A short research report  by Francis MASSEN

francis.massen@education.lu

 version 1.01  02 Nov 2009

Abstract:

Despite a recent scary article in the Haaretz newspaper, CFL lamps do not present any UV related dangers under normal conditions of usage. The biologically effective UV irradiance, as well the UVA irradiance diminish both in an exponential manner, approaching zero at a distance of 30 cm. Close contact should be avoided.

1.  The Israel and British reports on UV dangers

There are some strong sentences like this one: "... the lamps should be affixed to ceilings at a distance of more than 30 centimeters from the room's occupants, and should not be used as permanent lighting on desks or walls close to a person's body, and in rooms where the residents spend many hours of the day. "

The article also mentions a position paper issued by the British Health Protection Agency (HPA) a year ago. The paper stated that the bulbs may emit UV rays that under certain conditions may expose people to higher amounts of radiation than the recommended level. Look here  [2] for a discussion of these findings. In short, there is a warning that at a close distance of 3 cm from a single-walled CFL (the type usually found), UVB irradiation levels may be too high.

2. The BMB UV experiment

As the manager of the meteorological station of the LCD (http://meteo.lcd.lu) I am familiar with UV radiation and its measurement. So I made a very quick investigation on the UV problem, using a single walled 20W/1300 lumen IKEA lamp:

fig.1. The 20W/1300 lumen IKEA CFL used in the test.

To measure both UVB and UVA irradiances, I used a semi-professional instrument from Solar Light Company, the Ultraviolet Meter Model 3D:

fig.2. The Solar Light Co. Model 3D UV Meter

The sensor on the left measures the biologically effective UVB irradiance ( 280 - 320 nm); the unit of measurement is the MED/h (Minimal Erythema Dose per hour).

1 MED    = 210 Jm^-2 and 
1 MED/h = 58.3 mW effective UVB irradiance.

A relatively fair skinned person (phototype II) without prior exposition to the sun will get a sunburn (an erythema) if exposed for 1 hour to an irradiance of 1 MED/h. As an example: here in Diekirch, Luxembourg the maximum UVB intensity was 3.31 MED/h during the summer 2008.  This is equivalent to 9.2 UVI; you might look here [3] for an explanation of MED, J/m^-2 and UVI. Watch out: there is a disagreement on the definition of MED. The EU usually assumes 1 MED = 250 Jm^-2, in the USA 1 MED = 210 Jm^-2. In this report we will stick to the USA definition.

The sensor on the right measures the UVA radiation (320 - 400 nm); the readings must be multiplied by 10 to correspond to Wm^-2.  During summer 2008 the maximum UVA irradiance was 47.2 Wm^-2  in Diekirch.

A research done in 1998 (see report here [4] found that the Model 3D readings should be multiplied by 0.88 to correspond to those of a WMO grade UVB sensor (Biometer from Solar Light Co.).  In this report this calibration factor is assumed being still valid, and the raw readings of the Model 3D instrument will be adjusted by this multiplier.

Starting with touch-contact (distance = 0) the distance between the bottom of the lamp and the top of the sensor was increased up to 30 cm. The next pictures show the situation at d = 0 cm and d = 3 cm:

fig. 3. The raw readings of UVB irradiance at a distance 0 and 3 cm. Corrected results would be 2.29 and 0.59 MED/h.

Here are the results, MED210 = biologically effective UVB reading in MED/h units, MED210ok = 0.88*MED210 ; UVA readings are in Wm^-2 and distances in cm.

As can be seen, at a distance of 30cm both UVB and UVA irradiances are practically zero. At  touch-contact, the  UVB irradiance is high: touching the lamp for 1/2.29*60 = 26 minutes would possibly yield an erythema and give a not negligible UVA irradiance (as well as a serious burn!). Actually this is a rather unusal situation.
If the CFL is fitted to a desktop lamp,  a close-up distance of 20 cm is not impossible. For a 8 hour working day the effective UVB dose would be 0.4 MED, i.e. the dose one would have received in the above-mentioned summer in 7 minutes at midday!

A simple exponential model gives an exceptional good fit for both types of UV radiation: R² is practically 1, what means that the model is able to perfectly explain all variability.

fig.4.  Effective UVB irrradiance versus distance from lamp

The plot shows that a distance of 20cm is sufficient to avoid any noticeable UVB exposure. Close contact should be avoided; this will be the rule during intelligent usage, as the tubes become burning hot after less than 5 minutes!

fig.5.  UVA irrradiance versus distance from lamp

UVA irradiance diminishes exponentially in the same manner as does UVB. Here a distance of 20 cm is also sufficient to avoid any noticeable exposure.

A second test was done using an relatively inexpensive personal UV sensor manufactured by the SAFESUN Pty from Australia (now out of production). The reading of the instrument shows the UVI. A calibration test done in 1998 (here [5]) found that the raw readings should be multiplied by 0.74 to give the correct UVI. In this case close contact would correspond to a UVI of 4.4.
Expressed as MED/h (with 1 MED = 210 Jm^-2) this is 1.9 MED/h, reasonably close to the reading of 2.3 of the considerably more professional Model 3D instrument.

fig.5. Close contact UVI = 6 reading by the personal SAFESUN sensor

A quick test with a long-tube type of fluorescent lamp showed a much smaller UVB emission compared to the compact types: at close contact the irradiance is less than 0.2 MED for a 30W tube.

3. Conclusion

This short research confirms the warnings of the HPA and Israeli committee. Single walled CFL's do emit UV radiation, but the intensity rapidly falls of with distance. A longer close contact could be potentially risky; as this contact is very unpleasant due to the high temperature of the tube, this is a very unusual situation. So the potential danger should be seen in perspective: if the CFL's are used in a normal manner at a distance greater than 30cm, the risk from UV irradiance is close to zero.

REFERENCES

[1]        Haaretz.com: Ministers to warn: Exposure to fluorescents may cause skin cancer.  29 October 2009. http://www.haaretz.com/hasen/spages/1124163.html

[2]        Health Protection Agency:: Emissions from compact fluorescent lights. October 2008.   

            http://www.hpa.nhs.uk/webw/HPAweb&HPAwebStandard/HPAweb_C/1223534061375?p=1204186170287

[3]        Massen F., Harpes N.: The UVI poster. 1997. http://meteo.lcd.lu/papers/uv/uvi/uvi_03.html

[4]        Massen F., Harpes N., Breuskin P.: A comparison of three portable UV sensors. 1998. http://meteo.lcd.lu/papers/calib/uvcomp1/uvcomp1.html

[5]        Massen F., Harpes N.: A preliminary analysis of the Safesun personal UV sensor. 1998. http://meteo.lcd.lu/papers/calib/uvbcal03/uvbcal_03.html

History:

last modified: 02 Nov 2009

go back