Oil-based polyurethane-coated urea reduces nitrous oxide emissions in a corn field in a Maryland loamy sand soil

Journal Pre-proof Oil-based polyurethane-coated urea reduces nitrous oxide emissions in a corn field in a Maryland loamy sand soil Ricardo Bortoletto-Santos, Michel A. Cavigelli, Sheila E. Montes, Harry H. Schomberg, Anh Le, Alondra I. Thompson, Matthew Kramer, Wagner L. Polito, Caue Ribeiro PII:

S0959-6526(19)34199-X

DOI:

https://doi.org/10.1016/j.jclepro.2019.119329

Reference:

JCLP 119329

To appear in:

Journal of Cleaner Production

Received Date: 24 July 2019 Revised Date:

1 November 2019

Accepted Date: 14 November 2019

Please cite this article as: Bortoletto-Santos R, Cavigelli MA, Montes SE, Schomberg HH, Le A, Thompson AI, Kramer M, Polito WL, Ribeiro C, Oil-based polyurethane-coated urea reduces nitrous oxide emissions in a corn field in a Maryland loamy sand soil, Journal of Cleaner Production (2019), doi: https://doi.org/10.1016/j.jclepro.2019.119329. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Ltd.

1 1

Title: Oil-Based Polyurethane-Coated Urea Reduces Nitrous Oxide Emissions in a Corn

2

Field in a Maryland Loamy Sand Soil

3 4

Authors: Ricardo Bortoletto-Santos1; Michel A. Cavigelli2; Sheila E. Montes2; Harry H.

5

Schomberg2; Anh Le2; Alondra I. Thompson2; Matthew Kramer2; Wagner L. Polito3; Caue

6

Ribeiro1*

7 8

1

9

Instrumentação, Rua 15 de Novembro 1452, Centro, São Carlos, São Paulo, 13561-206, Brazil.

Laboratório

Nacional

de

Nanotecnologia

para

o

Agronegócio

(LNNA),

Embrapa

10

2

11

Agricultural Research Center, 10300 Baltimore Ave., Beltsville, Maryland, 20705, United States.

12

3

13

São Carlos, Av. Trabalhador São-Carlense, 400, Arnold Schimidt, São Carlos, 13566-590, SP,

14

Brazil.

USDA-ARS, Sustainable Agricultural Systems Laboratory, Henry A. Wallace Beltsville

Instituto de Química de São Carlos (IQSC), Universidade de São Paulo (USP), Campus de

15 16 17 18 19 20 21 Corresponding Author:

28 Dr. Michel A. Cavigelli

22 Dr. Caue Ribeiro

29 United States Department of Agriculture

23 Embrapa Instrumentação

30 Beltsville Agricultural Research Center

24 Rua XV de Novembro, 1452

31 10300 Baltimore Ave.

25 São Carlos, SP, 13560-970 - Brazil

32 Beltsville, MD, 20705 - USA

26 e-mail: [email protected]

33 email: [email protected]

27 Phone: +55 (16) 2107-2915

34 Phone: +1 (301) 504-8327

35

2 36

ABSTRACT

37 38

Urea is the most widely used N fertilizer due to its high N concentration (46%), cost

39

effectiveness and ease of handling. However, urea is susceptible to loss as N2O, a greenhouse

40

gas and catalyst of stratospheric ozone decline. Polymer-coated fertilizers may be effective in

41

reducing such losses but the appropriate coating thickness for effective field performance is

42

unknown. We prepared urea granules with polyurethane coating (PCU) based on castor oil at

43

2% to 8% by weight. We tested the fertilizer value of these materials for maize (Zea mays L.)

44

and measured field loss of N2O. Maize grain yield and N uptake were similar in treatments

45

fertilized with urea (uncoated) and each of the coated materials. Cumulative N2O-N per unit of

46

grain yield, however, was reduced by 80% with PCU8% compared with uncoated urea. Results

47

indicate that polyurethane-coated urea performs similarly to uncoated urea for maize production

48

while reducing soil N2O emissions up to 60-80%, with an efficiency factor twice as high as that

49

suggested by the Intergovernmental Panel on Climate Change (IPCC) for N fertilizers. Our

50

results indicate that urea with a polyurethane coating of 8% (PCU8%) had the best combined

51 52 53 54

agronomic and environmental performance in a field study. Keywords: N2O emission; maize; urea; castor oil; hydrophilic polymers.

3 55

1. INTRODUCTION

56

Fertilizers, including those containing nitrogen (N), have been fundamental to increasing

57

agricultural productivity per unit of land. However, fertilizer N can be lost from agricultural soils

58

due to many N transformation processes occurring in soil (Robertson and Vitousek 2009;

59

Trenkel 2010; Silva et al. 2017; Mariano et al. 2019). Fertilizer N losses reflect asynchrony

60

between N application and/or release and crop N uptake. Important loss mechanisms include

61

nitrate leaching, runoff and erosion, ammonia (NH3) volatilization, and gas emissions such as

62

dinitrogen (N2), nitric oxide (NO) and nitrous oxide (N2O) (Ruser et al. 2006; Firestone and

63

Davidson 1989; Raymond et al. 2016; Cancellier et al. 2016; Mariano et al. 2019). Notably, N2O,

64

which is produced largely during nitrification and denitrification in agricultural soils, has a direct

65

influence on the greenhouse effect, since it has a global warming potential 298 times higher

66

than that of CO2 (Forster 2007; Hyatt et al. 2010; Díaz-Rojas et al. 2014). It also serves as a

67

catalyst of stratospheric ozone degradation and its atmospheric concentration is increasing

68

(Ravishankara 2009; Parkin and Hatfield 2014). Agricultural soils are the primary source of N2O

69

and in the United States account for about 79% of anthropogenic N2O emissions (USEPA

70

2016).

71

Since N losses from agricultural soils impart substantial economic and environmental

72

costs, scientists have long sought ways to reduce the rate of N release from urea to better

73

match crop N uptake patterns (Azzem et al. 2014; Burton et al. 2008). While urea is not taken up

74

by crops, urea readily mineralizes to HCO3 and NH4+ in soils. Ammonium is then readily nitrified

75

to NO3- by soil nitrifying bacteria. Since plants take up N from soil in the form of NH4+ and NO3-,

76

the release of N in these forms is a critical step following application of urea to soil. Ammonium

77

and NO3-, however, also serve as substrates for N2O production by soil nitrifying and denitrifying

78

bacteria. Thus, the temporal pattern of urea release is critical to many subsequent N

79

transformations in soil.

80

One common approach to reducing the rate of N release from urea is to apply N fertilizers

81

as close to the time of high crop N demand as possible. For maize, this involves applying some

82

N fertilizer at planting with the bulk applied when maize is about 30 cm tall, just prior to peak

83

demand. This approach, “sidedressing,” has been shown to improve maize N use efficiency

84

(Keeney 1982; Phillips et al. 2009; Zebarth et al. 2008; Eagle et al. 2017) but it requires a

85

second application of fertilizer during a relatively narrow window of opportunity, which increases

86

application costs while increasing risk that timely application could be jeopardized if soil

87

conditions are not conducive to application at the optimal time. Another approach to better N

88

synchronicity is to modify urea diffusion kinetics with surface barriers, i.e., coatings, allowing a

4 89

single fertilizer application at planting. Polymer-coated urea, in particular, has received

90

considerable attention, reflected by the significant number of available products (Azeem et al.

91

2014; Shaviv 2000; Detrick and Hargrove 2002; Whittington 2005; Coogan and Damery 2003;

92

Ogle and Sims 2010; Chen et al. 2011; Halvorson et al. 2010; Chen et al. 2018).

93

However, few of these products are stable in warm environments, since most of them

94

(especially those based on polyacrylate polymers) control urea release via slow solubilization of

95

the polymer in water in a very uncontrolled manner, resulting in release that is often not different

96

than for uncoated fertilizers. In addition, many coated urea granules, including commercial

97

products, are reported to have a typical N content of 35 to 36 %, which indicates a material

98

containing about 20 % coating by weight (Trenkel 2010). Due to these limitations, we have

99

developed a bio-based hydrophobic polyurethane (PU) coating from vegetable oils that is

100

biodegradable and mechanically resistant. Coated granules with 7 to 9.5 % by weight (wt%) PU

101

show urea release below 50 % in water after 60 days (Bortoletto-Santos et al. 2016). Despite

102

the promising laboratory results, the effectiveness of these materials in improving nitrogen use

103

efficiency (NUE) and decreasing N losses in the field has not been tested.

104

Therefore, we evaluated these PU coated materials in the field to evaluate their impact on

105

maize N uptake and N2O emissions. Our hypothesis is that urea coated with bio-based

106

polyurethane (PU) derived from vegetable oils will substantially reduce N2O emissions while

107

maintaining maize grain yield since N release from the coated urea applied at planting will be

108

more temporally synchronous with maize N demand than uncoated urea applied at planting.

109 110

2. MATERIALS AND METHODS

111 112

2.1. Preparation of Fertilizer Materials

113 114

Urea granules (ranging between 2.8 and 3.2 mm in diameter) were coated with a

115

polyurethane resin system based on castor oil (referred to here as polyurethane coated urea,

116

PCU). The polyurethane formation consisted of a condensation reaction between castor oil

117

(generously provided by A. Azevedo Óleos, São Paulo, Brazil) and 4,4'-diphenylmethane

118

diisocyanate (MDI – Desmodur, Bayer). Castor oil was mixed with elemental sulfur (8.0 % by

119

weight) and then it was reacted with MDI in a ratio of 40:60 by mass (MDI:oil). The use of

120

elemental sulfur (Synth) in the polymer formation improved the coating’s filmogenic capacity and

121

provided better adhesiveness and flexibility to the coating.

5 122

The coating process was prepared from the dispersion of castor oil + elemental sulfur and

123

MDI over the granules, using a metal turntable coater rotating at 30 rpm, with 25 cm side shields

124

and air flow heated at 70 to 80 °C. Urea granules were prepared with polymer coatings at mass

125

ratios of 2 to 8 % by weight (in increments of 2 wt%), which did not alter N concentrations

126

substantially (Table 1). For example, 20 grams of polyurethane were applied to 1 kg of urea, in

127

the process employing 2 % of polymer.

128 129

2.2. Physical and Chemical Characteristics of the Coated Urea Materials

130 131

2.2.1. Elemental Analysis (CHN). Total nitrogen content of the materials (uncoated and coated

132

granules) was determined by elemental analysis using the CHNS / O 2400 Series II Elemental

133

Analyzer (PerkinElmer).

134 135

2.2.2. X-ray Microtomography. Analysis of the morphology of the urea and the coatings was

136

performed by X-ray microtomography (SkyScan 1172, Bruker). The images were acquired using

137

the following conditions: 3.94 µm spatial resolution (voxel size), 0.2° rotation step, 180° rotation,

138

and averaging of 10 frames. The reconstruction of the tomographic images was performed

139

using NRecon SkyScan software.

140 141

2.2.3. Water-Release Assay. A water-release assay was used to evaluate the efficacy of the

142

coating system, quantifying the urea release rate as a function of time at room temperature. The

143

assay consisted of immersing 0.5 g of urea (coated and uncoated) in a 250 mL beaker, under

144

constant stirring to promote homogeneous release of urea. In addition, 0.5 mL aliquots were

145

taken periodically: every 24 h over 10 days, then every 48 h until day 20, and every 120 h until

146

day 50. These aliquots were used to determine the concentration of urea released as described

147

in section 2.2.4. Uncoated urea was included as a control treatment. All assays were conducted

148

simultaneously to ensure identical laboratory conditions.

149 150

2.2.4. Urea Determination. Urea concentration in water was determined with a Shimadzu UV-

151

1601 PC using an adaptation of the methodology of With et al. (1961) and Bortoletto-Santos et

152

al. (2016). A sample (0.5 mL) of the solution was mixed with 2.5 mL of a 10 % trichloroacetic

153

acid solution (Sigma-Aldrich) and 0.5 mL of Ehrlich reagent (0.36 mol L-1 dimethylbenzaldehyde

154

in 2.4 mol L-1 hydrochloric acid) and absorbance was obtained in the 400–500 nm range.

155

6 156

2.3. Field Experiment

157 158

2.3.1. Experimental Site. A field experiment was established to evaluate PCU effects on maize

159

N uptake and soil N2O emissions at the Beltsville Agricultural Research Center (BARC), a

160

campus of the United States Department of Agriculture Agricultural Research Center (USDA-

161

ARS) in Beltsville, Maryland (39° 01' 54.4'' N, 76° 56' 11.6'' W). Properties of the soil, a Downer

162

loamy sand (Coarse-loamy, siliceous, semiactive, mesic Typic Hapludults), are provided in the

163

Supplementary Information (Table S1).

164 165

2.3.2. Field Management. The field was chisel plowed to a depth of 20 cm followed by one pass

166

of a disk, a field cultivator and a packer. Corn (Pioneer P9675AMXT) was planted 15 July, 2017

167

using a John Deere 7200 planter at 69,160 plants ha-1 in rows spaced 76 cm apart. Before

168

planting, potassium fertilizer, based on soil test results, was applied at a rate of 168 kg K2O ha-1

169

(based on soil test results for soil samples collected February 2017). A combination of pre-

170

emergent herbicides typical for the region was applied just prior to planting at the following

171

rates: atrazine (0.34 kg a.i. ha-1; 6-chloro-N-ethyl-N’-(1-methylethyl)-1,3,5-triazine-2,4-diamine),

172

simazine (1.22 kg a.i. ha-1; 6-chloro-N, N’-diethyl-1,3,5-triazine-2,4-diamine), S-metolachlor

173

(1.42 kg a.i. ha-1; 2-chloro-N-(2-ethyl-6-methylphenyl)-N-(2-methoxy-1-methylethyl) acetamide)

174

and mesotrione (0.186 kg ha-1; 2-(4-Mesyl-2-nitrobenzoyl)-1,3-cyclohexanedione).

175

Each fertilizer treatment was randomly assigned to one plot in each of four blocks.

176

Nitrogen fertilizer application rate (160 kg-N ha-1) was typical for the region to achieve

177

economically optimum grain yield (Coale 2010). The amount of material applied for each

178

fertilizer type was calculated based on the total N content of each material (Table 1). Fertilizer

179

materials were applied 18 July by hand in trenches (10 cm wide x 2 cm deep) dug using hand

180

hoes; trenches were parallel to and about 38 cm from each of two adjacent maize rows in each

181

plot. Fertilizers were distributed evenly along the entire 3 m of each of three trenches per plot

182

and then immediately covered with the disturbed soil to minimize any potential volatilization of

183

ammonia. In the no fertilizer control treatment, trenches were dug and then covered with soil but

184

no fertilizer materials were applied. In addition, another treatment (V6) was evaluated where

185

uncoated urea was applied 33 days after planting (18 July) when the maize was at the V6

186

growth stage. After maize harvest (Nov. 21), a cereal rye (Secale cereale L.) cover crop was

187

planted in rows spaced 19 cm apart at 125 kg seed ha-1 using a no-till drill.

188

7 189

2.3.3. Nitrous Oxide Measurements. Nitrous oxide measurements were based on the

190

methodology described by Parkin and Venterea (2010). Rectangular aluminum anchors (internal

191

dimension 0.640 m x 0.328 m; 0.210 m2 each) were installed about one meter from the end of

192

each plot between the two middle rows of maize, with the long side perpendicular to the maize

193

rows, as shown in Figure 1a. Anchors were hammered into the soil to a depth of about 10 cm on

194

18 July, 2017.

195

Each anchor (illustrated in Fig 1b) was fitted with a gutter along its outer perimeter. At

196

the time of sampling, the gutters were filled with water and a stainless-steel lid, fitted with a

197

sampling port and a vent, was seated in the gutter. The water formed a gas tight seal during

198

sampling. Prior to placing the lids on the anchors, a soil thermometer was placed on the soil

199

surface inside the chamber and the air temperature was recorded. Gas samples were taken

200

through the sampling port using a 10 mL syringe fitted with a 22 g, 2.54 cm needle. Samples

201

were injected into pre-labeled 12 mL vials (Labco Exetainer, Lampeter, UK) with butyl rubber

202

septa that had previously been flushed with 99.9% N2. Ten mL gas samples were withdrawn

203

from each chamber at 0, 7, 14 and 21 mins. Lids were removed immediately after sampling and

204

the chamber internal air temperature was recorded using the previously placed thermometer.

205

Gas samples were taken after each rainfall event of approximately 10 mm or more, usually for

206

three consecutive days (46 total sampling events during the 106 day maize growing season).

207

Samples were usually taken between 9 and 11 in the morning when soil and air temperature are

208

near mean daily values (Alves et al 2012).

209

The concentration of N2O in each gas sample was determined by gas chromatography

210

(VARIAN 450 gas chromatograph equipped with an electron capture detector - temperature 300

211

o

212

column. An auto sampler (Combi PAL System) injected 5 mL samples from each vial into the

213

GC. Nitrogen was used as the carrier gas and CH4/Ar as auxiliary gas, both with 30 mL min-1

214

flow rates.

C), with split / splitless injector (temperature 120 oC) and silica capillary Fused Poropak QS

215

Nitrous oxide concentrations in the vials were converted to N2O concentrations in the

216

anchor headspace using standard curves and the ideal gas law, corrected for mean

217

temperature in the anchor at sampling time (temperature after sampling minus temperature prior

218

to sampling) and for the dilution effect of injecting headspace samples into a preflushed vial.

219

Standard curves were constructed by preparing, in duplicate, vials with at least five different

220

N2O concentrations in the same types of vials used for samples by adding known volumes of

221

N2O from a Certified Standard tank (2.295 ppm N2O) to vials previously flushed with UHP N2.

222

Ten standards vials were analyzed on the GC for every 40 samples analyzed to ensure that

8 223

concentrations of the standards bracketed the concentrations of the samples. Anchor volumes

224

were calculated using the mean of ten measured values of the height above the soil surface for

225

each anchor. Chamber volumes were calculated as the sum of the volume defined by the lid

226

and that of the anchor.

227

Linear regression of N2O concentration over time (assessed using the coefficient of

228

determination, R2) was used to calculate the emission rate for each chamber for each sampling

229

day. In addition, linear interpolation between N2O emissions values on subsequent sampling

230

days was used to calculate N2O emissions for dates when samples were not collected. Daily

231

emissions (measured and interpolated) were summed for the 106 d sampling period.

232

Volumetric soil water content (m3H2O m-3soil) and soil temperature (°C) were measured

233

within one meter of each gas sampling anchor for the 0-12 cm depth using Campbell Scientific

234

CS655 sensors (Campbell Scientific, Inc., Logan, UT) installed vertically in the soil. The CS655

235

is a quasi-TDR sensor that generates and transmits an electromagnetic pulse along two 12 cm

236

long stainless-steel rods (wave guides) into the soil. The time required for the wave to return to

237

the probe oscillator correlates to the surrounding medium’s dielectric permittivity. The

238

permittivity is calculated from the travel time and inserted into Topp’s equation to calculate

239

volumetric water content (Topp et al. 1980; Campbell Scientific 2015). Data were collected

240

using a Campbell Scientific CR1000 data logger. Sensor measurements were made every five

241

minutes and averaged every ten minutes to provide detailed information about changes in soil

242

water content. Data were periodically downloaded directly from the data logger. Soil moisture

243

and temperature data were used to help interpret soil N2O emissions.

244 245

2.3.4. Determination of ammonium (NH4+-N) and nitrate (NO3--N) in soil. Release of urea in soil

246

was estimated by analyzing soil extracts for NH4+-N, the product of urea mineralization, and

247

NO3--N, the product of NH4+ nitrification. These processes occur rapidly in warm soils and plants

248

take up N as NH4+ or NO3-. Soil samples were taken in each plot at two different depths (0-15

249

cm and 15-30 cm), using a 19 cm diameter probe. Four cores were taken at each depth within

250

about 15 cm of the maize row in each plot. The four cores at a given depth were combined to

251

form one sample on each of nine sampling days (2, 10, 17, 24, 30 August; 11, 25 September;

252

and 13 and 27 October).

253

The soil samples were air-dried, ground to pass a 2-mm sieve and stored in pre-labeled

254

bags. After drying, 3 g soil were weighed and 30 mL of 1 M KCl were added. The samples were

255

shaken for 1 h at 200 rpm and allowed to settle for 30 mins. Then the extracts were filtered

256

using a Millipore Filter where a blank (KCl only) and a standard containing QAQC soil were also

9 257

prepared. Filtrate of the KCl extraction was analyzed for inorganic N (NO3- + NH4+)-N

258

colorimetrically using a SEAL Auto-Analyzer 3 (Mequon, WI). Concentration of inorganic N in

259

extracts was adjusted to reflect the weight of soil. Soil inorganic N results are presented in the

260

supplementary materials.

261 262

2.3.5. Statistical Analyses. Response variables were analyzed using R (R Core Team 2018).

263

The variables were first Box-Cox transformed to stabilize variances across each variable’s

264

range. For all variables except Yield (λ = 1.6), N2O emissions factor (λ = 0.3), and N uptake (λ =

265

0.1), the log (λ = 0) transformation was chosen. Because of the physical layout of treatments in

266

contiguous blocks, where treatments in two different blocks could be physically closer than two

267

treatments from the same block, we decided that an analysis using spatial position and soil

268

characteristics was more appropriate than that commonly used for a randomized complete

269

block. We thus started with the transformed data and a full model (treatment, linear and

270

quadratic spatial trends and their interaction, and soil composition (proportion clay and

271

proportion sand).

272

resulted in a final model for each variable. For each variable, p-values for pairwise comparisons

273

of treatments were adjusted using Tukey’s HSD, with α = 0.02 (since a repeated measures

274

adjustment [six dependent variables] also needs to be made, setting α = 0.02 gives the false

275

discovery adjustment).

A step-wise selection on all explanatory variables other than treatment

276 277

3. RESULTS AND DISCUSSION

278 279

Figure 2 shows that the polymeric coating exhibits good cohesion and uniformity on the

280

surface of the urea granule. Also, the contact region between the coating and the fertilizer

281

shows good adhesion (or interaction) between the materials. The SEM images (Fig 2a to 2e)

282

reveal that the coating has a variable thickness between 25 µm (PCU2% and PCU4%) and 50

283

µm (PCU6% and PCU8%).

284

Figure 3 shows the urea release curves in water for the various PCUs and the uncoated

285

urea. It is evident that the coating thickness is directly correlated to the release profile (the

286

release rate of the fertilizer) in that there is a positive relationship between coating thickness

287

and urea release times. While materials with thicker coatings released less than 100% of urea

288

after 40 days (e.g. 80% urea released in 42.5 d for PCU6%), urea release for PCU2% was

289

similar to that for uncoated urea, reaching ~90% release in less than 2 days. This indicates that

290

a thin polymer coating is deficient, resulting in poor controlled release performance.

10 291

Under field conditions, temporal N2O emissions patterns, shown in Figure 4, were also

292

impacted by the thickness of the polyurethane coating. Higher N2O emissions were observed for

293

U, PCU2% and PCU4% compared to the other materials soon after the application of the

294

fertilizers in the soil (0 to ~40 days after planting). For example, N2O emissions from 8 to 18 d

295

after fertilizer materials were applied (18 to 28 July) were 680 ± 101; 620 ± 78 and 466 ± 44 g

296

ha-1, respectively, for U, PCU2% and PCU4% (Table 2). Nitrous oxide emissions in the PCU6%

297

and PCU8% treatments were substantially lower (266 ± 61 and 177 ± 8 g ha-1, respectively;

298

Table 2) and not different than the two control treatments (SC and V6 Table 2 and Figure 4).

299

Since soil moisture patterns were very similar among all treatments, we present mean

300

values for all plots in Figure 5. Figures 4 and 5 together clearly illustrate that highest N2O

301

emissions coincide with periods of high soil moisture during the early maize growing season (i.e.

302

following rainfall events). N2O is produced by both nitrifying and denitrifying bacteria under low

303

oxygen conditions (Davidson et al. 1991), which occur during periods with high soil moisture

304

because the rate of oxygen consumption by soil organisms is greater than the rate of oxygen

305

diffusion into soil from the atmosphere (Davidson et al. 1991). However, it is also evident from

306

Figures 4 and 5 that soil moisture alone is not a sufficient condition leading to N2O emissions as

307

emissions are very low in all treatments (except V6) between 45 and 75 days after planting,

308

when soil moisture was relatively high on numerous occasions. Both nitrifying and denitrifying

309

bacteria also require the presence of N substrates — NH4+ for nitrification and NO3- for

310

denitrification (Davidson et al. 1991). Conditions of low soil oxygen and high mineral N occurred

311

in the Urea, PCU2% and PCU4% treatments during the early growing season. However, maize

312

N uptake rates increase considerably about 30 days after planting (Hanway 1996) such that soil

313

mineral N is usually very low for a number of months, which results in low N2O emissions

314

regardless of soil moisture conditions. The V6 treatment, however, shows a substantially

315

different N2O emission pattern with no substantial peaks until the relatively wet period beginning

316

29 August when mineral N would have been substantially greater than in the other treatments

317

since urea in this treatment was applied on 17 August. Although sidedressing is a strategy to

318

increase temporal synchrony between N application timing and crop N uptake, N2O emissions

319

can still occur when soils are wetted soon after urea application when much of the mineral N

320

released from urea is not yet taken up by the crop. In general, it is interesting to note that high

321

N2O emissions continued after each wetting event in almost all treatments even as soil moisture

322

declined rapidly in this sandy soil. The reason for this pattern is unclear.

323

To compare N2O emissions results statistically, we divided the 106 d experimental period

324

into six time periods defined by both soil moisture and N2O emissions patterns, as illustrated in

11 325

Figure 5 and defined in Table 2. Days that represent the beginning and end of each period had

326

low soil moisture and N2O emissions. Table 2 shows that during both the first and second

327

periods N2O emissions were greatest in the Urea, PCU2% and PCU4% treatments,

328

intermediate in the PCU6% treatment and lowest in the PCU8%, SC and V6 treatments. These

329

results reflect patterns described above in that the materials with the fastest urea release rates

330

produced the most N2O early in the season, suggesting that N release occurred before maize N

331

uptake was substantial. The more thinly coated urea products (PCU2%, PCU4%) did, however,

332

have significantly or near significantly lower N2O emissions than uncoated urea during the third

333

time period.

334

N2O emissions in the PCU6% and PCU8% treatments where substantially lower than for

335

Urea through the third sampling period, suggesting that N release was substantially reduced

336

prior to maize taking up substantial N. N2O emissions for PCU6% were similar to that for Urea

337

during period 4 and N2O emissions were greater for both PCU6% and PCU8% than all

338

treatments other than V6 during periods 5 and 6; however, absolute values of emissions during

339

these two periods were relatively low.

340

Cumulative N2O emissions during the maize growing season show a general pattern of

341

reduced N2O emissions with increasing thickness of the polyurethane coating. Emissions for

342

PCU2% and PCU6% were reduced by almost 50% while emissions for PCU8% were reduced

343

by about 70% compared to uncoated urea. The N2O emission factor (EF) for urea, which was

344

calculated after subtracting N2O-N emissions from the SC treatment from all other treatments

345

was 2.2, indicating that 2.2% of the N applied in the fertilizer was emitted as N2O-N. This is an

346

EF twice as high as that suggested by the Intergovernmental Panel on Climate Change (IPCC)

347

for N fertilizers but within the IPCC uncertainty range of 0.3-3% (Eggleston et al. 2006). PCU2%

348

and PCU6% reduced the EF by about half while PCU4% reduced the EF by only 29%. PCU8%,

349

however, reduced the EF by 82%. Thus, all PCU products except PCU4% reduced the N2O EF

350

by more than the V6 treatment, which reduced the EF by 35%. These results are consistent with

351

the water release curves presented earlier in that a thicker coating resulted in slower release of

352

urea which would limit excess soil mineral N and thereby decrease transformation to N2O. The

353

incongruous results for PCU4% might reflect an artifact created during the coating process;

354

nonetheless, the overall impact of coating thickness delaying and reducing N2O emissions is

355

consistent.

356

While these PCU materials, particularly PCU8%, were very effective at reducing soil N2O

357

emissions, this is a benefit only if they are also effective N fertilizers. Our maize yield, N uptake

358

and N recovery data show that all four PCU products were as effective as urea in supplying N.

12 359

Mean maize grain yield, N uptake and N recovery were the same for all PCU materials as for

360

Urea (Table 3). This indicates that the slower release of N from the coated ureas compared to

361

uncoated urea suggested by the N2O data did not hamper N uptake by maize, i.e. the N release

362

by these materials was at least as synchronous with maize N uptake as urea. The substantially

363

reduced N2O emissions and similar maize performance for PCU8% treatment indicate

364

substantially improved synchrony between N release and maize N uptake. This relationship can

365

be quantified by expressing N2O emission relative to grain yield as presented in Table 3. While

366

Urea produced almost 0.5 g N2O-N per kg maize grain yield, PCU6% and PCU8% reduced this

367

value by about 50% and 70%, respectively. The value for PCU8% was no different than that for

368

the SC treatment, indicating that its N2O emission from fertilizer is probably negligible. The

369

PCU2%, PCU4% and V6 treatments were not significantly different than the Urea treatment, but

370

the lower values suggest a trend confirmed by PCU6% and PCU8%.

371

Although sidedress N application can improve maize N use efficiency (Keeney 1982) and

372

reduce the production of N2O per kg of grain, according to an hierarchical model meta-analysis

373

of studies comparing sidedress and preplant application of nitrogen fertilizers (Eagle et al.

374

2017), we did not find this to be the case. Other individual studies have also shown no

375

consistent effect of sidedressing N on N2O emissions (Phillips et al. 2009; Zebarth et al. 2008).

376

In our study, the use of coated fertilizers outperformed sidedressing urea, providing an

377

alternative management option that could eliminate the need for two N applications in one year

378

for maize, thereby reducing fuel use and farmer labor. However, these results, from a year with

379

relatively high precipitation, need to be corroborated with data collected during additional years.

380

One concern about slow release fertilizers is that N release might occur after N uptake by

381

the cash crop, thus leading to potential N losses after the maize growing season. To test the

382

availability of soil N after maize harvest and into the spring of 2018, we planted rye after maize

383

harvest to assess soil N availability. Biomass of rye + weeds was lowest in SC and similar in all

384

other treatments except PCU8%, which had at least 40 % more biomass than the other

385

treatments (data not shown). Nitrogen uptake by rye + weeds, however, showed no statistical

386

differences among any treatments. These results suggest that N availability was similar in all

387

treatments after the maize growing season. However, since we were not able to quantify N2O

388

emissions during this period, it is uncertain whether any potential delayed release of urea from

389

the coated materials may have contributed to N2O emissions during the winter and spring. In

390

addition, an important caveat to our conclusions is that we did not measure NO3--N leaching in

391

this experiment. It is possible that leaching differed among treatments; that information will

392

provide a more complete picture of the sustainability of these materials. In addition, our

13 393

promising results indicate a need to broaden the conditions under which these PCUs are tested.

394

Additional work is needed under different soil types, climates and weather conditions. In

395

addition, since NH3 volatilization following urea application is not eliminated by subsurface

396

application of urea (Rochette et al. 2013), further studies investigating the effects of these PCUs

397

on ammonia volatilization are warranted.

398 399

4. CONCLUSION

400

Our results clearly show that these PCU materials can substantially reduce soil N2O

401

emissions without impacting maize grain yield, N uptake and N recovery. The importance of

402

these results is reflected in the widespread production of maize, a crop with high N demand.

403

While these results are specific to sandy soils in the mid-Atlantic region of the USA, additional

404

research is needed to test these materials on different soil types and climate regimes and over

405

longer time periods.

406 407

AUTHOR INFORMATION

408

Corresponding Author

409

Dr. Caue Ribeiro

410

Embrapa Instrumentação (CNPDIA), Rua XV de Novembro 1452, São Carlos, SP, 13560-970,

411

Brazil.

412

Phone: +55 16 2107 2915; email: [email protected]

413

ORCID - Caue Ribeiro: 0000-0002-8908-6343

414 415

Dr. Michel A. Cavigelli

416

United States Department of Agriculture—Agricultural Research Service,Beltsville Agricultural

417

Research Center, 10300 Baltimore Ave., Beltsville, MD, 20705 – USA.

418

Phone: +1 (301) 504-8327; email: [email protected]

419 420

14 421

ACKNOWLEDGMENTS

422

The authors are thankful for financial support given by CAPES (Finance Code 001,

423

grant#

424

(#458.763/2014-4), SISNANO/MCTI, FINEP and Embrapa AgroNano research network. This

425

publication is based upon work supported by the United States Department of Agriculture,

426

Agricultural Research Service. In addition, we gratefully acknowledge help from Chris Rasmann,

427

Amy Tran, Gwen Bagley, and Briana Otte who helped collect data and/or provide critical

428

logistical support. Trade and company names are given for the reader’s benefit and do not imply

429 430 431

endorsement or preferential treatment of any product by the author affiliations.

88881.131771/2016-01

and

CAPES-Embrapa

Program),

FAPESP,

CNPq

DECLARATION ON CONFLICT OF INTEREST

432 433 434

REFERENCES

435

Alves, B.J.R., Smith, K.A., Flores, R.A., Cardoso, A.S., Oliveira, W.R.D., Jantalia, C.P.,

436

Urquiaga, S., Boddey, R.M., 2012. Selection of the most suitable sampling time for static

437

chambers for the estimation of daily mean N2O flux from soils. Soil Biol. Biochem. 46, 129–135.

438

Azzem, B., Kushaari, K., Man, Z.B., Basit, A., Thanh, T.H., 2014. Review on materials &

439

methods to produce controlled release coated urea fertilizer. J. Control. Release 181, 11-21.

440

Bortoletto-Santos, R., Ribeiro, C., Polito, W.L., 2016. Controlled release of nitrogen-source

441

fertilizers by natural-oil-based poly(urethane) coatings: the kinetic aspects of urea release. J.

442

Appl. Polym. Sci. 133, 43790.

443

Burton, D.L., Zebarth, B.J., Gillam, K.M., Macleod, J.A., 2008. Effect of split application of

444

fertilizer nitrogen on N2O emissions from potatoes. Can. J. Soil Sci. 88, 2, 229–239.

445

Cancellier, E.L., Silva, D.R.G., Faquin, V., Gonçalves, B.A., Cancellier, L.L., Spehar, C.R.,

446

2016. Ammonia volatilization from enhanced-efficiency urea on no-till maize in brazilian cerrado

447

with improved soil fertility. Ciênc. Agrotec. 40, 2, 133-144.

448

Campbell Scientific, 2015 Instruction Manual: CS650 and CS655 water content reflectometers,

449

revision 5/15. https://s.campbellsci.com/documents/es/manuals/ cs650.pdf (accessed 21 August

450

2017).

The authors declare that they have no conflict of interest.

15 451

Chen, H., Wan, L., Fan, L., Xu, H., Cao, H., 2011. Water soluble alkyd resin-sulfur coated

452

controlled release fertilizer and preparation thereof. US n. 2011/0072871 A1 (accessed 21 April

453

2018).

454

Chen, J., Lü, S., Zhang, Z., Zhao, X., Li, X., Ning, P., Liu, M., 2018. Environmentally friendly

455

fertilizers: a review of materials used and their effects on the environment. Sci. Total Environ.

456

613–614, 829–839.

457

Coale, F.J., 2010. Soil Fertility Management: Agronomic Crop Nutrient Recommendations

458

Based on Soil Tests and Yield Goals. University of Maryland Extension Information Bulletin

459

SFM-1. https://extension.umd.edu/sites/extension.umd.edu/files/_images/programs/anmp/SFM-

460

1.pdf

461

Coogan, R.G., Damery, S., 2003. Aqueous resin dispersions. US n. 6.548.588 B1 (accessed 20

462

April 2018).

463

Davidson, E.A., Rogers, J.E., Whitman, W.B., 1991. Fluxes of nitrous oxide and nitric oxide from

464

terrestrial ecosystems. In: Microbial production and consumption of greenhouse gases.

465

American Society for Microbiology, Washington, pp. 219 235.

466

Detrick, J.H., Hargrove, G.L., 2002. Polymer-sulfur-polymer coated fertilizers. US6338746B1

467

(accessed 20 April 2018).

468

Díaz-Rojas, M., Aguilar-Chávez, A., Cárdenas-Aquino, M., Del, R., Ruíz-Valdiviezo, V.M.,

469

Hernández-Valdez, E., Luna-Guido, M., Olalde-Portugal, V., Dendoovena, L., 2014. Effects of

470

wastewater sludge, urea, and charcoal on greenhouse gas emissions in pots planted with

471

wheat. Appl. Soil Ecol. 73, 19-25.

472

Eagle, A.J., Olander, L.P., Locklier, K.L., Heffernan, J.B., Bernhardt, E.S., 2017. Fertilizer

473

management and environmental factors drive N2O and NO3 losses in corn: a meta-analysis. Soil

474

Sci. Soc. Am. J. 81, 1191-1202.

475

Eggleston, H.S., Buendia, L., Miwa, K., Ngara, T., Tanabe, K., 2006. IPCC: guidelines for

476

national greenhouse gas inventories, IGES, Hayama, Japan. Prepared by the National

477

greenhouse gas inventories programme.

16 478

Firestone, M., Davidson, E., 1989. Microbiological basis of NO and N2O production and

479

consumption in soil, in: Exchange of trace gases between terrestrial ecosystems and the

480

atmosphere. Chichester, United Kingdom, pp. 7-21.

481

Forster, P., Ramaswamy, V., Artaxo, P., Berntsen, T., Betts, R., Fahey, D.W., Haywood, J.,

482

Lean, J., Lowe, D.C., Myhre, G., Nganga, J., Prinn, R., Raga, G., Schulz, M., Van Dorland, R.,

483

2007. Changes in atmospheric constituents and in radiative forcing, in: Climate Change 2007:

484

the physical science basis. Contribution of working group I to the fourth assessment report of

485

the intergovernmental panel on climate change. Cambridge Univ. Press, Cambridge, United

486

Kingdom and New York, NY, pp. 129-234.

487

Halvorson, A.D., Del Grosso, S.J., Alluvione, F., 2010. Tillage and nitrogen source effects on

488

nitrous oxide emissions from irrigated cropping systems. Soil Sci. Soc. Am. J. 74, 2, 435–445.

489

Hanway, J.J., 1966. How a corn plant develops. Special Report 38.

490

Hyatt, C.R., Venterea, R.T., Rosen, C.J., Mcnearney, M., Wilson, M.L., Dolan, M.S., 2010.

491

Polymer-coated urea maintains potato yields and reduces nitrous oxide emissions in a

492

Minnesota Loamy Sand. Soil Sci. Soc. Am. J. 74, 2, 419-428.

493

Keeney, D.R., 1982. Nitrogen management for maximum efficiency and minimum pollution, in:

494

Nitrogen in agricultural soils. Agronomy Monograph, 22, ASA, CSSA, and SSSA, Madison,

495

Wisconsin, pp. 605-649.

496

Mariano, E., Sant Ana Filho, C.R., Bortoletto-Santos, R., Bendassolli, J.A., Trivelin, P.C.O.,

497

2019. Ammonia losses following surface application of enhanced-efficiency nitrogen fertilizers

498

and urea. Atmospheric Environ. 203, 242-251.

499

Ogle, J.M., Sims, J.M., 2010. Controlled release fertilizer composition. US n. 2010/0275665 A1

500

(accessed 20 April 2018).

501

Parkin, T.B., Hatfield, J.L., 2014. Enhanced efficiency fertilizers: effect on nitrous oxide

502

emissions in Iowa. Agron. J. 106, 2, 694-702.

503

Parkin, T.B., Venterea, R.T., 2010. Chamber-based trace gas flux measurements. In: Sampling

504

protocols,

505

www.ars.usda.gov/research/GRACEnet (accessed 10 July 2017).

USDA-ARS

GRACEnet

Project

Protocols,

R.F.

Follett

(Ed.).

17 506

Phillips, R.L., Tanaka, D.L., Archer, D.W., Hanson, J.D., 2009. Fertilizer application timing

507

influences greenhouse gas fluxes over a growing season. J. Environ. Qual. 38, 1569-1579.

508

R Core Team, 2018. R: A language and environment for statistical computing. R Foundation for

509

statistical computing, Vienna, Austria. https://www.R-project.org/ (accessed 05 August 2018).

510

Ravishankara, A.R., Daniel, J.S., Portmann, R.W., 2009. Nitrous oxide (N2O): the dominant

511

ozone-depleting substance emitted in the 21st century. Science 326, 123–125.

512

Raymond, J.E., Fox, T.R., Strahm, B.D., Zerpa, J., 2016. Ammonia volatilization following

513

nitrogen fertilization with enhanced efficiency fertilizers and urea in loblolly pine (Pinus taeda L.)

514

plantations of the southern United States. For. Ecol. Manag. 376, 247-255.

515

Robertson, G.P., Vitousek, P.M., 2009. Nitrogen in agriculture: Balancing the cost of an

516

essential resource. Annual Review Environ. Resour. 34, 97-125.

517

Rochette, P., Angers, D.A., Chantigny, M.H., Bertrand, N., Côté, D., 2004. Carbon dioxide and

518

nitrous oxide emissions following fall and spring applications of pig slurry to an agricultural soil.

519

Soil Sci. Soc. Am. J. 68, 1410-1420.

520

Rochette, P., Angers, D.A., Chantigny, M.H., Gasser, M.-O., MacDonald, J.D., Pelster, D.E.,

521

Bertrand, N., 2014. Ammonia volatilization and nitrogen retention: How deep to incorporate

522

urea? J. Environ. Qual. 42, 1635-1642.

523

Ruser, R., Flessa, H., Russow, R., Schmidt, G., Buegger, F., Munch, J.C., 2006. Emission of

524

N2O, N2 and CO2 from soil fertilized with nitrate: effect of compaction, soil moisture and

525

rewetting. Soil Biol. Biochem. 38, 263–274.

526

Shaviv, A., 2000. Advances in controlled release of fertilizers. Adv. Agron. 71, 1–49.

527

Silva, A.G.B., Sequeira, C.H., Sermarini, R.A., Otto, R., 2017. Urease inhibitor NBPT on

528

ammonia volatilisation and crop productivity: a meta-analysis. Agron. J. 109, 1-13.

529

Topp, G.C., Davis, J.L., Annan, A.P., 1980. Electromagnetic determination of soil water content:

530

measurements in coaxial transmission lines. Water Resources Research 16, 3, 574-582.

531

18 532

Trenkel, M.E., 2010. Slow-and controlled-release and stabilized fertilizers: An option for

533

enhancing nutrients use efficiency in agriculture. International Fertilizer Industry Association

534

(IFA), Paris. 163 p.

535

USEPA (United States Environmental Protection Agency), 2014. Inventory of U.S. greenhouse

536

gas emissions and sinks: 1990–2014, USEPA Climate Change Div., Washington, DC.

537

https://www.epa.gov/ghgemissions/inventory-us-greenhouse-gas-emissions-and-sinks-1990-

538

2014 (accessed 11 November 2017).

539

Whittington, A.A., 2005. Coating for fertilizer. US n. 2005/0076687 (accessed 21 April 2018).

540

With, T.K., Petersen, B., Petersen, T.D., 1961. A simple spectrophotometric method for the

541

determination of urea in blood and urine. J. Clin. Pathol. 14, 202−204.

542

Zebarth, B.J., Rochette, P., Burton, D.L., Price, M., 2008. Effect of fertilizer nitrogen

543

management on N2O emissions in commercial corn fields. Can. J. Soil Sci. 88, 189-195.

544

19 545

546 547 548 549 550 551

FIGURES

Figure 1. (a) Schematic diagram showing placement of gas sampling chamber anchors relative to maize rows in a maize plot and (b) photo of an anchor (background) with the lid in the foreground as well as the materials used for gas sampling (1 - stopwatch, 2 – syringe (inserted into sampling port), 3 – sample vials and 4 – thermometer, placed on soil surface).

20 552

553 554 555 556 557

Figure 2. SEM images of (a) uncoated urea and (b-e) the interface between urea and castor oil polyurethane at 2, 4, 6, or 8 wt%, respectively. In addition, microtomography image of (f) an uncoated urea granule and (g) a urea granule coated with 8 wt% polyurethane, which is highlighted in yellow.

21

558 559 560

Figure 3. Average curves of urea release in a water immersion. Vertical bars are standard deviations.

22

561 562

Figure 4. N2O emissions from soil planted to maize and fertilized at planting with urea, urea coated with

563

polymer at 2, 4, 6, or 8 wt% (PCU2% to PCU8%), a soil control (SC) treatment with no fertilizer applied,

564

and a treatment in which urea was applied at sidedressing when the maize was at the V6 growth stage

565

(V6). N application rate was 160 kg N ha for all but the SC treatment.

-1

23

566 3 H2O

-3 soil)

567

Figure 5. Mean volumetric water content (m

568

Numerical values above the soil moisture curves indicate six different time periods during which N2O

569

emissions data were summed to facilitate statistical comparisons, as presented in Table 2.

570

m

for all plots measured during the experiment.

24 571

TABLES

572 573 574

Table 1. Material, code, percentage of PU coating by weight and total nitrogen content of the coated and uncoated urea granules. Material

Code

PU coating (wt%)

Total N (%)

Urea (uncoated)

Urea

-

47.4 ± 0.2

Urea + PU 2%

PCU2%

2.1 ± 0.1

46.8 ± 0.2

Urea + PU 4%

PCU4%

4.2 ± 0.2

45.9 ± 0.2

Urea + PU 6%

PCU6%

6.0 ± 0.2

45.1 ± 0.3

Urea + PU 8%

PCU8%

8.1 ± 0.2

44.6 ± 0.3

575 576 577 578

Table 2. Cumulative soil N2O emissions over six sampling periods (see Figure 5) for a field planted to †

maize and fertilized with urea and urea coated with polymers . Sampling

time period

Dates

N2O-N emissions (g ha-1) Urea

PCU2%

PCU4%

PCU6%

PCU8%

SC

‡

V6

§

1

18 July - 28 July

680 a

¶

620 a

466 ab

266 bc

177 d

142 d

222 cd

2

28 July - 07 Aug

1402 a

900 a

1323 a

308 b

76 c

50 c

51 c

3

07 Aug - 15 Aug

1003 a

343 b

575 ab

376 b

108 d

52 e

45 e

4

15 Aug - 01 Sept

682 a

213 cd

380 ac

528 ab

233 bc

84 e

97 de

5

01 Sept - 04 Oct

165 c

127 c

156 c

410 b

316 b

93 c

2175 a

6

04 Oct - 03 Nov

54 c

52 b

54 b

208 a

177 a

48 b

106 a

Cumulative

106 d

3986 a

2254 b

2954 ab

2096 b

1087 c

469 d

2696 ab

579 580 581 582 583 584

† PCU2% to PCU8% indicates polymer concentration at 2, 4, 6, or 8 wt%. ‡ SC is the control with no fertilizer applied. § V6 is urea applied at sidedressing when the maize was at the V6 growth stage. ¶ Values followed by the same letter within a row are not statistically different at P=0.05.

25 585 586

Table 3. Maize N uptake and recovery, maize grain yield and N2O emissions factor following maize †

harvest in a field fertilized at maize planting with urea and urea coated with polymers . Treatment

587 588 589

N uptake (kg N ha-1) ‡

-1

Maize grain yield

Cum. N2O-N (g ha ) /

(Mg ha-1)

Grain yield (kg ha-1)

52.8 ± 12.3 ns

8.52 ± 0.92 a

0.47 ± 0.15 a

N recovery (%)

Urea

139 ± 19 a

PCU2%

135 ± 11 a

50.1 ± 6.9 ns

8.54 ± 0.49 a

0.26 ± 0.04 ab

PCU4%

145 ± 15 a

56.3 ± 9.5 ns

8.79 ± 0.90 a

0.34 ± 0.07 ab

PCU6%

155 ± 14 a

62.7 ± 9.2 ns

8.90 ± 0.93 a

0.24 ± 0.05 bc

PCU8%

148 ± 12 a

58.7 ± 7.5 ns

8.21 ± 0.72 a

0.14 ± 0.04 c

SC

66 ± 4 b

-

3.89 ± 0.46 b

0.13 ± 0.03 c

133 ± 33 a 49.2 ± 20.7 ns 6.56 ± 0.62 ab 0.42 ± 0.08 ab V6 Treatment designations as in Table 2. ‡ Values followed by the same letter within a column are not statistically different at P=0.05. †

26 590

SUPPLEMENTARY INFORMATION

591

Soil texture analyses were conducted by Brookside Laboratory (New Bremen, OH, USA)

592

on soil samples collected in the field December 2017. Chemical analyses, including soil test P

593

and K used to manage P and K fertilizers (data not shown), were conducted by Waypoint

594

Analytical Labs (Richmond, VA, USA) on samples collected in February 2017.

595 596

Table S1. Chemical and physical properties of Downer Loamy sand used in this experiment. †

pH

g kg 6.2

597 598 599 600 601

‡

CEC

-1

meq 100 g

OC

14

§

¶

Sand -1

3.4

Silt

¶

¶

Clay

-1

---------- g kg ----------845.8

80.7

73.5

† pH - using SMP buffer. ‡ OC - organic C content, measured by the loss on ignition method. § CEC - cation exchange capacity. ¶ Texture analysis - pipette method.

602

The soil NH4+-N data (Figure S1a and S1b) show that values were low in all treatments

603

on all dates except for 23 September. The lack of difference between the control (SC) and

604

fertilizer treatments indicates that conversion of urea to NH4+-N and then NO-3-N was fast in the

605

fertilized treatments and/or that maize N uptake was faster than NH4+-N release rates. The high

606

NH4+-N on 23 September coincides with very low soil moisture (Fig. 5). At this time the maize

607

showed visible signs of water stress (data not shown), which spurred use of irrigation to

608

alleviate this stress on 27 September. Due to water stress the maize was likely not able to take

609

up soil mineral N, such that N accumulated in the soil at both soil depths, both as NH4+-N and as

610

NO3--N (Figure S1a-d). After soil moisture levels increased again, mineral N uptake continued

611

and both NH4+-N and NO3--N levels were reduced. Finally, on the last sampling day, 27 October,

612

the maize had reached or was close to reaching physiological maturity when N uptake is very

613

low. Therefore, both NH4+-N and NO3--N levels increased at both soil depths.

614

Soil nitrate data (Figure S1c and S1d) show a more complex pattern than NH4+-N data,

615

showing some differences among treatments and depths. For example, NO3--N levels early in

616

the season (10 August) were relatively high in all treatments including SC at both depths. At 0-

27 617

15 cm depth NO3--N in treatments receiving fertilizer materials was different from that in SC only

618

on 25 September, suggesting that the source of most of the NO3--N in most treatments was soil

619

organic matter. Exceptions to this generality include the urea treatment at 0-15 cm depth early

620

in the season when NO3--N in urea treatment seems to be higher than in other treatments. This

621

is consistent with the quicker release of urea in water in the lab (Fig. 3). NO3--N in the PCU4%

622

treatment also seems higher than in SC on the third sampling date (17 August). It is not clear

623

why this is. On 25 September, NO3--N in all treatments except PCU4% and V6 were higher than

624

in SC. The higher values are consistent with low maize N uptake during this drought period as

625

discussed earlier. The low value in PCU4% can be explained by data for the 15-30 cm depth

626

where PCU4% showed substantially higher soil NO3--N than other treatments, indicating that

627

NO3--N in this treatment had leached to lower soil depths though it is not clear why leaching

628

would have been higher in this treatment than others. Low NO3--N in V6 on 25 September at 0-

629

15 cm depth may be due to the later application timing of urea in V6 resulting in less NO3--N

630

present in soil at that time. This is consistent with lower NO3--N at 15-30 cm in the V6 treatment.

631

In the V6 treatment, however, NO3--N appears later in the season than in other treatments as

632

indicated by NO3--N at both depths on 13 October. Finally, PCU8% was also lower than other

633

treatments on 11 September, suggesting there was less NO3--N available for leaching in this

634

treatment, likely due to slow release of urea from this material. Lack of synchrony between soil

635

mineral N and N2O emissions is consistent with the literature in which individual N2O emissions

636

peaks are generally not predictable from soil mineral N values (Rochette et al. 2004; Burton et

637

al. 2008). This is because N2O emissions reflect not just N2O production and consumption in the

638

soil but also the physical diffusion of N2O out of the soil into the atmosphere. In addition, N2O is

639

controlled by C:NO3--N ratio and other factors that are not reflected in soil mineral N data

640

(Rochette et al. 2004).

641

28 642

643 644

Figure S1. NH4+-N and NO3--N in the soil for all treatments with two different depths (0-15 cm

645

and 15-30 cm).

646 647

Highlights

1. Polyurethane-coated urea reduced soil N2O emissions up to 60-80%; 2. Control of urea release achieved using polyurethane coating derived from castor oil; 3. Rate of urea release was controlled by the thickness of the coating.

Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ☒The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: