Desiccant Activation Tests:

Activated Alumina Versus Silica Gel

Corinne M. O’Halloran
Slice Engineering, LLC

Abstract—Moisture absorbed by 3D printing filament can cause Relative humidity is defined as the amount of water vapor
a multitude of print quality issues. The Filament Drying Desiccant present in the air relative to the amount that would be present if
by Slice Engineering® presents a solution to removing humidity the air were saturated [6]. The residual relative humidity of the
during filament storage. A desiccant is a hygroscopic material used
air is one approach to measure the effectiveness of a desiccant.
to remove moisture from surrounding air to protect sensitive
products. The most common desiccant is silica. Desiccant Desiccant activation tests were conducted to observe and
activation tests were conducted to observe and compare the measure the drying behavior of two types of desiccant: silica
moisture-adsorption of both silica gel and activated alumina. The gel and activated alumina. The change in relative humidity over
change in relative humidity of the air within a vacuum-sealed a period of time was used to compare the ability of both
container after 18 hours was utilized to determine the effectiveness desiccants to adsorb surrounding moisture. The tests aimed to
of the desiccant. The tests were started at an approximate relative achieve and maintain a zero-percent relative humidity in a
humidity of 43% and temperature of 23 °C. The lowest relative
humidity reached by silica gel was 9.7% at the 17 hour mark.
vacuum-sealed environment. The characteristic differences
Activated alumina achieved a zero-percent relative humidity after were analyzed using the test results.
9.5 hours and maintained that level for the remainder of the
experiment. The effectiveness of the two desiccants were II. PROCEDURE
determined to be 77.9% for silica gel and 100% for activated
alumina. The data presented is used to determine how Slice The adsorbing ability of each desiccant was explored through
Engineering’s Filament Drying Desiccant behavior compares to an the following desiccant activation tests. These findings allowed
existing desiccant and whether it is a viable option for 3D printing for the comparison of the effectiveness of both types of
filament storage.
desiccants.
Index Terms—Activated alumina, desiccant, relative humidity, A. Test Method
silica
Two Slice Engineering Filament Drying Desiccant canisters
I. INTRODUCTION were filled with an approximately equal volume of desiccant,
one with silica gel beads and one with activated alumina

A vital step to achieving a great quality print is drying

filament. When printing with wet filament, it is likely
to experience bubbles, extruder jams, poor adhesion,
undesirable surface finishes, and more. All of these interfere
spheres. Due to the difference in weight per bead for both
desiccants, an equal volume of each was measured rather than
an equal weight.
To ensure no residual moisture was left over from previous
with the success of a print. A potential solution is to dry the
filament spools and properly store them in a sealed container exposure, the desiccant was restored before completing the
equipped with desiccant. A desiccant is a hygroscopic material drying test. First, the 1/8 in diameter silica gel was baked at a
that aims to eliminate humidity from the surrounding air to temperature of 120 °C for 1 hour. Next, the 1/8 in diameter
create and sustain a moisture-free environment [1]. The activated alumina desiccant was baked at a temperature of 200
desiccant that is most commonly used is silica. Slice °C for 1 hour. The temperatures at which the desiccants were
Engineering’s Filament Drying Desiccant is composed of 1/8 in baked were based off manufacturer recommendations. Both
diameter spheres of activated alumina. Activated alumina is an were baked directly in the canister in a Whirlpool® Free
extremely porous form of aluminum oxide with a high surface Standing Electric Range kitchen oven.
area [2]. Both of these desiccants work through adsorption, or The test took place in a vacuum-sealed container. The
the accumulation of molecules at the surface, rather than container used was a PrintDry Vacuum Sealed Filament
absorption into the bulk of the material [3,4]. Therefore, a high Container. Molykote® 111 Compound was utilized as a grease
surface area is ideal for adsorption allowing more molecules to to aid the sealing of the container.
attach to the surface. A hygrometer was used to measure and record the relative
Most filaments are hygroscopic materials, just like desiccant. humidity in the vacuum-sealed container throughout the tests.
That means they too naturally attract surrounding moisture. Most inexpensive hygrometers only measure down to a relative
When this happens, hydrolysis can occur where the water humidity of 10% or 20%, which is enough to saturate filaments
molecules start to degrade the polymer chains [5]. When over time. It is important to note a hygrometer with the ability
printing, the water in the filament heats up very quickly to the to measure low relative humidity was necessary. The device
point of boiling. A high moisture content corresponds to a low used here was an Omega™ Pen Size Temperature and Humidity
viscosity and therefore, a high flow rate, which in turn
USB Data Logger (OM-HL-SP-TH) with the ability to measure
contributes to over extrusion.

747 SW 2nd Ave, IMB 36, Suite 296

Gainesville, FL 32601
www.sliceengineering.com 1/3
relative humidity from 0% to 95%. The desiccant and A. Performance Comparison After 1 Hour Elapsed
hygrometer were inserted into the container. The hygrometer The first hour was observed to be the steepest decrease in
was set to start recording the time and relative humidity data. moisture for both of the desiccants’ behavior. After 1 hour had
Lastly, the lid was secured, and the vacuum-seal was created elapsed in the vacuum-sealed container, the silica gel desiccant
using the PrintDry hand pump that came with the container. had decreased the relative humidity by 23.6% resulting in an
After 18 hours had passed, the seal was released, and the initial adsorption rate of 0.39% per minute. The activated
hygrometer was set to stop recording. The data was transferred alumina desiccant had reduced the relative humidity a total of
to a computer to be analyzed. These steps were completed for 30.1% resulting in an adsorption rate of 0.50% per minute. This
both silica gel and activated alumina. reveals that activated alumina is 1.28 times faster-acting in the
first hour than silica gel.
III. RESULTS B. Overall Performance Comparison
The relative humidity versus time data was plotted using After the full 18 hours had passed, the silica gel had not
MATLAB for both desiccants (Fig. 1). Both tests started at reached zero-percent relative humidity. It fluctuated between
approximately the same percent relative humidity, 43%, and the 9.7% and 9.8% for the last hour indicating that the silica gel
temperature remained at approximately 23 °C. beads had reached their maximum limit of moisture-adsorption.
The activated alumina eliminated all moisture from the
environment after 9.5 hours, while silica gel was only at 10.8%.
The 0.0% relative humidity was maintained by the activated
alumina for the following 8.5 hours. Further testing should be
conducted to measure the maximum limit of adsorption of the
activated alumina.
The overall change in relative humidity created by the silica
gel was only 34.1% versus activated alumina at 43.5%, which
corresponds to a 77.9% effectiveness for silica gel and a 100%
effectiveness for activated alumina. Effectiveness is defined as
the ratio of the achieved change in relative humidity to the
desired change in relative humidity. A summary of the
performance of the desiccants, and the accompanying
uncertainty, is shown in Table I.

TABLE I
DESICCANT PERFORMANCE SUMMARY
Value
Metric
Activated Alumina Silica Gel
Adsorption rate (%/min) 0.50 ± 5 0.39 ± 5
Fig. 1. Relative humidity versus time test results for both desiccants: activated
alumina and silica. Lowest Achieved
0.0 ± 5 9.7 ± 5
Humidity (%)
A. Silica Gel Change in Relative
43.5 ± 5 34.1 ± 5
Humidity (Δ%)
In the first hour, the humidity dropped from 43.8% to 20.2%
Effectiveness (%) 100.0 ± 5 77.9 ± 5
when using the silica gel beads. After a period of 16.9 hours,
the silica gel beads had reached what would be its lowest
relative humidity of 9.7%. The relative humidity fluctuated
C. Limitations
between 9.7% and 9.8% for the remainder of the 18 hour
experiment. Environmental limitations are present when repeatability is
considered. Tests were conducted in the month of July in
B. Activated Alumina Gainesville, Florida. The average relative humidity under these
In the first hour, the humidity dropped from 43.5% to 13.4% conditions ranges from 60% to 89% throughout the day [7].
when using the activated alumina desiccant. The environment Performance of both desiccants may vary slightly depending on
in the container reached a relative humidity of 0.0% in 9.5 the climate they are being used in.
hours. This zero-percent humidity was maintained for the An equal volume of each desiccant, filling up the same size
remainder of the test. canister, was measured for the tests. This method has the
potential for accompanying error. Future tests should measure
IV. DISCUSSION out the same number of beads for each desiccant to eliminate
this error.
The desiccant activation tests allowed for the comparison
It is likely that both desiccants were exposed to moisture in
between silica gel and Slice Engineering’s activated alumina
the air during the transfer between the oven and the vacuum-
desiccant.
sealed container. The transfer time was assumed to be small

747 SW 2nd Ave, IMB 36, Suite 296

Gainesville, FL 32601
www.sliceengineering.com 2/3
compared to the overall length of the experiment. The moisture
that could have been adsorbed was determined to be negligible.
Finally, the vacuum-sealed atmosphere inside the container
was limited by the hand pump used to create the seal. This
introduces human error in removing air from the container. An
electric pump may have produced a tighter seal. Ultimately, this
limitation did not affect the data collection as the period before
the seal was created was not included in the analysis.

V. CONCLUSION
Activated alumina achieved and maintained a zero-percent
relative humidity in a vacuum-sealed environment. This
adsorbing behavior was compared to that of silica gel. Activated
alumina is a useful option for a desiccant when storing 3D
printing filament. Further testing could be done to explore the
ability of activated alumina to extract and adsorb additional
moisture from filament once the surrounding air has reached a
relative humidity of 0.0%. Research has been done on forms of
desiccant drying plastic pellets used for injection molding,
including the use of circulating air and the effect of applying
heat [8,9]. These methods could be followed to determine if
storing filament with activated alumina desiccant can further
dry the internal moisture and compare the results to active
drying methods. The impact that desiccant effectiveness has on
print quality could be investigated, as well.

REFERENCES
[1] Delta Adsorbents. Apr. 2019. What are Desiccants? [online] Available
at: <https://www.deltaadsorbents.com/what-are-desiccants-2/>
[2] Delta Adsorbents. 2021. Activated Alumina F200 Air/Gas Streams.
[online] Available at: <https://www.deltaadsorbents.com/activated-
alumina-f200-air-gas-streams/>
[3] ChemBAM. 2021. Adsorption vs Absorption. [online] Available at:
<https://chembam.com/definitions/adsorption-vs-absorption/>
[4] Britannica, The Editors of Encylcopaedia. 2013. Adsorption.
Encyclopedia Britannica. [online] Available at:
<https://www.britannica.com/science/adsorption>
[5] T. Landry. Beat Moisture Before It Kills Your 3D Printing Filament.
MatterHackers. July, 2016. [online] Available at:
<https://www.matterhackers.com/news/filament-and-water>
[6] National Oceanic and Atmospheric Administration. 2021. Relative
Humidity (RH). [online] Available at:
<https://graphical.weather.gov/definitions/defineRH.html>
[7] Florida Climate Center. 2021. Relative Humidity. [online] Available at:
<https://climatecenter.fsu.edu/products-services/data/other-
normals/relative-humidity>
[8] D. V. Stan. 2020. Considerations on the Drying of the Raw Material and
Consequences on the Quality of the Injected Products. Materiale
Plastice. Vol. 57. Issue 1. Pg. 46-56. [online] Available at: <
https://doi.org/10.37358/MP.0.0.5311>
[9] O. Kast, C. Bonten. 2019. Calculation of the drying process of
hygroscopic polymer pellets in desiccant dryers and with additional
microwave application. AIP Conference Proceedings 2065,
030037 [online] Available at: <https://doi.org/10.1063/1.5088295>

747 SW 2nd Ave, IMB 36, Suite 296

Gainesville, FL 32601
www.sliceengineering.com 3/3