WSU Tree Fruit Research & Extension Center

Postharvest Information Network

Sunday, March 24, 2019

WSU-TFREC/Postharvest Information Network/Matching Nitrogen Equipment to Your Needs

Matching Nitrogen Equipment to Your Needs

Introduction to Liquid Nitrogen

Interest in the use of liquid nitrogen and on-site generated nitrogen gas for use in CA storage has grown rapidly during the past years. There are basically three types of systems being tested and marketed: liquid nitrogen; on-site air separators such as pressure swing adsorption generators; and hollow fiber membrane separators. Applications of these systems have been reported.

Typical liquid nitrogen systems consist of a liquid storage tank, a vaporizer, and related controls to supply nitrogen gas, pressurized to about 40 psi, to the storage rooms. Popular tank sizes range from 6,000 to 11,000 gallons. Tanks and the vaporizer are leased from nitrogen suppliers; users provide a fenced concrete base for the tank and vaporizer, and install the supply lines and control system to the storage rooms. Injecting liquid nitrogen into the storage rooms provides some refrigeration as the liquid is vaporized, but the extra cost for insulating the supply lines to the room and potential nitrogen burns to the fruit may offset refrigeration savings.

Liquid nitrogen can establish very rapid CA with purge rates of up to 20,000 cubic feet per hour (CFH), but care must be taken to allow adequate venting to avoid a potentially catastrophic over-pressurization of the room. Additionally, very high purge rates may let purge gas escape through the vent before it is completely mixed with the storage air. Experience has shown that introducing nitrogen gas behind the refrigeration coils and venting low on the return wall assures high mixing effectiveness. Since liquid nitrogen costs substantially more than on-site generated nitrogen gas, there is a tendency to combine liquid nitrogen with dry lime or existing scrubbers, using the nitrogen only for rapid CA and lime or scrubbers for CO2 scrubbing.

Pressure Swing Adsorption Air Separators

One alternative to liquid nitrogen is the pressure swing adsorption (PSA) nitrogen generating system. Pressurized dried air is pushed through a bed of pelletized carbon material (molecular sieve) that adsorbs oxygen while allowing nitrogen to pass through. After a few minutes the bed material becomes saturated with oxygen and must be purged by releasing the pressure and venting the bed. Compressed air flow is directed through a second bed to continue the adsorption process while the first one is being regenerated. To prevent contamination of the beds, it is essential that the supply air to the beds be completely dry and free of contaminants such as oil.

A significant amount of oxygen-rich air is discharged to the atmosphere during the tank change-over period in the cycle. The amount of oxygen-rich vent air depends on the purity of the desired output stream, bed capacity, temperature, bed pressure and other factors. Typical nitrogen recovery rate is about 46% for 97% purity, 40% for 98% purity, and 32% when producing 99% nitrogen gas. A product receiver tank assures a constant product flow with a constant purity level. Typical bed pressure is 120 psi. Equipment output for CA application ranges from 3,000 CFH at 98% purity and 40-HP compressor, to 11,000 CFH at 98% purity with a 150-HP compressor.

Hollow Fiber Membrane Separators

The principle of hollow fiber membrane air separation is based on varying permeation rates for different gases, such as oxygen, carbon dioxide and nitrogen, through the membrane. The hollow fiber membrane-type gas separator consists of bundles of small hollow polymeric fiber tubes in a vessel (Figure 1).

Compressed air is fed into the separator, pressurizing the inside of the individual tubes. As the air flows through the tubes, oxygen and carbon dioxide permeate through the fiber walls faster than the nitrogen. This oxygen-rich stream is vented to the outside. The non-permeate gas (nitrogen) outflow stream is piped into the storage room. By changing the output flow rate, the purity of the output can be controlled. Nitrogen purity levels of up to 99.9% can be obtained, but for CA applications purity levels of 97% to 99% are usually used.

As is the case with PSAs, membranes have a substantial oxygen-rich permeate stream that is vented to the atmosphere. Operation of the system is very simple, and the separator itself has no moving parts. To prevent contamination of the membrane it is essential that the supply air be completely clean and absent of contaminants such as oil. As feed air temperature is increased, product output also increases, but feed air temperatures must be accurately controlled as high feed air temperatures shorten the life of the membrane.

Feed air conditioners are often used for membranes in order to control feed air temperatures and to maximize product flow. Carbon dioxide and water vapor also permeate through the membrane walls, thus it can also be used for scrubbing by recirculating air from CA rooms. This patented process by NICAP utilizes two membranes, one to recirculate room air to scrub CO2 from the gas and another one to produce nitrogen from fresh air to make up for the permeate stream lost from the first membrane.

Principles of a Purge System to Establish CA

Purging CA rooms with a nitrogen stream is a safe and reliable method for establishing a low oxygen atmosphere. One major advantage a nitrogen purge system has over "burners" is that it does not produce carbon dioxide which must be scrubbed from the purge stream going into the room. Use and optimization of nitrogen gas purge systems using simulation techniques have been reported. The principle of nitrogen purging is illustrated in Figure 2.

Nitrogen gas is introduced into the room where it immediately mixes with the room air. An equal volume of room air containing oxygen is vented out, thus the oxygen level of the room decreases constantly. The amount of reduction in the oxygen level is related to the number of exchanges of the void storage volumes and the nitrogen purity of the purge stream (% nitrogen in purge stream), as shown in Figure 3.

The oxygen pulldown rate is highest during initial purging when the difference between the oxygenlevel of the room air and the purge stream air is largest. As this difference becomes smaller, oxygen pulldown becomes slower and less efficient.

The higher the purity level of the purge stream, the faster the room oxygen level is pulled down for an equal purge stream volume. For example, with a 100% nitrogen purge stream and perfect mixing of the gases in a room before the mixture is vented out, the oxygen concentration is reduced to about 7% with the first exchange volume. After the second volume, the oxygen level will have decreased to less than 3%. On the other hand, with a 98% N2 purity (2% O2) purge stream the rooms oxygen level is reduced only to 9% with one volume changed and 4.5% with two volumes exchanged.

The output of nitrogen generators (Figure 4) is inversely related to nitrogen purity of the out-flow stream. For example, typical specifications for a 40-HP hollow fiber membrane separator (Membrane B) using 100% fresh air are: 3,070 ft3 per hour at 95% nitrogen, 2,200 ft3 at 97% nitrogen and 1,190 ft3 at 99% nitrogen. Another manufacturer claims significantly higher output rates (Membrane A). This equipment should first be operated at a high output setting, then as the oxygen level in the room is lowered the output stream flow rate must be reduced gradually to increase its purity.

Output rates for PSA generators in general do not vary as much with purity of the outflow stream as those for the hollow fiber membrane generators (Figure 4). Typical specifications for a PSA unit provided by one manufacturer are: 3,700 ft3 per hour at 95% nitrogen, 3,300 ft3 at 97%, and 2,570 ft3 at 99% nitrogen with an approximate power input of 40 HP.

Estimating Oxygen Pulldown Time

In a purge system the mathematical relationships between the purge gas flow rate, initial and final oxygen concentration, purge gas oxygen level, and void storage volume are well known and, therefore, can be used to calculate the time required to achieve oxygen pulldown to a specific level. The equation is:

T = (V/F)ln((Ci-Cp)/(Cf-Cp))


T = time, hours
V = void volume of storage, ft3 (approx. 65% of empty volume)
F = purge gas flow rate, CFH
Cp = purge gas oxygen concentration, %
Ci = initial storage room oxygen concentration, %
Cf = final storage room oxygen concentration, %

This equation does not take into account the oxygen used by fruit respiration and potential purge inefficiency due to incomplete mixing of purge gas in the room.

A quick way to estimate oxygen pulldown time is to use the nomograph shown in Figure 5.

For example, estimate the pulldown time to 3.5% oxygen for a 2,000 bin room using an air separator with an output of 8,000 CFH at 98% N2. On the nomograph, draw a horizontal line from the 3.5% final room oxygen level to intersect the 98% purge gas purity line. From this point go down vertically to the 2,000-bin line, then draw a horizontal line to the 8,000 purge gas flow rate line. From there, go down vertically to read a pulldown time of 24 hours on the bottom line.

The nomograph is based on zero respiration. A substantial reduction in pulldown time can be expected due to fruit respiration. Figure 6 illustrates the effect of fruit respiration on oxygen pulldown based on mathematical material balance method by Malcolm and Beaver (1989). Proportional time reduction can also be expected with void volumes lower than 38 ft3 per bin.

Scrubbing with a Purge Stream

The operating principles used for reducing oxygen in air also apply to scrubbing of carbon dioxide. In this case, accumulated carbon dioxide from fruit respiration is being removed instead of oxygen. Some operators scrub each room on an intermittent basis, such as 2 or 3 times a day. Others purge the rooms continually with a nitrogen stream, maintaining a slight pressure in the rooms. This reduces oxygen infiltration into the CA rooms. For example, if a desired CO2 level is 2.5%, the concentration may be permitted to accumulate to 2.6% before purging. Purging is then stopped when the CO2 level is reduced to 2.4%. If the purge stream contains no CO2, the reduction in CO2 in a storage room follows the 100% nitrogen curve (Figure 3).

If a purge stream comes from an air separator, PSA or membrane, it will contain some oxygen. To optimize equipment performance, operate at the highest oxygen concentration possible when purge-scrubbing CO2. If there is no oxygen infiltration into the storage room, it should be possible to operate with an oxygen concentration equal to the desired room air oxygen plus the CO2 concentration. For example, in a room with a 1% O2 and 2.5% CO2 concentration, a purge stream containing 3.5% oxygen could be used under ideal operating conditions. However, because of changing atmospheric conditions and "leaky" rooms, oxygen infiltration is common and, in general, oxygen concentration in the purge gas needs to be less than the sum of O2 and CO2 concentrations desired in the room, but it usually can be higher than the oxygen setpoint for the storage room.

Equipment Application and Selection

The basic question to answer when selecting a purge system for CA is what capacity, in cubic feet per hour, and related purge stream purity is needed for a specific storage capacity. Installing the correct system capacity is important. Not having enough capacity during oxygen pulldown and scrubbing during the first few months in storage can be disastrous. On the other hand, some excess capacity is desirable and not very costly. For any CA system, such as air separators (PSA, Membranes) alone or in conjunction with lime and/or scrubbers, a system analysis must be made that takes into consideration equipment needs for 1) initial oxygen pulldown and 2) scrubbing of carbon dioxide. The highest output demands are toward the end of the room-filling period when rooms still need to be pulled down and high respiration rates demand high scrubbing requirements. The following analysis pertains to purge systems only.

Oxygen pulldown. Major factors affecting purge gas requirements for oxygen pulldown are total void volume to be purged, desired final oxygen level and desired pulldown time. Other factors to consider are purge gas purity and fruit respiration. Required purge gas capacities are directly proportional to the room size, desired final oxygen level, and how fast you want to establish CA conditions. For example, pulling a room down to 3% oxygen takes about 1.6 times as long as lowering it to only 5% oxygen. The nomograph (Figure 5) can be used for a quick determination of the amount of nitrogen needed for oxygen pull-down. A more accurate estimate can be obtained using Table 1.

Table 1. Estimated purge volume needed per bin for oxygen pulldown to various oxygen levels for purge gas purities show

Purge gas required, ft3 per bin using purge gas nitrogen purities of the following
Desired oxygen level 100% 99% 98% 99% 98%
  without respiration effect with respiration effect

For example, estimate the purge gas flow rate needed to reduce the oxygen level in a 2,000 bin room to 3% in 20 hours using a nitrogen stream of 98% purity. The required purge gas volume is 111 ft3 per bin excluding respiration or 88 ft3 per bin including respiration (Table 1). Thus, the purge rates are: 2,000 bins x 111 ft3 per bin/20 hours = 11,100 CFH excluding respiration or 2,000 bins x 88 ft3 per bin/20 hours = 8,800 CFH including respiration. This example clearly illustrates the importance of fruit respiration. However, you must also consider factors such as purging inefficiency and high void volumes that tend to increase expected pulldown time. Scrubbing of carbon dioxide. Purge gas consumption rates f or various respiration rates and from CO2 levels are shown in Table 2.

Table 2. Estimated daily purge gas consumption at 98% N2 purity per bin (880 lbs fruit) for scrubbing CO2 to maintain the indicated CO2 concentration for respiration rates shown.

Respiration rate % CO2 increase/day Purge gas consumption, ft3/bin per day for room CO2 setpoint of

Daily purge gas requirements are directly proportional to the respiration rate of the fruit expressed in percent increase in the CO2 level in a 24-hour period in a tightly sealed room. Purge gas consumption is also greatly dependent upon the CO2 setpoint. For example, purge gas requirements double when reducing the CO2 setpoint from 2% to 1% and almost double again when lowering it from 1% to 0.5%. Varying the purge stream purity (within allowable limits) has a negligible effect on purge gas requirements. Total purge gas requirements, CFH, can be estimated by multiplying the indicated daily consumption times the number of bins and dividing by 24 hours. For example, to scrub a room with fruit having a respiration rate of 0.4% per day, maintaining a CO2 setpoint of 2.0%, you need 7.6 cu ft purge gas per bin per day. To scrub 20,000 bins, the required purge gas rate in ft3 per hour is:
CFH=20,000 bins X 7.6 ft3/bin/24 hours=6,330 CFH.

Sizing Equipment

An important factor to consider when selecting equipment is the need to establish CA for some rooms at the same time other rooms require scrubbing. If the rooms are sealed on alternate days and you can pull a room down in one day, then it is possible to scrub every other day. It is also important to consider the higher respiration rates during the first month of storage. Usually the highest purge requirements occur toward the end of the pulldown season when respiration rates are still high. Extra scrubbing needs could be met by putting some lime into the rooms and/or using existing scrubbers.

Respiration data for fruit in CA storages are not readily available. Survey data collected at Cornell University indicate that the maximum expected CO2 production in CA rooms would be about 0.7 to 0.8% per day. Grower experiences in Washington are somewhat lower, about 0.6% CO2 per day after initial pulldown, then reduce to 0.2%-0.3% after a few weeks of storage.

Example of equipment selection. For simplicity, let us assume that a CA storage complex includes ten 2,000-bin rooms with one room being sealed every Monday, Wednesday and Friday. Expected respiration rate of the fruit is 0.6% CO2 per day during the first 2-3 weeks in storage, and then drops to 0.3% in another 3-4 weeks. Oxygen pulldown to 3.5% is desired. We now can set up a room-filling schedule and calculate required output capacity needed for pulldown and scrubbing. Table 3 includes the room-filling schedule and equipment time requirements for oxygen pulldown and scrubbing.

Table 3. Room filling schedule, oxygen pulldown and scrubbing for a 20,000-bin storage complex, using a 10,000 CFH at 98% N2 output air separator and filling 3 rooms per week. (Pulldown to 3.5% O2; respiration rate of 0.6%/day, maintaining 2% CO2 setpoint).

DayPulldown bins (#)Scrubbing bins (#)Time for pulldown (hrs)Scrubbing (hrs)Operation (hrs/day)
12,000 19 19
2 2,000 2.32.3
4 4,000 4.64.6
6 6,000 6.96.9
7 6,000 6.96.9
98,000  1111
11 10,000 15.615.6
13 12,000 20.320.3
14 12,000 13.813.8
16 14,000 24*24
18 16,000 5*24
20 18,000 24*24
21 18,000 24*24
22 18,000195*+24
23 20,000 24+24
Next 22 days  2424 
Next 135 days 12**12  
* Scrubbing completed next day
+ Equipment output capacity limited: CO2 level in rooms may rise temporarily.
** Respiration rate reduced to 0.3% CO2 per day.

A shortage of purge gas may develop toward the end of harvest when scrubbing requirements are at a peak and rooms still need to be pulled down. By that time, the respiration rate of the rooms filled at the beginning may have substantially decreased, reducing the scrubbing requirement. Also, CO2 levels in the rooms may rise slightly which, in turn, reduces purge gas requirements for scrubbing until all the rooms are filled. Another possibility is to add some lime to the rooms to reduce overall scrubbing requirements or to operate existing scrubbers during the first few weeks. If two rooms are sealed on the same day, it may be possible to cascade from one room into the other to reduce total purge gas requirements for the two rooms. For long-term storage the equipment output is more than adequate.

System Costs

Blanpied and Bartsch (1988) reported a cost analysis for establishing low oxygen atmosphere and maintaining 3% CO2 for a 180-day storage season comparing four systems: 1) COB and lime, 2) COB plus carbon scrubbers, 3) air separator plus lime and 4) air separator plus carbon scrubber. Based on manufacturer data and operator information from the state of New York, estimated total cost per bin for establishing CA at 5% CO2 and maintaining 3% CO2 for a 180-day storage season were: $1.72/bin for COB and lime, $1.80/bin for liquid N2 plus lime, and $2.12/bin for air separator plus lime. Costs were slightly higher when a carbon scrubber was used instead of lime.

Data for cost calculations of pressure swing adsorption air separators for Pacific Northwest storages are shown in Table 4.

Table 4. Estimated fixed annual costs for on-site nitrogen generating equipment (PSA) used for establishing and controlling CA for a 180-day storage period.1
Equipment Output, CFH
at 98% N2
Fixed cost per year lease2 Total bin capacity3 Estimated cost/bin
FixedOperating at $.05/KWhTotal cost
PSA, 60 HP4,50021,6009,0002.400.673.07
PSA, 75 HP6,00024,00012,0002.000.672.67
PSA, 100 HP7,50026,00015,0001.760.672.43
PSA, 125 HP10,00028,20020,0001.410.672.08
PSA, 150 HP11,00031,,20022,0001.410.672.08
1Cost shows for $.05/KWh. Subtract $.40/bin with power cost of $.02/KWh; add $.67/bin with power cost of $10.KWh. Membrane with same cost and power, but with about 10% less output when operated in 100% purge mode would have about 10% higher total cost/bin. Not enough data are available for membranes operated in recirculate mode to estimate cost.
2Estimated typical 5-yr lease from manufacturer, repair and major maintenance included. Lessees must replace filers and oil as per manufacturer's recommendations.
3Based on rapid oxygen pulldown to 3% and scrubbing to maintain a 2% CO2 level, with a respiration rate of 0.6% per day during the first 6 weeks, then 0.3% for the remainder of the storage season. Room filling schedule: not more than 35% of storage total capacity per week.

Typical equipment sizes for this area range from 40 to 150 HP. Equipment output is based on the manufacturer's specifications. Fixed costs used are annual payments for equipment lease (approximately 20% of purchase price). Operating costs are based on KW rating of compressor and estimated number of hours of operation each day, similar to the schedule shown in Table 3. Total bin capacity is based on the room filling-schedule shown in Table 3, rapid oxygen reduction and scrubbing of rooms to maintain a 2% CO2 level with a respiration rate of 0.6% CO2 rise per day during the first 45 days, then 0.3% per day for the remainder of the storage season. Total estimated costs per bin vary from $3.07 for 60-HP units to $2.08 for the 125 to 150-HP units with an electric power rate of $.05 per KWh. Total costs are $0.40 per bin lower with a power cost of $0.02/KWh, or $0.67 per bin higher with a power cost of $0.10/KWh. Labor costs are not included.

One option available for reducing oxygen pulldown and maintaining CA conditions is to combine the use of an air separator with dry lime for some of the CO2 scrubbing. For example, it is possible to increase the number of bins in Table 3 to 30,000 by adding enough lime to the rooms to reduce the air separator scrubbing requirements for the first 6 weeks. This procedure reduces the fixed cost/bin (Table 4; 10,000 CFH Output) to $0.94 and reduces the operating cost to $0.55/bin but would add an estimated lime cost of $0.20/bin, for a total cost of $1.69/bin. Further cost reduction can be obtained by increasing the number of bins to 40,000. This will reduce the fixed cost to $0.70/bin. More lime will be required to do about 60-70% of the CO2 scrubbing during the first 6 weeks at an estimated cost of $0.30-0.40/bin. The air separator would run continuously during the entire storage season with a power cost of $0.51/bin and an estimated total cost of $1.56/bin.


The use of on-site nitrogen generating equipment for establishing and maintaining desired atmospheres in CA storages is rapidly increasing. This equipment offers increased operational flexibility and practices that may provide additional benefits to growers. Hollow fiber membrane and pressure swing adsorption (PSA) air separators offer an effective alternative to traditional hydrocarbon-fueled nitrogen gas generators.

These systems need to be properly engineered for a specific storage application taking into consideration oxygen pulldown and carbon dioxide scrubbing requirements and the rate of fruit being harvested and brought into storage.

These systems must be properly operated and managed to fully utilize the oxygen pulldown and carbon dioxide scrubbing capability of the equipment.

Oversizing equipment to prevent bottlenecks during the room filling time when pulldown and scrubbing requirements are highest requires little additional cost.

Research is needed to determine the effect of using a purge system on the storage environment and quality of the stored fruit because of volatile removal during purging.

Consult the authors of this article for further information.

The law requires that pesticides be used as the label directs. Uses against pests not named on the label and low application rates are permissible exceptions. If there is any apparent conflict between label directions and the pesticide uses suggested in this publication, consult your county extension agent.

Use pesticides with care. Apply them only to plants, animals or sites listed on the label. When mixing and applying pesticides, follow all label precautions to protect yourself and others around you. It is a violation of law to disregard label directions. If pesticides are spilled on skin or clothing, remove clothing and wash skin thoroughly. Store pesticides in their original containers and keep them out of the reach of children, pets and livestock.

Dr. Henry Waelti, Extension Agricultural Engineer and Dr. Ralph P. Cavalieri, Assistant Professor

Washington State University, Pullman, WA

Tree Fruit Postharvest Journal 1(2):3-13
August 1990

Tree Fruit Research & Extension Center, 1100 N Western Ave, Washington State University, Wenatchee WA 98801, 509-663-8181, Contact Us