Flax fibre quality and influence on interfacial properties of composites

Abstract

Flax fibre holds the potential to serve as an alternative to glass fibre as reinforcement in composite applications. To fully achieve this, the interaction between fibre and matrix must be improved and more consistently controlled. Only then will industry accept natural fibres as a sustainable engineering material choice. Traditionally, interfacial strength improvement has been accomplished through expensive and time consuming chemical surface modification(s). To achieve improved market potential and viability, new methods of developing composite ready flax fibre must be researched and developed through an assessment of the impact of fibre traits for unmodified fibre. Metal, fungal, bacterial, wax and glucose content were examined in this study to determine their correlative effects upon interfacial adhesion, as were fibre characteristics such as colour, density, fineness, fibreshape thickness, conductivity and pH levels. Composite performance was evaluated using fibre pullout and interfacial shear strength tests. These first attempts at correlating as-received flax fibre traits and resulting flax fibre composite properties contain the initial steps towards identifying key flax fibre characteristics that influence composite performance so that proper growth and fibre processing approaches can be developed.

Keywords:

• flax
• fibre
• quality
• linseed
• composite
• thermoset

1. Introduction

Natural fibres are being utilised due to their lower cost, low density, biodegradability, relatively high strength, low abrasiveness, abundance, renewability, non-hazardous nature, recyclability and low equipment requirements. In particular, flax fibre has seen emergence as a potential replacement for glass fibre within composite materials. The flax plant can be grown under subtropical conditions as a winter crop or under temperate conditions as a summer crop (Foulk et al. 2005). It is often considered as an environmentally responsible crop with no insecticides, low herbicides, good erosion control and low fertilisation rates.

Legislation on the recycling of end of life vehicles has helped promote interest in utilisation of flax as a fibre reinforcement in composites. Compared to most bast fibres, flax fibres exhibit some of the highest specific strength and modulus properties, lowest density and best non-abrasive and insulating qualities. Linen flax fibre has already found its way into replacing fibreglass for certain automotive applications in Europe and North America, such as interior dash and trim components as well as underbody shields and covers. DaimlerChrysler (2001) indicates that the natural fibre automotive components require 83% less energy and are 40% less expensive than glass fibre components.

However, natural fibres continue to exhibit drawbacks such as variable fibre quality and poor matrix interaction. It is well understood that proper fibre–matrix bonding is crucial for the performance of flax composites (Mustata 1997). This is due to the hydrophilic nature of flax caused by the hydroxyl groups in the fibre's cellulose and lignin (Nevel and Zeronian 1987) poorly interacting with polymers which tend to be hydrophobic in nature. To address this, a large body of research has been placed into the development of chemical surface treatments which allow for improved composite adhesion.

For natural fibres, a number of treatments have been explored to improve surface interaction with polymer matrices. Some of the most successful methods have been through the use of coupling agents, which are able to react with both the fibre surface and the polymer matrix forming covalent and hydrogen bonds. This has been accomplished through the use of isocyanates (Kokta et al. 1990, Joseph et al. 1996), triazine (Oksman 2001) and silanes (Culler et al. 1986, Rong et al. 2001, Cantero et al. 2003). Other treatment methods include acetylation (Mohanty et al. 2001, Baiardo et al. 2004), benzoylation (Manikandan Nair et al. 1996, Joseph et al. 2000) and peroxide treatment (Sreekala et al. 2000). However, perhaps the most common method of chemically modifying the surface of natural fibres has been mercerisation, or the use of alkalis such as NaOH to improve the surface interaction between matrix and fibre (Baley et al. 2006, Van de Weyenberg et al. 2006). Alkali treatments lead to surface roughening, fibrillation and removal of natural and artificial impurities (Mishra et al. 2001).

Despite the extensive work that has been accomplished in improving the interfacial interaction between natural fibres and polymer matrices, the focus has been upon the use of chemical treatments in order to modify the fibre surface. However, it is well known that there is a large variance within flax fibre due to growth, harvesting and processing. Basic research and development are still required in order to facilitate the use of natural fibres into new applications and to obtain optimal final product natural fibres. One of the main concerns for the composite industry's incorporation of flax fibres thus lies not in just the development of surface treatments to promote better interaction, but also in identifying the sources of variability that exist within flax fibre due to diverse crop and retting conditions as well as processing techniques. If flax traits which have correlation to surface interaction can be identified, the necessity for surface treatments in order to yield consistent quality becomes lessened thereby reducing cost. Only then will flax fibre become a truly sustainable engineering material choice.

Beyond the inherent variability from crop to crop and plant to plant along with variable climate, the location that flax fibre comes from on a plant can also play a role upon the fibre's properties. Charlet et al. (2007) conducted research on the characteristics of flax fibres as a function to their location in the stem and the properties of the derived unidirectional composites. It was concluded that the chemical composition and mechanical properties of the fibres were strongly influenced by their location in the stem. Therefore, for a single fibre, there are variation in diameter, surface roughness and defects along the length of individual fibre along with variable chemical composition. Hence, the interfacial strength of the composite can be assumed to correspondingly be a complex property arising from the complex nature of flax fibres.

The fibre extraction process has also been shown to play a role in the variation found in flax fibre strength, fineness, colour and trash content (Van de Weyenberg et al. 2006, Farag 2008, Huang and Liu 2008). This extraction process is normally done by some form of retting, a process of controlled decomposition that allows for fibre and shive to naturally separate from each other. However, care must be taken in retting fibres. Over retting can lead to thinning of bundles or microbial attack on fibre cellulose reducing fibre strength. While under-retted flax fibres can yield higher shive and cuticular fragments along with coarser fibre surfaces (Weyenberg et al. 2006). After retting, the bark and the xylem are removed by mechanical separation, also termed scutching. The fibres are then aligned by hackling (Huang and Liu 2008). Chemical and bio-technical separation with water or ethylene diamine tetraacetic acid (EDTA) can be performed instead of mechanical separation of the bark (Mohanty et al. 2005). It is well understood that retting flax stalks produce fibres with different surface components, while mechanically processing at different levels yields fibres with varying degrees of quality (Foulk et al. 2008). However, the impact of flax fibre quality and its variability on composite performance through interfacial bonding has not been established.

The main focus of this study was to examine any correlation that exists between various traits of flax fibre and the subsequent interfacial bonding these fibres have with a thermoset matrix (in this case a vinyl ester resin). By doing this, steps to achieving advantageous composites can be developed through focus on the growth, processing and selection of flax fibre. A standardised method to assess flax fibre composite performance would facilitate premiums for high quality fibres for composites, assist in maintaining quality control from crop to crop and provide a means to assess the performance of different processing techniques. An understanding of the fibre characteristics that most influence composite performance could also lead to the development of different tailored additives, coatings, binders or sizing suitable for natural fibre and thermosetting materials.

2. Materials

Flax bales were chosen qualitatively in order to yield a range of characteristically different properties and fibre traits. In this study, 14 linseed flax samples and four linen flax samples were examined, ranging in both degree of cleanliness as well as degree and method of retting. By utilising a wide variety of sample sets, ranges of each fibre trait to be studied were determined, and subsequent statistical correlation was able to be performed in order to determine correlations between these traits and interfacial bonding strength.

For the flax samples, two methods were used to decorticate the fibres: hammer milling and long line scutching. Hammer milling was utilised on 12 of the Canadian grown linseed flax samples. Hammer milling was performed in three plants. Two of the plants are operated by Schweitzer-Mauduit (Winkler, Manitoba, Canada) while a third was operated by the now defunct Ecusta Fibers Ltd. (Winkler, Manitoba, Canada). Both stationary and mobile mills were utilised depending on the sample. The major difference in the two mills, although using similar technology, is that a stationary mill removes more dust and produces a cleaner material with higher cellulose content than the mobile mill. After hammer milling, several of the samples were also subsequently processed through a series of cleaning processes as described by Akin et al. (2005). A further subset of these samples was also passed through an aggressive opener (Masterclean opener, John D. Hollingsworth on Wheels, Inc., Greenville, SC, USA) to cottonise the fibre. These methods were all taken in order to bring about a wide range of traits between the various sample sets.

The second method used to decorticate the fibres was long line scutching, with subsequent processing sections. This was performed on two Canadian grown linseed flax samples, a Canadian linen flax sample and three European grown linen flax samples. The three Canadian samples were processed by Biolin Research, Inc. (Saskatoon, Saskatchewan, Canada) using roller mill technology to maintain fibre length and reduce mechanical processing. The three commercial grades of European fibre flax tow were poor in quality but designated appropriate for composites.

The resin chosen for this study was a promoted, non-thixed modified vinyl ester resin, trade name Hydropel R037-YDF-40 (AOC LLC, Collierville, TN, USA). Hydropel is a low viscosity resin for use in vacuum infusion processes. Vinyl ester was chosen because it is a common, inexpensive and moderate performance resin that is utilised to a large extent in glass fibre composite components. With flax fibre being sought as an economic alternative to lower end synthetic fibres such as glass, it is important that studies of interfacial performance be performed using a resin that holds a strong potential for being used, rather than seek a higher end epoxy system which may provide better performance but at a cost that would make it impractical for natural fibre composite usage.

3. Methodology

3.1 Fibre trait measurements

A series of tests were conducted to determine the quality of the processed flax fibres. Long line fibres were cut in length to ∼25 mm for fibre testing. Fibre trait testing spanned two classifications. The first are traits corresponding to constituent content, including impurities. These are traits that are suspected of having direct influence on the interfacial bonding between flax and resin matrix. These included bacterial populations, which were determined for the flax samples according to Chun and Perkins (1996), as well as fungal populations were conducted according to Chun and McDonald (1987). Other constituent contents, such as fibre wax levels, percent glucose levels and metal content were also determined according to Foulk et al. (2007).

The second set of traits examined were those used to characterise fibres. These are traits that do not necessarily have an effect on the bonding, but potentially hold a correlation to interfacial strength and thus the ability to act as screening tools. These included physical properties such as colour, determined by CIE L*a*b* colour space analysis. These were obtained using a Jasco V-670 (Jasco, Inc., Easton, MD, USA) spectrophotometer, equipped with SpectraManager version 2.0 data handling software (ASTM International 2003). Fibre fineness levels were also examined, conducted according to ASTM International (2005). Image analysis of fibre thickness was subsequently determined, using image analysis conducted with Fibreshape version 5.2 (Schmid and Müssig 2002). Fibre density was examined according to ASTM International standards (2004) using Procedure A. For these measurements, fibres were conditioned prior to density testing at 60°C for 72 h to achieve their dry weight, and then immersed in canola oil and re-weighed. Chemical properties examined, such as fibre conductivity and pH, were also performed using tests specified by Foulk et al. (2007).

3.2 Pullout

This approach is commonly utilised for fibre–matrix interaction testing of synthetic fibres, such as ultra high molecular weight polyethylene (Jana et al. 2006), as well as with bundled natural fibres (Brahmakumar et al. 2005, Chen et al. 2006) and large natural fibres (Aquino et al. 2003). Fibres were collected from each sample set, with nine fibre bundles being created from each in order to ensure that statistical minimums are met. Flax bundles were created by hand picking technical fibres from the loose flax, and aligning them unidirectionally into bundles of varying size and fibre content. Bundles were pulled through a clean Neoprene rubber gasket using a small threading needle. The fibres were uniformly trimmed and pulled so only 4 mm was protruding from the rubber surface. The rubber gaskets with embedded fibre bundles were then placed on a level rack. Droplets of vinyl ester were applied to the top surface of the rubber to fully enclose the fibres.

Once cured, the specimens were removed from the rubber by carefully cutting away the gasket, leaving the fixed depth resin matrix. Testing was performed on an Instron model 5567 load frame (Norwood, MA, USA) with a 2 kN load cell, in displacement control at a rate of 0.5 mm/min. The cured matrix droplet was prevented from moving by passing the fibre bundle through a slotted plate, and with tensile load applied to the fibres, the fibre bundle will see interfacial failure once critical stress is achieved. Specimen results were discarded during testing for specimens that saw premature fibre breakage before interfacial failure.

Microscopy was used to determine the interfacial area of the pulled fibre bundle. With pullout perimeter, fibre embedded length and force of pullout established, a modified Kelly–Tyson equation can be applied to obtain the interfacial shear strength, τ _i :

where F _max is the max load at pullout, C is the fibre bundle perimeter, L _e is the fibre embedded length and A is the area of the top surface of the fibre, which is encapsulated in resin and thus must be included in the surface area calculations.

Microscopy work was performed using a ZEISS Axiovert 40MAT microscope (Peabody, MA, USA) equipped with a Jenoptik Progres C10 3.3 megapixel CCD digital camera (Easthampton, MA, USA) and iSolution DT digital imagining software (Victoria, MN, USA), to determine the contact perimeter between the fibre bundles and vinyl ester matrix. These data, along with the known impregnation depth, were then used to determine the total contact area between the fibres and matrix.

3.3 Interlaminar shear strength (ILSS) testing

An interlaminar shear strength (ILSS) test was conducted as a possible alternative to the fibre pullout test. This test, standardised under ASTM International (2006), is normally used to determine the interlaminar shear strength of unidirectional composites to determine the maximum shear stress between layers of a laminate. It has been demonstrated in literature, however, that the short beam stress can also be related directly to the interfacial shear stress between fibre and matrix (Polacek and Jancar 2008, Seki 2009, Sugihara and Jones 2009). ILSS testing is less arduous to perform than the fibre pullout due to specimen preparation. In addition, it is a better representative of actual composite performance parameters than fibre pullout.

To yield the sample specimens required, a modified vacuum assisted resin transfer molding (VARTM) infusion process was implemented using a caul plate. The caul plate was placed on the top surface of the fibres under the bagging film, allowing for a flat surface on both ends of the part. This allows for sample production without stress concentrations, because uniform thickness is achievable. The resultant short beam strength (F ^sbs) can then be calculated using the following equation:

where P _m is the maximum load, b is the specimen width and h is the specimen thickness. Since failure in ILSS testing should be by shear and not through compressive or tensile yield, the short beam strength can be directly related to the interfacial bonding strength of the fibres and matrix. During testing, any samples found to fail by compressive fibre buckling or tensile fibre breakage were discarded.

4. Results

For each trait, an average value was determined for each flax sample set. This was done in order to ensure that variances in a sample set were accounted for and their effects negated. Thus, for each trait, 18 averaged values were produced, each corresponding to a specific flax bundle. The composite interfacial properties, fibre pullout and ILSS testing, were also performed for the various flax fibre sample sets in this manner, with averages being determined corresponding to the 18 individual flax bundle sets.

A correlation study was performed to determine if linear trends exist between the various flax fibre characteristics and their subsequent interfacial bonding strength with vinyl ester. Correlation in this study was determined by examination of the Pearson product-moment correlation coefficient (r). This examination is not able to ascertain the difference between correlation and causation, only demonstrate where linear trends may exist. The closer the coefficient is to either − 1 (representing an inverse relationship) or 1 (direct relationship), the stronger the correlation between the variables. A correlation is considered significant if |r| ≥ 0.5 is established between interfacial measurements and specific traits being compared. Although Pearson's correlation method provides an excellent means to empirically filter out the importance of outliers, in cases where the assumption of normality between trait and interfacial bonding measurements is incorrect, outliers within the factor data still hold the potential to skew calculations of correlation. In these situations, the Pearson product-moment correlation coefficient was recalculated using data with the outlier point removed.

A Pearson's correlation study was also done to compare the two interfacial bonding strength test methodologies. The two methods yielded an r-value of 0.43, indicating that there is some general trend between the two methods, but not a statistically strong correlation. This is not surprising, however, as both methods approach the analysis of interfacial bonding in different ways and with flaws inherent in their design. Flax fibre to fibre consistency within a bundle or composite sample is difficult to completely control as a natural product. Coupled with the inability to achieve perfect linearity of fibres, these variations will have an impact on each interfacial bonding test in slightly different ways, which will lead to subsequent discrepancies in the correlation of results from each. However, there being evidence of a trend, albeit not strong, is an indication that the two methods do not pose complete opposition to each other and both hold usefulness in examining trait correlation to bond potential.

4.1 Constituent content

Impurities and amount of waxes and other surface substances affect the surfaces of fibres, and removal of these impurities has been demonstrated to improve composite performance in the literature (Eichhorn et al. 2001). Thus, the demonstration of a range of content occurring due to growth and processing of flax, as demonstrated in the range of averages shown in Table 1, allows for the determination of which factors are detrimental or helpful to interfacial adhesion.

Table 1 Range of constituent content.

Trait	Low	High
Bacterial population (log(CFU/gm +1))	5.3	8.7
Fungal population (log(CFU/gm +1))	3.4	6.1
Glucose content (%)	0.032	0.43
Wax content (%)	0.98	2.23
Ca (ppm)	45.4	180.5
Fe (ppm)	0.9	6.2
K (ppm)	20	327
Mg (ppm)	9	74
Na (ppm)	1.6	163.6
Zn (ppm)	0.1	0.38

A Pearson's product-moment correlation of the relationship between these traits and interfacial bonding strength by both fibre pull out and ILSS testing is found in Table 2. Positive coefficients indicate that an increase in interfacial bonding was found with increasing the value of the particular component, while negative values indicate that interfacial bonding strength decreased with an increase in the trait. Through this analysis, it was found that two metal contents, calcium and magnesium, have a negative effect on interfacial bonding strength. Furthermore, increased bacterial population and increased wax content were also both found to play detrimental roles in the ability of the flax to form a strong bond with the vinyl ester matrix.

Table 2 Pearson's correlation between interfacial bond strength and constituent content.

Trait	Fibre pullout	ILSS
Bacterial population (log(CFU/gm +1))	− 0.507	− 0.506
Fungal population (log(CFU/gm +1))	− 0.422	− 0.206
Glucose content (%)	− 0.336	0.108
Wax content (%)	− 0.569	− 0.618
Ca (ppm)	0.451	0.532
Fe (ppm)	0.065	− 0.135
K (ppm)	− 0.182	0.155
Mg (ppm)	− 0.658	− 0.133
Na (ppm)	− 0.067	− 0.194
Zn (ppm)	− 0.438	− 0.389

4.2 Characterisation methods

A means of characterising flax fibre for the purpose of screening useful versus less adequate fibre bundles for composite usage is a second important concern to be addressed in the understanding of raw fibre potential. The range of averaged sample set values for the various traits that hold the potential to differentiate potentially higher end flax fibres is found in Table 3. The CIE L*a*b* colour space analysis indicates that distinctive colour representations for the flax fibres exist, while the physical and chemical characteristic measurements also indicate that distinctive variations occur within flax fibre due to some result of growth and processing.

Table 3 Range of characteristic measurements.

Trait	Low	High
L*	53.4	64.5
a*	1.29	1.52
b*	8.2	20.6
Density (g/cm^3)	1.29	1.52
Fineness (SSI)	1.54	5.11
Fibreshape thickness (μm)	15.2	74
Conductivity (Ω^− 1 cm^− 1)	197	837
pH	5.86	8.68

A Pearson's product-moment correlation of the relationship between these traits and interfacial bonding strength by both fibre pull out and ILSS testing is found in Table 4. Through this analysis, it was found that colour does not hold any significant correlation to interfacial bonding potential, and thus cannot serve as a means of screening potentially advantageous fibres for composite usage. However, fibre density was found to yield a positive correlation, indicating that fibres of higher density show a better interfacial bonding performance. Fibreshape thickness was found to have the opposite correlation, with increased fibreshape thickness signifying fibres which yielded poorer interfacial bond strength.

Table 4 Pearson's correlation between interfacial bond strength and characteristic measurements.

Trait	Fibre pullout	ILSS
L*	− 0.195	0.172
a*	0.28	0.296
b*	− 0.378	0.209
Density (g/cm^3)	0.404	0.557
Fineness (SSI)	− 0.432	− 0.017
Fibreshape thickness (μm)	− 0.885	− 0.465
Conductivity (Ω^− 1 cm^− 1)	− 0.238	0.119
pH	0.341	0.348

5. Discussion

5.1 Constituent content

Surface contaminants such as metals, waxes and bacteria all further exacerbate the difficulties of interaction between the already hydrophilic fibre surfaces with hydrophobic matrix polymers such as vinyl ester by acting as disconnects between the two systems. Thus, it is important to determine the causes of these contaminants so that they can be properly subdued to their minimal levels, thus yielding better composite ready fibres.

Metal content was found to be an important factor due to both increased magnesium and calcium content showing a correlation to decreases in interfacial strength. These metals are introduced in the growth of the fibres due to soil and growing conditions (Angelova et al. 2004, Foulk et al. 2008). Calcium is known to stabilise pectin molecules found within the plant stem. These results indicate that to yield the most advantageous fibre for composite usage, active measures must be taken to minimise both Ca and Mg.

Wax content can also be tied to the growth of the flax, however, like bacteria populations; wax content is most actively controlled through the retting of the fibre. It was shown by Morrison et al. (2000) that the more efficient the retting process, the less wax will be ultimately yielded on the surface of the fibre. Although water retting traditionally results in higher quality flax for the textile industry, it was found that dew retted fibres yielded significantly lower values of wax then water retted samples.

Although bacteria were shown to correlate to a detrimental effect on fibre surface interfacial strength, bacteria play a fundamental role within the retting process. In water retting, anaerobic bacteria are considered the primary agent responsible for fibre release (Sharma et al. 1992), which produce the pectinase enzymes that degrade pectins and allow the release of the fibre. In dew retting, rather than bacteria, fungus is mainly responsible for the production of pectinace enzymes (Sharma et al. 1992).

Although both the wax content and bacteria content's detrimental effect upon interfacial adhesion would seem to indicate that dew retting is more advantageous than water retting for composite flax fibre, interfacial adhesion is not the sole factor affected by the retting process. Morrison et al. (2000) found that fibre with higher wax, that were subsequently water retted, also had higher total aromatics, cutin, xylose, finer fibres and greater strength. Thus, work is still necessary to develop an optimum retting methodology which will optimise fibre strength and fineness, as well as interfacial bonding strength through surface wax removal and decreased residual bacteria populations.

5.2 Characterisation methods

The most important aspect to note from the fibre characteristics measured is that colour was shown to hold no useful correlation to interfacial bonding strength. As such, simple colour screening of fibres has been eliminated as a means of distinguishing useful flax fibres for composite usage based on interfacial bonding strength. The surface quality of the fibre is not accurately represented by any significant colour distinction. However, it was determined that density measurements can potentially act as screening tools, with denser fibres correlating to improved surface interaction with vinyl ester resin.

It has also been found that fibreshape thickness correlates with interfacial bonding strength, with thicker flax samples yielding greater interfacial bonding. But as was found for bacteria population to some extent, and wax content to a great extent, the correlation of fibreshape thickness and interfacial bonding is a direct result of retting. Fibreshape thickness (which is closely related to fibre fineness, which was not found to be significantly correlated to interfacial bonding) is found to be greater with less efficient retting processes, like that obtained by dew retting (Morrison et al. 2000). This further establishes that the degree of ret, as well as the retting process, play a significant role in the fibre–matrix adhesion.

6. Conclusions

Through the analysis of various constituent contents which can act as surface defects, as well as various characterising traits, correlations have been made with untreated flax fibre to determine optimal properties for the greatest fibre–matrix interfacial interaction between flax and vinyl ester resin. It was found that surface contaminants dictated by growing conditions, such as metals like magnesium and calcium, as well as constituents whose content is driven by the retting process, such as wax and bacteria, are detrimental to proper fibre–matrix interaction. These factors need to be addressed by growers and processors of flax fibre in order to develop controlled batches of flax if the fibre is to become a sustainable engineering material of choice in the composite market.

However, this will not be achieved just through better control of the flax growth and processing, but also through the ability to properly select the most ideal fibres which allow consistency to be maintained in an operation. By studying fibre characteristics which hold potential to act as screening mechanisms, it has been found that colour holds no useful role in determination of the potential of a fibre to achieve good bond with a thermoset matrix. However, density and fibreshape thickness were both found to hold correlation, with increases in either correlating to better interfacial adhesion.

Overall, replacement of glass fibre by natural fibres such as flax will not be based just on comparable specific composite mechanical properties, but also on fibre consistency. While chemical surface treatments have been proven to be extremely beneficial when properly utilised, the growth of the natural fibre sector will be dictated by the ability to achieve consistent, low cost improvements to mechanical bonding achieved by variation and control of the growing and harvesting of the fibres.

Acknowledgements

We gratefully acknowledge and appreciate the help provided by producers, processors and others including Agriculture and Agri-Food Canada, and Schweitzer-Mauduit International. We gratefully acknowledge Nancy Carroll, Martha Duncan, Pat Fields, Don Gillespie, Robert Harrison, Curtis Heaton, Linda James, Jimmy Lewis, Judy Marcus, Mattie Morris, Brad Reed and Debbie Sewell from USDA ARS CQRS for assisting with testing and set-up. We gratefully acknowledge Katie Schalkoff from Clemson University along with Luke Gibbon and Eric Kerr-Anderson from North Dakota State University for their assistance.

Disclaimer. Mention of a trade name, proprietary product or specific equipment does not constitute a guarantee or warranty by the US Department of Agriculture; information is for information purposes only and does not imply approval of a product to the exclusion of others that may be suitable.

Additional information

Notes on contributors

Michael A. Fuqua

1. 1. [email protected]

Chad A. Ulven

2. 2. [email protected]

Mercedes M. Alcock

3. 3. [email protected]

Notes

^1. [email protected]

^2. [email protected]

^3. [email protected]