Monopolar and bipolar — what is different

Physically, at the small electrode of a monopolar device the current spreads out spherically, so the heating falls off sharply in inverse proportion to the fourth power of the distance. According to the literature that actually sets out this calculation, the penetration of an electrosurgical electrode of 0.5 mm radius is no more than about 140 µm. Conversely, with a broad plate-shaped electrode the current spreads far more widely and goes deeper.

Let us restate the key point here. In both cases what sets the depth is the geometry of the electrode. Not “deep because it is monopolar” but “deep because that device's electrode is large and broad” is the correct way to put it. The clinical impression that “monopolar = deep” arose because devices intended to heat deeply have conventionally adopted a monopolar arrangement, not because of any property of the polarity itself.

There is data that calculated heating depth by finite element analysis and verified it in tissue experiments.

The bipolar tip came out deeper than the monopolar tips — against the spherical monopolar electrode (4.37 mm) 4.6–5.7 times, and even against the broad flat square electrode (10.58 mm) 1.9–2.4 times. Because it is a large-area applicator with widely spaced electrodes. In the same data, the time to reach 42°C was about 30 seconds for monopolar and about 100 seconds for bipolar.

What you should know when reading this data — all five authors belong to the manufacturer of the device in question, and it is a paper evaluating their own product. And the paper does not state which temperature isotherm was used to define “penetration depth”, so the figures cannot be compared directly with other studies. The reason we cite this data is not that we believe the absolute values, but because it is a counter-example showing that “bipolar is shallow” is not a necessity.

This rule is cited everywhere. We traced the citation lineage back.

Our assessment — it is closer to custom. This rule is an analytical approximation that assumes an idealised geometry of infinitely wide parallel plate electrodes. Real clinical applicators are not parallel plates but curved · multi-row · multi-electrode arrangements, so the premise does not hold. We did not find a study that varied the electrode spacing in human skin, measured the heating depth and confirmed this proportional relationship. And the calculated results in the preceding section do not agree with this rule — the bipolar tip came out at 20–25 mm, and that tip's electrode spacing is not that wide.

There is a similar rule of thumb for monopolar devices — “the penetration depth is about one third of the electrode radius”. A 2024 paper cites this rule, but we did not reach the original measurement behind it.

The aim of radiofrequency is to heat dermal collagen so that it denatures · contracts, and to draw out the regenerative response that follows. But at the same temperature the result differs completely depending on how long the exposure lasts. There is data that measured directly, under the microscope, the disappearance of the crimp structure of collagen.

The last two cells of this table are the key. At 50°C, heating for two hours barely changes the collagen at all. Raise the temperature by 5–10 degrees and the time required falls several-fold each time. The authors wrote that this behaviour is consistent with the Arrhenius law. That is, asking only “how many degrees does it reach” is half the question.

The temperatures actually aimed at in clinical practice

This gap between “dermis 65–75°C” and “tissue 40–44°C” is explained in two ways. First, the surface temperature and the deep dermal temperature are different. Second, a good many modern non-invasive devices have moved towards stimulating without damage rather than contracting collagen immediately.

What has actually been observed in human skin

In a 2004 study, changes in collagen fibrils were confirmed by electron microscopy in bovine tendon down to a depth of 6 mm. And yet in human skin no immediate significant change was visible by light microscopy — electron microscopy showed only scattered fibril hypertrophy in the mid dermis. At the level of gene expression, however, it was clear — type I collagen mRNA increased 2.4-fold at day 2 and 1.7-fold at week 1.

This study was supported by the device manufacturer with direct costs and equipment. And the rule “every 5-degree fall in temperature requires ten times the time for the same contraction” is widely cited, but we did not reach the original experimental paper for it, and it does not match the measured table above quantitatively either — in the measurements the ratio across a 5-degree band varies greatly from band to band, from 1.2-fold to 5.6-fold (skin 70→65°C 5.6-fold, 65→60°C 1.2-fold; tendon 70→65°C 4.2-fold, 65→60°C 1.7-fold). In this piece we have used only the numbers from the measured table as evidence.

The point of highest current density is directly beneath the electrode, that is, the epidermis. So every surface-type radiofrequency device has some means of protecting the epidermis.

Which cooling protects the epidermis better has no answer yet. We confirmed that a paper directly comparing the two methods came out in 2025, but we did not reach its text and result figures. And a 2026 paper on circulating water cooling expressly states as a limitation that “skin surface temperature was not objectively monitored during the procedure” — that is, there is as yet no data measuring the difference in epidermal temperature between the two methods.

For monopolar devices active cooling is effectively essential — because directly beneath the electrode is the most dangerous point. Bipolar devices need less energy to reach the same area, so the cooling burden is relatively smaller. Several reviews repeat the statement that bipolar is less painful, but we did not find a study directly comparing pain scores under the same conditions.

Effectively there is not. We found only one, and even that does not answer the question this piece is about.

There are three reasons why no conclusion can be drawn from this study. (1) What it assessed are epidermal hydration · barrier indices, not the “how far into the dermis” that this piece asks about. (2) It is beauty-device output of the maximum-40°C kind, physically a different domain from high-output medical devices. (3) Being unblinded · a single session · not split-face, there is almost no control of confounding.

In the fields of body contouring and cellulite too, we did not find a randomised study comparing the two modes directly. The comparative trials that exist are either comparisons between different energies such as radiofrequency and focused ultrasound, or comparisons of tip size within the same polarity. Recently, if anything, case reports have been appearing in the direction of combining monopolar and bipolar in sequence.

It differs fundamentally in principle from surface-type radiofrequency.

The relationship between energy and coagulation volume has been measured too — in porcine skin the coagulation volume was 0.033 ± 0.012 mm³ at 20 mJ per needle and 0.353 ± 0.173 mm³ at 100 mJ, with a correlation coefficient of 0.976. In 30 humans as well, the correlation between total energy and volumetric change was confirmed at 0.676 (p < 0.001).

Multipolar is not a separate family but an extension of bipolar. Several pairs of electrodes are placed and switched between, and the purpose is not “deeper” but “more evenly” — the aim is to homogenise the problem of heating concentrating in one place in conventional radiofrequency. The depth rule remains dependent on the electrode spacing.

Each device advertises a different frequency. The explanation that “the lower the frequency, the deeper it penetrates” is repeated in the literature too — and what is offered as the basis is the relationship that penetration depth is inversely proportional to the square root of the frequency.

What that relationship refers to is the skin depth of electromagnetism. For “frequency sets the depth” to hold, this value has to be on the scale of millimetres. So we did the calculation ourselves — we put in the coefficients of the standard model describing the dielectric properties of tissue (Gabriel et al., 1996) and obtained the conductivity and skin depth at each frequency.

In this band the skin depth is 39 cm to 3.4 m. More than 200 times the thickness of the dermis (about 2 mm). That is, “inversely proportional to the square root of the frequency” is true as an equation but decides nothing at the scale of skin. In this frequency range the electromagnetic field is effectively quasi-static, and the current distribution inside tissue is determined entirely by the geometry of the electrodes and the impedance contrast between tissues. An explanation of the form “2 MHz goes a few millimetres deeper than 6.78 MHz” has no physical basis.

So does frequency change nothing — no, it does not

Pull the conductivity contrast between tissues out of the same calculation and a different picture appears.

The higher the frequency, the wider the conductivity gap between dermis and fat. In a structure where the current can divide along several paths — for example the fibrous septa running through the fat layer — the larger this gap, the more strongly the current concentrates into the more conductive channel.

And this calculation matches the actual animal experiment exactly. In a 2025 study in which monopolar radiofrequency was delivered to minipigs at different frequencies, 2 MHz produced a broad, deep thermal response within the fat layer and 6.78 MHz produced heating localised along the fibrous septa. Using the two frequencies in sequence gave the largest temperature rise. On biopsy at 30 days, the dermis and fibrous septa were thickened in all three conditions, and the morphology of the adipocytes was preserved.

The accurate statement is this — frequency changes not “how deep” but “where inside the tissue the current concentrates”. The depth is set by the electrode.

We note one piece of conflicting data. In the finite element analysis above, the same bipolar tip came out at 25.00 mm at 0.5 MHz and 20.15 mm at 1 MHz — 24% deeper at the lower frequency. This difference is not explained by the skin depth calculation above. It may be the proportion of the displacement current component as the permittivity changes, or the effect of electrode-skin impedance matching, but the paper did not explain the reason, all the authors belong to the manufacturer, and the definition of penetration depth is not stated either. With no independent replication, we have not taken this as evidence.

We hear the worry that “radiofrequency makes your face lose its fullness”. It has in fact been reported. But you should know the context precisely.

The title of this paper is itself “rare, preventable, and correctable”. The point is not that “radiofrequency melts fat” but that “it arises with excessive delivery, and operator training actually reduced it”. The fact that the incidence fell to below a third after training was introduced is the evidence for that.

The 2025 FDA safety communication

On 15 October 2025 the FDA issued a safety communication about RF microneedling. The risks listed in the original text are “burns, scarring, fat loss, disfigurement, and nerve damage, and the need for surgical repair or medical intervention to treat injuries”. The FDA's recommendation is clear — “RF microneedling is a medical procedure, not a cosmetic one, and should not be used at home”, and it should be received from a licensed healthcare provider with training and experience.

The FDA did not state a number of reports in the communication. There is a paper analysing the adverse event database that underpins this communication, but we did not reach its original text and so have not put its figures in this piece. For reference, you should also know that such spontaneous reporting data, because the total number of procedures is unknown, gives not an incidence but a proportion of reports.

Adverse events with surface-type radiofrequency

Whether the adverse events of monopolar and bipolar differ cannot be answered. There is no direct comparison data. That reports of burns · fat atrophy appear relatively more numerous on the monopolar side may well be because the cumulative number of procedures and the length of market exposure for high-output monopolar devices are far greater — a possibility that cannot be excluded. The denominators differ.

The contraindications common to all are implanted medical devices such as pacemakers · defibrillators, collagen vascular disease and autoimmune disease. A history of radiotherapy and recurrent herpes call for caution. Aftercare is set out in aftercare following RF treatment.

The conclusion of this piece is not “which one is better”. No basis exists at present for saying that. What we can say instead is this — if in a consultation you hear depth explained on the basis of “how many MHz” or “because it is monopolar”, that is an explanation with no physical support. What actually sets the depth is the size and arrangement of the electrodes, and how that device is used. The differences between the three devices we hold are set out in Density · XERF · Oligio — all radiofrequency, so what differs, and which suits whom in lifting treatment should change with skin thickness and fat volume.

Written and reviewed by Lee Chi-Hak, MD, medical director of Miso Clinic in Daegu, South Korea. Every study cited above is given together with its design, its size and the limitations the authors themselves recorded, and where we could not find data, we have said that we could not find any.