Influence of Scale on Biomass Growth and Nutrient ... - ACS Publications

Subscriber access provided by Gothenburg University Library

Article

Influence of scale on biomass growth and nutrient removal in an algal-bacterial leachate treatment system Kaitlyn D Sniffen, Jacob Ryan Price, Christopher M Sales, and Mira Olson Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b03975 • Publication Date (Web): 20 Oct 2017 Downloaded from http://pubs.acs.org on October 21, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31

Environmental Science & Technology

1

Influence of scale on biomass growth and nutrient removal in an algal-bacterial leachate

2

treatment system

3 4

Keywords: algae, wastewater, leachate, scale-up, biomass growth, nutrient removal

5 6

Authors: Kaitlyn D. Sniffen, Jacob R. Price, Christopher M. Sales, Mira S. Olson*

7

Affiliation of all: Drexel University, 3141 Chestnut St., Philadelphia, PA, 19104

8

Corresponding author*: [email protected]

9 10

Abstract: Data collected from experiments conducted at a flask scale are regularly used as input

11

data for life cycle assessments and techno-economic analyses for predicting the potential

12

productivities of large scale commercial facilities. This study measures and compares nitrogen

13

removal and biomass growth rates in treatment systems that utilize an algae-bacteria consortium

14

to remediate landfill leachate at three scales: small (0.25L), mid (100L) and large (1000L). The

15

mid- and large-scale vessels were run for 52 consecutive weeks as semi-batch reactors under

16

variable environmental conditions. The small-scale experiments were conducted in flasks as

17

batch experiments under controlled environmental conditions. Kolomogov-Smirnov statistical

18

tests, which compare the distributions of entire data sets, were used to determine if the ammonia

19

removal, total nitrogen removal, and biomass growth rates at each scale were statistically

20

different. Results from the Kolmogov-Smirnov comparison indicate that there is a significant

21

difference between all rates determined in the large-scale vessels compared to those in the small-

22

scale vessels. These results suggest that small-scale experiments may not be appropriate as input

23

data in predictive analyses of full scale algal processes. The accumulation of nitrite and nitrate

ACS Paragon Plus Environment


24

within the reactor, observed midway through the experimental process, is attributed to high

25

relative abundances of ammonia- and nitrite-oxidizing bacteria, identified via metagenomic

26

analysis.

27 28 29

Introduction: Algae are frequently investigated for use in nutrient removal from wastewater sources, as

30

the presence of nutrients can cause costly economic and ecological damage to receiving water

31

bodies [1, 2]. While it is most commonly suggested that algae-based nutrient removal be utilized

32

in domestic wastewater treatment, the direct treatment of landfill leachate may also be favorable.

33

Landfill leachate is a high strength liquid waste, containing concentrations of ammonia regularly

34

exceeding 800mgN/L, and is most commonly trucked to and treated at domestic wastewater

35

treatment facilities, costing landfills hundreds of thousands to millions of dollars each year.

36

Treating the leachate directly may reduce these costs to landfills. Few studies have examined the

37

treatment of landfill leachate by algae, and even fewer have done so on a long-term and large

38

scale [3-9].

39

Data from long-term, large scale studies should be used as input for predictive analyses

40

such as life cycle assessments (LCAs) and techno-economic analyses (TEAs) which are used to

41

evaluate if a commercial investment in a technology are environmentally favorable and

42

economically feasible, respectively. However, many of these assessment tools utilize input data

43

generated from small-scale studies, carried out under conditions that are significantly different

44

than those that occur on a commercial scale [10-14]. Studies conducted by Moody et al. (2014)

45

and Quinn et al. (2015) have shown that the wildly variable results from LCAs and TEAs are due

46

to inconsistent input data and system boundaries, which define the processes included in the


Page 2 of 31

Page 3 of 31


47

assessments. These studies emphasize that it is crucial to use input data from studies that mimic

48

the environmental conditions and type of growth vessel of the proposed commercial scale

49

process in these predictive analyses [15, 16].

50

It has repeatedly been shown in studies using algae [15, 17], E.coli [18-21], and yeast

51

[20, 22] that production efficiencies from small-scale studies using microorganisms do not scale

52

proportionally to those occurring on a commercial-scale. For example, changes in vessel

53

characteristics, specifically surface area to volume ratios, [20, 23, 24], mixing dynamics, such as

54

types of mixing used, [20, 24, 25], and environmental conditions (e.g., temperature, light

55

exposure) [26-28], can significantly affect the activity of microorganisms [20, 23, 24]. Despite

56

this discrepancy between results from small-and large-scale studies, the vast majority of LCAs

57

and TEAs for evaluating the overall environmental impact and technical feasibility of large-scale

58

algae production systems use algal growth rates and densities from small-scale studies as input

59

values for their estimation of algal growth technologies [15, 16, 29].

60

The study presented here uses a system that couples algae biomass production with the

61

treatment of landfill leachate. Specifically, this study compares algal growth and nitrogen

62

removal using untreated leachate as a nutrient source in liquid cultures of a mixed algae/bacteria

63

consortium at three different scaled systems: 0.25L, 100L, and 1000L. Algal growth and

64

nitrogen removal rates measured in small-scale (0.25L) systems are statistically compared to

65

rates to observed at mid- (100L) and large-scales (1000L) to determine if results from small-

66

scale experiments are representative of larger-scale systems.

67

The small-scale (0.25L) vessels were operated as 7-day batch experiments [30-32] in an

68

environmental chamber under controlled laboratory conditions. The flask-scale and controlled

69

growth conditions were deliberately chosen as the comparison scale for this study as this



70

experimental set up is similar to those most frequently used to generate input data for LCAs and

71

TEAs. In contrast, the mid- (100L) and large- (1000L) scale vessels were operated on a 7-day

72

semi-batch cycle but under the uncontrolled environmental conditions of a greenhouse. These

73

operating conditions were chosen as the environmental conditions of a full-scale outdoor system

74

will not be controlled.

75 76

Materials and Methods:

77

Nutrient feed and inoculum:

78

The nutrient feed used in this study was raw, untreated landfill leachate collected from the

79

Sandtown Landfill in Felton, DE. Leachate was refrigerated at 2ºC from collection until use.

80

Over the duration of the study, the ammonia content and pH of the leachate ranged from 320 to

81

935 mgN/L and 6.48 to 7.56, respectively. A more detailed analysis of the leachate composition

82

at each collection time point, data provided by the Sandtown Landfill, is described in Table S1 in

83

the Supplemental Information. The algae-bacteria culture used in this study was first collected

84

from a fish pond on the University of Pennsylvania campus. The inoculant was grown in the

85

urban greenhouse using leachate as a nutrient feed.

86 87

Sampling, preparation, and analysis of metagenome – The microbial population’s composition

88

and structure determines a bioreactor’s kinetic behavior, which is commonly observed through

89

chemical analysis. To assist in understanding these behaviors, metagenomic sequencing was

90

applied to a sample collected from one of the growth vessels. Total DNA was extracted from

91

four biological replicates collected prior to the beginning of the experimental period of this

92

study. DNA extraction was carried out immediately after sample collection following the


Page 4 of 31

Page 5 of 31


93

methods described in detail within Price et al 2016 [33]. The four replicate DNA extractions

94

were then assessed for quality and concentration via NanoDrop 2000 and QuBit 2.0 and then

95

pooled and submitted for 2x250 paired-end sequencing on an Illumina HiSeq sequencer. The

96

Metagenomics RAST pipeline [34] was used to analyze the taxonomic and functional

97

composition of the resulting sequences. Raw reads were uploaded to the MG-RAST server and

98

paired-end reads were joined using the join-paired-ends function. Quality control measures were

99

applied including the removal of artificially induced sequencing artifacts [35], the removal of

100

sequences derived from H. sapiens [36], and the removal of low quality sequences (minimum

101

Phred score of 15, and all sequences were trimmed such that they contained a maximum of 5 low

102

quality base calls) [37]. Metagenome identification and quality control statistics are provided in

103

Table S2 in the Supplemental Information.

104 105

Mid & large scale experimental design:

106

Setup – This set of experiments used two 100L Plexiglas aquarium tanks (ATs) and two 1000L

107

raceway ponds (RWPs) from MicroBio Engineering (San Luis Obispo, CA). The 100L ATs and

108

the 1000L RWPs are referred to herein as mid- and large-scale vessels, respectively. The

109

working volumes of the ATs and RWPs were 60L and 600L, respectively. Examples of the mid-

110

and large-scale vessels used in this study can be found in Figure 1 A-B [38]. All vessels were

111

inoculated with an algae-bacteria culture, at the beginning of this study in February 2016, and

112

housed in a greenhouse on the roof of an academic building at Drexel University in Philadelphia,

113

PA [5, 38]. These vessels were exposed to semi-ambient conditions; no additional lighting or

114

heating/cooling was provided. Daylight hours ranged from 9hr:20min to 15hr:1min throughout

115

the year, though the actual amount of light to reach the cultures in the greenhouse was



116

substantially less due to weather and urban obstructions. Temperatures of cultures ranged from

117

7.8 to 41.7°C.

118 119

Operation – The mid and large scale vessels were operated as semi-batch reactors, on a seven –

120

day batch cycle. Initial samples were taken at the beginning, and final samples were taken at the

121

end of each weekly cycle, where ammonia-N, nitrate-N, nitrite-N, and biomass density were

122

measured and analyzed. To prepare the system for the next weekly cycle, the entire volume of

123

the mid-scale AT was pumped into the large-scale RWP. After mixing, one third of the liquid

124

and biomass volume was removed and then replaced with water and leachate. A portion of this

125

mixture was then transferred back into the AT and the next week’s cycle would begin. The

126

volumes of the mid- and large-scale vessels were mixed together in between each week in order

127

to maintain equivalent initial nutrient and biological conditions at the two scales. The volume of

128

leachate added each week was increased slowly throughout the study, ranging from 5 to 40L per

129

week. Exact weekly leachate additions can be found in Figure S1 of the Supplemental

130

Information. These vessels were operated for 52 consecutive weeks, February 2016 – February

131

2017. A more detailed description of this method can be found in Sniffen et al (2017) [38].

132 133

Environmental Monitoring – Water temperature, pH, and dissolved oxygen within the tanks and

134

raceway ponds were collected at five-minute intervals throughout the yearlong study using

135

Neptune System’s APEX controller and probes (Morgan Hill, CA). Weekly maxima, minima,

136

and averages were calculated for each of these parameters.

137 138

Oxygen production analysis – The total suspended solids contained in the weekly samples was


Page 6 of 31

Page 7 of 31


139

made up of live and dead cellular biomass as well as particulates from the leachate additions.

140

The rate of oxygen production was measured to determine the oxygen production activity

141

relative to the total suspended solids density [39, 40]. Oxygen production rates of the biomass

142

samples were measured at the beginning and end of each week. Rates from each sample were

143

measured using 450mm Unisense O2 probes (Aarhus N, Denmark) over 10 minute intervals,

144

which were run in biological triplicates. Each sample was run at 100% and 25% dilution of the

145

initial samples. The 25% dilution samples were found to give a better representative oxygen

146

production rate than samples with high biomass densities which may have experienced

147

significant self-shading (data not shown).

148 149

Mixing dynamics – The flow rate through the raceway pond was determined using the EPA float

150

method [41]. The mixing efficiency of the ponds was determined using a salt tracer test. Five

151

liters of a 250 g/L sodium chloride salt tracer solution was poured into the raceway pond. Water

152

samples were taken at 6 locations around the pond at 10-second intervals. The conductivity of

153

these samples was measured and the concentration of salt tracer was determined using a standard

154

curve.

155 156

Small-scale experimental design:

157

Flask-scale studies were performed using the same algae culture and leachate as those used in the

158

mid- and large- scale vessels. The ranges of nutrient and biomass concentrations used in these

159

small-scale studies mimicked those used in the mid- and large-scale studies. These small-scale

160

studies used 250 mL flasks with 100mL working volume, shaken continuously in an Innova 44

161

environmental chamber (Eppendorf, Hamburg, Germany), under constant 25°C temperatures and



162

continuous light, as depicted in Figure 1 C. Each nutrient and biomass concentration condition

163

was tested using biological triplicates. The duration of all small-scale experiments was seven

164

days. Ammonia-N, nitrate-N, nitrite-N, and biomass density were measured and analyzed at the

165

beginning and end of each of the experiments.

166 167

Analytical methods:

168

Samples collected from experiments run at all scales were analyzed using the following

169

analytical methods: Ammonia-N, nitrate-N, and nitrite-N were measured using Hach test

170

methods 10031, 10020, and 8153, respectively, with a DR2400 spectrophotometer. Removal

171

rates of each nitrogen species along with total dissolved nitrogen were calculated using

172

Equations 1-2. In Eqn. 1-2, CoN and CfN are the initial and final concentrations of the nitrogen

173

species [mgN/L], respectively, where t is time [days] [5], NT is the total inorganic nitrogen

174

concentration [mgN/L], and Ri is the removal rate of species i [mgN/L/day] for each batch

175

period. These equations were developed based on preliminary studies which monitored nitrogen

176

removal and biomass growth every 1 to 2 days during week-long studies. These preliminary

177

studies found that the removal and growth rates were linear throughout the week.

178 179

(Eqn. 1)

NT= CNH3 + C NO2- + C NO3-

(Eqn. 2)

Ri= (CoN – CfN)/t

180 181 182 183

Biomass density was measured at the beginning and end of each week by standard total

184

suspended solid protocol using a 0.45µm filter and 20mL of sample [42]. Biomass density


Page 8 of 31

Page 9 of 31


185

samples were run using biological duplicates. Weekly biomass growth rates were calculated

186

using equation 3, where Co,Biomass and Cf, Biomass are the initial and final biomass concentrations

187

[g/L], RB is the biomass growth rate [g/L/day], and t is time [days].

188 189

(Eqn. 3)

RB= (Cf, Biomass – Co, Biomass)/t

190 191

Statistical analysis – Weekly rates (nutrient removal and biomass growth) and environmental

192

data were pooled for regression and correlation analysis to identify general trends. Weekly

193

nutrient removal and biomass growth rates were then binned by scale for rate comparisons.

194

Kolmogov-Smirnov statistical tests are used to determine if the distributions of two data sets are

195

significantly different. Data binned by scale was compared using Kolmogov-Smirnov tests to

196

determine if the weekly rates calculated at each scale were statistically different. Regressions and

197

correlation data categories included: initial and final weekly ammonia-N, nitrate-N and nitrite-N

198

concentrations and removal rates; initial and final weekly biomass concentrations and growth

199

rate; initial and final weekly oxygen production rates; as well as maximum, minimum, and

200

average pH, water temperature, dissolved oxygen, and number of daylight hours. SPSS 24[43]

201

and Microsoft Excel were used to analyze this data.

202 203

Results and Discussion:

204

Environmental conditions – Seasonal trends of the temperature, pH, and dissolved oxygen (DO)

205

in the tanks and raceway ponds, over the course of the yearlong study can be seen in Figure 2A-

206

C. Over the course of a day, the pH, temperature, and dissolved oxygen in the tanks and raceway

207

ponds rise and fall on a diurnal cycle. This daily trend is shown over a representative two-week



208

span, depicted in Figure 3. As confirmed by observed data, the ATs and RWPs were subjected

209

to identical environmental conditions due to their colocation.

210

(Figure 2A-C)

211

(Figure 3)

212

Mid- and large-scale experimental results – The ranges in initial ammonia, total nitrogen and

213

biomass concentrations over the course of the study were 0.2-161 mgN/L, 8.12- 187.8 mgN/L,

214

and 40-1780 mgBiomass/L, respectively. A breakdown of initial conditions by scale is presented

215

in Table S3 in the Supplemental Information. Leachate input volumes were approximately 20

216

L/week during the first half of this study, while the input volumes during the second half were

217

approximately 30-40L/week. During times of high nitrite concentrations in the systems, no

218

leachate was added. Weekly leachate input volumes during this study can be found in Figure S1

219

of the Supplemental Material.

220 221

The dissolved nitrogen content of the leachate was 99% in the form of ammonia. Ammonia and

222

total nitrogen removal rates along with biomass growth rates varied over the course of the study;

223

the average ± standard deviation of these rates for the large-, mid- and small-scales are presented

224

in Table 1. The ammonia and total nitrogen removal rates for each of the vessels are presented

225

in Figure 4 A-B.

226 227

Table 1: Average ± standard deviation of ammonia removal, total nitrogen removal, and biomass growth rates, at each scale. Large-scale Mid-scale Small-scale Total N Removal Rate 1.63 ± 2.95 1.82 ± 3.97 3.16 ± 2.73 (mgN/L/day) Ammonia Removal Rate 2.33 ± 3.05 2.68 ± 3.90 2.96 ± 2.71 (mgN/L/day) Net Biomass Growth Rate 12 ± 24 4 ± 38 18 ± 21 (mg/L/day)

228


Page 10 of 31

Page 11 of 31

229


(Figure 4 A-B)

230 231

The maximum biomass density measured over the course of this study was 2100 mg/L. Biomass

232

growth rates ranged from -106 to 199 mg/L/day. A negative growth rate means that there was

233

biomass loss over the course of the week. Biomass loss was seen in 70 out 205 weeks’ worth of

234

data from all vessels. Negative biomass growth rates did not statistically favor any scale, season,

235

or environmental condition. Detailed discussion of this point can be found in Sniffen et al. (2017,

236

submitted). When accounting for only weeks with positive growth rates the average and standard

237

deviation was 28 ± 32 mg/L/day. This range of growth rate is similar to rates reported for algae

238

growth fed with domestic wastewater [44-47]. Weekly biomass growth rates for all vessels can

239

be found in Figure S2 of the Supplemental Information.

240 241

Metagenomics results – Alpha diversity within the metagenomic sample was estimated to be 376

242

taxa. Annotations from the NCBI Reference Sequence Database [48-50] and the M5nr protein

243

sequence database [51], acquired via the MG-RAST interface, were used to analyze and interpret

244

the taxonomic composition of the microbial community. The vast majority of reads were

245

attributed to the domains of Bacteria and Eukaryota (Table S4 in the Supplemental Information).

246

Proteobacteria and Bacteroidetes were the first and second most abundant phyla respectively

247

(Table S5 in Supplemental Information). Under the annotations for both reference databases,

248

Chlorophyta and Cyanobacteria comprised about 2% of the annotated reads (Table S5).

249

Chlamydomonas reinhardtii, Volvox carteri, and Scenedesmus obliquus accounted for the bulk

250

of the reads attributed to Chlorophyta (Table S6 in the Supplemental Information). The

251

distribution of reads amongst species falling within the phyla Cyanobacteria was much more



252

uniform at the species level (Table S7 in the Supplemental Information); agglomerating taxa by

253

Genus indicates that the genus Synechococcus accounts for roughly 30% of the reads within

254

Cyanobacteria, followed by Cyanothece spp. at 19%, and Nostoc spp. at 12% within this phyla.

255

AOB and NOB were observed to comprise roughly 12% and 1% of the total microbial

256

population (Table S8 in Supplemental Information). The anammox bacteria Candidatus Kueneia

257

was also discovered in the M5nr annotations. This indicates that anammox may be a potential

258

nitrogen transformation pathway within this system (Table S8).

259

Low abundances for eukaryotes, such as Chlorophyta, within this study, are common in

260

metagenomic analyses, arising from their being underrepresented within reference databases,

261

including those used by MG-RAST [52]. This induces bias towards the identification and

262

annotation of bacterial sequences, while sequences from eukaryotic sources may fail to be

263

annotated. Because of this, the abundances of Chlorophyta reported here should be viewed as a

264

floor or minimum and should not be directly compared to the relative abundance of the kingdom

265

Bacteria, or the taxa therein.

266 267

Statistical analysis – A Pearson’s correlation matrix was calculated for all of the categories listed

268

in the statistical methods section. Strong, statistically significant correlations were found

269

between: initial ammonia concentration and ammonia removal rate (r=0.854, p

Influence of Scale on Biomass Growth and Nutrient ... - ACS Publications

Recommend Documents