Content uploaded by Luca Chiaraviglio
Author content
All content in this area was uploaded by Luca Chiaraviglio on Nov 11, 2019
Content may be subject to copyright.
5G Technology: Which Risks From the Health
Perspective?
Luca Chiaraviglio, Marco Fiore and Edouard Rossi
Abstract The deployment of a new generation of mobile communication networks
requires the installation of a dedicated radio access infrastructure. In the case of 5G,
this unavoidable practice is creating a controversy about the potential issues for the
public health that new radio base stations may entail. In this chapter, we discuss five
major health risk allegations against 5G, namely: (i) the links between insurgence
of tumors and exposure to ElectroMagnetic Fields (EMFs) generated by 5G; (ii)
the increase of EMF levels due to an uncontrolled proliferation of 5G sites; (iii) the
health risks associated to emissions in the new mm-Wave spectrum adopted by 5G;
(iv) the uncertainty about the actual 5G EMF emission levels caused by the absence
of dedicated measurements; (v) the impossibility to remove the previous uncertainty
determined by the lack of measurement tools suitable for 5G technologies. We exa-
mine these arguments from an engineering perspective, by tacking into account the
outcome of state-of-the-art scientific studies, the current relevant regulations and the
technical features of 5G technologies. Our review indicates that there is no incon-
trovertible scientific evidence supporting any of the five claims. While we second
the need for further investigations, we also remark a factual fabrication of fake news
on the risks of 5G for the public health, which may severely distort the perception
of this technology by the population at large.
Luca Chiaraviglio and Edouard Rossi
Department of Electronic Engineering, University of Rome Tor Vergata, Rome, Italy, and Con-
sorzio Nazionale Interuniversitario per le Telecomunicazioni, Italy e-mail: luca.chiaraviglio@
uniroma2.it
Marco Fiore
Institute of Electronics, Computer and Telecommunication Engineering, National Research Coun-
cil of Italy, Turin, Italy, e-mail: marco.fiore@ieiit.cnr.it
1
2 Luca Chiaraviglio, Marco Fiore and Edouard Rossi
1 Introduction
An excerpt from a January 1900 article of a reporter based in Portsmouth, New
Hampshire [1] reads: “Electric, electric, electric! [...] The untapped electrical fluid
leaking from these outlets and wires, we are told, may cause serious bodily damage
and –with prolonged exposure– possibly death. We are certainly gaining momentum
in this Modern Day, but can anyone tell us where we are headed?”. Almost 120
years later, electricity is a globally widespread technology that is instrumental to
the daily activities of billions and is considered completely secure –as long as it is
handled abiding by minimum safety standards.
Today, a similar controversy arises concerning the fifth generation (5G) of wire-
less communication systems. As 5G heads into commercialization after a decade
of research and development, arguments are made about the potential risks of the
technology for the public health. The fear of 5G, propagating via social media and
fuelled by an apparently diffused belief that the economic interests of industries and
governments may be the sole drivers for the deployment of the new technology, has
grown to the point of pushing politicians to publicly advocate for experimental trials
to be withhold in major cities in Europe [2].
Are these worries justified, or are we facing a similar unnecessary anxiety as that
recorded for electricity early into the twentieth century? To help shedding light on
this matter, we present and discuss five main allegations against the installation of
5G radio base stations, stemming from the potential negative effects that the new
infrastructure may have on public health. Our analysis is carried out from a purely
technical perspective, by leveraging multiple sources of information that include the
applicable Italian regulations on EMF emissions, the technological characteristics
of 5G radio base stations, and the results of scientific studies on the effect of EMF
exposure on living beings. As such, we offer a different viewpoint from that adopted
by works in the literature that focus on clinical or medical considerations about EMF
exposure (see, e.g., [3]). Differently from the medical/clinical field, our goal is in fact
to better contextualize potential health risks with respect to the actual engineering
of 5G networks.
We remark that our examination concerns the possible health dangers associated
with the deployment of new 5G radio base stations. In other words, we aim at bet-
ter understanding if there are evidences that the EMF that will be generated by the
5G radio access infrastructure may jeopardize the health of citizens. This choice
is motivated by the consideration that distress in the public opinion is generally
connected to the installation of new base stations. However, the reader should be
aware that this chapter only explores one facet of the problem: other potential risks,
which are less criticized, may exist, in particular with regards to the personal con-
sumption of new communication technologies. Indeed, a more frequent utilization
of smartphones determined by the growing compulsion for mobile services causes
an inherent higher exposure of individuals to the EMFs generated by the end termi-
nals themselves –which, in fact, are often the main source of EMFs in the proximity
of users already today [3, 4, 5, 6].
5G Technology: Which Risks From the Health Perspective? 3
In the remainder of the chapter, we first review each of the five allegations on
health dangers associated with the installation of 5G radio base stations in Sections 2
to 6. Based on these reviews, we draw conclusions and discuss opportunities for
future research in Section 7.
2 Allegation I: Exposure to EMF generated by 5G radio base
stations increases the risk of developing tumors
The first and most prominent dispute is that exposure to the EMFs of 5G radio
base stations (and in fact to radio-frequency transmissions in general) induces a
higher risk to develop specific classes of tumors. Claims in this sense are typically
supported by citations to the following documents.
1. In two different studies released in 2018, the USA-based National Toxicology
Program (NTP) experimentally proved that exposure to high levels of radiation
from radio base stations, like those deployed for 2G and 3G mobile networks,
is associated with the emergence of heart tumors in male rats [7, 8]. Also, the
studies evidence a possibility that a higher incidence of brain and adrenal gland
tumors is also correlated with radio base station emissions.
2. A study of the Ramazzini Institute, an Italian non-profit private organization, re-
leased shortly after those carried out by the NTP, corroborated the health risks
associated with radio base station emissions [9]. Specifically, the research reports
an increase in the incidence of brain and heart tumors in Sprague-Dawley rats ex-
posed to EMF generated by a radio base station. In fact, a statistically significant
increase in the incidence of heart Schwannomas was observed in treated male rats
already at exposure levels much lower than those considered in the NTP studies,
and compatible with those experienced by people on a daily basis.
3. The International Agency for Research on Cancer (IARC) classified in 2011 non-
ionizing waves (including the waves generated on frequencies adopted by radio
base stations) as possibly carcinogenic to humans. The IARC classification will
be actually revised in the near future, also based also on the outcomes of the re-
ports mentioned above. Non-ionizing waves are likely to be classified as probably
carcinogenic to humans after such a revision.
To provide a sound analysis of the results above and subsequent claims, we first
present the current regulations on EMF exposure limits. We then discuss the out-
comes of the NTP and Ramazzini studies, as well as the IARC classification, also in
the light of the EMF limits enforced by aforementioned applicable laws.
4 Luca Chiaraviglio, Marco Fiore and Edouard Rossi
Table 1 EMF limits comparison of Italian [10, 11] and ICNIRP [12] regulations for general public
areas, residential areas and proximity of a radio base station site to a sensitive place. The Italian
limits are more stringent compared to the ICNIRP-based ones.
Regulations
Areas ICNIRP Italian
28 [V/m], f∈(10,400] [MHz] 60 [V/m], f∈(0.1,3] [MHz]
1.375·f1/2[V/m], f∈(400,2000] [MHz] 20 [V/m], f∈(3,3000] [MHz]General Public
61 [V/m] f∈(2−300] [GHz] 40 [V/m], f∈(3,300)[GHz]
28 [V/m], f∈(10,400] [MHz]
1.375·f1/2[V/m], f∈(400,2000] [MHz]Residential
61 [V/m], f∈(2−300] [GHz]
6 [V/m] for all frequencies
Site None Minimum distance constraints
2.1 Regulations on EMF emissions
The fact that high levels of EMFs are hazardous for human health has been very
well known for decades, as proven by a vast scientific literature on the subject.
Exactly for this reason, there exist strict regulations that limit the maximum amount
of EMFs from radio-frequency and electrical sources that citizens shall be exposed
to. The allowed EMF levels are determined based on the present knowledge we
have about the health effects of radio emissions, with the goal of ensuring the public
safety.
EMF limits are in fact not uniform globally, and the related legislation vary across
countries. In our analysis, we will consider the EMF limits enforced in Italy. Where
relevant, we will also refer to the limits defined by the International Commission
on Non-Ionizing Radiation Protection (ICNIRP), which are adopted in many coun-
tries worldwide, including most European countries. We summarize the two sets of
limits in Table 1. Since the objective risks of EMFs, mainly involving heating of
the radiated tissues, strongly depend on the frequency at which the EMF is gen-
erated, the limits vary with frequency. In addition, while the ICNIRP limits do not
differentiate among the different portions of the territory, Italian regulations pursue a
different strategy, and distinguish between general public areas where people do not
stay for long amounts of time, and residential areas where people tend to live and
work. Moreover, different municipalities in Italy, such as the city of Rome, adopt
additional restrictions, typically by prohibiting the installation of radio base stations
in proximity of sensitive places like public parks, hospitals, or schools.
The frequencies concerned with 5G and other mobile communication technolo-
gies are all above 400 MHz. Therefore, by looking at Table 1, it is evident that
Italian laws impose limits that are already stricter than those recommended by IC-
NIRP for general public areas (at 20 [V/m]), and substantially lower in residential
areas (at 6 [V/m]). In general, the ICNIRP limits are such that a lower EMF ex-
posure does not produce any known hazardous effect on health. The Italian limits
are much more stringent on the basis of a precautionary principle, so as to consider
the potential impact of still unknown health effects. In the following, we will only
consider the nationwide limits applied in Italy, and disregard further constraints that
5G Technology: Which Risks From the Health Perspective? 5
Table 2 Tests performed in the NTP studies [7, 8], reporting the test name, the adopted frequency,
the values of EMFs measured during each test, the average EMF over 24 hours, the positioning of
the EMF and the 24-hour EMF with respect to the Italian limits.
Test Frequency Test EMF 24h EMF EMF ≥20 [V/m] limit? 24h EMF ≥6 [V/m] limit?
GSM High 900 [MHz] 291 [V/m] 111 [V/m] Yes, 15 times higher Yes, 19 times higher
GSM Med 900 [MHz] 206 [V/m] 78 [V/m] Yes, 10 times higher Yes, 13 times higher
GSM Low 900 [MHz] 147 [V/m] 56 [V/m] Yes, 7 times higher Yes, 9 times higher
GSM High 1900 [MHz] 257 [V/m] 98 [V/m] Yes, 13 times higher Yes, 16 times higher
GSM Med 1900 [MHz] 178 [V/m] 68 [V/m] Yes, 9 times higher Yes, 11 times higher
GSM Low 1900 [MHz] 126 [V/m] 48 [V/m] Yes, 6 times higher Yes, 8 times higher
may be imposed by local administrations. These are highly heterogeneous, hence
too involved to discuss –although their application may further reduce the EMF
exposure of inhabitants due to radio base station emissions.
It is also important to mention that the EMFs values to be compared against a
limit are computed in a different way by two models. In both cases, these values
are computed as an average of the EMFs measured over a time interval of a fixed
duration. However, the ICNIRP limits consider an averaging interval of 6 [minutes]
(currently under revision, with proposals to increase the interval up to 30 minutes),
whereas the Italian limits are enforced on an average EMF computed over 24 hours
for residential areas and over 6 minutes for general public areas, respectively. There-
fore, when analyzing the EMF levels reported by different studies in the literature,
we will apply suitable scaling factors to align them with the EMF measurement con-
ditions prescribed by law. For instance, an EMF measured at a specific location of
10 [V/m] over 18 hours plus 0 [V/m] over the 6 subsequent hours correspond to an
average EMF of 7.5 [V/m] over 24 hours. This last value is the one that must be
used to verify adherence to the limits set by the relevant Italian laws.
2.2 A review of the NTP studies
As mentioned before two studies carried out by NTP showed a higher emergence
of heart tumors in male rats exposed to EMF similar to those generated by 2G and
3G radio base stations [7, 8]. Table 2 summarizes the settings of the experiments re-
ported in the NTP studies. The table details: the test name, the operating frequency,
the EMF exposure of patients considered in the experiments (computed as the aver-
age EMF over the different chambers used in the test), the 24-hour average EMF,1
the comparison of the EMFs against the Italian limit of 20 [V/m] for general public
areas, and the comparison of the 24-hour EMF against the Italian limit of 6 [V/m]
for residential areas. Based on the table, several considerations are in order.
1We assume that the average EMFs reported in [7, 8] are computed from the EMF values recorded
during the exposure window (which is set to 9 hours and 10 minutes per day). Therefore, the 24-
hour average EMF is scaled by 9[hour]10[minutes]
24[hours] times the value of EMF reported in the NTP studies.
This is a conservative assumption to compare the 24-hour EMF against the 6 [V/m] Italian limit.
6 Luca Chiaraviglio, Marco Fiore and Edouard Rossi
First, the technology taken into account in both studies is the GSM one2, which
was part of second generation (2G) mobile networks. Newer generations of radio
base stations (implementing 3G, 4G and the forthcoming 5G networks) are demon-
strably more effective than 2G base stations in reducing the radiated power, and
consequently their induced EMF levels. Therefore, the outcome of both studies is
not representative of 3G, 4G or 5G technologies, but just of higher-radiating 2G
base stations whose usage is today generally limited in developed countries.
Second, the EMF exposure levels in both studies are inflated at values that are
orders of magnitude higher than the EMF levels measured in proximity to users from
cellular networks under operation, as shown, e.g., by the recent survey of Sagar et
al. [6]. Actually, both NTP studies place the 2G radio base station very close to the
radiated animals, at a few meters of distance. Instead, in operational networks, the
closest base stations are typically tens or hundreds of meters away from the user, as
they are placed on top of poles and above rooftops of soaring buildings. Moreover,
the zone in proximity to the base station is generally also made not accessible to the
public. This makes the exposure conditions assumed by the NTP studies unrealistic
from an engineering viewpoint.
Third, the aforementioned settings adopted by NTP result in EMF levels that
are largely beyond the Italian EMF limits, i.e., between 6 and 15 times higher than
the general public area limit of 20 [V/m], and between 8 and 19 times higher than
the residential area limit of 6 [V/m]. In fact, such EMF levels are well above the
international limits defined by ICNIRP, i.e., 41.25 [V/m] for the 900 [MHz] band
and 59.93 [V/m] for the 1900 [MHz] band, respectively. As a result, the conditions
reproduced in these experiments are not encountered in an operational network in
Italy or Europe. In fact, the levels of EMFs considered in the studies are so high that
they would even trigger tissue heating effects.
Fourth, as also reported by a note of ICNIRP [13], the NTP measured a survival
rate of male rats not exposed to any source of EMF significantly lower than the ones
exposed to EMFs, for all the tests. This issue may have also introduced a bias in
the presented results, as the different lifespan of rats may have impacted the tumor
emergence probability.
Based on the previous observations, we conclude that the NTP studies [7, 8] do
not report any evidence of carcinogenicity if the EMFs generated by the radio base
stations are below the limits imposed by the Italian law or recommended by ICNIRP.
We remark that such limits are enforced for all cellular network technologies under
operation, including the forthcoming 5G ones.
2.3 A review of the Ramazzini Institute study
The study of the Ramazzini Institute reports a higher incidence of brain and heart
tumors in rats exposed to EMF generated by a radio base station [9]. Table 3 illus-
2The NTP studies cover also CDMA-based 2G radio base stations, for which similar conclusions
hold, and not reported here for the sake of brevity.
5G Technology: Which Risks From the Health Perspective? 7
Table 3 Tests performed in the study of the Ramazzini Institute [9], reporting the test name, the
adopted frequency, the values of EMFs measured during each test, the average EMF over 24 hours,
the positioning of the EMF and the 24-hour EMF with respect to the Italian limits.
Test Frequency Test EMF 24h EMF EMF ≥20 [V/m] limit? 24h EMF ≥6 [V/m] limit?
GSM 5 1800 [MHz] 5 [V/m] 4 [V/m] No No
GSM 25 1800 [MHz] 25 [V/m] 20 [V/m] Yes, 1.3 higher Yes, 3 times higher
GSM 50 1800 [MHz] 50 [V/m] 40 [V/m] Yes, 3 times higher Yes, 8 times higher
trates the settings of the tests performed by the Ramazzini Institute, by detailing:
the test name, the considered frequency, the rat EMF exposure measured during the
experiment, the 24-hour average EMF,3and the positioning of such EMF with re-
spect to the Italian limit of 20 [V/m] for general public areas and of 6 [V/m] for
residential areas.
In fact, the study found a statistically significant increase of one disease (i.e.,
the Intramural Schwannoma) only for male rats exposed to an EMF of 50 [V/m],
labelled as test GSM 50 in the table. In addition, no statistically significant increase
of diseases has been found for female rats, as well as when the male and female
subsets are considered together, for the same test. By analyzing the values reported
in Tab 3, we can note that the test GSM 50 considers exposure levels that are largely
beyond the maximum EMF limits alowed in Italy. More precisely, the EMFs in
this test are 3 times higher than the general public limit and 8 times higher than the
residential area limit. All the other tests, considering lower EMFs (namely test GSM
5 and test GSM 25), did not find any evidence of health risks associated to the EMF
emissions by the radio base station.
In all cases, similarly to the NTP studies, these tests were conducted in condi-
tions not applicable to users served by a real-world modern cellular networks. These
include (i) rats being placed at a very short distance of a few meters from the radio
base station, (ii) the adoption of an outdated and power-inefficient technology, i.e.,
the 2G one. We believe that such settings severely limit the generalization of the re-
sults to base stations deployed in cities in developed countries, which are generally
located far from the users and radiate a lower EMF compared to 2G.
In the light of these considerations, we conclude that, also in the case of the
Ramazzini Institute study [9], there is no proof of carcinogenicity of the EMFs gen-
erated by radio base stations that operated within the Italian limits.
2.4 Comment on the IARC classification
As reported by a dedicated note of ICNIRP [13] and outlined above, the studies from
NTP and the Ramazzini Institute have limitations that do not allow substantiating
the carcinogenicity of radio-frequency EMFs generated by radio base stations. Con-
3The 24-hour average EMF is computed by scaling the EMF imposed during the experiment (i.e.,
5 [V/m], 25 [V/m], 50 [V/m]) by the factor 19 [hours]
24 [hours], since the experiment duration of [9] is set to
19 [hours] per day.
8 Luca Chiaraviglio, Marco Fiore and Edouard Rossi
sequently, it is not expected that the IARC will change the classification of EMFs
from possibly carcinogenic (level 2B) to probably carcinogenic (level 2A).
In any case, we remind that, even for carcinogens at the maximum level (level
1), the dose plays a crucial role in determining the carcinogenicity of the sub-
stance/mixture/exposure. For example, a low dose of a carcinogen may have no
impact on health. So far, there are no evidences that exposure to EMFs generated
by radio base stations under realistic conditions (e.g., distance from the radio base
station in the order of dozen meters and more) and received EMFs below the Italian
EMF limits are hazardous for the public health.
3 Allegation II: 5G will bring an uncontrolled proliferation of
radio base stations and EMF levels
A second dispute related to health risks of 5G is that the deployment of this tech-
nology over the territory will result in an uncontrolled proliferation of 5G radio
base stations and consequently of EMF levels. This allegation is somehow linked to
speculations that the Italian EMF limits will be increased from the current 6 [V/m]
limit to 61 [V/m] defined (for the 2-300 GHz band) by ICNIRP, in order to ease the
installation of such a large number of new radio base stations.
We first consider the impact of deploying a large number of radio base stations
in a portion of territory to formally demonstrate that, contrary to a common belief,
this condition allows to steadily reduce the transmitted power of each base station,
compared to the case in which few base stations are installed. In addition, we show
with a simple numerical example that the average power radiated over the territory
is not increased when the number of base stations is increased.
Let us assume a simple scenario, where the goal of the operator is to ensure that
the power received from a radio base station in any served portion of the territory
is above a minimum level, which is needed, e.g., to guarantee the connectivity to
users. Let us formally denote the received power and the minimum power level as
PRand PR
min, respectively. In our scenario, the operator has to ensure that PR≥PR
min
over the whole target geographical region. The received power PRdepends on the
power emitted by each radio base station, denoted as PE. In order to compute PR
from PE, we need to consider the attenuation affecting the electromagnetic wave
that traverses the air on the path between the radio base station and the portion of
territory served by the base station. More formally, we need a propagation model
that defines how PEis reduced with distance, and use that to retrieve PR. In our
illustrative example, and for the sake of clarity, we will consider a simplistic model
where the emitted power is scaled by the distance dfrom the serving base station,
elevated to a path loss exponent γ.4The relationship between PRand PEis then
expressed as:
4More complex models integrate also other features, such as the antenna gain, the fading effect
and the working frequency. These terms are intentionally omitted here, as they are not needed to
prove our point.
5G Technology: Which Risks From the Health Perspective? 9
PR=PE
dγ(1)
The value of γtypically depends on the attenuation conditions (e.g., line of sight or
non line of sight with respect to the serving radio base station).
Under this model, the condition that the minimum power level PR
min be received
in each portion of the territory is equivalent to the following equation:
PE
dγ
max
=PR
min (2)
where dmax is the maximum distance between a portion of territory served by a
given radio base station and the serving base station. Intuitively, dmax represents the
maximum coverage distance of one radio base station over a portion of the territory.
Now, let us consider two cases with a different number of radio base stations.
In the first one, our portion of territory is served by a set of uniformly distributed
radio base stations, with maximum coverage distance and emitted power denoted
as dmax (1)and PE(1), respectively. In the second case, let us consider a set of uni-
formly distributed radio base stations, each of them with coverage distance dmax(2)
and emitted power PE(2). Let us assume that the number of radio base stations in
(2) is higher than in (1). Under this assumption, the maximum coverage distance
dmax (2)is lower than dmax (1), due to the fact that, with a larger number of radio
base stations, each of them has to ensure a lower coverage radius to serve the por-
tion of territory under consideration. Let us denote with kthe ratio between dmax (1)
and dmax (2):
k=dmax (1)
dmax (2)(3)
By applying Equation (2), we get:
PE(2)
dγ
max (2)=PR
min =PE(1)
dγ
max (1)(4)
By considering the equivalence of the first and the third term and by adopting Equa-
tion (3) to express dmax (2)=dmax (1)
k, we get:
kγPE(2)
dγ
max (1)=PE(1)
dγ
max (1)(5)
which can be simplified to:
PE(2)=PE(1)
kγ(6)
The previous condition states that, when the number of base stations is increased
(case (2) compared to case (1)), the emitted power from each base station is reduced
by a factor kγ. For example, if we consider γ=2 (a common setting in the literature
in case of free space propagation), a circle portion of territory, and circle coverage
provided by the radio base stations, it holds that k=4 when passing from one radio
10 Luca Chiaraviglio, Marco Fiore and Edouard Rossi
(a) 1 base station (b) 16 base stations
(c) 256 base stations (d) 1024 base stations
Fig. 1 Power radiated over the territory for different number of base stations - in dBm scale. A
minimum power level of -90 [dBm] is ensured. The close-in propagation model of [14] with a
5G base station frequency of 3700 [MHz] is adopted. The power radiated by each base station is
reduced when the number of base stations is increased (Figures best viewed in colors).
base station in deployment (1) to 16 radio base stations in deployment (2). There-
fore, the emitted power of each radio base station in deployment (2) is reduced by
kγ=16 compared to deployment (1). This is beneficial for the received power PR
over the territory, which tends to be better distributed and in general not increased
in case (2) compared to case (1).
To give more insight, Figure 1 reports the power radiated over the territory for dif-
ferent numbers of deployed base stations, by assuming a minimum received power
of −90 [dBm], a 5G base station frequency of 3700 [MHz], the close-in propagation
model of [14] (which is more complex and realistic for 5G compared to a model
solely based on distance),5an isotropic radiation pattern of each base station (uni-
form in all the directions), and coverage of each BS corresponding to a Voronoi
tessellation (which is computed from the positions of the base station sites).
5The shadowing component in the original model of [14] is not included in our work for the sake
of simplicity.
5G Technology: Which Risks From the Health Perspective? 11
In particular, when there is a single base station (Figure 1.(a)), the received power
is clear higher than -90 [dBm] for a vast zone in the surroundings of the base sta-
tion. Interestingly, as soon as the number of base stations is increased (Figures 1-
(b),(c),(d)), it is possible to dramatically reduce the amount of power radiated by
each base station, with a reduction of the zone where the received power is higher
than −90 [dBm]. These outcomes are in accordance with the model presented in
Equation (1)-(6). In addition, we compute the average power radiated over the ter-
ritory, which is equal to: i) −77.03 [dBm] for 1 base station, ii) −77.76 [dBm] for
16 base stations, iii) −79.26 [dBm] for 256 base stations, and iv) −80.17 [dBm]
for 1024 base stations. Therefore, we can note that the average power (and conse-
quently the EMF) radiated over the territory is not increased when the number of
base stations deployed over the territory is increased.
Summarizing, we report evidences that adding more base stations to serve a ter-
ritory does not introduce an increase in the received power (and consequently of the
EMFs) compared to the case in which few base stations are installed.
Clearly, in order to support the 5G service, new radio base stations will be in-
stalled over the territory. The 5G network will be run in parallel to the 2G, 3G and
4G radio base stations that are already deployed over the territory. In addition, each
of these technologies is supported by a set of operators. These facts, coupled also
by the very strict regulations that are enforced in Italy, allow us to hypothesize that
5G will not result into a large proliferation of new radio base stations. This is espe-
cially true in urban areas, where most of the sites hosting base stations are already
saturated, i.e., it is not possible to install new radio base stations without violating
the 6 [V/m] limit. In any case, however, the installation of the new 5G radio base
stations will be done in accordance to the EMF regulations. As a result, the EMF
radiated over the territory will be strictly controlled.
Finally, concerning potential relaxations of the 6 [V/m] EMF limits in Italy, we
recall that there is not a single EMF limit in Italy: as reported in Table 1, the EMF
regulations include multiple EMF limits, depending on the portion of territory under
consideration and the adopted frequency. Anyway, we could not find any supporting
evidence that the regulator will increase the 6 [V/m] limit for residential areas.
4 Allegation III: 5G radio base stations will exploit mm-Waves,
which are dangerous
The third dispute brought to demonstrate the danger of 5G is that this technology
will exploit new frequencies, called mm-Waves, which have not been used before
and whose impact on the health is not known. In this regard, we recall that 5G will
not use exclusively the mm-Wave frequencies. Table 4 reports the outcome of the
5G frequency auction that was completed in Italy in 2018: the frequencies in the
5G auction include sub-GHz (at 700 [MHz]), sub-6 GHz (at 3700 [MHz]) and mm-
Wave (at 26 [GHz]). In particular, the 700 [MHz] and the 3700 [MHz] frequencies
are close to frequencies already used by 2G/3G/4G networks under operation. Actu-
12 Luca Chiaraviglio, Marco Fiore and Edouard Rossi
Table 4 Outcome of the 5G frequency auction in Italy [15, 16]. The costs per MHz for the sub
GHz and the sub 6 GHz frequencies are two orders of magnitude higher than the cost per MHz for
mm-Wave frequencies.
Type Frequency Offered Band Auction Base Assigned Band Final Cost Cost per MHz
sub GHz 700 [MHz] 75 [MHz] 2110 [Me] 60 [MHz] 2039 [Me] 34 [Me]
sub 6 GHz 3700 [MHz] 200 [MHz] 396 [Me] 200 [MHz] 4346 [Me] 22 [Me]
mm-Wave 26 [GHz] 1000 [MHz] 162 [Me] 1000 [MHz] 164 [Me] 0.16 [Me]
22
14
9 9
2 0,48 0,20 0,17 0,20 0,12
5
10
15
20
25
Italy
Germany
South Korea
UK
Spain
Austria
Finland
Ireland
Czech Republic
Latvia
Cost per MHz [MEUR / MHz]
Fig. 2 Cost per MHz in the 3400-3800 [Mhz] spectrum in different countries in the world. The
Italian cost per MHz is the highest one.
ally, the 700 [MHz] was used by another technology - the TV broadcasting - which
will dismiss such frequency by 2021.
In fact, mm-Wave frequencies will only accommodate communication for niche
services in 5G, and most traffic (hence EMF emissions) will occurr on bands that
are largely used by previous generations of mobile networks. Let us analyze in more
detail the outcome of the auction reported in Tab 4. The table reports for each fre-
quency: (i) the total bandwidth offered in the auction, (ii) the total auction base, (iii)
the total bandwidth that have been assigned to the operators, (iv) the total costs for
the operators at the end of the auction, (v) the ratio between (iv) and (iii) (called
“cost per MHz”). Several considerations are in order. First, the auction base was
clearly higher for the 700 [MHz] frequency compared to the other ones, despite the
fact that this frequency includes the lowest portion of bandwidth. Second, the most
expensive frequency at the end of the auction resulted in the 3700 [MHz], whose
cost passed from 396 [Me] when the auction was opened to 4346 [Me] when the
auction was closed. Third, not all the bandwidth available at the 700 [MHz] was
bought by the operators. However, the final cost incurred in this frequency was in
the same order of magnitude compared to the 3700 [MHz] frequency. Fourth, the
operators bought all the bandwidth offered at 26 [GHz] (i.e., the mm-Wave). How-
ever, the final cost (164 [Me]) was very close to the auction base (162 [Me]).
Fifth, the final costs per MHz resulted in 34 [Me] and 22 [Me] for 700 [MHz] and
5G Technology: Which Risks From the Health Perspective? 13
3700 [MHz], respectively, and only 0.16 [Me] for 26 [GHz]. By comparing the cost
per MHz in the 3400-3800 [MHz] spectrum in different countries in the world (re-
ported in Fig. 2), we can note that Italian operators have paid a cost per MHz much
more higher compared to the one sustained in other countries (e.g., Germany, South
Korea, UK, Spain).
It is thus clear that the largest investment of the operators for 5G was not on
the mm-Waves, but on the 700 [MHz] and 3700 [MHz] frequencies. In particular,
the 700 [MHz] frequency will allow to cover large portions of territory, while the
3700 [MHz] will bring a mixture of coverage and capacity. These two options appear
to be the most promising ones for the operators compared to the 26 [GHz] frequency,
which is instead tailored to the maximization of the capacity of users.
However, a natural question is then: Why are the operators not performing the
same investment in the mm-Wave compared to the other frequencies? To answer
this question, we need to remind that the government has imposed to the operators
tight coverage constraints for the 700 [MHz] and 3700 [MHz] frequencies, while no
coverage constraint is imposed on the 26 [GHz] frequency. In addition, mm-Waves
are subject to very strong attenuation effects compared to lower frequencies. For
example, mm-Waves are largely attenuated when passing through obstacles (e.g.,
walls, buildings), resulting in poor signal levels (and consequently low capacity) in
indoor environment. Therefore, we expect that the 5G technology will be realized
mainly through the exploitation of the 700 [MHz] and 3700 [MHz] frequencies. In
this scenario, the 26 [GHz] frequency will be exploited to cover selected portions of
territory (not the whole one), where it will be possible to deploy, e.g., small cells to
limit the attenuation effects and bring high capacity to users.
Last but not least, we also remind that the potential impact of mm-Waves on
health has been extensively studied in the past (see, e.g., the survey by Zhadobov et
al. [17]). Actually, the body scanners used in many airports in the world already use
mm-Waves. Different studies were conducted to evaluate the impact of these waves
on the health, as reported by the ICNIRP note [18]. Compared to lower frequencies,
high levels of EMFs from the mm-Waves involve only heating of the skin and not
of the inner tissues, due to the weak penetration properties of such frequencies [19].
However, the heating effects from mm-Waves are already excluded if the EMFs are
below the maximum limits, which include all the frequencies up to 300 [GHz], and
hence also the mm-Waves ones.
We can thus safely classify as fake news the fairly popular story of a storm of
birds dying by flying in proximity of a mm-Wave 5G base station. By considering
the power levels radiated by 5G base stations, which are in the range 100 [W]-
200 [W] of maximum power [20], and current EMF limits, which are already en-
forced for 5G base stations, we can easily conclude that the power levels at which
the 5G radio base stations operate are several orders of magnitude below the ones
capable of producing any heat effect on a bird flying in proximity of the base station.
14 Luca Chiaraviglio, Marco Fiore and Edouard Rossi
5 Allegation IV: There is a lack of experimental studies
regarding the emissions of 5G radio base stations
The fourth dispute states that there is a threat on the population’s health since no
experimental studies showing the emissions of 5G radio base stations are currently
available. In this scenario, the citizens are seen as lab-rats exposed to untested EMF
field levels, especially in the towns currently hosting the 5G experimental trials.
In order to discuss such claims, we need to explain how power is radiated from a
radio base station over the territory. The pre-5G technologies typically employ an-
tennas which were sectorized to cover portions of territory falling inside a cone of
radiated power, which starts from the base station site. The amplitude of the cone
(for both its horizontal and the vertical components) depends on the technology
as well as on the features of the antenna that is deployed. The 4G technology (in
its latest revisions) integrates the possibility of placing more antennas working in
coordination to serve the same portion of territory. Such technique, called Multiple-
Input Multiple-Output (MIMO) (and deeply analyzed in subsequent Chapters of this
book) allows an increase in the capacity provided to the users. Focusing now on the
5G, the antennas adopted by this technology will employ the MIMO communica-
tion, coupled with the possibility to concentrate beams of power over the territory,
in order to increase the offered capacity in the zones where the users are actually
located. Compared to the previous generations, therefore, the radiated power over
the territory will be less uniform, as it will be varied over space and time.
However, it is expected that such smart antennas will be realized mainly for ra-
dio base stations running on the 26 [GHz] and the 3700 [MHz] frequencies (i.e.,
those guaranteeing medium and large capacity). Given that such radio base stations
are already being deployed on the territory (especially the ones on 3700 [MHz],
see, e.g., [21] for the network deployed in the city of Bologna), the Italian regional
environmental protection agency (ARPA) is currently authorizing the installation
of such radio base stations by simulating their EMFs under very conservative con-
ditions (e.g., maximum radiation pattern over all the directions) [20]. This choice
provides safety to the population, since the actual EMF levels are in general lower
than the ones assumed during the authorization phase. However, we also advocate
the need to include in the Italian regulations exact procedures describing how to
perform EMF value calculations to meet the limits for 5G radio base stations. Fi-
nally, as soon as the 5G network will become operative in Italy, we expect that
measurements of EMFs generated by 5G base station equipment will be made pub-
licly available by ARPA, as normally happens nowadays for other generations of
mobile networks [22].
5G Technology: Which Risks From the Health Perspective? 15
Table 5 Range and measured frequencies by two probes used to perform wideband measurements.
These probes can be used to measure the EMFs generated by 5G radio base stations.
Type Frequency Range EMF Range Sensitivity Measurable 5G Frequencies
A [23] 100 [kHz] - 8 [GHz] 0.2 [V/m] - 130 [V/m] 0.2 [V/m] 700 [MHz], 3700 [MHz]
B [24] 20 [MHz] - 40 [GHz] 1 [V/m] - 1000 [V/m] 1 [V/m] 700 [MHz], 3700 [MHz], 26 [GHz]
6 Allegation V: It is impossible to measure the EMF levels of 5G
radio base stations
The last dispute is that, since this technology will be run on new frequencies and will
involve novel technology features (e.g., MIMO and beamforming), it is not possible
to realistically measure the EMF generated by a 5G radio base station.
We start by recalling some basics on the measurement of the EMF generated
by a radio base station. First of all, EMF measurements performed with the many
applications available for smartphones are to be considered unreliable, since end ter-
minals have a very limited capability to assess physical layer properties in general
and EMF levels in particular. Proper EMF measurements from radio base stations
must be told apart into (i) measurements performed over a range of frequencies,
called wideband measurements, and (ii) measurements performed on a specific fre-
quency (excluding the other ones), which are typically referred to as narrowband
(or selective) measurements. The choice of measurement approach (narrowband or
wideband) severely impacts the type of equipment that has to be employed. For ex-
ample, narrowband measurements typically require spectrum analyzers, while wide-
band measurements generally adopt less complex devices, i.e., an EMF meter unit
and an EMF probe. However, in both cases, the measurements need to be performed
by qualified personnel, as these devices are not conceived to be used by the general
public.
In the remaining of this section, we will assume to perform wideband measure-
ments of 5G radio base stations (although narrowband measurements are being per-
formed by ARPA, see, e.g., [20]). In this scenario, Table 5 reports the type of probes
currently sold by a producer of EMF meters and EMF probes. The table reports also
for each type of probe the measurable frequency range, the minimum EMF sensitiv-
ity, the measurable EMF range, and the measurable 5G frequencies. Interestingly,
the probes that are actually on sale are already compatible with the frequencies of
5G radio base stations. For example, the first probe is able to measure all the EMFs
generated over the range 100 [kHz] - 8 [GHz] (thus including both the 700 [Mhz]
and the 3700 [Mhz] 5G frequencies), while the second probe is able to extend the
range up to 40 [GHz], which is well above the maximum 5G frequency of 26 [GHz]
in use in Italy. Clearly, we expect that the minimum sensitivity will be improved in
the near future (e.g., the second probe has a minimum sensitivity of 1 [V/m]).
Focusing then on the impact of MIMO and beamforming on the measurements
of EMFs, these features introduce a variation of the received EMF over space and
over time components. The fact that the EMFs vary over time is not new, as cur-
rent 4G networks already perform a time-variant scaling of the radiated power, in
16 Luca Chiaraviglio, Marco Fiore and Edouard Rossi
accordance to the traffic requested by users. The element of novelty is that in 5G the
positioning of the radiating beams will be not known a-priori, as these beams may
be used, e.g. to follow a user or a set users moving over the territory. Such features
introduce an additional level of complexity in measuring the EMF radiated by a 5G
base station, as the EMF measured in one point is impacted by the presence or the
absence of beams focused on the measurement point during the measurement inter-
val. For this reason, we expect that the regulators will introduce in the near future
specific guidelines on how to measure the EMFs generated by the 5G radio base sta-
tions. For example, one way to face the unpredictability of EMF variations would
be to employ a large set of EMF meters, by performing the measurements in paral-
lel over the coverage area of the 5G base station under consideration. Clearly, such
measurements should be repeated over time, in order to track the time-variability
of the beams. Eventually, a second solution (currently requested by ARPA to differ-
ent operators [20]) is to force the operator to generate beams that are fixed and are
oriented towards the measurement locations.
7 Conclusions and Future Works
In this chapter, we discussed five main allegations regarding the potential health ef-
fects due to the EMFs generated by 5G radio base stations. By reviewing the relevant
scientific documentations, we could not find evidences of carcinogenicity associated
with an EMF exposure below the limits set by the applicable Italian laws. In addi-
tion, we clarified that an increase in the number of 5G radio base stations would
only allow to steadily decrease the power (and consequently the EMFs) generated
by each base station, while preserving the received signal strength (hence quality
of communication) over the territory. In all cases, given the current regulations on
EMF limits and the presence of pre-5G sites, we do not expect that 5G will bring to
a huge proliferation of new radio base stations.
We have also shed light on the potential danger of mm-Waves, by: (i) providing
evidence that 5G will be mainly based on frequencies that do not belong to the mm-
Waves class; and, (ii) reporting the absence of scientific works that demonstrate
health effects associated with an EMF exposure from mm-Waves below the limits.
Also, the presumed lack of experimental studies of EMF emissions from 5G radio
base stations does not seem factual, since it is possible to take into account the new
features brought by 5G during the authorization and measurement steps.
While our work allows to conclude that most of the disputes against 5G radio
base stations are not supported by scientific evidence, we would like to conclude our
discussion by stressing the importance of continuing medical and clinical research.
This is needed to evaluate any potential health impact of low EMFs generated by all
devices, including the ones in close proximity to users, e.g., 5G smartphones, tablets
and laptops.
5G Technology: Which Risks From the Health Perspective? 17
References
1. Electricity Sparks Fears in 1900? Available at http://www.seacoastnh.com/
electricity-sparks- fears-in- 1900/, last accessed on 24th October 2019.
2. Celine Fremault Tweet: “I will not allow the people of Brussels to be transformed into guinea
pigs: no raising of standards until I have the guarantee that these antennas comply with
them.”. Available at https://twitter.com/i/web/status/1112707642051514370, last
accessed on 29th October 2019.
3. R. Baan, Y. Grosse, B. Lauby-Secretan, F. El Ghissassi, V. Bouvard, L. Benbrahim-Tallaa,
N. Guha, F. Islami, L. Galichet, and K. Straif, “Carcinogenicity of radiofrequency electro-
magnetic fields,” The lancet oncology, vol. 12, no. 7, pp. 624–626, 2011.
4. W. Joseph, P. Frei, M. Roösli, G. Thuróczy, P. Gajsek, T. Trcek, J. Bolte, G. Vermeeren,
E. Mohler, P. Juhasz, et al., “Comparison of personal radio frequency electromagnetic field
exposure in different urban areas across europe,” Environmental research, vol. 110, no. 7,
pp. 658–663, 2010.
5. G. Dürrenberger, J. Fröhlich, M. Röösli, and M.-O. Mattsson, “EMF monitoring-concepts, ac-
tivities, gaps and options,” International journal of environmental research and public health,
vol. 11, no. 9, pp. 9460–9479, 2014.
6. S. Sagar, S. Dongus, A. Schoeni, K. Roser, M. Eeftens, B. Struchen, M. Foerster, N. Meier,
S. Adem, and M. Röösli, “Radiofrequency electromagnetic field exposure in everyday mi-
croenvironments in Europe: A systematic literature review,” Journal of Exposure Science and
Environmental Epidemiology, vol. 28, no. 2, p. 147, 2018.
7. National Toxicology Program et al., “Toxicology and carcinogenesis studies in Hsd: Sprague-
Dawley SD rats exposed to whole-body radiofrequency radiation at a frequency (900 MHz)
and modulations (GSM and CDMA) used by cell phones,” NTP Tech. Rep., vol. 595, pp. 1–
384, 2018.
8. National Toxicology Program et al., “NTP technical report on the toxicology and carcino-
genesis studies in B6C3F1/N mice exposed to whole-body radio frequency radiation at a fre-
quency (1,900 MHz) and modulations (GSM and CDMA) used by cell phones,” NTP Tech.
Rep., vol. 596, pp. 1–260, 2018.
9. L. Falcioni, L. Bua, E. Tibaldi, M. Lauriola, L. De Angelis, F. Gnudi, D. Mandrioli,
M. Manservigi, F. Manservisi, I. Manzoli, I. Menghetti, S. Montella, R. Panzacchi, D. Sgargi,
V. Strollo, A. Vornoli, and F. Belpoggi, “Report of final results regarding brain and heart tu-
mors in Sprague-Dawley rats exposed from prenatal life until natural death to mobile phone
radiofrequency field representative of a 1.8 GHz GSM base station environmental emission,”
Environmental research, vol. 165, pp. 496–503, 2018.
10. DECRETO DEL PRESIDENTE DEL CONSIGLIO DEI MINISTRI 8 luglio 2003 (In Italian).
Available at https://www.gazzettaufficiale.it/eli/id/2003/08/28/03A09711/
sg, last accessed on 4th November 2019.
11. DECRETO-LEGGE 18 ottobre 2012, n. 179 (In Italian). Available at https:
//www.normattiva.it/uri-res/N2Ls?urn:nir:stato:decreto.legge:
2012-10- 18;179!vig=, last accessed on 4th November 2019.
12. A. Ahlbom, U. Bergqvist, J. Bernhardt, J. Cesarini, M. Grandolfo, M. Hietanen, A. Mckinlay,
M. Repacholi, D. Sliney, J. A. Stolwijk, M. L. Swicord, L. D. Szabo, M. Taki, T. S. Tenforde,
H. P. Jammet, and R. Matthes, “Guidelines for limiting exposure to time-varying electric,
magnetic, and electromagnetic fields (up to 300 GHz),” Health physics, vol. 74, no. 4, pp. 494–
521, 1998.
13. ICNIRP and others, “ICNIRP note: Critical evaluation of two radiofrequency electromagnetic
field animal carcinogenicity studies published in 2018,” Health physics, vol. on press, 2019.
14. K. Haneda, J. Zhang, L. Tan, G. Liu, Y. Zheng, H. Asplund, J. Li, Y. Wang, D. Steer, C. Li,
et al., “5G 3GPP-like channel models for outdoor urban microcellular and macrocellular en-
vironments,” in 2016 IEEE 83rd Vehicular Technology Conference (VTC Spring), pp. 1–7,
IEEE, 2016.
18 Luca Chiaraviglio, Marco Fiore and Edouard Rossi
15. Bando di gara per la procedura di assegnazione di diritti d’uso delle frequenze nelle Bande
694-790 MHZ, 3600-3800 MHZ e 26.5-27.5 GHZ (In Italian). Available at http://www.
gazzettaufficiale.it/eli/gu/2018/07/11/80/s5/pdfg, last accessed on 29th Oct.
2019.
16. Assignment of 5G Frequency in Italy: Call and Outcome. Available at https:
//www.mise.gov.it/index.php/it/comunicazioni/servizi-alle- imprese/
tecnologia-5g/bando- 5g, last accessed on 29th Oct. 2019.
17. M. Zhadobov, N. Chahat, R. Sauleau, C. Le Quement, and Y. Le Drean, “Millimeter-wave
interactions with the human body: state of knowledge and recent advances,” International
Journal of Microwave and Wireless Technologies, vol. 3, no. 2, pp. 237–247, 2011.
18. ICNIRP and others, “ICNIRP statement - health issues associated with millimeter wave whole
body imaging technology,” Health physics, vol. 102, no. 1, pp. 81–82, 2012.
19. D. Nelson, M. Nelson, T. Walters, and P. Mason, “Skin heating effects of millimeter-wave
irradiation-thermal modeling results,” IEEE Transactions on Microwave Theory and Tech-
niques, vol. 48, no. 11, pp. 2111–2120, 2000.
20. “5g is coming: are we ready? (in italian),” Ecoscienza - Arpae Emilia Romagna, vol. 4, pp. 24–
66, 2019.
21. 5G Deployment Inside the Bologna City (In Italian). Available at https://www.arpae.it/
dettaglio_generale.asp?id=4149&idlivello=2145, last accessed on 2nd Nov. 2019.
22. EMF measurement in Emilia Romagna region. Available at https://www.arpae.it/
dettaglio_generale.asp?id=2618&idlivello=1534, last accessed on 2nd Nov. 2019.
23. WaveControl WPF8 Field Probe Datasheet. Available at https://www.wavecontrol.com/
rfsafety/images/data-sheets/en/WPF8_Datasheet_EN.pdf, last accessed on 29th
Oct. 2019.
24. WaveControl WPF40 Field Probe Datasheet. Available at https://www.wavecontrol.
com/rfsafety/images/data-sheets/en/WPF40_Datasheet_EN.pdf, last accessed on
29th Oct. 2019.