Chinese (Simplified)EnglishFrenchItalianPortugueseSpanish

Ocean Optics
Worldwide Headquarters
Largo, Florida, USA

+1 727-733-2447

[email protected]

Sales, Service
& Support Facility
Duiven, The Netherlands

+31 26-319-0500
+33 442-386-588

[email protected]

Ocean Optics GmbH Sales,
Service & Support Facility
Ostfildern, Germany

+49 711-34-16-96-0

[email protected]

Sales Support
for the
United Kingdom

+44 1865-819922

[email protected]

Sales, Service
& Support Facilities
Shanghai, PRC – Beijing, PRC

+86 21-6295-6600

[email protected]


Squeezing the Most Out of Soybeans

Rapid NIR Measurement of Oil and Protein Content in Soybeans

The vegetable oils produced for human consumption, industry, animal feed and fuels come from many crops. As the needs of those industries have changed over the years, so too have the preferred sources of vegetable oil. In 1990s, sunflower was one of the primary players in the vegetable oil market, offering 5 tons of oil for every 10 tons of sunflower seeds pressed. In fact, 90% of the oil production plants in Argentina at that time were dedicated to sunflower alone. It was considered a “healthy” oil for our diets, and was a key ingredient in animal feed to speed weight gain and maximize flavor.

Since that time, however, protein has begun to play a much bigger role in the public’s perception of nutrition, while even “healthy” fat consumption has been discouraged. With this change in thinking came a greater demand for protein, favoring crops that could produce high levels of protein for animal feeds as well as oils. Enter the soybean. For every 10 tons of soybeans, 8 tons of high-protein vegetable meal and 2 tons of oil can be produced, as compared to 5 tons of meal and 5 tons of oil from sunflower.

Component  Sunflower
(Average Composition)
(Average Composition)



Protein 13% 36%
Moisture 10% 12%
Other 27% 32%

As a result, there has been a global shift in production from sunflower to soybean to meet the growing demand for protein-rich animal feed over oil-rich formulations. Soybeans are grown to extract their fats for vegetable oil and their solids for high-protein meals. Measurement of oil and protein content with NIR spectroscopy at harvest and during processing maximizes byproduct output from this valuable crop.

World Soybean Meal Production 2015-16 Projected


Protein in Soybeans – Regional at Best

The protein content of soybeans is heavily dependent on the nitrogen content of the soil in which they are grown, which in turn varies by region. Just as some regions are known for their distinctive wines, so too are certain regions in South America known for the characteristics of their soybeans. This is primarily a result of variance in soil quality, though farming methods can also impact protein content of soybeans.

  • Santa Cruz de la Sierra, Bolivia: This area is home to a famous valley for the farming of soybeans, one of the richest in the world. Yield and quality here are extremely good, with soybean protein levels roughly 10% above the average. The soil is black and rich in organic materials and nitrogen.
  • Chaco, Paraguay: The soil here is very special. Nitrogen and organic content is very high, as are iron levels. Paraguayan soybeans have traditionally yielded protein levels at least 20% higher than average, but are somewhat “red” in color due to the iron.
  • Brazil: This country is one of the main soybean producers worldwide, boasting a territory so big that all the possible soil and weather conditions for this crop are available. This includes ultra-high protein soybeans, approaching 30% higher protein than the average. Brazil also has some varieties that are black.

Sorting out Soybeans

Soybeans are harvested during a fairly short season, and must be very rapidly sorted by protein level and stored to keep up with the influx of volume being harvested. Thousands of combines, trucks, and train wagons make their way from the field to the silos, storage and processing plants. To ensure efficient use of the crop, the highest protein grains must be diverted to special silos destined for processing the grains into high protein meal.

Traditional chemical analysis of protein content must be done by trained technicians operating within a lab, using specialized equipment and many reagents. Its greatest downfall, however, is the time needed to determine protein content. NIR spectroscopy, in contrast, offers:

  • Fast response
  • Automatic procedures
  • Easy to use instruments
  • Clean operation
  • Minimal space requirements
  • Multi-parameter analysis

Ocean Optics partner TechnoCientifica has developed the InLab NIR-512 system for automated measurement of protein, oil and moisture in soybeans using NIR spectroscopy and chemometric analysis. This system’s dynamic sampling device allows the sample to be poured into a funnel through a top hopper, building a column of grain approximately 10” high. Collection of NIR reflectance spectra is performed through a special optical sensor that illuminates through a quartz window, dynamically and continuously while the grain falls down to a bottom drawer. The sample may be processed more than once, allowing a large quantity of spectra to be collected from the whole sample before averaging and predicting final results. Automated software controls data collection and processing, delivering protein, oil and moisture data in under 6 seconds. Internal thermal stabilization of the system ensures that consistent answers are delivered, regardless of the ambient temperature of the measurement environment.

InLab NIR-512

Understanding the Analysis

To develop the chemometric models used for analysis in the InLab NIR-512 system, hundreds of samples of different soybeans were processed and sent for analysis of protein, fat and moisture content at an accredited external lab. Each sample was measured many times from 900-1700 nm, building up a database of spectra from which to develop a model.

Initial NIR spectra of soybeans used in model development

Initial NIR spectra of soybeans used in model development

Partial Least Squares (PLS) analysis was performed using GRAMS software, specifically the PLS/IQ module, using the following parameters:

  • PLS1
  • Cross Validation – Files Out=3
  • Preprocessing: SNV + Detrend
  • Wavelength regions: Specially selected for each parameter
  • Multi Model Calibration File (simultaneous prediction of all the parameters)

The model developed enabled the spectra to be correlated to protein, oil and moisture content, with excellent prediction agreement with measurements.

Soybean protein model prediction performance

Soybean fat model prediction performance

Soybean Protein Performance

Soybean protein model prediction performance

Soybean Moisture Performance

Soybean moisture model prediction performance

Extending Measurements to the Plant

The utility of NIR spectroscopy does not end at harvest. It is also an important method for online process monitoring of soy meal and oil production to ensure that maximum oil content is extracted. It simultaneously allows soybean meal producers to ensure that they are meeting the regulated limitations on maximum water and residual oil content in the soybean meal, as well as the minimum required levels of protein content. While up to tens of thousands of dollars can be lost per hour in delayed feedback from conventional or even “rapid” onsite lab testing, online NIR spectroscopy allows instant, continual process optimization and quality validation.


Rapid measurements of protein content are key to high-throughput sorting of soybeans at harvest, and for cost-efficient process feedback during soybean meal production. NIR reflectance measurements offer the speed and accuracy to measure protein, fat and moisture at both points in the supply chain, enabled by chemometric analysis and automation. As the food industry moves increasingly toward instant answers to improve efficiency and output, Ocean Optics and our partners are providing solutions to do so.

Making Your Own NIR Measurements of Grains and Seeds

While the InLab NIR-512 system offers an automated system complete with calibrated models for use in industry, these same type of measurements and model development could be performed with an Ocean Optics modular system consisting of the following components:

Download this Application Note