45th International Conference on Environmental Systems ICES-2015-220

12-16 July 2015, Bellevue, Washington

Evaluation of Thermal Insulation Performance of a New

Multi-Layer Insulation with Non-Interlayer-Contact Spacer

Takeshi Miyakita1, Ryuta Hatakenaka2 and Hiroyuki Sugita3

Japan Aerospace Exploration Agency, Tsukuba city, Ibaraki prefecture, 305-8505, Japan

Masanori Saitoh4
Orbital Engineering Inc., Yokohama city, Kanagawa prefecture, 221-0822, Japan

Tomoyuki Hirai5
Toska-Bano’k Co., Ltd., Bunkyo ward, Tokyo, 112-0014, Japan

Efficient insulation for tanks of the cryogenic propulsion stage or for infrared
astronomical satellites is one of the crucial key requirements. Multi-layer insulation (MLI)
blankets are used for spacecraft as excellent insulations. For conventional MLI blankets
using net spacers or embossed films however, it is difficult to control the layer density and
the thermal insulation performance degrades due to the increase in conductive heat leak
through interlayer contacts. At low temperatures, the proportion of conductive heat transfer
through MLI blankets is large compared to that of radiative heat transfer, hence the decline
in thermal insulation performance is significant. A new type of MLI blanket using new
spacers; the Non-Interlayer-Contact Spacer MLI (NICS MLI) has been developed, which
uses small discrete spacers and can exclude uncertain interlayer contact between films. The
insulation performance is measured with a boil-off calorimeter. Because the NICS MLI
blanket can exclude uncertain interlayer contact, the test results correlated well to
estimations. Furthermore, the NICS MLI blanket shows significantly good insulation
performance (effective emissivity is 0.0008 between 300K and 77K boundary temperature),
particularly at low temperatures, due to the high thermal resistance of this spacer.

Nomenclature
 = the Stefan-Boltzmann constant
eff = effective emissivity
s = layer surface emissivity
qtotal = total heat flux through MLI
qrad = radiative heat flux
qcond = conductive heat flux
qconv = convective heat flux
Thot = hot boundary temperature
Tcold = cold boundary temperature
N = number of layers
R = thermal resistance of a spacer
lp = pitch between spacers
m = evaporation rate of liquid nitrogen
hlg = latent heat of evaporation
SBT = heat transfer surface area of the boil-off tank

1
Engineer, Thermal Systems Group, Aerospace Research and Development Directorate, miyakita.takeshi@jaxa.jp
2
Engineer, Thermal Systems Group, Aerospace Research and Development Directorate, hatakenaka.ryuta@jaxa.jp
3
Head, Thermal Systems Group, Aerospace Research and Development Directorate, sugita.hiroyuki@jaxa.jp
4
Engineer, Engineering Group, saitoh@orbital-e.co.jp
5
General Manager, Product Development Division, hirai0006@toska-banok.com
 I. Introduction

M ulti-Layer Insulation (MLI) blankets are the most efficient thermal insulation element. Conventional MLI
blankets comprise multiple layers of low-emissivity films and netting spacers, the latter of which prevent
direct contact between films to reduce conductive heat leaks (Fig. 1 (a)). Instead of netting spacers, embossed films
are sometimes used to decrease the contact area between films. However, neither netting spacers nor embossed films
can exclude interlayer contact between film and the netting spacer or films, which has a major impact on conductive
heat leak through the MLI blankets, depending on the degree of contact. The thermal insulation performance of MLI
blankets is thus strongly dependent on the mounting arrangements. One of the parameters of thermal insulation
performance is the effective emissivity of an MLI blanket, as defined by the following equation:

qtotal
 eff 
 T  Tcold 
4 4
(1)
hot

Effective emissivity of 0.0050 or lower for a conventional MLI is achievable in well-controlled laboratory tests.
However, experience has shown that when a blanket is configured for spacecraft applications, effective emissivity
closer to the range 0.0150 to 0.0300 is representative of the current design, manufacturing and installation methods
for medium-area applications3.
Extremely high thermal insulation performance is particularly important for exploration spacecraft such as lunar
surface rovers, cryogenic science platforms, and long duration, in-space, cryogenic propulsion systems. The Japan
Aerospace Exploration Agency (JAXA) is planning one such mission: the Space Infrared Telescope for Cosmology
and Astrophysics (SPICA). The SPICA satellite has a cryogenically-cooled (< 6 K), large telescope (in the 3 m
class). Most heat radiation from the Sun to the spacecraft is blocked by sun shields and thermal radiation shields
covered with MLI blankets. It is important to maintain the insulation performance of MLI to retain the extremely
low temperature of its sensor.
For long-term propulsion systems, including cryogenic propellant tanks, insulation performance is also important
as a key technology to reduce heat entering the propellant storage tank and minimize boil-off losses of propellant
during long-term missions such as Mars exploration. The performance of conventional MLI blankets declines
significantly in cryogenic systems due to the low proportion of radiative heat transfer through MLI relative to that of
conductive heat transfer at low temperatures. The temperature of SPICA’s thermal radiation shields is 55 K to 121
K13, while that of the hydrogen tank is 20 K, which makes it challenging to maintain the thermal insulation
performance of MLI blankets installed in such cryogenic space missions. In this study, a new type of MLI, Non-
Inter layer-Contact Spacer MLIs (NICS MLIs) are experimentally designed and assembled and the insulation
performance is measured by a boil-off calorimeter.

(a) Conventional MLI with netting spacers (b) MLI with discrete spacers
Fig. 1 Schematics of a conventional MLI (a) and MLI with discrete spacers (b).

II. Non-Interlayer-Contact Spacer MLI development

Outside a well-controlled laboratory, e.g. for an installation configuration on a cuboid body like a satellite or a
cylindrical side structure like the rocket propellant tank, the insulation performance of conventional MLIs declines.
Our team studied the proportion of heat leak through conventional MLIs10 (Fig. 2). The edges of conventional MLIs
are seamed to combine films, but the degradation of conventional MLI is primarily attributable to machine sewing
and secondarily to the contact between films and netting spacers. Conductive heat leak through MLIs increases with
the contact area of films and spacer or with contact pressure. For ideal MLI the layer density is well controlled in the
laboratory, the performance is high and easier to predict. However, for conventional MLI when actually used on

International Conference on Environmental Systems

spacecraft, the degree of interlayer contact is difficult to
control or predict which makes it difficult to quantify the
performance before thermal balance testing of spacecraft.
Many equations are proposed to estimate the thermal
performance of conventional MLI blankets4-7. However,
since these must take the kind of spacer, sewing condition
and so on into account, it is difficult to apply them to
each MLI blanket.
A new type of spacer, non-interlayer-contact spacers
(NICS) has been developed14. While conventional spacers
such as netting spacers are inserted in the whole surface
layer, these new spacers are intermittently arranged and
hold up the film to exclude any incidental interlayer Fig. 2 The proportion of heat leak through
contact (Fig. 1 (b)). This section describes the NICS conventional MLI10
design in detail as well as the assembly method for NICS (The layer number is 12, layer density is 2.3
MLI. layer/mm and the outermost and innermost layer
The NICS is designed to meet the SPICA satellite temperatures are 300 K and 77 K respectively)
requirement for thermal insulation14 and the schematics of
an NICS and NICS MLI are shown in Figs. 3 and 4 respectively. NICS controls the gaps between films to be
intermittently arranged on the film and prop the upper film. The contact points of the films are only the top face or
subsurface of the discrete spacers and the film surfaces do not come into contact. Consequently, conductive heat
leaks can be easily estimated because only conduction through the spacer need be considered, rather than interlayer
contacts of films and thermal performance can be more precisely quantified. Such spacers have also been developed
by S. Dye et al.2 Their MLI has also been tested using a discrete type spacer, revealing a lower heat leak than
conventional MLI, while obtaining high thermal resistance improves insulation performance.
The detailed thermal resistance design of NICS is described in Ref. 14. As shown in Fig. 3, a NICS comprises
three stages. The role of the second stage is to ensure high thermal resistance with an elongated path. The diameter
is 10 mm and the height between the first and third stages is 4 mm. The thermal conductive path is 40.9 mm long,
with a cross-sectional area ranging from 0.36 mm2 to 0.51 mm2. If the thermal conductivity of the spacer material is
0.30 W/m K, its thermal resistance is calculated at 3.32 × 105 K/W, which is extremely high in a small body. The
NICS specifications are shown in Table 1.
Films are held between the upper and lower spacers and interconnected, while the first NICS stage includes a
locking pin to join the joint hole at the third stage. The film has small holes at equivalent intervals to the spacer pitch,
through which the locking pin passes and joins the joint hole of another spacer. Accordingly, the NICS MLI
eliminates the need for sewing to integrate films.
NICSs are made by injection molding, which allows them to be produced in bulk, economically and with high
dimensional accuracy. Polyetheretherketone (PEEK) was selected as the spacer material with heat tolerance,
injection moldability and the tolerance of electron beam radiation in mind.

Fig. 3 Schematics of Fig. 4 Schematics of the non-interlayer-contact spacer MLI

the non-interlayer-contact spacer (assembling)

Table 1 Specifications of a non-interlayer-contact spacer.

Cross-sectional area to
Thermal resistance Diameter Height Mass Material
length ratio
3.32 × 105 K/W 1.0 × 10-5 m 10 mm 6.9 mm 0.101 g PEEK

International Conference on Environmental Systems

 III. Materials and Method

3.1. Measurement equipment

The thermal insulation performance of
MLI blanket samples was measured by a
double-guarded calorimeter with liquid
nitrogen. A boil-off calorimeter is one of the
most famous methods used to measure
insulation performance of MLI. In particular,
the vertical-type boil-off calorimeter has an
advantage whereby gravity does not apply in
the direction to press the layers. Two types of
tank shape are used, one of which is
rectangular, to simulate a spacecraft body and
evaluate the degree of degradation due to the
corner. The other is cylindrical, to simulate a
curved surface shape like the sun shield of an
astronomy satellite or the propellant tank of a
rocket. As shown in Fig. 5, the inner tanks
comprise the main boil-off tank and the upper
and lower guard tanks, all of which are filled
with liquid nitrogen. This configuration
ensures uniform temperature distribution
around the measurement tank along the axial
direction. The size of the rectangular boil-off
tank is 300 mm (height) × 236 mm (width) ×
Fig. 5 Schematics of the double-guarded calorimeter
236 mm (depth) while the cylindrical tank is
300 mm (height) × 300 mm (diameter). These two boil-off tanks have equivalent heat-receiving areas. The top
flange suspends these inner tanks and is equipped with a cryogen exhaust port from each tank, an evacuation port, a
safety valve, a vacuum gauge installation port, a feed-through terminal for thermocouples and a water tube for
thermostat circulation. An aluminum shroud surrounds the test piece, while the temperature of its inner surface is
maintained by thermostat circulation, from 276 to 353 K. The temperature distribution in the inner surface of the
shroud is within 1 K. The inner surface of the shroud and the outer surface of the liquid nitrogen tanks are finished
in black anodized aluminum as high-emissivity surfaces to minimize radiative thermal resistance between MLI
surfaces to the shroud or tank and minimize the impact of uncertain contact heat transfer between the innermost MLI
surface and the tank surface.
The heat flux through the MLI blanket is measured by the evaporation rate of liquid nitrogen from the main tank.
Equation (2) is used to calculate the heat flux q.

m  hlg
q  (2)
S BT

The flow rate of nitrogen gas having evaporated from the measurement tank is measured using three types of
flow meter as follows:
(a) Wet-type volume flow meter W-NK-0.5 (Shinagawa Corp.)
(b) Mass flow meter CMS-9500 (Azbil Corp.)
(c) Soap-film volume flow meter (50 ml type, time measured manually, for cross-checking)
The wet-type flow meter allows us to measure wide-ranging flow rate (from 0.016 to 5000 mL) with a single flow
meter, but has the disadvantage of yielding a fluctuating flow rate due to the difficulty in rotating the measurement
wheel when the flow rate is small, whereas the averaged flow rate calculated from the integrated flow rate is
considered reliable owing to its measurement principle. Conversely, the mass flow meter is suitable for this
experiment since the final desired value is not a volume flow rate but a mass flow rate, despite the relatively narrow
measurement range. In this paper, some of the data is taken with (a) and the rest with (b). It was confirmed in
another experiment that the averaged flow rates of these two flow meters match.

International Conference on Environmental Systems

 The temperatures of the tanks and MLI surface are measured using Type-T thermocouples and the measurement
points are individually chosen per test piece. The diameter of the copper and constantan wire is 100 m or 50 m
and the wires of the thermocouple for the MLI surface are routed on the surface for a certain length. The tip of the
thermocouple is electrically insulated with Kapton tape and covered with a piece of vapor-deposited polyimide tape
(UTC-025R-NANA, Ube Industries, Ltd.) to form the same surface as samples.

3.2. Test piece

A list of the test pieces for thermal performance tests is shown in Table 2. Two configurations of NICS MLIs
were evaluated, one of which is installed on a rectangular tank and the other on a cylindrical tank. For comparison,
‘reference MLI’ and ‘conventional MLI’ were also evaluated. As the detail of the each test pieces is described later,
average areal density of each test pieces is calculated and shown in Fig.6. NICS MLI is heavier than conventional
MLI and reference MLI with same number of layers, because NICS MLI uses thicker film.

Table 2 List of test pieces

No. Name Calorimeter Layers Connection Perforation Inner layer film Outer layer film
Cylindrical
1 Reference MLI (1)
tank Film ×24
/Net ×23 Interleaved
Rectangular
2 Reference MLI (2) lapping Vapor-deposited Polyimide surface
tank
×1 polyester side of aluminized
Cylindrical Film ×12 (a)Dia.=2mm 6μm film 25μm film
3 Reference MLI (3)
tank /Net ×11 Pitch=50×50mm
Butting joint Random pattern KF-6B UTC-025R-NANN
4 Conventional MLI (1) ×4 (Both sides (Single side
Rectangular Film ×12 (edges of tank) aluminized) aluminized)
tank /Net ×11 Simple overlap
5 Conventional MLI (2) ×4
(edges of tank)
(c)Via Spacer: Vapor-deposited polyimide 50μm film
Interleaved KC-50B (Both sides aluminized)
Dia.=3mm on film
Non-Interlayer-Contact Rectangular lapping
6-8 Film 6/5/4 Min Dia. of spacer Only outermost layer is KC-50S
Spacer MLI (1) tank with spacer ×4
pin=0.6mm (Single side aluminized)
(edges of tank)
Pitch≒50 mm×50mm Spacer pitch = 50 mm × 50 mm
(b)Via Spacer:
Interleaved Dia.=3mm on film Vapor-deposited polyimide 125μm film
Non-Interlayer-Contact Cylindrical Film UTC-125S-AANN (Both sides aluminized)
9-13 lapping Min Dia. of spacer
Spacer MLI (2) tank 12/10/8/6/4
with spacer ×1 pin=0.6mm Spacer pitch ≒ 90 mm × 100 mm
Pitch≒90 mm×100mm

Fig. 6 Calculated average areal density

International Conference on Environmental Systems

3.2.1. Reference MLI
The reference MLIs were prepared as samples which were ideally mounted
on a spacecraft. Concerning the connection method, interleaved lapping is
performed for one layer after another when test pieces are attached to the boil-
off tanks, as shown in Fig. 7. Vapor-deposited polyimide tape (UTC-025R-
NANA, Ube Industries, Ltd.) is used to fix the outermost layer, while another
vapor-deposited polyimide tape (UTC-025R-ANNA, Ube Industries, Ltd.) is
used to fix the inner layer and maintain the emissivity of each surface
equivalent to the original surface. This connection method reduces the heat leak
on lapping but is almost impossible in practice because of the immense time
and effort involved. Vapor-deposited polyimide film (UTC-025R-NANN, Ube
Industries, Ltd.) is used for the outermost and innermost layers and the
polyimide sides are directed toward the outside of the blanket. Thin vapor- Fig. 7 Interleaved lapping
deposited polyester film (KF-6B, Kaneka Corp.) is used for the inner reflective
shields while the net spacer (KN-20, Kaneka Corp.) is also used.

3.2.2. Conventional MLI

The conventional MLIs mimic samples which are
ordinarily mounted on spacecraft and the material of
each layer is the same as the reference MLI. All edges
are sewn with a sewing machine and taped with vapor-
deposited polyimide tape (UTC-025R-NANA, Ube
Industries, Ltd.), which has optical properties equivalent
to the outermost layer of the blanket, for reinforcement.
A hook and loop fastener (A8693Y-71/B2790Y-00, 16
mm wide, Kuraray Fastening Co., Ltd.) is sewn to
blankets with a polyester string, also using a sewing
machine. The stitches for the fastener are also taped
with vapor-deposited polyimide tape (UTC-025R-
Fig. 8 Simple overlap (left) and butting joint (right)
NANA, Ube Industries, Ltd.) on the outer side of the
blanket. At the corner, sample No. 4 is connected by a
butting joint, while sample No. 5 is connected by simple overlapping, as shown in Fig. 8.

3.2.3. Non-Interlayer-Contact Spacer MLI

Two types of non-interlayer-contact spacer MLI were made. Sample Nos. 6 to 8 include NICS MLI (1), which
are installed on the rectangular tank to demonstrate the ease of installation on the cubed spacecraft body. (Fig. 9).
NICS MLI (1) was made for each of the tank sides and a tuck was made on the film, which was bent at the corner.
The spacer on the corner fixes the right and left films together as shown in Fig. 10. The spacer works for tapes,
meaning NICS MLI needs no seams or tapes.
Spacers are arranged with a 50 mm grid and 50 µm films are used to maintain the film flatness. There are a total
of six film layers. Vapor-deposited polyimide film (KB-50S, Kaneka Corp.) is used for the outermost layers, while
thin vapor-deposited polyimide film (KB-50B, Kaneka Corp.) is used for the inner reflective shields. Sample No. 7
is formed by removing the outermost layer from sample No. 6, while sample No. 8 is made by removing the
outermost layer from sample No. 7, meaning the outer layer surface of sample No. 6 is the polyimide side and those
of sample Nos. 7 and 8 are aluminized surfaces.
Samples Nos. 9 to 13 are NICS MLIs (2), which are installed on a cylindrical tank to determine the ease of
installation on the rocket propellant tank or astronomy satellite shield (Fig. 11). NICS MLIs (2) have one joint line
on the side and the spacer on the corner fixes the films together just like the corner of NICS MLI (1). Spacers are
arranged by 90 mm in an axial direction by (about) 100 mm in a circumferential direction and vapor-deposited
polyimide 125 µm films (UTC-125S-AANN, Ube Industries, Ltd.) are used. Using thick films for NICS MLI (2),
the spacer pitch is extended twice beyond NICS MLI (1). Sample No. 9 includes 12 film layers, while sample Nos.
10 to 13 are formed by removing the outermost layer from sample No. 9.

International Conference on Environmental Systems

 Fig. 9 Picture of the non-interlayer-contact Fig. 10 Schematic of the overlapping of
spacer MLI (1) installed on the rectangular tank the non-interlayer-contact spacer MLI

Fig. 11 Picture of the non-interlayer-contact spacer MLI (2) installed on the cylindrical tank
(Left: overhead view, Right: bottom view)

3.3. Test Procedure

To prevent MLI blankets from becoming contaminated with dust, the test MLI blanket is prepared in a clean
room. Using thermocouples, the temperatures of the outermost layer of the MLI film and inner layers of NICS MLI,
the surface of the inner tanks and the surface of the shroud were all measured. The evaporation rate and pressure of
the vacuum tank were also measured using a flow meter and ionization vacuum gauge. The vacuum chamber was
evacuated by a turbo-molecular pump while the vacuum chamber was warmed to 353 K over two days for moderate
baking. After baking, liquid nitrogen is transferred into the inner tanks. The pressure of the vacuum tank after the
liquid nitrogen is transferred is less than 10-4 Pa. To investigate the temperature dependency of the thermal
insulation performance of MLI blankets, the shroud temperature was controlled to 276, 300 and 353 K.

IV. Results and Discussion

4.1. Results of the thermal performance test and data correction method.
The heat fluxes measured by the boil-off calorimeter are shown in Fig. 12. Reference MLI and NICS MLI use
the left axis and conventional MLI uses the right. The values of the mass flow rate to assign equation (2) are
corrected for the influence of atmospheric pressure. Liquid nitrogen in the measurement tank is under saturated
vapor pressure and affected by ever-changing atmospheric pressure. If the measurement is performed when the
atmospheric pressure declines by 1 Pa, the temperature declines by 8.6×10-5 K. To compensate for the removed heat
capacity of liquid nitrogen and the decreased measurement tank temperature, the mass flow of evaporated nitrogen is
increased and overestimated and vice versa15. The pressure at the mass flow meter outlet and the volume of liquid
nitrogen in the tank are measured by a pressure manometer and capacitance-type level gauge, respectively. The error
bar sums up the accuracy of the flow meter and level gauge. The measurement heat fluxes are also corrected by
deducting heat which is not through the center MLI sample.

International Conference on Environmental Systems

 Some experimental equations to predict the temperature dependency of MLI thermal performance are proposed
in literature [4-7]. The coefficients of the equation depend on many parameters such as the number of layers, spacer
material, film material, pre-treatment method, layer density and so on. Although the general form of the equation
should resemble those of literature, the coefficients should be evaluated individually per MLI system. Basically, the
equation proposed by Cunnington et al.6 is adopted in this study, while the coefficient of each term is simplified as
follows using three free-parameters A, B and C:
TH  TC
q MLI  A   (TH4.67  TC  C )  TH  TC 
4.67
)  (B  (3)
2
where TH and TC denote the temperatures
of the hot- and cold-side boundaries
respectively. Fig 12 shows both the raw test
data and fitting curves generated using
equation (3) and Fig.13 shows effective
emissivity.

4.2. Reference MLI and conventional MLI.

No. 1 reference MLI (1) and No. 3
reference MLI (3) are installed on a
cylindrical tank with interleaved lapping and
these two samples differ in terms of the
number of layers, No. 1 has 24 layers, while
No. 3 has 12. Comparing the two, Reference
MLI (1), with many layers, shows superior
performance at high temperatures, but the
performance of reference MLI (1) and
reference MLI (3) rarely differs at low
temperature. At a high temperature, when Fig. 12 Experimental results of boil-off calorimeter and
radiation dominates, increasing the number fitting curves
of layers to improve thermal performance is (Temperature of cold boundary is 77 K.)
effective. At a low temperature however,
when conduction dominates, increasing the
number of layers has little effect on
thermal performance.
No. 2 reference MLI (2) comprises 24
layer films as with No. 1 reference MLI (1),
but reference MLI (2) is installed on a
rectangular tank and reference MLI (1) on
a cylindrical tank. Both comprise the same
materials and are applied to the same
connection method, interleaved lapping.
However, since it is difficult to install one
blanket of reference MLI on a rectangular
tank with a corner, the average layer
density of reference MLI (2) on a
rectangular tank may become smaller than
reference MLI (1) on a cylindrical tank.
The thermal performance of MLI samples
using conventional netting spacers like
reference and conventional MLIs are more
severely affected by the installation
methods due to the change in layer density
or contact pressure between films.
Despite the fact conventional MLIs Fig. 13 Effective emissivity calculated from experimental results
comprise the same materials as reference

International Conference on Environmental Systems

MLIs except for the installation method and edge sewing, the thermal performance of conventional MLIs is much
lower than that of reference MLIs. As discussed in chapter II, about 70 % of the heat leaks through conventional
MLIs are attributable to conductive heat through stitching and overlap or joint lines. In this test, the thermal
performance on edge stitching and the joint line of conventional MLIs was severely degraded.

4.3. Non-interlayer-contact spacer MLI

As shown in Fig. 11, the thermal performance of NICS MLIs is far superior to conventional MLIs. The thermal
performance of No. 9 NICS MLI (2) is 39 times than that of No. 4 conventional MLI (1), despite both having 12
layers. Furthermore, the thermal performance of No. 13 NICS MLI (2), which has only 4 layers, is 14 times that of
No. 4 conventional MLI (1).
The differences between Nos. 6-8 NICS MLI (1) and Nos. 9-13 NICS MLI (2) emerge in the spacer pitch and
boil-off tank configuration. When NICS MLI (1) is compared to NICS MLI (2), the latter shows superior thermal
performance with an equivalent number of layers. When the thermal performances of NICS MLIs were measured,
the temperature distributions of the outermost films were also measured and the temperature difference between the
center and corner of the NICS MLI (1) sample was less than 2 °C. Therefore, the difference in thermal performance
between NICS MLI (1) and NICS MLI (2) is primarily attributable to the spacer pitch rather than the installation
configuration. As will hereinafter be described in detail, when the spacer pitch increases, not only does the
conductive heat leak through the spacer decrease but also radiative heat exchange between the spacer and film,
meaning improved thermal insulation performance.
The temperature dependency of thermal insulation performance is noted. The gradient of heat flux over the
temperature of conventional and reference MLIs is small and the relationship seems linear. Conversely, the gradient
of heat flux over the temperature of NICS MLIs is large at high temperature and small at low temperature. Because
NICS MLI only comprises a few layers, radiative heat leak increases at high temperature, but the thermal resistance
of the spacer is very high, which, in turn, means minimal conductive heat leak through the spacer and total heat leak
at low temperature, where conduction dominates.

4.4. Analysis of the thermal insulation performance and proportion of heat leak
As already mentioned in section II, the thermal performance of NICS MLI can be easily estimated because the
conductive heat of NICS MLI only passes through the spacer and is very simple. To estimate the thermal
performance of the NICS MLI precisely, heat transfer through the NICS MLI is modeled using a Thermal Desktop
as shown in Figs. 15 and 16. The spacer is
modeled in three parts, namely the first, second
and third stages. The area and figure of the
horizontal surface of the spacer model is the
same as the real thing and radiative heat flux
between the spacer and film is taken into
consideration. The temperature dependency of
film emissivity is considered using Eq. (4) 9,

 i  6.69  10 4  Ti 2 3                        (4) 

The three parts of the spacer are connected by

thermal conductors and the total thermal
resistance of the spacer corresponds to 3.32 × 105
K/W, which is the design value. As boundary
conditions, the temperatures of the shroud are
specified from 77 K to 400 K and that of the boil-
off tank at 77 K. The horizontal dimension of the
model is the spacer pitch and the spacer is
arranged in the center. To eliminate the edge Fig. 14 Comparison of experimental and calculated
effect, all sides are covered by surfaces that act as results of heat flux through NICS MLI.
reflectors/blockers in the radiation calculations. (Cold boundary temperature is 77 K)
The calculated heat fluxes of NICS MLIs are shown in Fig. 14 with the measured heat flux. When the calculated
and measured heat fluxes are compared, the calculations agree very well with the experiment, with an error rate of
less than 17 %. Reducing the uncertainty of conductive heat leak due to interlayer contact and reducing conductive

International Conference on Environmental Systems

heat leak due to the high thermal resistance of the spacer allows the thermal performance of NICS MLI to be
precisely estimated.

Fig. 15 Schematic of the heat transfer model through NICS MLI

Subsequently, the proportion of heat leak for each sample is

analyzed, whereas for NICS MLIs, the radiative heat leak
between films and the film and spacer and the conductive heat
leak through the spacer are analyzed using the thermal
mathematical model as referred to above. With regard to
conventional and reference MLIs, the following thermal
mathematical model is conceivable:
The total heat flux qtotal is expressed as the sum of radiative
heat flux qrad, conductive heat flux qcond and convective heat
flux qconv as follows:

qtotal  q rad  q cond  q conv                         (5) 

Here, it is assumed that radiation and conduction are not

coupled and the principle of superposition can be applied4.
Under the condition of applicability to spacecraft, convective
heat transfer via rarefied gas is negligible and radiative heat flux
is expressed in the form of radiation exchange between large
parallel planes as follows:
 

q rad 
1
(1  i )  (1  j )  1
 
 Ti 4  T j4               (6)  Fig. 16 Schematic of the thermal
mathematical model of NICS MLI

 i is the emissivity of the hotter layer surface,  j is the emissivity of the colder layer surface and both follow by equ
ation (4). The conductive heat flux is expressed as follows: 

 
qcond  h Ti  T j                                                                               (7) 

Here, h is the total conductive heat transfer coefficient between two layers, which means the sum of the spacer
thermal conductance and contact thermal conductance between the film and netting spacer. The total conductive
heat transfer coefficients between each layer are assumed to be constant. The temperatures of the outermost and
innermost layers are given as boundary conditions and the heat flux between each of the films is repeatedly
calculated until the calculated heat flux matches the measured value by changing the total conductive heat transfer
coefficient and the temperature of each layer.
The proportion of heat leak of No. 6 NICS MLI (1) and No. 12 NICS MLI (2), which comprise 6 layers when the
outermost layer temperature is 300 K, are shown in Fig. 17. Moreover, the proportion of heat leak of No. 3 reference

10

International Conference on Environmental Systems

 MLI (3) and No. 4 conventional MLI (1), which comprise 12 layers, are shown in Fig. 18 when the outermost
layer temperature is 300 K. The calculated temperature of each layer is shown in Fig. 19.
Shown as Figs. 17 and 18, the conductive ratio of heat leak through conventional MLI (1) exceeds 89 %, while
the conductive ratio of heat leak through NICS MLIs is less than 8 %, less than a tenth of that of conventional MLIs.
Alternatively the conductive ratio of heat leak through reference MLI (3) is 28 % at the hot boundary layer and 79 %
at the cold boundary layer respectively. As seen above, the ratio of conduction to radiation is changing in each
interlayer of the same sample. As the
temperature declines, the radiation ratio does the
same. Due to the varying proportion of heat leaks,
the temperature dependency of thermal
insulation performance shows different features
for each of the samples. The temperature
dependency of the thermal insulation
performance of conventional MLIs, which is
conduction dominated, seem linear, with a
significant degradation of performance at low
temperatures. Conversely, the conductive heat
leak of NICS MLI is quite small, meaning
thermal performance can be maintained, even at
low temperatures. The calculated conductive
heat transfer coefficient in each layer is 3.03×10-
6
W/K for NICS MLI, 3.45×10-2 W/K for Fig. 17 The calculated proportion of heat leak of No. 6
reference MLI (3) and 7.10×10-1 W/K for NICS MLI (1) and No. 12 NICS MLI (2) at the hot
conventional MLI. boundary temperature is 300 K.

Fig. 18 The calculated proportion of heat leak of No. 3 reference MLI (3) and No. 4 conventional MLI (1)
with a hot boundary temperature of 300 K.

As shown in Fig. 17, the radiation via spacer in NICS MLI is considerable. This is because the emissivity of the
spacer surface is relatively higher than that of the film surface, so despite the small spacer area ratio, the contribution
of the spacer radiation is relatively high for the heat exchange in each film. Accordingly, increasing the spacer pitch
improves the thermal insulation performance of NICS MLI.
Finally, the effective emissivity is calculated. The effective emissivity between 77 K and 300 K is 0.0326 for No.
4 conventional MLI (1), 0.0027 for No. 3 reference MLI (3) and 0.0008 for No. 9 NICS MLI (2), respectively.

11

International Conference on Environmental Systems

 Fig. 19 The calculated layer temperature of No. 3 reference MLI (3), No. 4 conventional MLI (1), No. 6
NICS MLI (1) and No. 12 NICS MLI (2) at a hot boundary temperature of 300 K.

V. Conclusion and future work

A new type of MLI using new spacers, the non-interlayer-contact spacer MLI, was developed and thermal
performance tests using a rectangular and cylindrical boil-off calorimeter were performed. The rectangular boil-off
tank simulates a cubic spacecraft body, while the cylindrical tank has a curved surface resembling the sun shield of
an astronomy satellite or a rocket propellant tank, and the degree of degradation due to the mounting method is
evaluated using two types of boil-off tank. Accordingly, the thermal insulation performance of the NICS MLI is
independent of the mounting procedure on the tanks. Conversely, reference MLI, which is mounted idealistically
and conventional MLI, both of which use conventional netting spacers, are significantly influenced by the mounting
procedure given the difficulty of controlling the layer density or interlayer-contact. Because the NICS MLI excludes
any uncertain interlayer contact between films due to use of discrete spacers, heat passes of NICS MLI are very
simple. Accordingly, the estimations of thermal insulation performance of NICS MLI using a simple thermal
mathematical model effectively matches the experimental results with an error rate of less than 17 %.
According to the experimental results, the thermal insulation performance of NICS MLIs is far superior to
conventional MLIs, particularly at low temperature. Moreover, when the spacer pitch increases, thermal insulation
performance improves, because not only does the conductive heat leak through the spacer decline, the radiative heat
exchange between the spacer and film is also reduced. Thermal analysis shows that the conductive ratio of heat leak
through conventional MLIs exceeds 89 %, conversely, the conductive ratio of heat leak through NICS MLIs is less
than 8 %; less than a tenth of that of conventional MLIs. For the above reasons, NICS MLI performs well,
particularly at low temperatures, where conduction dominates. The effective emissivity of the NICS MLI, with a spacer
pitch of about 100 mm and a film layer of 12, achieved 0.0008 at Thot = 300 K.
Because of the features of the previously predictable and quite high insulation performance at low temperature,
the benefit of the NICS MLI in space missions like cryogenic astronomy satellites or long-term propulsion systems
has been shown. This unique feature should be confirmed by low-temperature testing in the future. The cryogenic
calorimeter to measure insulation performance using Gifford-McMahon coolers have been designed, and we have
the plan to manufacture the new calorimeter during the current year and measure thermal performance of the NICS
MLI at lower than 20 K.
Variety of issues for the NICS MLI should be examined like perforation hole specification, fastener to a
spacecraft and durability for launch load and so on. The thickness of the NICS MLI is thicker than conventional
MLI so the NICS MLI must evacuate more air inside the layers. Vent hole configuration identification and
depressurization testing have been initiated and will be reported on in a future report. For rocket propellant tank,
insulation is subjected to force at low temperature during launch. The stand-alone tensile and compaction tests at
ambient temperature have already completed. We should also look at tensile and compaction property of the spacer
at low temperature and evaluate the NICS MLI blanket under vibration and acoustic loading. For long-term

12

International Conference on Environmental Systems

propulsion systems, radiation tolerance will be an important aspect of the spacer. The above considerations for the
proposed MLI will be addressed prior to its application to specific missions.

References
1
Leslie Buchanan and Steve Buerger, “MLI Layup Optimization Test,” Proceedings of Space Thermal Control Workshop,
2010.
2
S. Dye, A. Kopelove and G. L. Mills, “Integrated and Load Responsive Multilayer Insulation,” Cryogenic Engineering
Conference, Vol. 55, 2010, pp. 946-953.
3
David. G. Gilmore, ‘Spacecraft Thermal Control Handbook Volume 1: Fundamental Technologies,’ The Aerospace
Corporation, 2002, pp. 161-168.
4
Moshfegh, B. et al., “Thermal Analysis of Multilayer Insulation at Room Temperature,” Advanced Computational Methods
in Heat Transfer, Vol. 3, 1990, pp. 357-368.
5 Jochen, Doeneche, “Survey and Evaluation of Multilayer Insulation Heat Transfer Measurement,” SAE technical paper

series, No. 932117, 1993.

6
C. W. Keller, G. R. Cunnington and A. P. Glassford, ‘Thermal Performance of Multilayer Insulations,’ NASA CR-134477,
1974.
7 L. D. Stimpson and W. Jaworskis, “Effect of Overlaps, Stiches, and Patches on MLI,” AIAA paper, No. 72-285, 1972.
8 C. L. Tien, G. R. Cunnington, “Radiation Heat Transfer in Multilayer Insulation having perforated Shields,” AIAA Paper,

pp. 73-718, 1973.

9
N. Inai, Japan Society of Mechanical Engineers, Vol. 43, No. 365(1977-1), 217, 1977.
10 S. Okazaki, M. Murakami, H. Kawasaki, T. Yabe, H. Sugita, and Y. Kanamori, “Experimental Study of the influence of

processing on MLI performance for Space Use,” AIAA-2008-01-2067, 2012.

11
H. Kawasaki, S. Okazaki, H. Sugita and M. Murakami, “Temperature dependence of Thermal Performance in Space using
Multilayer Insulation,” AIAA-2012-3407, 2012.
12 Ryuta Hatakenaka, Takeshi Miyakita, Hiroyuki Sugita, Masanori Saitoh, Tomoyuki Hirai, “Thermal Performance and

Practical Utility of a MLI Blanket using Plastic Pins for Space Use”, American Institute of Aeronautics and Astronautics, 2013,
pp. 1-11
13
Y. Sato, H. Sugita, et al., “Conceptual design of cryogenic system for the next-generation infrared space telescope SPICA”,
Proc. of SPIE Vol. 7731, 2012.
14
Takeshi. Miyakita, et al., “Development of a New Multi-Layer Insulation Blanket with Non-Interlayer-Contact Spacer for
Space Cryogenic Mission”, Cryogenics 64 (2014) pp. 112-120.
15
Ryuta Hatakenaka, Takeshi Miyakita, et al., “Development and testing of a zero stitch MLI blanket using plastic pins for
space use”, Cryogenics 64 (2014) pp. 121-134.

13

International Conference on Environmental Systems