PROJECT RESULTS

Development of Hemp Food Products and Processes

Applicant: 

Shaun Crew
Hemp Oil Canada Inc.
Ste. Agathe, Manitoba  R0G 1Y0  Canada

Table of Contents:

• Background and Objectives
• Objective 1
• Objective 2
• Objective 3
• Food Chemistry and Food Safety Analysis

ARDI Project:

#98-209

Project Status:

Completed September, 2000

Background and Objectives:

The theory is that hemp seeds do not contain THC, but rather that the seed coats are contaminated by the leaf matter surrounding the seed bud. Our project purpose was to develop processes that would allow us to eliminate and/or reduce the levels of THC presently found in hemp food products and develop a process to sterilize the whole seeds and render them non-viable. Development of various processes would be the key to moving hemp food products from "niche" markets into the main stream for use by value added manufacturers and food processing companies looking to boost the nutritional content of their products.

Our approved project involved researching three particular processes to eliminate and/or reduce the levels of THC:

1. Develop a process to pre-wash the whole hemp seeds with a chemical or non-chemical agent which would breakdown the THC component and wash it away.
2. Develop and improve existing impact hulling processes to accommodate hemp seeds and completely eliminate the contaminated hulls and abnormal seeds through screen separation and air-washing, resulting in a "THC Free" hemp nut (meat).
3. Develop various "dry heat" processes to effectively sterilize hemp seeds by utilizing and improving existing grain drying, roasting and microwave technologies.

*Research note: In a separately conducted project funded in part by the Manitoba Rural Adaptation Council (MRAC) and Websar Laboratories entitled "Laboratory Analysis of THC content in Industrial Hemp Seed", it was determined that:

"THC is intrinsically found in all parts of the hemp seed, albeit at far lower levels in the hemp seed nut than that found on the seed coats which receive the highest degree of contamination."

With this knowledge, we were able to best focus on methods of reducing these levels of THC versus the complete elimination of such contaminants.

Objective 1:  Development of a process to pre-wash the whole hemp seeds with a chemical or non-chemical agent which will breakdown the THC component contaminating the hulls and wash it away.

Procedure and Project Activities:

The development of this process (and the other two processes researched) required that the D ⁹-tetrahydrocannabinol (D ⁹-THC) baseline data be collected first from six varieties of industrial hemp seed being grown in Manitoba. The varieties selected included Fedora 19, USO 14, Felina 34, Fin 314, Fasamo and Ferimon 12. The seed analyzed above was first commercially cleaned and conditioned in a seed cleaning plant.

The D ⁹-THC analytical expertise of Websar Laboratories in Ste. Anne, MB was enlisted to prepare the samples for analysis and determine the D ⁹-THC levels of the six whole seed varieties according to the Health Canada approved protocols for testing D ⁹-THC. The results are summarized in Table 1.

With the baseline data in place, the next step was to select one of the six varieties to utilize in the pre-washing process. The German seed variety Fasamo was selected for this purpose.

Numerous cleaning agents were identified, discussed and selected for testing in this study including:

• a food grade detergent
• a food grade degreaser
• a super-chlorinated solution called Oxilink used for cleaning contaminated water sources and a second double strength Oxilink solution
• a hydrogen-peroxide solution
• a food grade ethanol/water solution
• water only

Preparation of the cleaning agent solutions tested above were prepared according to the various manufacturers specifications/instructions. Triplicate samples (10 grams) of Fasamo seed were weighed out for each of the washing trials. The 10 gram sample portion of seed was stirred for 2 minutes, using a stainless steel spatula, in 250 ml of each of the test solutions. The seed was then removed from the cleaning solution and rinsed in 100 ml of HPLC grade water by stirring for 2 minutes with a stainless steel spatula. The seed was then drained and dried in a dehydrator for 30 minutes at 35ºC. The individual samples were then stored in an amber storage bottle until the D ⁹-THC analysis was performed. Duplicate analysis of each sample replicate was performed with the results indicated in Table 2.

Results and Discussion:

Table 1.  D ⁹-THC in Whole Hemp Seed – Baseline Data

* Fasamo seed was selected to utilize in the pre-washing solution tests. Note: the allowable limit by Health Canada for any hemp seed derivative is 10 m g/g (parts per million). All varieties tested for the baseline data were within this range.

Table 2. Results of D ⁹-THC Levels After Cleaning in Test Solution

n = number of replicates. Each replicate consists of a duplicate analysis of the sample.

Bar chart of the result in Table 2.

Note: The results from the double strength Oxilink test were not included above. It was postulated that this solution may have caused oxidation of the CBD also present on the seed coat which turns it into D ⁹-THC, thus increasing the level of D ⁹-THC as compared to the baseline data for the whole seed.

Conclusions:

The results of pre-washing the whole seeds in the various solutions proved that the levels of D ⁹-THC contaminating the seed coats could indeed be significantly reduced from 20% up to 88%.

The most interesting observation is that a simple pre-wash of the seeds in water only was also relatively as effective as any of the other solutions tested. The use of water as a pre-wash would obviously be the most economical cleaning agent to utilize. However, the study did not address the advantages of each test solution being utilized also as an agent to rid the seed of any microbial food contaminants.

From purely a cost perspective to reduce the levels of D ⁹-THC, the use of a food grade detergent or degreaser was most effective.

Objective 2.  Develop and improve existing impact hulling processes to accommodate hemp seeds and completely eliminate the contaminated hulls and abnormal seeds through screen separation and air-washing, resulting in a "THC Free" hemp nut (meat).

Procedure and Project Activities:

This step of the study involved the greatest amount of time and research to achieve the desired results. As discussed in the interim report, the scope of work involved in determining what processing equipment would be best suited to the hulling and cleaning of hemp seed has been no less than a "trail blazing" experience.

Processing equipment available on the market for seed hulling and seed cleaning had never been designed for working with a seed such as hemp. Industry experience dealing with such an "oily" product was limited, at best, to such seeds as sunflower or flax. The meat of a hemp seed is almost 45% oil content, which creates numerous problems in handling this sticky, oily product.

The shortcomings of presently available processing equipment became clearly evident when tested with commercial volumes of seed. Lab or pilot scale tests offered promise until a commercial volume was processed and the system clogged or gummed from a buildup of this oily product. In addition, the capacity of product processed effectively through any equipment tested was much lower than the particular equipment was rated for with other traditional seeds or grains.

The hulling of an oil seed was discussed extensively with more than a dozen seed processing equipment manufacturers including Forsberg, Carter Day, Can-Seed Equipment, Lewis M. Carter to name a few and seed hulling companies such as SnowFire Seeds, Kyle Enterprises, Northern Sales and Mikada Corporation. Assistance was also solicited from Manitoba Agriculture, Agriculture & Agri-Food Canada and the Canadian Grains Commission.

Seed processing equipment such as hullers, cleaners, aspirators and gravity tables were tested both at the manufacturers test facilities and "in house". The tests conducted included:

• Seed variety tests – 6 different hemp seed varieties were tested
• Seed moisture levels – seed varieties tested had moisture levels from 7-13.5%
• Processing temperatures & humidity – ranged from 2°C up to 20°C
• Hulling equipment chambers & hulling rings – types of impact impellers and rings
• Processing speeds – the effect of changing RPM processing speed
• Seed sizing – utilizing sized hemp seed for processing
• Screen types & sizing – various types of cleaning screens tested for sizing & separation of hulled material
• Aspiration & air-washing – effective aspiration of hulled material was tested
• Product handling – movement of hulled material through the processing line

Results and Discussion:

• The best seed varieties tested were either the French varieties (Fedora 19, Felina 34 or Ferimon 12) or the Ukrainian varieties (USO 14 or 31). The common denominator was these varieties offered larger, plumper seeds that provided the highest hit ratio in the impact hullers selected for use. Smaller seeds tend to be smashed or they would "peanut butter" going through the huller.
• The highest hit ratio through the huller was achieved with seed varieties that had a low moisture level (7-9%). Seed tested with moisture levels above 9.5% had reduced hit ratios through the huller due to the seed coat having more flexibility versus a drier seed coat that would crack open more readily. The low moisture level seeds produced hit ratios in excess of 75%, where hit ratios of seed with higher moisture levels produced hit ratios of 55-60% at best.
• The best environment for hulling hemp seed was at processing temperatures below 20° C in conditions of low humidity. Higher temperatures and humidity tend to draw more of the oil content in the hemp seed to the surface of the seed coat or seed meat making the handling of the product very difficult.
• The best equipment found for hulling hemp seed were small impact hullers that had the capability to control the processing RPM and the feed of whole seed into the impact impeller. The less aggressive the impact hulling system, the higher the hit ratio would be. Commercial scale sunflower seed hulling equipment tended to be far too aggressive for utilizing with hemp seeds. Teflon coated steel hulling rings and chambers provided the best results for hulling hemp seeds. Results from trying to hull hemp seeds with a scarifier did not produce any positive results
• Depending on the equipment being tested, a slower, less aggressive processing RPM produced the best results and hit ratio through the impact huller.
• The precision sizing of the raw input hemp seed prior to hulling produced the best results, particularly further down the processing line for cleaning and separating the hulls from the meats
• A multi-screen seed cleaner and precision sizers produced the best results for cleaning and separation of the hulls, fines and meats. Both wire screens and flat perforated steel screens were tested. The type of screen selected was dependent upon what size and part of the hulled seed was being scalped at that step in the process. The screen type and size could vary from one seed variety or seed size as compared to another.
• Effective air-washing and aspiration was one of the keys to cleaning and handling this product. This form of cleaning fines and slivers of hull from the meats provided the most gentle means of handling the product without causing degradation to the desired hemp seed meats. Screen cleaning and separation, on the other hand, caused the most problems with clogging and gumming of the product, requiring that screens be cleaned off regularly during processing
• Product handling, as with cleaning and separating the hulled seed, was best handled with air through a pneumatic air system, which treats the product in a less abrasive, gentle manner

Conclusions:

The overall best means of processing hulled hemp seed was accomplished by using smaller, less aggressive seed cleaning equipment and maximizing the use of air-washing and aspiration.

The best hit ratios through the impact huller were achieved with seed varieties that had a low moisture level between 7-9%. The drier the seed, the greater the cracking of the seed coat in the hulling chamber.

In all cases, hit ratios through the huller were determined by hand sieving hulled seed samples and separating out any whole seed that had not cracked or hulled. Weights for the whole seed (un-hulled) were compared to the weight of the entire seed sample as a percentage of un-hulled seed.

It should be further noted that none of the equipment tested was able to produce the capacity of product that the system was rated for handling with other traditional seeds or grains. The best results were always achieved utilizing equipment that offered RPM control of the processing speed and where the feed rate of product into a huller, screen cleaner, aspirator or gravity table could be controlled and reduced below the manufacturers suggested rate for feed capacity.

The use of Teflon coated screens or precision sizers made the cleanup and maintenance of these parts much easier. However, regularly scheduled cleaning and maintenance is a fact of life in handling this oily product, which can quickly buildup on machine parts, screens, feed lines, hoppers and exhaust fans.

There is much room and opportunity for the further development of cleaning and separation equipment to produce hulled hemp seed. Short of hand sifting the final product, the best percentage of clean product produced in a mechanical seed cleaning system is 95-98% clean pure hemp meats.

Although several other unconventional methods and means of producing hulled hemp seed were researched, none proved as effective in producing either a higher hit ratio or clean hulled seed product, as well as standard seed cleaning equipment with some performance modifications.

Objective 3.  Develop various "dry heat" processes to effectively sterilize hemp seeds by utilizing and improving existing grain drying, roasting and microwave technologies.

Procedure and Project Activities:

In order to sell and distribute whole hemp seed products, the Industrial Hemp Regulations require the seed to be sterilized. At present, Health Canada has only approved steam sterilization as a means to render a hemp seed non-viable and incapable of germination. The steam heating of hemp seed to render it non-viable is perhaps an acceptable method to utilize for producing hemp bird seed, but was not considered acceptable for the production of a food grade hemp product for edible consumption.

In this step of the study, alternative "dry heat" methods to sterilize hemp seeds and render them non-viable were researched including batch dryers, continuous flow dryers, roasters, microwave and e-beam technologies.

The research was conducted both "in house" and at several laboratories and processing plants including Sunny Day Products, Western Grain Dryer, Mikada Corporation, InfraReady Products and Acsion Corporation.

Control samples and processed samples of the hemp seed were submitted to the Valley Seed Laboratory to determine the seed viability before and after applying each process. Various tests of differing temperatures and processing time periods were analyzed. Health Canada will allow up to 5% abnormal and "no normal" within a tested lot. Our research attempted to determine the processing procedure required to render 100% of the tested seed sample "dead" and non-viable.

In some cases, the oil composition of the seed variety utilized was also tested before and after the processing to determine possible degradation of the essential fatty acids.

Results and Discussion:

All of the processing methods tested in this study resulted in producing sterilized seed with the exception of standard batch and continuous flow grain dryers that produced a more inconsistent roasting of the seed. Seed closer to the plenum of a grain dryer was heated much more than the seed travelling along the outside of a dryer near the air exhaust ports. Therefore, the sterilization results were also inconsistent.

Processes such as microwave or e-beam technology, although successful in rendering the seed sterilized, would in fact not be the most economically viable processing method.

The best results for sterilizing hemp seed were achieved utilizing either a coffee or nut roaster, with the processing of smaller batches (up to 300 lbs.)

Dry Roasting Table of Results

Various samples of hemp seed were roasted at a constant temperature for varied lengths of time to determine the optimum temperature and time required to render the seed non-viable. Replicate batches were completed (3) at each roast temperature and roasting time from which a sample was drawn to make up a composite test sample as follows:

* No explanation was sought to determine why only 99% of the seed was dead when 100% of the seed was dead when roasted for 5 minutes less than this sample. Both results were within the Health Canada guidelines for non-viable seed.

Various samples of hemp seed were passed under an electron beam sterilizer at the Acsion plant in Pinawa, MB to determine the optimum treatment dose required to sterilize the hemp seed. The e-beam dosage is measured in kilo-Guys (kGy). The dosage strength was selected by Acsion technicians based on work performed on similar types of seed. Test results are shown on the following table:

E-beam strengths above +25 kGy was determined to be the point at which sterilization of the seed would be most effective.

Conclusions:

The best means of sterilizing hemp seed from both a cost perspective and process perspective is by utilizing a dry heat coffee or nut roaster modified to handle the smaller sized hemp seed.

A drum roaster provides greater control and consistency of the seed temperature and the roasting period. In addition, most roasters are equipped with a cool down tray to rapidly cool the seed after roasting and avoid further cooking of the delicate proteins and oil within the seed. A roaster affords the opportunity to heat the seed up to a temperature that will render it non-viable without over heating the seed to a point where the nutritional value is "cooked".

Microwave and e-beam technologies may offer a suitable alternative process if the cost of utilizing such a process can be further reduced. One advantage of both the microwave and e-beam technologies is the much higher capacity of seed that can be sterilized at one time. However, the processing cost is still three to fours times higher than with a drum roaster, not to mention the much higher capital expenditure to procure this type of equipment.

Samples obtained from subjecting the seed to roasting in a standard grain dryer showed both damage to the seed tested (from aggressive handling) and inconsistent heating of the grain. For this reason, no viability tests were performed on seed that was subjected to dry heat sterilization in a standard grain dryer.

Research note: Since we began researching alternative sterilization methods, Health Canada has now approved the use of microwave technology as a method for sterilizing hemp seed. Roasted hemp seed has fallen into the category of a hemp seed derivative versus a sterilized hemp seed, having a process applied to it that inherently renders the seed non-viable.

Food Chemistry and Food Safety Analysis

Throughout the course of conducting this research project, numerous lot samples of products produced were analyzed to determine composition, fatty acid analysis, peroxide value, free fatty acids and food microbiology. Testing was performed both before and after processing.

These tests were conducted at the Food Development Centre in Portage la Prairie and at Norwest Labs (University of Manitoba location).

Food composition and fatty acid analysis was conducted before and after various processing to determine if the product had changed the composition and/or caused any degradation to the product (rancidity). Peroxide value and free fatty analysis were conducted to determine the quality of whole hemp seed and hemp oil’s oxidative stability.

Changes in the level of peroxidized fatty acids and free fatty acids, either through processing or over a period of time, help to determine the overall quality and stability of the product and shelf life expectations.

Food Microbiology analysis tests were also conducted to determine food safety levels of the various hemp products. Testing conducted included standard plate counts (SPC), coliforms, yeasts, moulds, escherichia coli (E. coli) and salmonella.

Data collected from these tests will be utilized for nutritional product labeling and consumer information.

Acknowledgements:

This project was made possible due to funding from the Governments of Manitoba and Canada through the Canada-Manitoba Agri-Food Research and Development Initiative (ARDI).

ARDI Home - Contact ARDI