J Acupunct Meridian Stud 2009;2(4):273−279

RE S E A RCH A RTI CL E

Thermal Properties of Direct and Indirect

Moxibustion
Seung-Ho Yi*
Acupuncture and Meridian Science Research Center, Kyung Hee University, Seoul, Korea

Received: Jun 2, 2009 Abstract

Accepted: Sep 24, 2009 Moxibustion therapy chiefly utilizes heat generated by the combustion of moxa.
Therefore, understanding the thermal properties of moxibustion is essential when
KEY WORDS: studying the mechanisms involved in moxibustion therapy. Therefore, we mea-
airflow; sured temperatures of direct and indirect moxibustion. For indirect moxibustion,
direct moxibustion; moxibustion on garlic was used. To determine the influence of the environment on
moxibustion, we applied airflow of 0.0−0.8 m/s. An increase in the airflow caused
moxibustion on garlic;
a concomitant increase in the maximum temperature of direct moxibustion, from
temperature;
160 to 300ºC and the time duration was reduced by half. However, the maximum
thermal effect
temperature of indirect moxibustion demonstrated the opposite effect, with the
temperature decreasing from 45 to 40ºC. This is attributed to the upstream airflow,
which indicates the importance of the air inside moxa. For indirect moxibustion
using garlic slices of different thicknesses, we found the optimum condition for the
buffer layer of a garlic slice. The maximum number of consecutive moxibustions
using one garlic slice was three. These results are consistent with traditional meth-
ods. This observation illustrates that the importance of garlic slices in modulating
the combustion heat and proper thermal stimulus to the patient.

1. Introduction efficacy of moxibustion. However, moxibustion has

In traditional medicine from the Western Pacific
region, moxibustion is one of the main therapeutic
modalities, delivering heat to specific areas such
as meridian points. Heat stimulation [1−3] and
some pharmacological effects from moxa (Artemisia
spp.; mugwort) [4,5] have been attributed to the
therapeutic efficacy of moxibustion. The efficacy
of the chemicals in moxa is enhanced by the com-
bustion heat of the moxa, promoting the hypoder-
mic transmission of beneficial chemicals. Therefore,
it is apparent that heat plays a major role in the
some disadvantages, including the possibility of burns
from excessive heat [6]. To resolve this issue, it is
imperative to study the thermal characteristics of
moxibustion.
Moxibustion therapy incorporates direct moxi-
bustion and indirect moxibustion. In direct moxi-
bustion, the moxa contacts the skin surface directly
and for indirect moxibustion, buffer substances such
as ginger, garlic or salt are placed between the
skin and moxa. Air or paper can be used as buffer
layers for modern indirect moxa devices [7]. The
presence of additional buffer layers gives indirect

*Corresponding author. Acupuncture and Meridian Science Research Center, Kyung Hee University, B101, 1 Hoegi-dong,
Dongdaemoon-gu, 130-701 Seoul, Korea.
E-mail: shyi@khu.ac.kr

©2009 Korean Pharmacopuncture Institute

274
moxibustion unique thermal characteristics. Chiba
et al reported the thermal and antiradical proper-
ties of indirect moxibustion and noted that the gap
between moxa and the skin was more important
than moxa weight [8]. They also stated that the
maximum temperature induced by indirect moxi-
bustion was about 65ºC on the skin surface, and
45ºC in the subcutaneous layer [9]. A research group
measured the radiation curve of direct and indirect
moxibustion to develop moxibustion instruments
[10]. A study of infrared radiation (IR) from indirect
moxibustion suggested that there is a difference in
IR between non-traditional and traditional thermal
materials [11].
In spite of its many clinical applications, few
studies have been reported on the efficacy of mox-
ibustion, partly due to the lack of a reliable sham
moxibustion. To further study themoxibustion
effect, a robust sham moxibustion was recently
introduced [12].
In this report, we measured the temperature of
both direct and indirect moxibustion on garlic at
different time course. The optimum condition for
indirect moxibustion was determined and was in
accordance with the traditional method on the con-
secutive number of cones and thickness of a garlic
slices. To examine the environmental effect on the
combustion of moxa, the airflow dependency of
both types of moxibustion was also studied. Some
guidelines for the practicing clinical environment
were suggested.
2. Materials & Methods
2.1. Materials
In this study, we used moxa cones with a diameter,
height and weight of 14.6 ± 0.6 mm, 15.5 ± 0.7 mm
and 0.28 ± 0.02 g, respectively. The main ingre-
dient of the moxa was dried mugwort leaf, pro-
duced in Kwangwha Island, South Korea. The moxa
cones were prepared by Oriental Hospital, Kyung
Hee University Medical Center, Seoul, Korea, and
have been applied in treating patients in the
hospital.
For indirect moxibustion, garlic slices were used.
Garlic was purchased at a local market and a clove
was cut into slices of similar thickness using a sur-
gical knife. A slice with the largest diameter was
selected for each experiment. Most of the slices
were derived from the central part of the cloves,
allowing for a higher degree of conformity in the
ingredients. The slices were then cut into a cir-
cular shape of 17 mm diameter. Clinically, garlic
slices of 1−2 mm have been generally used for indi-
rect moxibustion by practitioners [13]. In our study,
S.H. Yi
Airflow
Computer
Acquisition
system
Garlic slice
Moxa
Hot plate
Figure 1 A schematic diagram illustrating the experi-
mental set up.
the garlic slices of seven different thicknesses
(0.5, 1.0, 1.3, 1.5, 1.7, 2.5, and 3.0 mm) were used
during measurements. All garlic slices were individ-
ually prepared just before each measurement to
reproduce consistent conditions, such as moisture,
between experiments.
2.2. Experimental set up
The measurement set up is depicted in Figure 1.
The set up was installed on a stainless steel lab
bench with a forced ventilation system. An auto-
matic temperature acquisition module and 0.08 mm
thermocouple were used. As shown in Figure 1, a
thermocouple was placed between an aluminum
hot plate and the moxa cone for direct measure-
ment, and between a garlic slice and the hot plate
for indirect moxibustion. Detailed procedure is
described in Kim et al 2009 [14].
2.3. Procedure
2.3.1. Direct moxibustion
A thermocouple was positioned at the bottom
center of the moxa cone, which was located on the
hot plate. The acquisition software was set up and
ambient temperature was monitored. As the cone
top was ignited using a lit wooden toothpick, data
acquisition started. According to Traditional Chinese
Medicine practitioners, the fire must be minimized
just enough to light the moxa apex to ensure its con-
sistency and efficacy. After ignition, a window near
the bench was almost completely closed. The open-
ing at the bottom of the window supplied fresh air
for maintaining combustion and to prevent uncon-
trolled excessive airflow that might affect burning.
These procedures were repeated twice for each
airflow at 0.0, 0.3, 0.5, 0.6, 0.7 and 0.8 m/s.
Thermal properties of direct and indirect moxibustion 275

340 0.0 m/s 260 320

320

Maximum temperature (°C)

0.3 m/s 240 300
300 0.5 m/s 280
280 0.7 m/s 220 260

Duration time (s)

260
Temperature (°C)

240 200 240

220 220
200 180 200
180
160 160 180
140 160
120 140 140
100 120 Duration Maximum 120
80
60 temperature
40 0 0
20 0.0 0.2 0.4 0.6 0.8 1.0
0 Airflow (m/s)
0 50 100 150 200 250 300 350
Time (s) Figure 3 Duration time and maximum temperature of
direct moxibustion with all airflows. A clear anti-corre-
Figure 2 Temperature curves of direct moxibustion lation between the two properties can be observed.
with four different airflows.

Table 1 Time taken to reach the maximum temper-

2.3.2. Indirect moxibustion
ature of direct moxibustion with various air flows
All procedures were performed as for direct moxi- Airflow 0.0 0.3 0.4 0.5 0.6 0.7 0.8
bustion, except that the thermocouple was placed (m/s)
indirectly under the center of the moxa with the Time (s) 224 223 151 118 108 117 104
garlic slice in between. In our experiment, moxi-
bustion was performed four times for each garlic
slice before it was replaced by a fresh one. The 9
Increase rate
garlic slice was changed by the heat of combustion
Temperature change rate (°C/s)

and, as a result, its thermal properties were altered
during the course of the experiment. We considered
a temperature between 40−45ºC as the therapeutic
temperature window and defined the therapeutic
time as being when the temperature was in this win-
dow [15]. Therefore, by measuring the time course
of moxibustion temperature with each slice, we de-
termined the optimum conditions for the indirect
moxibustion. With the selection of the optimum slice
thickness, we acquired the thermal characteristics
of indirect moxibustion with various airflows and
compared these with the thermal characteristics
of direct moxibustion.
3. Results
3.1. Direct moxibustion
We measured the temperature at the lower center of
the moxibustion during various airflows (0−0.8 m/s),
as shown in Figure 2. No abrupt change in the tem-
perature was observed. With an increase in air-
flow, peak temperature increased and the curve
became gradually narrower. The shape of the curve
became considerably more symmetrical.
To quantify the broadness of the temperature
curve, we introduced duration time and maximum
time. The duration time was defined as the time
when the moxa temperature was higher than the
8 Decrease rate
7
6
5
4
3
2
1
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Airflow (m/s)
Figure 4 Comparison between increase rate and
decrease rate of direct moxibustion temperature.
hot plate temperature (34ºC) during its combus-
tion. The maximum time was determined as the
time taken to reach the maximum temperature.
Figure 3 shows the duration time and maximum
temperature of each moxibustion. There is an anti-
correlation between the duration time and maxi-
mum temperature. The maximum time is presented
in Table 1 and clearly demonstrates the amount of
peak shift. To find the change in rate on a tem-
perature curve, we simply calculated the slopes of
both sides at each temperature curve, as shown
in Figure 4. To obtain the values in the figure, two
points on the same side of a curve were selected,
which corresponded to 20% and 80% of its peak value.
276 S.H. Yi

A 55 B 55
1st 1st
2nd 2nd
50 3rd 50 3rd
4th 4th

Temperature (°C)
Temperature (°C)

45 45

40 40

35 35

30 30

0 0
0 100 200 300 400 0 100 200 300 400
Time (s) Time (s)

C 55 D 55
1st 1st
2nd 2nd
50 3rd 50 3rd
4th 4th
Temperature (°C)

Temperature (°C)
45 45

40 40

35 35

30 30

0 0
0 100 200 300 400 0 100 200 300 400
Time (s) Time (s)

E 55 F 55
1st 1st
2nd 2nd
50 3rd 50 3rd
4th 4th
Temperature (°C)

Temperature (°C)

45 45

40 40

35 35

30 30

0 0
0 100 200 300 400 0 100 200 300 400
Time (s) Time (s)
Figure 5 Results of four consecutive moxibustion on a single garlic slice of (A) 1.0 mm, (B) 1.3 mm, (C) 1.5 mm,
(D) 1.7 mm, (E) 2.5 mm and (F) 3.0 mm.

As shown in Figure 4, an anti-correlation is observed 3.2. Indirect moxibustion

between the increase rate and decrease rate of
each temperature curve. The change in rate var-
ied rapidly at airflows of 0.3−0.4 m/s. According
to these results, there seems to be a demarcation
level between the airflows of 0.3 m/s and 0.5 m/s.
The peak position, peak value and shape of the
temperature curves clearly show the presence of
the level.
3.2.1. Determination of the optimum garlic
thickness
We measured the temperature of indirect moxi-
bustion on garlic slices of various thicknesses.
Some results are shown in Figure 5. Results of four
consecutive moxibustion tests with each single
Thermal properties of direct and indirect moxibustion 277

garlic slice are depicted in the same figure for easy 46

comparison. 0.0 m/s
44 0.3 m/s
With 1.0−1.7 mm slices, therapeutic temperature 0.5 m/s
42 0.7 m/s
was achieved up to the third moxibustion and the

Temperature (°C)
curves showed broad therapeutic windows in this 40
condition. The maximum temperature of the fourth 38
moxibustion was higher than 50ºC, which was out- 36
side the therapeutic window. By contrast, the tem-
perature using the slice thickness of 2.5 mm and 34
3.0 mm never reached 50ºC. However, these indi- 32
rect moxibustions revealed relatively small thera- 30
peutic windows. For the 3.0 mm slice, the first two 5
moxibustions did not even reach 40ºC. These re- 0
sults implied that there is little therapeutic effect 0 100 200 300 400
with slice thickness of 2.5 mm and 3.0 mm and Time (s)
these indirect moxibustions are almost of no use. Figure 6 The temperature time course of airflow
Therefore, based on these results, we selected the dependency measurement with the optimum garlic
1.7 mm thickness as the optimum condition for in- thickness of 1.7 mm.
direct moxibustion and executed airflow dependency
measurement with it.
In Figure 5, the temperature curve of a 0.5 mm 46 200
slice was omitted as the maximum temperature
Maximum temperature (°C)
180
reached 90ºC, which was obviously much higher than 160

Therapeutic time (s)

44
the other thicknesses. This result clearly demon- 140
strated that there is a lower limit for the thickness 120
of a buffer layer for insuring efficacies of indirect 42
100
moxibustion. 80
40 60
3.2.2. Airflow dependency 40
4 Maximum temperature
2 Therapeutic time 20
Figure 6 represents typical temperature curves ob- 0 0
tained by an airflow dependency experiment. As 0.0 0.2 0.4 0.6 0.8 1.0
airflow increases, the curves shift to the left and Airflow (m/s)
their peak temperature decreases from 45ºC to Figure 7 Maximum temperature and duration time of a
below 40ºC. The decrease in temperature is exactly 1.7 mm garlic slice, extracted from the results in Figure 6.
opposite to that of direct moxibustion where tem-
perature goes from 130ºC up to higher than 300ºC.
The curve became narrower with increased airflow,
Table 2 Duration time of indirect moxibustion using
demonstrating that both the duration and thera-
a 1.7 mm garlic slice with various airflows
peutic time were shortened. Airflow dependency
of the maximum temperature and therapeutic time Airflow 0.0 0.3 0.4 0.5 0.6 0.7
is summarized in Figure 7 and the duration time in (m/s)
Table 2. Time (s) 231 228 238 199 152 149

4. Discussion
Our studies have focused mainly on the primary
factors that affect the thermal characteristics of
moxibustion and the optimum therapeutic condi-
tions of indirect moxibustion. We believe that these
studies are necessary to understand the mecha-
nisms involved in moxibustion therapy and to de-
velop effective moxibustion devices.
Our data showed the importance of airflow con-
trol, regardless of the nature of moxibustion. Dif-
ferences in the maximum temperature of both types
of moxibustion were attributed to the presence of
a garlic slice. However, the decrease of the maxi-
mum temperature of indirect moxibustion with air-
flow requires explanation. The transport rate of
thermal energy to the thermocouple is reduced for
a reason.
We hypothesized that the difference in the spe-
cific heats of air inside the moxa and water in the
garlic slice as well as the upflow might be responsi-
ble for the result. Ignited moxa heated the air sur-
rounding moxa tissues and this hot air transferred
278
thermal energy to the bottom of the moxa by con-
vection, where a thermal detector was installed.
However, for indirect moxibustion, hot air must first
warm a garlic slice, which has a very high specific
heat compared with air. As airflow increased, more
hot air moved to the top of the moxa and less hot air
delivered its thermal energy to the slice. Since the
thermal detector was located below the slice, the
temperature acquired should be lower with airflow.
If this explanation is valid, both the importance of
the amount of air contained inside moxa and the
affect of convection as a heat transport mechanism
need to be reconsidered.
According to the results of direct moxibustion,
environment control, such as ambient airflow,
seems to be crucial to ensure reliable and repeat-
able moxibustion. Results in Figures 3 and 4 show
that airflow dependency of the duration time and
maximum temperature in the slower flow regime
is less severe than that in the faster flow regime.
Any action increasing the airflow around moxibus-
tion might cause unwanted heat stimulus to
patients during the treatment. Practically, control-
ling temperature and heat of moxibustion as a
function of time is not easy in clinical application
and this may cause some inconsistency in moxibus-
tion efficacy.
The results of indirect moxibustion on garlic
slices with different thicknesses provided valuable
clues on the mechanism of efficacy in moxibustion
treatment. For moxibustion on garlic or ginger,
there have been certain rules in Traditional Korean
Medicine on the usage of the buffer layers for mul-
tiple moxibustions at a single position. The thick-
ness of the buffer layer is approximately a half
of 10th cun (1 cun = 33 mm) and the layer must be
replaced after three cones (the number of moxi-
bustion) have been used [16]. The recommended
thickness of the layer is about 1.6−1.7 mm, which
is in agreement with our optimal thickness judged
by its thermal properties. During consecutive mox-
ibustions with slices of 1.0−1.7 mm thickness, the
fourth moxibustion showed a peculiarly high tem-
perature of more than 50ºC, which called for
replacement with a new slice. These two facts
strongly suggested that the reason for introducing
these buffer layers is to control or modulate the
heat from ignited moxa. The potential for utilizing
the chemicals from moxa seems to be unlikely as
the garlic slice prevents them from reaching the
skin surface.
Our experiments have several limitations. We
used forced upstream airflows for our measure-
ments as the simplest possible case. In reality, the
direction of airflow is arbitrary and therefore, fur-
ther study is necessary to reflect the effect of arbi-
trary airflow occurring in the clinic. Measurement
S.H. Yi
results would be affected by the aluminum hot
plate. Aluminum and the skin have very different
thermal properties such as specific heat and thermal
conductivity and, in addition, the skin has its own
fluid cooling system to cope with excessive external
heat. Therefore, in clinical application, the moxi-
bustion temperature needs to be higher to stimulate
the skin.
Acknowledgments
The author wishes to sincerely thank the volunteer
students of Hansung Science High School for as-
sistance with some measurements. Gratitude is
also extended to the Oriental Hospital, Kyung Hee
University Medical Center for providing moxa
cones, and to Dr Dawn Nowlin for correcting the
manuscript.
This work was partly supported by the R&E Pro-
gram (2008-02-005) and Basic Science Research Pro-
gram through the National Research Foundation,
Ministry of Education, Science and Technology (R11-
2005-014).
References
1. Greenberg RS. The effects of hot packs and exercise on
local blood flow. Phys Ther 1972;52:273−8.
2. Petrofsky JS, Laymon M. Heat transfer to deep tissue: the
effect of body fat and heating modality. J Med Eng Technol
2009;33:337−48.
3. Touchberry C, Le T, Richmond S, Prewitt M, Beck D, Carr D,
et al. Diathermy treatment increases heat shock protein
expression in female, but not male skeletal muscle. Eur J
Appl Physiol 2008;102:319−23.
4. Kobayashi K. Organic components of moxa. Am J Chin Med
1988;16:179−85.
5. Sakagami H, Matsumoto H, Satoh K, Shioda S, Ali CS,
Hashimoto K, et al. Cytotoxicity and radical modulating
activity of moxa smoke. In Vivo 2005;19:391−7.
6. Yamashita H, Tsukayama H, Tanno Y, Nishijo K. Adverse events
in acupuncture and moxibustion treatment: a six-year sur-
vey at a national clinic in Japan. J Altern Complement Med
1999;5:229−36.
7. Dongbang Acupuncture Inc. Available at: http://www.
dbneedle.com [Date accessed: May 29, 2009]
8. Chiba A, Nakanishi H, Chichibu S. Thermal and antiradical
properties of indirect moxibustion. Am J Chin Med 1997;25:
281−7.
9. Chiba A, Nakanishi H, Chichibu S. Effect of indirect moxi-
bustion on mouse skin. Am J Chin Med 1997;25:143−51.
10. Hong W, Jing J, Song J, Xu Y, Cai J. Research on the Mea-
suring and Duplication Techniques of Curve of the Moxibus-
tion in Traditional Chinese Medicine. In: Engineering in
Medicine and Biology Society, 2005. IEEE-EMBS 2005. 27th
Annual International Conference of the IEEE, September
1−4, 2005; Shanghai 2006:4267−70.
11. Shen X, Ding G, Wei J, Zhao L, Zhou Y, Deng H, et al. An
infrared radiation study of the biophysical characteristics
of traditional moxibustion. Complement Ther Med 2006;14:
213−9.
Thermal properties of direct and indirect moxibustion
12. Zhao B, Wang X, Lin Z, Liu R, Lao L. A novel sham moxibus-
tion device: A randomized, placebo-controlled trial. Com-
plement Ther Med 2006;14:53−60.
13. Aung S, Chen W. Clinical introduction to medical acu-
puncture. Thieme New York: Medical Publishers, 2007:164.
14. Kim H, Shim I, Yi SH, Lee H, Lim HS, Hahm DH. Warm needle
acupuncture at Pungsi (GB31) has an enhanced analgesic
279
effect on formalin-induced pain in rats. Brain Res Bull
2009;78:164−9.
15. Schepers RJ, Ringkamp M. Thermoreceptors and thermosen-
sitive afferents. Neurosci Biobehav Rev 2009;33:205−12.
16. Xiao F, Mu J. Acupunctu and Moxibustion. Advanced TCM
series Volume 6. Science Press Beijing. Printing House of
the China Academy of Science, 2000:237.