ARGONNE NATIONAL LABORATORY

Study of a Dry Room in a Battery

Manufacturing Plant using a Process Model

Shabbir Ahmed (Corresponding Author)

Chemical Sciences and Engineering Division, Argonne National Laboratory, Bldg. 200,
9700 S. Cass Avenue, Argonne IL 60439 USA
[email protected], Tel: (630) 252-4553, Fax: (630) 972-4553

Paul A. Nelson
Chemical Sciences and Engineering Division, Argonne National Laboratory, Bldg. 200,
9700 S. Cass Avenue, Argonne IL 60439 USA
[email protected]

Dennis W. Dees
Chemical Sciences and Engineering Division, Argonne National Laboratory, Bldg. 200,
9700 S. Cass Avenue, Argonne IL 60439 USA
[email protected]

© 2016. This manuscript version is made available under the Elsevier user license
http://www.elsevier.com/open-access/userlicense/1.0/
Abstract
The manufacture of lithium ion batteries requires some processing steps to be carried
out in a dry room, where the moisture content should remain below 100 parts per
million. The design and operation of such a dry room adds to the cost of the battery.
This paper studies the humidity management of the air to and from the dry room to
understand the impact of design and operating parameters on the energy demand and
the cost contribution towards the battery manufacturing cost. The study is conducted
with the help of a process model for a dry room with a volume of 16000 cubic meters.
For a defined base case scenario it is found that the dry room operation has an energy
demand of approximately 400 kW. The paper explores some tradeoffs in design and
operating parameters by looking at the humidity reduction by quenching the make-up
air vs. at the desiccant wheel, and the impact of the heat recovery from the desiccant
regeneration cycle.

Keywords
Dry room, lithium ion battery, battery manufacturing, humidity control

Page 1 of 35
Table of Contents
1. Introduction ..................................................................................................................... 3

2. Dry Room Process Model ................................................................................................ 4

3. Results and Discussion..................................................................................................... 6

3.1. Moisture Entering the System and Its Removal .......................................................... 9

3.2. Effectiveness of the Heat Exchanger ......................................................................... 11

3.3. Cooling the Make-Up Air ........................................................................................... 12

3.4. Ambient Air Conditions ............................................................................................. 15

3.5. Purge Rate ................................................................................................................. 16

3.6. Other Parametric Sensitivities ................................................................................... 18

3.7. A More Favorable Case.............................................................................................. 19

4. Conclusions ......................................................................................................................... 21

5. Abbreviations ...................................................................................................................... 23

Acknowledgments................................................................................................................... 23

Bibliography ............................................................................................................................ 25

Page 2 of 35
 1. Introduction
Increasing demand for lithium ion batteries (LIB) [1] and the resultant demand for

reliable dry rooms [2] have spurred numerous suppliers offering dry room technologies.

The trend is toward larger manufacturing plants to benefit from economies of scale [3].

These plants seek to squeeze out inefficiencies and cost factors. The dry room

represents a step in the manufacturing process where the energy demand is very high

because of the large volume of air that needs to be temperature controlled and dried.

A dry room is an essential part of the manufacturing plant for lithium ion batteries

[4] [5] [6]. Here the cells are filled with the electrolyte which is very sensitive to

moisture (e.g., lithium hexafluoride reacts with water) and sealed in an environment

with moisture concentrations below 100 parts per million by volume (ppmv). Small

variations in the moisture content can affect the capacity and/or the cycle life of the

cells produced [7].

The dry room environment is maintained by designing a leak-proof contained

volume, where the incoming air contains very low moisture content (say 15 ppm or

0.066 grains per pound (gpp)), such that the exit gas does not exceed the specified

upper limit (say 100 parts per million volume (ppmv) or 0.44 gpp). The air flow rate

through the dry room is controlled with sensors to maintain the exit gas at a moisture

content of 100 ppm or less. Management of ventilation through a combination of vent

placement and air flows ensures rapid resolution of any occasional spikes in moisture at

key spots in the room [8]. The room enclosure is kept leak-proof which includes use of a

Page 3 of 35
vapor barrier seal on the concrete floor followed by an epoxy finish coat. Some key

parameters in the design of the dry room include the volume of the room, the allowable

moisture concentration, and the amount of moisture released from the materials

entering and personnel working in the room [9].

The objective of this paper is to estimate the energy consumption and the cost

contribution to the battery manufacturing plant from operating a dry room. A generic

system diagram has been proposed to enable a reasonable estimate of the cost and

energy demand of operating a dry room.

2. Dry Room Process Model

This study was conducted for a dry room in a battery manufacturing plant that will

produce 100,000 packs of automotive lithium ion batteries (LIB). The plant equipment is

amortized over 6 years. The dry room is assumed to have a volume of 16,000 m3.

Page 4 of 35
 Make-Up Air
9°C, 50 gr/lb H2O 24°C, 3.0 gr/lb 426 kW
33°C, 50% RH A6
152 kW
Cool
Pre-Cool Blend Filter A5
A1 A2 A3 A4
57 kW 10°C, 3.0 gr/lb
25°C A11
8 kW
100 ppmv, 0.44 gr/lb 41,000 scfm

Dry Room 250 kW

14°C
A10 5% B7 A7 95%
0.066 gr/lb
63 kW Post-Heat 12°C Desiccant
0.066 gr/lb Wheel
A9 10°C
A8
Post-Cool
0 kW
10°C B8 16°C
Discharged Air
146°C B11 3.2% H2O
Regen 122°C B12 B14
Heat B10 142°C 36°C
30 kW
B13
B9 122°C 8 kW

Ht
.E
xc
h.
Figure 1. Schematic of the air management for the Dry Room

Figure 1 is a schematic of the management of the air supply for the dry room,

including some assumptions, input parameters, and calculated results for a base case

system. The make-up air at 33°C, 50% relative humidity (RH) or 2.5 vol.% moisture, is

pre-cooled to 9°C to drop out some of the moisture and then blended with the return

air from the dry room. The combined stream emerges at a temperature of 24°C and a

moisture content of 0.07 vol.% (3 gpp). This stream is filtered and cooled down to 10°C,

and then split with 95% going towards the dry room (A7-A10), the balance (purge

stream) to be eventually discharged (B7-B14) from the system. A low purge rate (5%) is

advantageous in that it lowers the heating and cooling loads in the system, and is

Page 5 of 35
discussed in Section 3.5. The dry-room stream is passed through a desiccant wheel

where the moisture content is reduced to 15 ppmv (0.066 gpp). This gas is then heated

or cooled as needed such that the dry room exit air is at 25°C. For this base case

scenario, the dry room inlet temperature is 14°C. Within the dry room the air picks up

moisture from the personnel, from the negative electrodes that come in with moisture

content, and the opening of the airlock doors. The air flow rate and its inlet temperature

are calculated such that the return air is at 25°C and with a moisture content of 100

ppmv (0.44 gpp), or less. For the base case set of conditions, the inlet air is at 14°C and

the air flow rate is 20 m3 s-1 (41,000 ft3 min-1). This air flow rate corresponds to a

residence time of 13.6 minutes within the dry room.

The desiccant wheel is regenerated by passing the discharge air stream heated to

146°C. The discharge air stream flows multiple times through the desiccant wheel to

heat and cool the desiccant wheel and to allow for recovery of some of the sensible

heat, as shown in the figure. Table 1 lists the assumptions and input parameters used in

calculating the results shown in Figure 1.

Table 1. Assumed / Input Parameters in Dry Room Process Model (Base Case)

3. Results and Discussion

With the dry room inlet and exit air streams stipulated to contain 15 ppm (0.066

gpp) and 100 ppm (0.44 gpp) of moisture, respectively, the air flow needed through the

Page 6 of 35
dry room is calculated to be 19.6 m3 s-1 (41,600 ft3 min-1). This flow represents a

residence time or volumetric air turnover every 13.6 minutes. With the heat generated

in the dry room assumed to be 250 kW, the inlet air temperature needs to be at 14°C to

ensure that the air leaving the dry room is at the specified 25°C.

Table 2. Moisture Content and Energy Demands in the Process (Base Case)

Table 2 shows the heating and cooling loads for the steps in the process. The largest

cooling load of 426 kW is at the Cool station (A5-A6). The combined cooling load at

three stations (Pre-Cool, Cool, and Post-Cool) is 483 kW. Assuming a coefficient of

performance of 3.5, the electric power needed for the heat removal is 138 kW. A

heating load of 30 kW is needed to regenerate the desiccant wheel. Electric power is

needed for the blowers (167 kW) and refrigeration (138 kW), adding up to 305 kW. The

thermal energy requirements are at Post-Heat (63 kW) and for the discharge air heating

for zeolite wheel regeneration (30 kW). The total energy requirement for the operation

(the sum of the thermal and electrical) is 398 kW. If we consider that the electric power

from the grid is generated from natural gas (NG) with an efficiency of 40%, then the

energy required for the system becomes ((138+167)/0.4 + 93 =) 856 kW. Using the EPA

greenhouse gas equivalencies calculator [10], the greenhouse gas emissions from the

generation of the thermal and energy usage for the dry room air management converts

to 5.4 MT of CO2 Equivalent. These estimates do not include the inefficiencies of heat

transfer in the dry room system.

Page 7 of 35
 The modeling results were used to estimate the cost of the capital equipment and

raw materials, while the labor requirement for the baseline plant was estimated from

private communications with knowledgeable individuals. These are shown in Table 3.

Table 3. Costs for the Base Case System

The cost of the dry room operations was based on an algorithm defined by Peters and

Timmerhaus [11], which also provides a range of percentages. Table 4 lists the

estimated percentages for the cost assessment in column 2. The annual cost of

operations was then calculated by adding the Total Operating Cost to the annualized

total capital investment.

The cost of the dry room operations was estimated to be $3.5 M per year, or $35 per

pack, or $1270 per m3 of the dry room, for the plant producing 100,000 battery packs

per year. For reference, Argonne’s BatPaC model [12] [13] estimates the cost of a 10

kWh battery for a Plug-in Hybrid Electric Vehicle (PHEV) producing 100K packs per year

is $3100 per pack.

Table 4. Assumed Input Parameters in Cost Estimate

To gain perspective on the effect of inaccuracies in the assumed values in

preparing the estimate, an alternative estimate was obtained by using a range of the

percentages for the parameters in Table 4, which are shown in column 4. A random

Page 8 of 35
number generator was used to generate a percentage value within the range for each

cost item to calculate the cost of dry room operations. This was repeated over 1000

times, to yield an average value of $38 per pack, with a standard deviation of $2. The

results were spread between a maximum and minimum of $46 and $31, respectively,

with a frequency distribution as shown in Figure 2.

Figure 2. Frequency distribution of cost calculations using randomized percentages from Table 4, column 4.

3.1. Moisture Entering the System and Its Removal

Page 9 of 35
 The moisture mainly comes into the system with the make-up air. At 33°C and

50% RH, water vapor enters the system at a rate of 19.4 g s-1 (1.08 mol s-1, 2.5 vol.%).

The Pre-Cool station cools the gas to 9°C and condenses out 10.6 g s-1 of moisture. The

second source of moisture into the system is with the materials and personnel entering

the dry room, adding 1.23 g s-1. Blending the make-up air with the return air from the

dry room (containing 100 ppmv H2O or 0.066 gpp), results in the mixture with a

moisture content of 682 ppm (3 gpp). The Cool station reduces the mixture temperature

to 10°C (saturated vapor pressure is 1.2 vol.%) but because of the already low moisture

content, no further condensation occurs at this station. The main air stream (A7)

undergoes the final moisture removal at the desiccant wheel, where 9.5 g s-1 (0.53

mol s-1) H2O is adsorbed, dropping the moisture level in the air to 15 ppmv (0.066 gpp).

The air emerging from the desiccant wheel is warmed up to ~12°C because of the

heat of adsorption that is generated when the moisture is adsorbed onto the desiccant.

This air stream is then heated to 14°C before entering the dry room.

Within the dry room, moisture is released by the personnel and negative electrodes,

while some moisture enters during the opening of the air lock doors. This breakdown is

shown in Table 5.

Table 5. Moisture Source in Dry Room

Page 10 of 35
 The discharge stream from the splitter carries 0.51 g s-1 (0.028 mols s-1) of moisture

and extracts the 9.5 g s-1 of moisture that was adsorbed on the desiccant wheel. The

discharge stream exits the system at ~36°C with 10.0 g s-1 of moisture (1.3 vol.%). The

saturated vapor pressure at 36°C is approximately 6%. The material balance for the

moisture entering and leaving the system is shown in Table 6.

Table 6. Moisture Entering and Leaving the System

3.2. Effectiveness of the Heat Exchanger

Figure 1 shows a heat exchanger being used to recover part of the energy needed to

regenerate the desiccant wheel. With an approach temperature of 20°C, the heat

exchanger requires a heat transfer area of 2200 m2. Eliminating the heat exchanger

would simplify the process but it would increase the heat load on the air heater by a

factor of more than 5, from 30 to 168 kW. The feasibility of a smaller heat exchanger

was explored by studying the effect of the approach temperature by varying it from

20°C to 132°C (no heat exchanger needed). Figure 3 shows that with increasing

approach temperature, the heat load needed to raise the discharge air temperature to

146°C (for desiccant regeneration) increases linearly. The larger approach temperature

allows for a smaller heat exchanger, as shown by the non-linear decrease in the heat

transfer area requirement. The cost of the dry room operations ($ per pack) decreases,

and is lowest without the heat exchanger.

Page 11 of 35
 Figure 3. The heat exchanger size and cost of natural gas determines the cost of the dry room operations.

-1
Natural gas = 2 cents kWh .

3.3. Cooling the Make-Up Air

The Pre-Cool station cools the incoming make-up air to condense out a significant

portion of the moisture in the make-up air. For the base case scenario, the heat removal

load for cooling to 9°C in this step is 57 kW. The extent of cooling at this station has

ramifications at points downstream. A look at the effect of the temperature setting at

the Pre-Cool station was conducted by varying the temperature from 1° to 33°C. The

effect is shown in Figure 4 and Table 7.

Lowering the Pre-Cool temperature from 33°C (no precooling) changes the cooling

load at the Pre-Cool station from zero to 76 kW at 1°C, and at the Cool station from 457

to 416 kW. The combined cooling load increases by 8%, from 457 kW at 33°C to 492 kW

at 1°C.

Page 12 of 35
 As the Pre-Cool setting is lowered, the amount of heat needed at the Post-Heat

increases from 20 to 78 kW. The total thermal energy load increases by 96% (54 to 106

kW) when the Pre-Cool temperature is reduced from 33°C to 1°C. The Total Energy

Need represents the sum of the thermal energy load and the power needed for

refrigeration (=Cooling Load / Coefficient of Performance). The Total Energy Need

increases 18%, from 351 to 414 kW as the Pre-Cool setting is reduced from 33 to 1°C.

There appears to be a small increase in the cost of operations as the Pre-Cool

temperature is lowered. However, the lower Pre-Cool temperature settings reduces the

burden of removing moisture at the zeolite wheel, which can then turn slower and also

benefit from a smaller temperature differential as a result of the heat effects of less

adsorption and desorption.

Figure 4. Effect of the Pre-Cool Temperature Setting on the Heating and Cooling Needs

Page 13 of 35
 Table 7. Effect of the Pre-Cool Temperature on Heating and Cooling

The distribution of water removal from the system as a function of the Pre-Cool

temperature is shown in Table 8. At 9°C, 55% of the water entering with the make-up air

(51.5% of the total water entering the system) is removed at the Pre-Cool station. The

balance is removed with the purge stream (2.5%) and via the desiccant wheel (46.0%),

ultimately leaving with the discharge stream (B14). With a setting of 33°C, there is no

water removal at the Pre-Cool station. Thus all the water is removed via the desiccant

wheel (94.9%) and the purge stream (5.1%). The trends of the water distribution at the

different temperatures are shown in Figure 5.

Table 8. Mass balance for water at two Pre-Cool temperature settings.

Figure 5. Distribution of Water into and out of the process as a function of the Pre-Cool temperature.

Page 14 of 35
 3.4. Ambient Air Conditions
The main function of the operations supporting the dry room is to ensure that the

moisture content brought in with the ambient make-up air and the moisture that is

introduced / generated within the dry room is reduced to very low levels. For the Base

Case conditions, the moisture coming in with the make-up air is overwhelmingly higher

(by 2 orders of magnitude) than from inside the dry room. Furthermore, the

temperature of the ambient air also varies with the plant location and season.

The effect of the ambient air conditions was studied by varying the temperature

from 40°C to -17.8°C (104°F to 0°F). The RH was held constant at 50% for all

temperatures. With lower temperatures and lower moisture content, the cooling load at

the Pre-Cool station diminishes and becomes zero at 9°C (the Pre-Cool set temperature).

At temperatures below 9°C, the load at the Cool station is also reduced. The drier air

entering the desiccant wheel generates less heat of adsorption and so the gas exits (A8)

at a lower temperature. The effect of the ambient air temperature on the heating and

cooling demands, as well as the overall energy demand for the dry room operation are

shown in Figure 6 and Table 9. Over the range of temperature studied, the Total Energy

Need is found to pass through a minimum value of 385 kW when the ambient air is at

20°C; at still lower temperatures more energy is needed at Post-Heat to meet the dry

room inlet temperature requirement.

The effect on cost is also shown in Figure 6, where the cost of electricity decreases,

while the cost of natural gas increases with decreasing ambient temperatures. The total

Page 15 of 35
cost of operations ($ per pack) is seen to decrease sharply from 40°C to 20°C, and then

more gradually, reflecting the slopes of the other energy and cost curves.

Figure 6. Effect of the Ambient Air Temperature on the Energy Demands and Costs (RH=50%).

Table 9. Effect of Ambient Air Temperature Heating and Cooling Needs, and Costs (RH=50%).

3.5. Purge Rate

Purging is necessary in a recirculating system to limit the buildup of some species. In

the case of the dry room, moisture and carbon dioxide are generated. Both these

species are removed effectively by the molecular sieve used in the desiccant wheel, as

these molecular sieves are also used for drying and CO2 removal from natural gas [14].

The purge flow ensures that the concentrations of any other chemical species that are

not removed via condensation or the molecular sieve do not build up too much.

The purge rate determines the amount of make-up air needed and, therefore, is a

significant determinant in the energy needs and the cost of operations. The effect of the

Page 16 of 35
purge rate was studied by varying it from the base case of 5% to 20%. A higher flow

associated with increasing the purge / make-up air increases the heat removal loads.

The larger moisture content entering the system increases the heat effects of the zeolite

wheel, and the heat required for regeneration. Thus, increasing the purge rate increases

the energy demand (cooling and heating) and equipment sizes, all of which lead to

higher costs of operation. Raising the purge rate from the base case of 5% to 20% purge,

increases the energy demand by 62% to 645 kW and increases the cost per pack for the

dry room operations by 10% to $38.37. These results are summarized in Figure 7 and

Table 10.

Figure 7. Effect of the purge rate on the heating and cooling needs, and the cost of operations.

Table 10. The heating and cooling needs, and the cost of operating the dry room at different purge rates.

Page 17 of 35
 3.6. Other Parametric Sensitivities
The model includes a large number of assumed input values and the sensitivity of

the output values to some of these parameters has been studied. The effects of plant

life, the volume of the dry room, the prices of natural gas and electricity, and the

coefficient of performance are shown in Figure 8. Changing the amortization rate from

the base case of 6 years to 4 or 8 years changes the cost of the dry room operation from

$34.85 per pack to $43.45 or $30.55, respectively. Changing the size of the dry room by

10% changes the per pack cost by a dollar or approximately 3%. Changing the price of

electricity from the base case 0.5 cents per kWh to 0.3 or 0.7, changes the cost from

$34.85 to $34.11 or $35.59, respectively. A 40% (100 kW) change in the amount of heat

generated in the dry room changes the temperature setting of the dry room inlet air and

therefore the heat load, thereby changing the cost by ~60 cents per pack. Changing the

COP from 3.5 to 3 or 4, changes the cost by approximately 10 cents per pack, while a 1

cent per kWh change in the price of natural gas changes the operations cost by about 11

cents per pack.

Page 18 of 35
 Figure 8. Effect of plant life (amortization period), dry room volume, price of electricity, heat generated in the

dry room, the Coefficient of Performance, and the price of Natural Gas on the operating cost of the dry room in

dollars per pack. The numbers in the plot represent the differential from the base case cost of $34.85 per pack.

3.7. A More Favorable Case

After reviewing the effects of various design and operating conditions, a more

favorable set can be selected such that it lowers both the energy consumption and cost

of operations. The following explores some options:

 Lowering the purge rate will reduce both the cost and energy demand (Figure 7).

The savings has to be balanced against the risk of potential for impurity buildup

at low purge rates.

Page 19 of 35
  Sizing the heat exchanger in the desiccant regeneration loop showed a reduction

in operating costs, where the differential savings diminish at the larger approach

temperatures (Figure 2). However, a smaller heat exchanger will increase the

total energy demand of the plant and therefore the greenhouse emissions

footprint.

 Increasing the temperature setting at the Pre-Cool station indicates a savings in

total energy demand and the cost. However, this also transfers the burden of

moisture removal to the zeolite wheel, which is likely to increase the cost of

maintenance and renewal, a feature that is not adequately captured in these

calculations. Removing the moisture from the small volume of make-up air is

perhaps more effective than removing it from the large volume of air going to

the dry room.

 Improvements in the battery manufacturing process to reduce the moisture

removal load should have a marked effect on the cost of the dry room operation.

The cumulative effects of the combination of eliminating the Pre-Cool station and

the heat exchanger are shown in Table 11. The former reduces the energy demand by

12% (relative to the base case), and then the elimination of the heat exchanger

increases the energy demand up to 487 kW, or 22% above the base case. The cost of

operations for the combination of changes is reduced to $32 from the base case of $35

per pack.

Page 20 of 35
 Table 11. Cumulative effects of progressive changes in the design and operating conditions on energy demand

and costs.

The scenarios simulated for this study are based on steady state conditions. Real

systems are dynamic and are therefore typically sensor driven, where the air quality

(humidity, temperature, concentrations of key components e.g., CO2, etc.) in the dry

room and at various points within the system are monitored continuously. When the

ambient air has low humidity or the moisture released/introduced into the dry room is

lower than the rated capacity, the automated controller with an appropriate algorithm

can reduce the energy demand and cost by a combination of one or more of the

following: (i) lowering the purge / fresh feed rate, (ii) raising the temperature set points

at Pre-Cool and Cool stations, (iii) slowing the revolution rate of the desiccant wheel.

The combination of sensors and a smart controller enable the detection of or prediction

of off-design conditions and are beneficial in that the system will not need to be

designed for excessive large air flow rates to recover from upset conditions, such as

frequent or prolonged opening of the air lock door, introduction of a batch of materials

with higher moisture content, number of workers present in the room being more than

designed, or some system level degradations such as flow controllers or component

failures. The development of the algorithm for these controllers would be helped with a

risk analysis that anticipates the probability of various upset conditions.

4. Conclusions

Page 21 of 35
 The results from the process model provide an understanding of the interactions of

the design and operating parameters. Such a model (or its results) can be incorporated

in the process control algorithm of a dry room so that the operation is optimized for a

given ambient air condition.

Sensitivities of the various input parameters indicate that the amount of air needed

has the biggest impact on the energy need and the cost of operations.

 The mass of air that flows through the dry room affects the energy needs and

the size of the equipment, and directly affects the cost of operations. Thus,

smaller rooms operating with low turnovers can reduce the energy demand

and cost.

 The heat exchanger plays a very important role in recycling heat. The cases

studied suggest that eliminating the heat exchanger would reduce the dollar

cost. However, that would sharply increase the total energy demand of

operating the dry room.

 Even though low purge rates reduce the cost of operations, certain minimum

purging is necessary to avoid excessive concentration increases of byproduct

gases generated within the dry room.

 Lowering the Pre-Cool temperature reduces the moisture removal load at

the zeolite wheel, but increases the overall cooling load and the energy

demand.

Page 22 of 35
 Ultimately, the energy demand and cost of operations is dependent on the quantity

of air that has to be managed, which is directly dependent on the volume of the dry

room, the air turnover rate in the dry room, and the amount of purge necessary.

5. Abbreviations
COP Coefficient of Performance

gpp grains per lb

kW kilowatt

LIB Lithium ion batteries

NG Natural gas

PHEV Plug-in Hybrid Electric Vehicle

ppmv Parts per million by volume

RH Relative Humidity

scfm Standard cubic foot per minute

Acknowledgments
The authors wish to acknowledge Kevin G. Gallagher and Gary Henriksen for their help in

preparing this manuscript. Support from David Howell and Peter Faguy at the Vehicle

Technologies Office, Office of Energy Efficiency and Renewable Energy, U.S. Department of

Energy, , is gratefully acknowledged. The submitted manuscript has been created by UChicago

Page 23 of 35
Argonne, LLC, Operator of Argonne National Laboratory (“Argonne”). Argonne, a U.S.

Department of Energy Office of Science laboratory, is operated under contract no. DE-AC02-

06CH11357. The U.S. Government retains for itself, and others acting on its behalf, a paid-up

nonexclusive, irrevocable worldwide license in said article to reproduce, prepare derivative

works, distribute copies to the public, and perform publicly and display publicly, by or on behalf

of the Government.

Page 24 of 35
Bibliography

[ D. Chung, E. Elgqvist and S. Santhanagopalan, "Automotive Lithium-Ion Battery (LIB)

1] Supply Chain and U.S. Competitiveness Considerations," 2015.

[ J. Pinho, "Dry Rooms: Why they are needed and a Review of the complex issues

2] associated with building them - particularly the large ones required to mass produce lithium

batteries," Advanced Battery Technology, vol. 46, no. 10, 2010.

[ Tesla Motors, [Online]. Available: http://www.teslamotors.com/gigafactory.

3]

[ C. Barlow, "Reaction of Water with Hexafluorophosphates and with Li

4] Bis(perfluoroethylsulfonyl)imide Salt," Electrochemical and Solid-State Letters, vol. 2, no. 8,

pp. 362-364, 1999.

[ L. Yang, M. Furczon, A. Xiao, B. Lucht, Z. Zhang and D. Abraham, "Effect of impurities

5] and moisture on lithium bisoxalatoborate (LiBOB) electrolyte performance in lithium-ion

cells," Journal of Power Sources, vol. 195, pp. 1698-1705, 2010.

[ M. Koltypin, D. Aurbach, L. Nazar and B. Ellis, "On the Stability of LiFePO4 Olivine

6] Cathodes under Various Conditions (Electrolyte Solutions, Temperatures)," Electrochemical

and Solid-State Letters, vol. 10, no. 2, pp. A40-A44, 2007.

[ R. S. Smith, "Advanced Dry Room Concepts," in First Battery Technology Symposium on

Page 25 of 35
7] Advanced Secondary Batteries, Seoul, Korea, 1996.

[ K.-S. Lee and S.-H. Choi, "Effect of geometric parameters on ventilation performance in

8] a dry room," Drying Technology, vol. 20(7), pp. 1445-1461, 2002.

[ V. Manek, K. Mindjov, G. Staikov and A. Nassalevska, "Battery dry room facility.

9] Relationship between the dry room parameters and the moisture control unit.," Journal of

Power Sources, vol. 26, pp. 603-605, 1989.

[ "Equivalency Results," U.S. EPA, [Online]. Available:

10] http://www.epa.gov/energy/greenhouse-gas-equivalencies-calculator. [Accessed 24

November 2015].

[ M. Peters, K. Timmerhaus and R. West, Plant Design and Economics for Chemical

11] Engineers, Third ed., McGraw-Hill HIgher Education, 2003.

[ P. Nelson, K. Gallagher and I. Bloom, "BatPaC (Battery Performance and Cost)

12] Software," Argonne National Laboratory, 2012. [Online]. Available:

http://www.cse.anl.gov/BatPaC/.

[ P. Nelson, K. Gallagher, I. Bloom and D. Dees, "Modeling the Performance and Cost of

13] Lithium-Ion Batteries for Electric Vehicles," 2011.

[ Hengye USA, [Online]. Available:

14] http://www.hermescatalyst.com/images/stories/pdf/Hermes_application_brochure.pdf.

Page 26 of 35
Page 27 of 35
Table 1. Assumed / Input Parameters in Dry Room Process Model (Base Case)
Assumption / Input Parameter Value
Ratio of Dry-Room-Air to Make-Up-Air 19
Make-Up Air Temperature 33 °C
Make-Up Air Relative Humidity 50 %
Temperature after Pre-Cool (A2) 9 °C
Temperature rise through Filter and Blower (A5) 3 °C
Temperature after Cooler (A5-A6) 10 °C
Moisture Level in Dry Room Inlet Air 15 ppmv (0.066 gpp)
Desiccant Capacity (MS-4A) 0.15 gH2O/gsolid
Desiccant in Solid Support 1 wt.%
Heat of Adsorption (MS-4A) 4,200 kJ/kgH2O
No. of Persons in Dry Room (30 workers x 24 hours) 720 man-hours/day
Moisture Contribution from Personnel 1500 gr/man-hour [7] [9]
Moisture Content in Negative Electrodes 0.05 wt.%
Door Opening / Closing 120 per day
Moisture Level of Dry Room Exit Air 100 ppmv (0.44 gpp)
Heat Generated in Dry Room (Equipment+Personnel) 250 kW

Desiccant Regeneration Temperature 146 °C

Heat Exchanger Approach Temperature 20 °C
Heat Transfer Coefficient for Heat Exchanger 10 Btu/h-ft2-°F
1.36E-3 cal/s-cm2-°C
COP for Refrigeration System 3.5

Page 28 of 35
Table 2. Moisture Content and Energy Demands in the Process (Base Case)
Calculated Results
Moisture Content at A2 1.14%
(50 gpp)
Moisture Removed at Pre-Cool (A1-A2) 10.6 g/s
Cooling Load at Pre-Cool (A1-A2) 57 kW
Temperature of Blended Air Stream at A3 24 °C
Moisture Content at A3 659 ppm
2.9 gpp
Moisture Removed at Cool (A5-A6) -
Cooling Load at Cool (A5-A6) 426 kW
Moisture Removed at Desiccant Wheel (A7-A8) 9.5 g/s
Cooling Load at Post Cool (A9-A10) -
Heating Load at Post-Heat (A9-A10) 63 kW
Dry Room Inlet Temperature (A10) 14 °C
Air Flow through Dry Room 19.6 m3/s
Desiccant Regeneration Heat Load (B10-B11) 30 kW
Total Cooling Load 483 kW
Electric Power for Refrigeration 138 kW
Electric Power for Blowers 167 kW
Total Thermal + Electric Load 398 kW

Page 29 of 35
Table 3. Costs for the Base Case System
Cost of Delivered Capital Equipment $741,000
Cost of Electricity ($0.05/kWh, $367/day) $134,000 per year
Cost of Natural Gas ($0.02/kWh, $44/day) $16,000 per year
Cost of Labor (14 man-hours/day) $78,000 per year
Cost of Utilities $180,000 per year

Page 30 of 35
Table 4. Assumed Input Parameters in Cost Estimate
Assumption / Input Parameter Estimated % Of Range %
1. Direct Costs
A. Purchased Equipment Model Model
Installation Model Purchased Equipment 25-55%
B. Buildings, process and auxiliary 30% Purchased Equipment 20-60
C. Service facilities and yard improvements 40% Purchased Equipment 40-60
D. Land 0% Purchased Equipment 0
2. Indirect Costs
A. Engineering and supervision 5% Direct Costs 5-10
B. Construction expense and contractor's fee 10% Direct Costs 5-15
C. Contingency 5% Fixed Cap. Investment 5-10
3. Fixed Capital Investment = I + II
4. Working Capital 5% Fixed Cap. Investment 5-10
5. Total Capital Investment = III + IV

I. Operating Cost, $/year

A. Direct Product Costs
Raw Materials Model Model
Operating labor Model Model
Direct supervisory and clerical labor 15% Operating Labor 10-20
Utilities Model Model
Maintenance and Repairs 5% Fixed Cap. Investment 2-8
Operating supplies 0.5% Fixed Cap. Investment 0.5-1
Laboratory charges 10% Operating Labor 10-15
Patents and royalties 0% Total Product Cost 0
B. Fixed Charges
Depreciation Model Model
Local taxes 3% Fixed Cap. Investment 1-4
Insurance 1% Fixed Cap. Investment 0.6-1
Rent 0% Fixed Cap. Investment 0
C. Plant Overhead Costs 10% Total Product Cost 5-15
II. General Expenses
A. Administrative costs 2% Total Product Cost 2-6
B. Distribution and selling costs 0% Total Product Cost 0
C. R&D costs 1% Total Product Cost 1-3
D. Financing 6% Fixed Cap. Investment 3-8
III. Total Operating Cost, $/year = I + II

Page 31 of 35
Table 5. Moisture Source in Dry Room
Moisture Source in Dry Room g/s %
Personnel 0.81 66
Door 1.3x10-4
Negative Electrodes 0.42 34
Total 1.23 100

Table 6. Moisture Entering and Leaving the System

Moisture Source in Dry Room In Out
Entering with Make-Up Air, g/s 19.4
Entering with Personnel & Materials in Dry Room, g/s 1.2
Removed at Pre-Cool, g/s 10.6
Leaving with Discharge Air (B14), g/s 10.0
Total, g/s 20.6 20.6

Page 32 of 35
Table 7. Effect of the Pre-Cool Temperature on Heating and Cooling
Pre-Cool Temperature 33°C 25°C 15°C 9°C 1°C

Cooling Load at Pre-Cool, kW 0 10 38 57 76

Cooling Load at Cool, kW 457 447 434 426 416
Cooling Load at Post-Cool, kW - - - - -
Total Cooling Load, kW 457 457 472 483 492
Heating at Post-Heat, kW 20 20 45 63 78
Regen Heat Load, kW 34 34 32 30 28
Total Thermal Energy Load, kW 54 54 77 93 106
Total Energy Need, kW 351 351 379 398 414
Cost, $ per pack 34.36 34.36 34.64 34.85 35.04

Table 8. Mass balance for water at two Pre-Cool temperature settings.

Pre-Cool Temperature 9°C 33°C

H2O Entering with Make-Up Air, g/s 19.4 19.4

H2O Entering with Personnel & Materials in Dry Room, g/s 1.2 1.2
10.6 0
H2O Removed at Pre-Cool, g/s (%)*
(51.5%) (0.0%)
9.5 19.6
H2O Removed at Zeolite Wheel, g/s (%)*
(46.0%) (94.9%)
0.5 1.0
H2O Removed with Purge Stream at B7, g/s (%)*
(2.5%) (5.1%)
H2O Removed at Discharge, g/s 20.6 20.6
*% of Total H2O entering the process

Page 33 of 35
Table 10. The heating and cooling needs, and the cost of operating the dry room at different
purge rates.
Purge Rate, % 5 10 15 20
Pre-Cool, kW 57 119 190 269
Cool, kW 426 429 432 435
Post-Cool, kW - - 6.3 40.5
Total Cooling, kW 483 548 628 745
Post-Heat, kW 63 28 - -
Regen Heat, kW 30 70 120 178
Total Heating, kW 93 99 120 178
Total Energy Need, kW 398 448 521 645
Cost per Pack, $/pack $34.85 $35.76 $36.85 $38.37

Table 11. Cumulative effects of progressive changes in the design and operating conditions on
energy demand and costs.
Total
Energy Cost per
Demand Pack
Base Case 398 kW $34.85
1. Eliminate Pre-Cool 351 kW $34.36
2. Eliminate Heat Exchanger from Regeneration Loop 487 kW $32.01

Page 34 of 35
Table 12. Effect of Ambient Air Temperature Heating and Cooling Needs, and Costs (RH=50%).

Ambient T, °C 40 33 20 10 0 -17.8
Ambient T, °F 104 91.4 68 50 32 0
Pre-Cool Load, kW 89 57 14 1 0 0
Cool Load, kW 426 426 426 426 415 393
Post-Cool Load, kW - - - - - -
Post-Heat Load, kW 63 63 63 79 89 96
Total Energy Need, kW 407 398 385 396 401 400
Cost of Electricity, K$ 138 134 128 127 125 122
Cost of Natural Gas, K$ 16.2 16.2 16.2 18.8 20.2 21.3
Cost per Pack, $/pack 35.11 34.85 34.51 34.53 34.50 34.38

Page 35 of 35