructure as WAIAn, only the microstructure of WAIAn was shown and discussed

23
568 hereafter. Furthermore, for easy comparison, the 4 non-algae samples are grouped into one tier
569 while the 12 algae samples are grouped into next 3 tiers (4 samples each) based on their WAIA
570 contents in a decreasing order. Their micrographs were presented in Figs. 3-6, respectively. It
571 should be noted here that the micrographs show only the microstructure of insoluble particulates
572 after wet acid digestion for a given sample. The density of these particulates shown in a
573 micrograph, unfortunately, could not correlate with the WAIA content for the given sample,
574 since how dense these particulates were packed depended on 1) how a WAIA sample was diluted
575 and spread over a microscope slide with a cover slip before microscopic examination, and 2)
576 which view field a picture was taken and recorded as a micrograph even though all were at the
577 same magnification of 400X.
578
579 WAIA from hulled oat grains and forage showed similar microstructure of amorphous siliceous
580 material having irregular shapes (Fig. 3A & B). WAIA of soy meal had some irregular
581 particulates but notably lacked in tubular structure, as compared to those of oat grain and forage
582 (Fig. 3C). Fine sand, on the other hand, consisted of mainly crystalline particulates of different
583 sizes and shapes (Fig. 3D).
584
585

24
586
587
588
589 Fig. 3. Micrographs of wet acid indigestible ash of oat grain (A), oat forage (B), defatted soy
590 meal (C), fine sand (D). Magnification: 400X for all. Bar = 20 µm.
591
592 For the tier with WAIAn contents in the range of 16.1-25.6%, WAIA of Algae 6 (diatoms)
593 consisted of countless translucent filaments and broken pieces, indicating biogenic silica nature
594 of diatom cell walls (Fig. 4A). The two dots on the top left appeared to be small pieces of sand
595 (that is, inorganic silica). In contrast, Algae 7 & 11 contained a lot of crystalline particulates of
596 different sizes, in addition to cell walls of diatoms, mostly in elliptical shapes (Fig. 4B & C).
597 Algae 3 showed another spectrum, having large and irregular particulates as a major portion and
598 some translucent cell wall structure as a minor one (Fig. 4D).
599

25
600
601
602 Fig. 4. Micrographs of wet acid indigestible ash of Algae 6 (diatoms) (A), Algae 7 (B), Algae 11
603 (C), and Algae 3 (D). Magnification: 400X for all. Bar = 20 µm.
604
605 For algae samples having WAIAn in the range of 0.9 to 2.9%, the next tier, varying
606 microstructures of WAIA were also observed (Fig. 5). Algae 1 showed sandy particulates and
607 translucent cell walls of diatoms in two distinct shapes, round and elliptical (Fig. 5A). For
608 Algae 2, circular diatoms cell walls as well as sandy particulates were predominant (Fig. 5B).
609 Algae 8‟s WAIA consisted of amorphous and crystalline particulates, large or small, and cell
610 walls of diatoms, mostly round shape (Fig. 5C). Algae 12 exhibited clusters of particles, but lack
611 of distinct diatom cell wall structure (Fig. 5D).
612
613

26
614
615
616 Fig. 5. Micrographs of wet acid indigestible ash of Algae 1 (A), Algae 2 (B), Algae 8 (C), and
617 Algae 12 (D). Magnification: 400X for all. Bar = 20 µm.
618
619 The 4 algae samples of the last tier contained very low but still measurable amount of WAIAn
620 (0.1 to 0.6% dm). Algae 4, with ash content at 1.9%, showed very minor contamination with
621 small particulates (Fig. 6A). Algae 9 had both diatom cells and sandy structure as its WAIA
622 components (Fig. 6B). Algae 5 (Chlorella) and 10 (Spirulina), on the other hands, both showed
623 amorphous siliceous cellular structure and sandy particulates, but diatom cell walls were absent
624 (Fig. 6C & D).
625

27
626
627
628 Fig. 6. Micrographs of wet acid indigestible ash of Algae 4 (A), Algae 9 (B), Algae 5 (Chlorella)
629 (C) , and Algae 10 (Spirulina) (D). Magnification: 400X for all. Bar = 20 µm.
630
631 The above observations indicate that: 1) WAIA of crop materials (grains or vegetative tissues)
632 consisted of mostly amorphous siliceous cell structures, with sandy material as a minor
633 component. 2) WAIA of algae samples consisted of three distinct silica-containing particulates
634 or structural substances: diatom cell walls, amorphous cellular structures (non-diatoms), and
635 sandy particulates of geologic origin. The amount and the composition of these three types of
636 silica substances varied with algae samples, which were further determined by species, growing
637 methods and conditions, harvest methods, and post-harvest handling. 3) Among the 4 algae
638 samples with high ash content, diatom sample consisted mostly of cell walls of diatoms, with
639 sandy particles as a minor component. For Algae 7 & 11, both sandy particles and diatom cell
640 walls were major components of their WAIA. Algae 3 sample was mostly contaminated with
641 sand and other silica materials of geologic origin, with diatom contamination playing a minor
642 role. 4) For algae samples with total ash content less than 15% and WAIA less than 2.9%, minor
643 contamination with sandy materials was mostly responsible for the amount of WAIA measured.

28
644 5) For all the samples, siliceous cell structures of non-diatoms might also account for some of
645 WAIA measured since early workers reported the presence of silica in non-diatom cell walls [16,
646 30].
647
648 Beside the matrix of WAIA, the 16 algae and non-algae samples were also examined under the
649 microscope in two other forms: the original form and the dry ashed form. The original form is
650 the form of the sample matrix upon receiving from providers. Since 12 algae samples came from
651 different sources, they might be harvested and dried in different ways. To reduce the number of
652 figures, under each of 4 tiers, two samples were chosen. Each chosen sample had two
653 micrograms, one for the original matrix and the other for the ash after dry ashing. They are
654 presented in Figs. 7-10.
655

656
657
658 Fig. 7. Micrographs of original matrix and ash by dry ashing of defatted soymeal (A, B) and fine
659 sand (C, D). Magnification: 400X for all. Bar = 20 µm.
660

29
661
662
663 Fig. 8. Micrographs of original matrix and ash by dry ashing of Algae 6 (diatoms) (A, B) and
664 Algae 11(C, D). Magnification: 400X for all. Bar = 20 µm.
665

30
666
667
668 Fig. 9. Micrographs of original matrix and ash by dry ashing of Algae 1 (A, B) and Algae 8 (C,
669 D). Magnification: 400X for all. Bar = 20 µm.
670
671 When samples were examined under the microscope in their original matrix, algae cells were
672 seen individually and/or in clusters, but their size, shape, and color varied greatly with sample
673 types. Therefore, the microstructure of the original matrix generally reflected the cellular nature
674 of a specific strain. Among the 12 algae, Algae 10 (Spirulina) was the only algae that consisted
675 of multiple cells. It was seen as broken pieces under the microscope (Fig. 10C). Again, all cell
676 structures were opaque, rather than translucent as seen for WAIA. A few large particulates were
677 seen in some viewing fields for some samples. They appeared to be sandy contaminants.
678

31
679
680
681 Fig. 10. Micrographs of original matrix and ash by dry ashing of Algae 4 (A, B) and Algae 10
682 (Spirulina) (C, D). Magnification: 400X for all. Bar = 20 µm.
683
684 Upon dry ashing in a furnace, the remaining ashes had similar microstructures for all samples
685 (Fig. 7-10 B &D). Although particulates were presented individually and/or in clusters,
686 individual particulates had similar size, shape, and color (Note: one exception was with diatom
687 ash (Fig. 8B) which was darker than ash from other algae samples). The ash particulates were not
688 translucent either. Apparently, during dry ashing, regardless of their ash contents, chemical
689 nature and original matrix, all biological samples were burnt and oxidized into inorganic
690 minerals. Thus, examination of ashed samples under a microscope not only failed to distinguish
691 any differences among algae samples but also failed to characterize the algae ash component.
692
693 By comparing micrographs of original samples and ashed samples (Figs. 7-10) with micrographs
694 of WAIA (Figs. 3-6), it is clear that the fine sand sample was the only one that showed a similar
695 microstructure in whatever forms (original matrix, dry ashed, and WAIA). For the rest of
696 samples, different microstructures were observed for the same sample, when examined in three

32
697 different forms. More importantly, for distinguishing algae samples and for identifying materials
698 that constitute the ash component, WAIA was the best form to be examined under a microscope
699 for two main reasons. First, the microstructure of WAIA was mostly translucent and the original
700 shape and size of silica-containing cell wall structures (such as diatom cell walls) and
701 containments (such as sandy particulates) were maintained. Distinction among three types of
702 silica-containing materials: diatom cell walls, sandy particulates, and non-diatom/amorphous
703 cellular structures, was evident in WAIA micrographs of algae samples. Second, WAIA
704 consisted of only siliceous materials. Other compounds (organic or inorganic) were digested,
705 vaporized, dissolved and removed into the supernatant. Thus it was the most concentrated form
706 of cell structures and contaminants that are siliceous in nature but nevertheless contributed to the
707 ash component. The concentration factor of WAIA over ash by dry ashing for a given sample
708 can easily be calculated by dividing the ash content with the WAIA content (since both were
709 based on % dry matter). For example, Algae 1 contained 6.1% ash and 1.0% WAIAn. So the
710 concentration factor for WAIAn over ash for Algae 1 was 6.1/1.0 = 6.1. The concentration factor
711 of WAIA over the original matrix for a given sample can also be calculated by dividing 100 with
712 the percent value of WAIA content. Thus, the concentration factor for Algae 1 was 100/1.0 =
713 100. In other words, by converting to WAIA, siliceous materials in Algae 1 sample were
714 concentrated 100 times. Because of this powerful concentration effect, under the microscope,
715 any contaminants or cellular structures that contained silica, even at very tiny amounts, could be
716 seen when WAIA was examined under a microscope for a given sample. A good example would
717 be the tier 4 algae samples shown in Fig. 6. These samples contained very low amounts of
718 WAIA, but contaminants can still be seen when their WAIA were examined under the
719 microscope. This effect was difficult to achieve when the original matrix of samples were
720 examined under a microscope, particularly for those samples that had relatively low ash contents.
721
722 3.7. Factors contributing high ash content in algae and proposed measures to mitigate them
723 Nagao et al. [36] found that net-plankton samples collected from shallow coastal water of central
724 Japan contained considerable amount of ash, varying from 19 to 80%, with a mean of 53%.
725 Because this value was much higher than that of algae grown in an oceanic area (about 11% ash),
726 they reasoned that the high ash contents from the net-plankton samples would not have
727 originated solely from zooplankton. Particulate inorganic materials, such as re-suspended

33
728 sediment and terrestrial runoff, are the most likely cause of the high ash content. Contamination
729 of diatoms is another possible contributor. However, they did not conduct studies to confirm
730 their hypothesis. In the present study, although the algae samples studied were not net-plankton
731 and did not contain zooplankton in general, the microscopic examination of WAIA, the unique
732 way of examining the ash components of algae biomass, clearly showed that the high ash content
733 of algae biomass is due in part to the presence of silica or silicates, resulting from: 1)
734 contamination with diatoms, cell walls of which have biogenic silica deposition [25], 2)
735 contamination with re-suspended sediments, terrestrial runoff, and open air dust, all containing
736 silicate or sand, and/or 3) non-diatom cellular structure that contains silica, whereas the presence
737 of silica in non-diatom algae cell walls has been reported [16, 30]. To some degree, the present
738 study also confirms the reasoning of Nagao et al [36] regarding their observed high ash content
739 of net-plankton samples.

740 There are two main methods for cultivating photoautotrophic algae: open pond systems (mostly
741 raceway ponds) and PBRs [37]. A typical raceway pond comprises a closed loop oval channel,
742 open to the air, and mixed with a paddle wheel to circulate the water. In contrast, the culture
743 medium in PBR is enclosed in a transparent array of tubes or plates and the micro-algal broth is
744 circulated from a central reservoir. Because of relatively low cost and easy adaption for a large
745 scale production, raceway ponds are widely used in commercial production of algal biomass
746 [38]. However, algae grown in open aqueous systems are prone to organic and inorganic
747 contaminants in the water, resulting from exposure of open ponds to rain, wind-blown materials
748 (dust), terrestrial runoff and other debris. Some of the contaminants become sediments while
749 others are dissolved or turned into suspended particles. Upon depositing on or contacting with
750 the growing biomass, they introduce silica into algae. Contamination by unwanted microalgae
751 (particularly diatoms) can also occur easily. Based on findings of the present study, diatom
752 contamination is another common way of introducing silica into algae biomass. Regardless
753 where it comes from, because silica is non-combustible (i.e. the boiling point of silicon dioxide is
754 2230°C), it remains as ash upon combustion. The observation that Algae 4, the only sample that
755 was grown under PBR, had the lowest ash content and almost zero amount of WAIA indicates
756 that PBR systems could greatly reduce contamination of silica-containing organic or inorganic

34
757 materials. It also serves as a piece of evidence to validate the findings of the present study
758 regarding major channels leading to high ash content in algae.

759 Results also show that WADA in the 12 algae samples ranged from 1.8 to 19.0% dm, with a
760 mean of 9.4% dm. It represented 29.6 to 98.6% of total ash, with a mean of 74.2 (Table 5). As
761 discussed earlier, although WADA represented all the ash dissolved in the supernatant after nitic
762 acid digestion, it consisted of two types of minerals: these that could be measured by subsequent
763 ICP-OES analysis (calculated as TM content) and those that could not be measured by the ICP-
764 OES analysis. Regardless, the observation that the 12 algae had higher average content of
765 WADA than WAIA (9.4 vs. 7.3, Table 5) indicates that absorption/deposits of and/or
766 contamination with non-silica compounds was also responsible in part for the high ash content of
767 algae. Algae calcification [28] mentioned earlier serves as an example. Since diatom and sand
768 samples contained 16.8% and 17.0% WADA, respectively, representing 49.0 and 17.8% of total
769 ash in the two samples, respectively, their entries into open pond systems also introduce non-
770 silica minerals in addition to siliceous compounds, and thus further raise the ash content of algae.
771 Additionally, marine water naturally contains sea salts, while waste water contains dissolved and
772 non-dissolved minerals. When algae are grown in marine water or waste water, their ash content
773 is expected to be high. Furthermore, microalgae are grown in very dilute cultures (mass
774 concentration < 1g/L) with densities close to that of the water. This feature plus small size and
775 negatively charged surface of microalgae keep them stable in their dispersed state. Therefore,
776 algae have to be harvested by mechanical, chemical, biological and/or electrical based methods.
777 Among them, chemical coagulation/flocculation is the main approach by which multivalent
778 metal salts, such as alum (potassium aluminum sulfate) and ferric chloride, are typically added.
779 Dissociation of these salts in the culture medium lowers electrostatic repulsion between the
780 negatively charged cell surfaces, enabling formation of algae cell aggregates [39]. Addition of
781 salt-based coagulants/flocculants can increase the ash content of algae as well.

782 From practical points of view, based on findings of the present study, several measures can be
783 used to produce algae with low ash content (or to prevent algae from having high ash content).
784 These include 1) using PBR to grow algae when its relatively higher cost over open pond
785 systems is justified, 2) placing open ponds within a greenhouse to prevent exposure to rain, dust
786 and other debris if a production scale is not too large, 3) building open ponds with concrete or

35
787 lining the soil base with a tarp or plastic material to prevent re-suspension of soil sediments and
788 other geologic particulates, 4) building open ponds in an elevated location to reduce run-off from
789 rain, 5) limiting introduction of silicon mineral (dissolved Si) into the water system to retard
790 growth of diatoms, 6) using fresh water instead of marine water or waste water, 7) using purified
791 cultures as much as possible, and 8) avoiding harvesting methods that involve addition of
792 multivalent salts. It is expected that any of these measures will be more effective than any efforts
793 to remove ash content directly from algae biomass, such as physical pretreatments [11].

794 3.8. Acid insoluble ash

795 Unlike WAIA, acid insoluble ash (AIA) has been known for a long time. Procedures for
796 determining AIA content was available to measure the presence of silica in medicinal plants,
797 such as Chinese herbs [40]. Since the 1960‟s, AIA has been routinely used as a natural internal
798 marker for animal nutrition studies on the basis that it consists of indigestible mineral
799 components, mainly silica [41, 42]. Unlike WAIA which is measured by wet ashing, AIA
800 content is measured gravimetrically after dry ashing a sample in a muffle furnace, boiling of ash
801 in a dilute HCl solution, filtering, washing of the hot hydrolysate, and re-ashing in the furnace
802 [43]. As a part of study on algae ash characterization, further work has been carried out in the
803 author‟s lab on method development for AIA analysis, measurement of AIA content in the same
804 12 algae and 4 non-algae samples, and its relationship to WAIA. Experimental data showed that
805 for the same biological or geologic samples, WAIA and AIA, although measured by the
806 procedures that differed significantly, had the same values of % dry matter. Therefore, both
807 WAIA and AIA were the same siliceous materials present in algae and other samples. They were
808 insoluble in HCl after dry ashing, indigestible by strong acids in vitro or by animals in vivo when
809 present in the original matrix. Results of this follow-up study are covered in a separate report.
810
811 4. Conclusions
812 With 12 algae and 4 non-algae samples, it was found that when using a common wet acid
813 digestion procedure of strong attacks for sample preparation prior to mineral analysis, there was
814 incomplete digestion for most algae samples due to the presence of silica materials. It was
815 proposed that algae contain two types of ash: wet acid indigestible (WAIA) and wet acid
816 digestible. The method to accurately measure WAIA content, which was equivalent to the

36
817 amount of siliceous materials, was developed, based on their digestibility in a wet acid digestion
818 system of strong attack. High correlation between WAIA content and ash content strongly
819 indicates that WAIA was an important contributor of the algae ash component. Microscopic
820 examination of algae samples in three forms (original matrix, ash by dry ashing, and WAIA)
821 showed that WAIA was the best matrix for identifying three types of silica materials: cellular
822 structures of non-diatoms, diatom cell walls and sandy particles of geologic origin. Among them,
823 contamination of diatoms and contamination of sandy particulates were the two major
824 contributors towards high ash content of algae. Other factors included deposits of silica in non-
825 diatom algae cells, and deposition or contamination of minerals that were not siliceous, such as
826 calcium carbonates.
827
828 In short, the present study was among the very few to simultaneously achieve several important
829 objectives, including 1) to document that silica-containing materials are important contributors of
830 the ash component for algae, particularly those of high ash contents, 2) to develop a reliable and
831 easy to adapt methodology for quantitatively measuring silica containing materials in biological
832 (including algae, forage, grain and grain by-product) and geological (sand) materials, 3) to
833 characterize the chemical nature of algae ash, and 4) to provide evidence and explanation on why
834 some algae had high ash content. Therefore, the study basically addressed several important
835 questions raised earlier. These include: 1) why algae have higher ash contents, some extremely
836 so, compared to other biological materials, 2) why algae have a large variation in ash content, 3)
837 do ash components in algae differ from other biological materials, 4) what is the chemical nature
838 of algae ash, and 5) how to prevent ash built-up during algae production. The study also
839 recommends that WAIA content be established as an important quality parameter for algae.
840
841 Acknowledgements
842 The author gratefully acknowledges Mike Woolman, biological science technician of U.S.
843 Department of Agriculture, Agricultural Research Service (USDA-ARS), for his assistance in
844 conducting the experiments, Rick Barrows, Ph.D., retired research fish physiologist of USDA-
845 ARS, and other individuals for helping and providing algae samples, and J. Michael Bonman,
846 Ph.D. of USDA-ARS, for critically reviewing the manuscript and providing valuable comments
847 for improvement

37
848 References
849 [1] O. Pulz, W. Gross, Valuable products from biotechnology of microalgae, Appl. Microbiol.
850 Biotechnol. 65 (2004) 635-648.

851 [2] P. Spolaore, C. Joannis-Cassan, E. Duran, A. Isambert, Commerical applications of

852 microalgae, J. Biosci. Bioeng. 101 (2006) 87-96.

853 [3] S.U. Kadam, B.K. Tiwari, C.P. O‟Donnell, Application of novel extraction technologies for
854 bioactives from marine algae, J. Agric. Food Chem. 61 (2013) 4667−4675.

855 [4] E. Stephens, R. de Nys, I.L. Ross, B. Hankamer, Algae fuels as an alternative to petroleum. J.
856 Pet. Environ. Biotechnol. 4 (4) (2013) 1-7.

857 [5] M.M.R. Fuentes, G.G.A. Fernandez, J.A. S. Perez, J.L.G. Guerrero, Biomass nutrient profiles
858 of the microalga Porphyridium cruentum, Food Chemistry 70 (2000) 345-353.

859 [6] L.M. L. Laurens, E. J. Wolfrum, High-throughput quantitative biochemical characterization

860 of algal biomass by NIR spectroscopy; multiple linear regression and multivariate linear
861 regression analysis, J. Agric. Food Chem. 61 (2013) 12307−12314.

862 [7] L.Yao, J. A. Gerde, S.-L. Lee, T. Wang, K.A. Harrata, Microalgae lipid characterization, J.
863 Agric. Food Chem., 63 (6) (2015) 1773–1787.

864 [8] K. Hampel, Silica, ash and metal content in algae biomass, Ch. 3 in “The characterization of
865 algae grown on nutrient removal systems and evaluation of potential uses for the resulting
866 biomass”. Ph.D. dissertation, Western Michigan University. Kalamazoo, MI. 2013, pp 74-116.

867 [9] R.E Austic, A. Mustafa, B. Jung, S. Gatrell, X. G. Lei. Potential and limitation of a new
868 defatted diatom microalgal biomass in replacing soybean meal and corn in diets for broiler
869 chickens, J. Agric. Food Chem. 61 (30) (2013) 7341–7348.

870 [10] M. Hupa, Ash-related issues in fluidized-bed combustion of biomasses: recent research
871 highlights, Energy Fuels 26 (2011) 4–14.

872 [11] W.T. Chen, J. Ma, Y. Zhang, C. Gai, W. Qian, Physical pretreatment of wastewater algae to
873 reduce ash content and improve thermal decomposition characteristics, Biores. Technol. 169
874 (2014) 816-820.

38
875 [12] X. Hou, X. Yan. Study on the concentration and seasonal variation of inorganic elements in
876 35 species of marine algae, the Science of the Total Environment 222 (1998) 141-156.

877 [13] A.B. Ross, J.M. Jones, M.L. Kubacki, T. Bridgeman, Classification of macroalgae as fuel
878 and its thermochemical behavior, Bioresource Technology 99 (2008) 6494–6504.

879 [14] M. Tuzen, B. Verep, A. O. Ogretmen, M. Soylak, Trace element content in marine algae
880 species from the Black Sea, Turkey, Environ Monit Assess. 151 (2009) 363–368.

881 [15] J.M.M. Adams, A.B. Ross, K. Anastasakis, E.M. Hodgson, J.A. Gallagher, J.M. Jones, I.S.
882 Donnison, Seasonal variation in the chemical composition of the bioenergy feedstock Laminaria
883 digitata for thermochemical conversion, Bioresource Technology 102 (2011) 226-234.

884 [16] D.J. Lane, M. Zevenhoven, P.J. Ashman, P.J. van Eyk, M. Hupa, R. de Nys, D.M. Lewis,
885 Algal biomass: occurance of the main inorganic elements and simulatioin of ash interactions
886 with bed material, Energy fuels 28 (2014) 4622-4632.

887 [17] S.A.; Benson, Holm, P.L. Comparison of inorganics in three low-rank coals, Ind. Eng.
888 Chem. Prod. Res. Dev. 24 (1985) 145–149.

889 [18] A. Pettersson, M. Zevenhoven, B.-M. Steenari, L.-E. Åmand, Application of chemical
890 fractionation methods for characterisation of biofuels, waste derived fuels and CFB co-
891 combustion fly ashes, Fuel 87 (2008) 3183–3193.

892 [19] M. Hoenig, Preparation steps in environmental trace element analysis -- facts and traps,
893 Talanta 54 (2001) 1021–1038.

894 [20] K.A. Anderson, G. Farwell, P. Gibson, B. Ricks, Total recoverable elements in biological,
895 plant, and animal tissue and feed samples. Analytical Sciences Laboratory Standard Methods,
896 SMM.57.070.03, 2015,Regents of the University of Idaho, Moscow, ID, USA.

897 [21] R.M. Twyman, Wet digestion. Reference Module in Chemistry, Molecular Sciences and
898 Chemical Engineering, from Encyclopedia of Analytical Science (2nd Ed.) Elsevier, New York,
899 NY, 2005, pp 147-152.

900 [22] AOAC Official Methods of Analysis, Association of Official Analytical Chemists, (2002)
901 Arlington, VA.

39
902 [23] AOCS Approved Procedure, Am 5-04, Rapid Determination of Oil/Fat Utilizing High
903 Temperature Solvent Extraction, American Oil Chemists‟ Society, Urbana, IL, 2005.

904 [24] L.M. L. Laurens, T. A. Dempster, H.D.T. Jones, E.J. Wolfrum, S. Van Wychen, J.S.P.
905 McAllister, M. Rencenberger, K. J. Parchert, L.M. Gloe, Algal biomass constituent analysis:
906 method uncertainties and investigation of the underlying measuring chemistries, Anal. Chem. 84
907 (2012) 1879−1887.

908 [25] A.M. Vieillard, R.W. Fulweiler, Z. J. Hughes, J.C. Carey. The ebb and flood of Silica:
909 Quantifying dissolved and biogenic silica fluxes from a temperate salt marsh, Estuarine, Coastal
910 and Shelf Science 95 (2011) 415-423.

911 [26] J.N.C. Whyte, Biochemical composition and energy content of six species of phytoplankton
912 used in mariculture of bivalves, Aquaculture 60 (1987) 231-241.

913 [27] R. Saidur, E.A. Abdelaziz, A. Demirbas, M.S. Hossain, S. Mekhilef. A review on biomass
914 as a fuel for boilers, Renewable and Sustainable Energy Reviews 15 (2011) 2262-2289.

915 [28] V. Kerkar, A.G. Untawale, Studies on structure and organization of calcium carbonate
916 deposits in algae, Current Science 68 (1995) 843-845.

917 [29] M.R. Marshall, Ash analysis. Ch. 7. In Food Analysis, S. S. Nielsen (Ed.). Springer, New
918 York, NY. 2014, pp 105-115.

919 [30] B.C. Parker, Occurance of silica in brown and green algae, Can J. Botany 47 (1969) 537-
920 540.

921 [31] M. Totland, I. Jarvis, K. E. Jarvis, An assessment of dissolution techniques for the analysis
922 of geological samples by plasma spectrometry, Chemical Geology 95 (1992) 35-62.

923 [32] P.F. Reay, W.D. Bennett, Determination of amorphous silica and total silica in plant
924 materials, Analytica Chimica Acta. 198 (1987) 145-152.

925 [33] D.J. Conley, An interlaboratory comparison for the measurement of biogenic silica in
926 sediments, Marine Chemistry 63 (1998) 39–48.

927 [34] EPA, Environmental Protection Agency, Method 3050B. Acid digestion of sediments,
928 sludges, and soils, Washington D.C. Revision 2, December 1996.

40
929 [35] E.H. Kennair, B. Railsback, Beach and shoreline sands from around the world: an
930 introduction to beach and shoreline sands. UGA Geology Department website,
931 http://www.gly.uga.edu/railsback/sands/sandsintro.html Accessed 11/2/2016.

932 [36] N.T. Nagao, K.Toda, K. Takaha, K. Hamaska, High ash content in net-plankton samples
933 from shallow coastal water: possible source of error in dry weight measurement of Zooplankton
934 biomass, Journal of Oceanography 57 (2001) 105-107.

935 [37] P.L. Gupta, S-M. Lee, H-J. Choi, A mini review: photobioreactors for large scale algal
936 cultivation, World J Microbiol Biotechnol. 31 (2015) 1409–1417.

937 [38] Y. Chisti, Large-scale production of algal biomass: raceway ponds. Ch. 2, in Algae
938 Biotechnology, Products and Processes. Bux, F and Chisti, Y. (Eds), Springer, Switzerland,
939 2016, pp 21-40.

940 [39] A.I. Barros, A.L. Gonçalves, M. Simões, J.C.M. Pires, Harvesting techniques applied to
941 microalgae: A review, Renewable and Sustainable Energy Reviews 41 (2015) 1489–1500.

942 [40] Anon. Pharmacopoeia of the People‟s Republic of China. The State Pharmacopoeia
943 Commission of People‟s Republic of China, Vol. 1, 2005 Ed. Chemical Industry Press, Beijing.

944 [41] V.S. Shrivastava, S.K. Talapatra, Pasture studies in Uttar Pradesh. II. Use of some natural
945 indicators to determine the plane of nutrition of a grazing animal, Indian J. Dairy Sci. 15 (1962)
946 154-160.

947 [42] J. Sales, G.P.J. Janssens, Acid-insoluble ash as a marker in digestibility studies: a review, J
948 Anim. Feed Sci. 12 (2003) 383-401.

949 [43] J. Van Keulen, B.A. Young, Evaluation of acid-insoluble ash as a natural marker in
950 ruminant digestibility studies, J. Anim. Sci. 44 (1977) 282-287.

951

952

953

954

41