The purpose of this article is to give a brief overview of safety considerations due to multiple wavelengths of laser light being used with optical fibers. This subject is becoming more relevant as time goes on, since the need to increase throughput through optical fiber communications systems (OFCS) has increased. To increase the throughput through an optical fiber, manufacturers of this equipment can do one of two things:
a. Increase the data rate; or
b. Add multiple wavelengths onto the same fiber.
Increasing the data rate is worthwhile up to a point, but will ultimately be limited by dispersion within the fiber. Dispersion causes the input pulses to spread out, which may cause bit errors at the receiver end. To compensate for this, repeater elements would have to be added. Faster data rates per channel also places demands on the speed of the electronics. Therefore, to further the throughput through OFCSs, multiple laser wavelengths are launched onto a single fiber.
Due to this phenomenon of multiplexing wavelengths onto a single fiber, it becomes confusing to determine what the resultant hazard will be. In terms of skin hazards (that is, laser light being directly exposed upon the skin), the total hazard becomes added together. However, the overall hazard to the eye is a more complicated scenario, since it ultimately depends on the individual wavelengths involved. This is because different wavelengths are diffracted differently by the eye lens, and will be more or less of a hazard when the total cumulative effects of all the wavelengths are considered. As it will be seen, the existing laser safety standard, IEC 60825, has a method of dealing with this.
In any dialog concerning laser safety, we must consider the Class of the laser. The Class refers to the laser component operated at full strength irrespective of the protection afforded by the embedding product. The “Class” of the laser is the classification system that IEC 60825-1 employs to describe a laser in terms of its potential harmful effects. Class 1 lasers would be considered as the safest, while Class 4 would constitute as the most potentially harmful. It should be noted that for the equipment with which this article is concerned (Optical Fiber Communication Systems), the classification system looks at the potential hazard at any accessible location within the entire optical fiber communication system of the particular equipment in question.
The other parameter that comes up while discussing laser safety is the accessible emission limit (AEL). The AEL is the maximum limit that a person is allowed to have access to in that particular Class of laser. The values of AEL are usually expressed in W (although they may often be expressed in units of Joules of energy for cases of small duration exposures or smaller wavelengths). The AEL of the different classes appear in several tables with the IEC 60825-1 standard.
The overall hazard level of the multiplexed signal depends on the individual power levels, and whether or not the hazards from the individual hazard levels are additive or not. Skin hazards (for those wavelengths that are used in OFCSs), the hazards are always additive. The so-called “magic number” is 1400 nm where the conditions change as follows:
a. If all of the wavelengths are either more or less than 1400 nm, then the combined hazard is higher (i.e., the resultant hazard is greater)
b. If the wavelengths are on different sides of the 1400 nm mark, then the retinal hazards do not add (i.e., the resultant hazard is not affected).
Before we begin to look at specific examples of how this works, let’s review some basic definitions. We will need to be able to calculate the beam diameter of a divergent beam that is at some distance from the end of an apparent source. Since we are dealing with fiber optics, we could say that we wish to calculate the beam diameter at some distance “r” from the end of a fiber face. This divergence is often specified by the Numerical Aperture of the particular fiber in question. This is demonstrated in Figure 1.
Figure 1
As shown in Figure 1, the numerical aperture is equal to the sine of 1/2 of the divergent angle. It is also related to the difference between the indices of refraction of the cladding and the core of the particular fiber that is used. (The classical definition of numerical aperture is the sine of the half angular aperture of the objective lens; for fiber optics, we will stick with our Equation as NA = sin q/2.) It is defined at the 5% of peak maximum point; that is, if the detector is placed within a circle created by the numerical aperture of the fiber, then the detector only misses 5% of the total power coming out of the fiber.
Figure 2
This means that 95 % of the total power lies within this circle (refer to Figure 2). This diameter in which 95 % of the power resides may be found by the following expression:
Eq. 1 D95 = a + 2 r tan q/2 = a + 2 r tan (arcsin NA),
where “r” is the distance from the apparent source to the point where you are observing the beam. Since “a” is very small (in the order of tenths of a mm), this term can be neglected, and Equation 1 becomes:
Eq. 2 D95 = 2 r tan (arcsin NA).
In the safety calculations, it is more mathematically convenient to consider the case of a Gaussian beam, where the diameter of interest is at 63 % of the total maximum power. Since the ratio between the 95% diameter and the 63% diameter is defined as 1.7, then:
Eq. 3 D63 = D95/1.7 = (2 r /1.7 ) tan (arcsin NA)) = 2r NA/1.7
Now most of the time, optical fibers are not specified in terms of NA, but rather in terms of the Field Mode Diameter, w. The field mode diameter can be described as the power distribution over the plane at the end of the fiber (see Figure 2). As seen in the figure, not all of the power is contained within the core of the single mode fiber; some of it “leaks out” into the cladding. It can be shown that since w = (1.7 l) / (p NA), then Equation 3 can be expressed in terms of w as:
Eq. 4 D63 = (2 r l) / (p w).
Note that the diameter becomes a function of r, l, and w. As the distance from the end of the fiber increases, so does D63 (much like water squirting from the end of a water hose). It also increases as the wavelength l increases; one could imagine that, since light at higher wavelengths have less energy than it would at lower wavelengths, it would have a greater tendency for divergence from the end of a fiber than shorter wavelengths would.
Figure 3
In order to predict how the power profile looks as you go a distance “r” from the end of a fiber, we need to come up with a coupling parameter. This parameter allows us to predict how power is passed through a particular aperture diameter, when we know what the total power is of the source. (Recall that when any wave passes through an aperture, that aperture acts as a source. See Figure 3.) This parameter is:
Eq. 5 h = 1 – exp –(Da/D63)2
where Da is the diameter of the aperture.
The power passing through the aperture becomes modified as follows:
Eq. 6 Pa = h Po
where Po is the total power that is being applied to the aperture.
Now in the case of making classifications to IEC 60825-1, there are specific sized apertures that are used. These are specified in Table 7 of IEC 60825-1 (shown here as Table 1). In our evaluation (which is focusing on lasers used in fiber optic applications which typically do not pose a hazard to the skin), we will consider the case of exposure to the eye.
Table 1: Aperture diameter applicable to measuring laser irradiance and radiance exposure (Table 7 in IEC 60825-1)
As shown, most of the wavelengths for our consideration lies in the band of 400 to 1400 nm, in which Da = 7 mm (which is the diameter of the fully dilated pupil). The other bands of consideration are those wavelengths over 1400 nm. Repetitively pulsed lasers use an aperture of 1 mm when evaluating individual pulses. When the user can be expected to be exposed for 10 seconds or more, then Da 3.5 mm. This aperture becomes some other value for exposures of between 0.35 seconds to10 seconds. It should be noted that, in case of doubt, the most extreme duration should be considered (10 seconds) for determining this aperture size.
Once Da and D63 are known, then the coupling factor m is known. If the total power of the laser is known, then the power passing through that aperture can be calculated. Then, from a laser safety viewpoint, the next question is whether this is an acceptable exposure limit in accordance to the standard. As shown in previous articles, this concept can be quantified by referring to the Accessible Exposure Limits (AELs), that are defined within the tables of IEC 60825-1. This quantity has units of either radiant energy (in Joules), or in radiant power (in Watts). The AEL is defined as the maximum level that a particular class is limited to. This quantity is dependent on the wavelength of the laser (henceforth referred to as l), the duration in time of the exposure, the tissue type which is exposed, and the retinal image size.
Once the proper AEL is determined, it is then possible to calculate what the maximum allowed emitted power of the laser should be before it is passed though an aperture (such as coupled into an optical fiber). The power passing through the aperture can be calculated in terms of the AEL as follows:
Eq. 7 Po = Pa / h = AEL / h.
The result of this gives us the maximum allowed emitted power of the laser source that is to be launched into the fiber.
This then brings us to the main point of the article, which is what happens to the hazard when more than one wavelength is launched onto a single fiber. As brought up in the beginning of this article, the hazard either adds or does not add together (the skin hazard and the eye hazard are evaluated separately). The hazard level for a multi wavelength case becomes equal to the power level of each particular wavelength divided by the AEL of each wavelength, summed over the total of each of the wavelength power terms divided by the AEL terms.
An excellent example of this is given in the IEC 60825-2 standard, example D.4.1.1. In this example, there is a multimode fiber with a 50 mm core fiber diameter, and a numerical aperture of 0.2 +/= 0.02 (note that in this example, we will see if the uncertainty of +/- 0.02 has any bearing on the result). Onto this fiber are launched six different optical signals at the following wavelengths:
-
840 nm
-
860 nm
-
1290 nm
-
1300 nm
-
1310 nm
-
1320 nm
In this problem, each wavelength has the same maximum time averaged power of -8 dbm. The example asks the extent of the hazard at the other end of the fiber.
It is interesting that the problem never takes into account the length of the cable; after all, there will be attenuation along the length of this fiber. The reason this is not a factor is that the fiber may break at any place along its length and, therefore, the worst case condition should apply.
The example’s first step toward a solution actually appears in the problem’s description: convert the power into units of mW. This is because all of the AEL calculations will be done in Watts. The power is calculated as follows:
Eq. 8 P(in mW) = 10dBm/10 = 10 -8 dBm/10 = 0.16 mW
The example provides a good demonstration on how to calculate the AELs and what assumptions to make. The first assumption is that viewing is not intentional. Therefore, as pointed out in sec. 8.4 e of IEC 60825-1, where the wavelength is beyond 400 nm but is not intended for intentional viewing (such as in some medical applications where a laser is purposely fired through the pupil for an extended period of time), assume the time of exposure to be 100 seconds.
Now the example describes how the hazards of the individual wavelength must be added. As pointed out earlier, this is because all of the wavelengths are beneath 1400 nm. We will first look at the emission levels for the lowest (and safest) class, which is class 1. To determine the AEL, refer to Table 1 of IEC 60825-1 (shown here as Table 2).
Table 2: Accessible emissions limits for Class 1 and Class 1M laser products (Table 1 in IEC 60825-1)
The table shows us that, for an exposure time of 100 seconds (which lies within the columns of 10 to 102 and 102 to 103 ), then for wavelengths between 700 up to 1400 nm, there is a choice of expressions. The expression we choose will be determined by the angular subtense, a. The angular subtense is the angle of the light source coming out of the fiber, when viewed at a distance of 100 mm away from the fiber face. This is defined in Equation 9:
Eq. 9 a = 2 arctan (D63/r) = D63/r, in radians.
Since D63 is at most 50 x 10-6 m (the premise of our example says that the core is 50 mm), and since r is given as 100 mm = 0.1 x 10-3 m, then:
Eq. 10 a = (50 x 10-6)M/(1 x 10 -1)M = 50 x 10-5 Rad = 0.5 mRad.
According to the notes to Tables 1 through 4 (see Table 3), the constant T2 is 10 seconds for a < 1.5 mrad. Since our answer of 0.5 is less than 1.5, then T2 is 10 seconds.
Table 3: Notes to Tables 1 - 4 (of the standard)
Since we have previously determined that t = 100 seconds, then t > T2 (refer to Table 2). Therefore, our choice of equations is limited to two; one for a < 1.5 mrad, and one for a > 1.5 mrad. Therefore, the proper choice is as follows:
Eq. 11 3.9 x 10-4 C4 x C7 Watts
Again, according to the notes of Tables 1 to 4 (see Table 3), we can find the values of C4 and C7. These are found as follows:
Eq. 12 C4 = 10 [0.002(l-700)] for the wavelengths of 840 nm to 870 nm we are considering
And:
Eq. 13 C4 =5 for wavelengths of 1290 1300, 1310, and 1320 nm
Likewise, the values of C7 are found as follows:
Eq. 14 C7 = 1 for wavelengths 840, 870 nm, and C7 = 8 for wavelengths > 1050 nm
Once each of these constants has been found, then each of the AELs for each wavelength may be calculated from Equation 11. They are found as follows:
P840nm = 0.74 mW
P870nm = 0.85 mW
P1290nm = P1300nm =P1310nm =P1320nm = 15.6 mW
The example next determines if the correction factor needs to be applied for passing the power into the aperture of the fiber we are considering. As stated in the beginning of the problem, the numerical aperture of the fiber was 0.2 +/- 0.02, and we purposely stated the uncertainty of 0.02. This means that the smallest NA for this fiber is 0.18.
As stated previously, 63 % of the power lies within a diameter D63, or from Equation 3, D63 = 2rNA/1.7. Since we know that the smallest NA is 0.2 - 0.02 = 0.18, all we need to know is the measurement distance, r. This can be found in Table 10 of IEC 60825-1, condition 2 (shown here as Table 4). Since we know that a < 1.5 mrad, then r = 14 mm.
Table 4: Diameters of the measurement apertures and measurement distances (Table 10 in IEC 60825-1)
Upon calculation, D63 = 3.0 mm. Since the measurement aperture is diameter 7.0 mm (the diameter of the fully dilated pupil), we see that all of the power passes through this aperture. Therefore, there is no need to use the correction factor.
As a further demonstration of the fact that aperture correction need not be done, let’s calculate h. It should be noted that as D63 increases, then h decreases (see Equation 5, which shows that the minimum h can be is 1). Therefore, when h is calculated using 7 mm as the diameter of the aperture and 3.0 mm as the diameter of the 65% point, then h = 0.996, which is obviously very close to 1. So, Pa = h Po is approximately Po, and no correction is needed.
The last step in the example is to set up the power to AEL ratio for each of the wavelengths and to sum them:
Eq. 15 S [(powerl)/AELl] = 0.16/0.74 + 0.16/0.85 + 4 x 0.16/15.6 = 0.45
Since the sum of the ratio is less than 1, then the original assumption is correct: the total emission is less than the Class 1 limits, and may thus be classified as Class 1.
Conclusion
The effect of multiplexing multiple wavelengths onto an optical fiber can add additional optical hazards. This hazard level can be calculated, and should be considered when creating an optical fiber network. Consideration should also be given to accurately determine the effect when laser light passes though an aperture, such as an optical fiber. g
References
1. IEC 60825-1:1993 +A1:1997 + A2:2001, International Electrotechnical Commission, “Safety of Laser Products. Part 1: Equipment Classification, Requirements and User’s Guide.”
2. IEC 60825-2:2004 (E), International Electrotechnical Commission, “Safety of Laser Products. Part 2: Safety of Optical Fibre Communication Systems.”
3. Conformity, October 2003, “Optical Safety Requirements for Lasers and LED Products,” Thomas M. Savino.
4. Jeff Hecht, Understanding Fiber Optics, Fourth Edition, Prentice Hall, 2001.
5. Dennis Derickson, Editor, Fiber Optic Test and Measurement, Prentice Hall, 1998.
About the Author
Tom Savino is a product safety engineer at Curtis-Straus LLC, a Bureau Veritas CPS Electrical unit. He can be reached at tsavino@us.bureauveritas.com.
