The problems of manufacturing in core fission cham bers and ion chambers small enough to fit the allowable space, rugged enough to withstand the in core environment, and inexpensive enough to permit extensive coverage of a power reactor core have led to efforts to develop other types of devices to measure neutron flux at fixed locations in a power reactor One device developed to solve these problems is the self powered neutron detector

The self powered neutron detector uses the basic radioactive decay process of its neutron activated material to produce an output signal As the name implies, no external source of ionizing or collecting voltage is required to produce the signal current The construction of the detector is quite different from that of an m-core fission chamber There is no gas filled region where ionization takes place, instead, the detector is of solid construction with a neutron-sensitive material connected to a lead wire, both being separated from the detector outer sheath by hard packed ceramic insulation The resulting detector is like a mineial-insulated coaxial cable, small m size and rugged

The simplicity of construction and operating principle lead to a number of advantages, including low cost, simplicity of readout equipment, low burnup rate, long lifetime, and ease of reproducing sensitivity

(a) Operating Principles. As shown in Pig 3 10, a

typical self-powered neutron detector consists of four parts emitter, insulation, lead wire, and sheath (or collec­tor) The emitter is a material of reasonably high thermal-
neutron activation cross section which, after activation, decays b emission of high energy betas with a reasonably short half life The insulation is a solid that must maintain high electrical resistance m the in-core temperature and nuclear-radiation environment, it should ideally emit no betas or electrons (e g, from neutron activation) Both the lead wire and sheath, or collectoi must emit few betas or electrons compared to the emitter so that undesirable background signals are minimized

When the self-powered neutron detector is placed in a neutron flux and connected to ground through an electrical current meter, the current measured is proportional to the net beta escape from the emitter If the detector remains in the neutron flux until equilibrium beta emission is reached, the current is directly proportional to the incident-neutron flux

When the activity induced by thermal neutron absorp­tion m the emitter is from a single radionuclide with a single half life, the detector output after exposure to the neutron flux for a time t is (3 1) where I(t) = detector current (amp) at time t (sec)

К = a dimensionless constant determined by detec­tor geometry and materials aatt = thermal neutron activation cross section of emitter material (cm2)

Q = charge emitted by emitter per neutron ab sorbed (coulombs)

N = number of emitter atoms T^ = half life of radionuclide generated in emitter (sec)

ф = thermal-neutron flux (neutrons env2 sec 1)

Steady state is reached when the exposure time is several times the half-life, T^, of the radionuclide

I0(t) = Kaact QN ф

where I0(t) is the steady state detector current at time t The inclusion of t is made necessarv by the fact that the number of emitter atoms, N, is decreasing with time

N = N(t) = N0 exp (-Кффах) (3 3)

where Кф is the fraction of the detector constant К which accounts for the neutron self shielding m the emitter and о is the thermal neutron-absorption cross section of the emitter material Introduction of the factor Кф is necessary since atoms of the emitter are exposed to a neutron flux that has been attenuated by the intervening atoms of the sheath, insulation and the emitter itself Substitution of Eq 3 3 into Eq 3 2 gives the steadv-state detector current

I0(t) = K<7actQN0 exp (—Кффог) ф (t>l^) (3 4)

The sensitivity of a self powered detector is defined as the change in the steadv state detector current per unit change in the thermal-neutron flux

s(t) [KaactQN0 exp (—Кффог)]

x (1 — Кффаг) (3 5)

The detector sensitivity is seen to decrease (because of burnup of the emitter material) approximately exponen tially with time The rate of sensitivity decrease is deter mined by the thermal-neutron flux that the detector is exposed to, the thermal neutron-absorption cross section of the emitter material, and the neutron self shielding m the emitter

If the detector life is defined as the time during which the sensitivity decreases to a fraction f of its initial value when the detector is in a constant thermal-neutron flux, then Eq 3 5 shows that

f=e1/T(l-7) (3 6)

where T = detector life in constant thermal neutron flux (ф) f = S(t = T)/S(t = 0) = relative sensitivity at t = T t = /Кффа = time to complete burnup of emitter material assuming and ф arc constant

The lifetime T can be calculated as a function of f from Fq 3 6 The values of T corresponding to f = 0 9, 0 8 0 7, 0 6, and 0 5 are found to be 1 = 0 05r 0 llr 0 17т 0 24r and 0 31r respectively Thus for example, if the detector lifetime is defined as the time for the sensitivity to decrease to 60% of its initial value then the lifetime is 24% of its “complete burnup’ time or 1 = 0 24r = 0 24/Kффа

The constant К in the basic equation, Eq 3 1, takes into account several effects associated with the detector structure and materials Specifically, the detector current is reduced by a factor Кф because the emitter atoms are partly shielded from the neutron flux (incident on the detector) by the atoms nearer the detector surface The detector current is also reduced by a factor because some of the betas emitted by the radionuclides in the emitter are unable to escape from the emitter Finally, the detector current is reduced by a factor Kg because the geometry, particularly the insulation thickness, may not permit some betas to reach (or traverse) the detector sheath Thus the constant К = K^K^Kg

(b) Construction and Materials Self-powered neutron detectors can be manufactured in several ways using different construction materials and different manufactur­ing techniques Table 3 3 shows the various neutron — activated beta emitters that have been considered as potential candidates for the emitter material Of the materials listed, only 103Rh and 51 V have been used in commercial applications Each of the others has been

rejected for one or more undesirable characteristics Be cause each has a low thermal neutron-activation cross section, 7Li, 1 1 B, and 27A1 yield unacceptably low signal-to noise ratios With a 2 58-hr half life, 55Mn results in too long a time constant if used in a detector Since 99Tc does not occur naturally, it is not readily available,11 5 In is unsatisfactory because 76% of its beta decay has a 54 min half-life Silver has both an acceptable cross section value and acceptable half-lives, but it would be difficult to compensate for burnup with the 24-sec 1 9 Ag burning up three times as fast as the 2 4-min 10 7 Ag

Neither 103Rh nor 51 V has any of these undesirable characteristics Because of its larger thermal neutron- activation cross section, 103Rh is used where short detec tors are needed for measuring local flux Many 103Rh detectors distributed throughout the reactor core can provide three dimensional power distribution information The relatively lower signal level of 5 1 V is more suited to long detectors designed to average the neutron flux over the full core height Such detectors cannot be used to deter mine power distribution along the axis of the reactor core, although several 5 1 V detectors dispersed throughout the core can provide data on the radial power distribution

Figures 3 11 and 3 12 show the two types of construc­tion commonly used for self powered neutron detectors in commercial use today

The earliest, and in many respects the simplest, type of detector construction is shown in Fig 3 11 The 0 020-in — diameter rhodium wire emitter is fastened to the Inconel lead wire of a standard 0 040-in magnesium oxide in­sulated, Inconel sheathed, coaxial cable All parts are baked out before assembly and are maintained scrupulously clean during assembly The section of the detector sensitive to neutrons has a larger diameter than the coaxial cable The construction of a neutron-detector assembly containing several of these detectors can be complicated by the change in diameter

Figure 3 12 shows an alternate construction The neutron-sensitive emitter is fastened to the Inconel lead wire before the insulation is installed Magnesium oxide insulators are threaded over the emitter and lead wire The Inconel sheath is slid over the insulators, and the entire assembly is swaged down to a finished diameter of 0 062 in Bakeout before assembly and rigorous cleanliness during

Fig 3 11—Self powered neutron detector

• INCONEL LEAD WIRE 0 062 IN DIA

RHOD’UM WIRE (EMITTER) •

Fig 3 12—Self powered neutron detector

assembly are important procedures The resulting detector has a coTistant diameter over its length, and the magnesium oxide insulation is compacted tightly around the emitter

Only three insulation materials have been considered for self powered neutron detectors aluminum oxide, beryl­lium oxide, and magnesium oxide Aluminum oxide has been most frequently used when the detector is assembled as shown in Fig 3 11 Aluminum oxide is not suitable for detectors assembled as shown in Fig 3 12 because of the danger of damaging the emitter or lead wire during the swaging operation In addition, aluminum is activated by thermal neutrons and emits high-energy betas (see Table 3 3) which contribute to background and reduce the signal-to-noise ratio The characteristic 2 З-mm beta decay of aluminum has been observed during irradiation of cables insulated with aluminum oxide Whether or not this is tolerable depends on the accuracy desired in the detector output signal Beryllium oxide has no significant advantage other than its extremely low thermal-neutron cross section This advantage, however, is more than offset by its high cost and toxicity Magnesium oxide is the most satisfactory of the three insulation materials because of its low cost, high resistivity, workability, and low noise potential

The two most commonly used sheath materials in detectors for pressurized-water and boiling-water reactors are type-304 stainless steel and Inconel because both are compatible with the reactor coolant Inconel is now standard in all commercial detectors The use of 30 r stainless steel in the sheath of detectors intended for high-accuracy applications is problematical since manga­nese, which constitutes about 0 5 wt % of most stainless steels, is activated by thermal neutrons and emits betas (see Table 3 3 )

Inconel is used universally as the lead wire material, although Nichrome was used successfully in earlier detec­tors

(c) Sensitivity The initial sensitivity of a self powered detector is given by Eq 3 5, with t = 0

S(0) = KaattQNo0 (3 7)

The number of emitter atoms at t = 0 is

where p = density of emitter material (g/cm3)

A0 = Avogadro’s number = 6 02 X 1023 A = atomic weight of emitter d = diameter of emitter (cm)

/ = length of emitter (cm)

The emitter is assumed to be cylindrical and made of a pure element It is further assumed that the emitter decays with the emission of a single beta, і e , Q = 1 60 X 10 1 9 cou lomb, the electron charge

As noted at the end of Sec 3 3 3(a), the constant К is the product of Кф, Kp, and Kg, the constants that take into account the neutron self-shielding, beta self-shielding, and geometric effects Figure 3 13 shows the values of these constants as a function of emitter diameter for rhodium with 10 mils (0 25 mm) of MgO insulation Figure 3 14 shows the same constants for vanadium emitters

The initial sensitivity of a rhodium detector with a 20-mil (0 51-mm) emitter and 10 mils of MgO insulation can be calculated from Eqs 3 7 and 3 8 using К values from Fig 3 13 and the rhodium cross section from Table 3 3 (The density of rhodium is 12 4 g/cm3 ) The result is

у = 1 З X 1 (Ґ11 amp/(neutrons cnrf2 sec 1 )

A similar calculation for a vanadium detector with a 20 mil emitter and 10 mils of MgO insulation yields (vanadium density is 6 0 g/cm3)

C

— = 7 1 X 1CU2 3 amp/(neutrons cm"2 sec 1 )

Fig 3 14—К factors for vanadium detectors with 10 mils of MgO

This value, as well as the one for the rhodium detector, is in good agreement with experimental values

Because of the rapid decrease in К with increasing emitter diameter, detector sensitivity depends more on emitter surface area than on emitter volume This is evident from Fig 3 15, where the detector sensitivity is shown as a function of emitter diameter The relation is nearly a straight line

Figure з 16 shows the change in detector sensitivity when the detector length and diameter are varied but the emitter mass is kept constant

(d)

Emitter Burnup. The depletion of the neutron sensitive emitter caused bv its absorption of thermal

Fig 3 16— Rhodium detector sensitivity vs length at con stant emitter mass

neutrons reduces the sensitivity of the detector (see Eq 3 5) In the derivation of Eq 3 5, it was assumed that all factors were constant except t However, the neutron self-shielding factor, K^, is not constant during prolonged exposure to neutrons it tends to increase as the emitter burnup increases because burnup removes (transforms) some of the self shielding atoms The resulting change m is not significant over a short period of operation, but it must be taken into account in am readout system that automatical^ compensates for detector burnup

From Eq 3 5 the relative detector sensitivity is

S(t)

^y=exp( K00ot)(l K00at)

= et/T(l (3 9)

where r = 1 /Кффо

For a rhodium detector with a 20 mil emitter and 10 mils of MgO insulation, the characteristic burnup time r is 0 913 X lO22/0 In an acerage thermal neutron flux 0 = 4 X 1013 neutrons cm”2 sec1 the value of r is 23 X 107 sec, or 7 18 years When the detector is exposed to this flux for one year, its sensitivity relative to its initial value decreases to exp (-1/7 18) X [1 —(1/7 18)1 =0 75 In other words, its sensitivity is now 75% of its initial value In two years, the sensitivity decreases to 55% of its initial value From this it follows that the emitter burnup time is longer than the normal reactor refuelling cycle The emitter burnup rates given in Table 3 3 are based on the assumption that Кф = 1 As noted earlier, Кф is less than 1, so the actual burnup rates will be lower than those given in the table

The product NQ in Eqs З 1 through 3 4 is equal to the total charge (betas) generated in the emitter before it is used up (all atoms activated) Since and Kg are the factors that identify how many betas escape to become useful detector output current, the total useful charge generated bv the detector is

q = K^KgQN (3 10)

If we use the К values from Fig 3 13 and calculate N from Eq 3 8, we find the total useful charge generated by a 20 mil rhodium detector with 10 mils of MgO insulation to be

q = 12 1 coulombs/cm of emitter length which is m good agreement with experimental values

(e) Response Characteristics The response character istics of a self powered neutron detector are directly related to the radioactive scheme of the radionuclides formed in the emitter

Figure 3 17 shows the decay scheme for self-powered detectors that use vanadium as the emitter material All neutron absorptions in the emitter material, 51 V, which has a thermal neutron-activation cross section of 4 9 barns, result in the creation of 52V The latter decays by beta emission to four excited states of 52Cr which then immediately go to the ground state by gamma emission The half life of this decay scheme is 3 76 min (226 sec) The 2 4 MeV beta accounts for almost 99% of the beta emission and provides energetic betas for a good signal to- noise ratio

Because the emitter radionuclide has a single beta decay period, the time response of the vanadium detector to a step change in thermal neutron flux from 0 to zero is

I(t) = I0e^ 69 31/2 2 6 =I0e 1/326 (3 1 1)

ind the time response to a step change in thermal neutron flux from zc ro to ф is

I(t) = I0(l — e t/32 6 ) (3 12)

where t is in seconds l(t) is the detector signal current, and I0 is the steady state signal given in Eq 3 2 The mean life 326 sec, is the half life divided by 0 693, it is the time for I(t) to decrease by 1/e Figure 3 18 shows the response of a vanadium self powered detector to a step decrease m neutron flux (Eq 3 11) on a semilogarithmic scale T e vanadium detector response follows an exponential with a characteristic 3 76 min half-life (time constant = 3 76/0 693 = 5 4 min) and reaches 90% of the step change in 12 5 min Figure 3 19 shows the response of a vanadium self powered detector to a step decrease and a step increase in neutron flux, the curves are plotted on a linear scale

Figure 3 20 shows the decay scheme for rhodium emitters Neutron absorption by rhodium creates two radioisotopes of 104Rh The total neutron activation cross section of rhodium is 150 barns The cross section for the creation of the ground state 104Rh is 139 barns (92 7% of the 150 barns), and, for the creation of the metastable і 0 4 m Rh, tross section is 11 barns (7 3% of 150 barns) The metastable state decays by gamma emission to the ground state with a characteristic half life of 4 4 min (264 sec) The ground state 1 04 Rh decays by beta emission to the ground state 104 Pd with a half-life of 42 sec In this final beta decay process, 98% of the emitted betas have an

energy of 2.44 MeV, sufficiently high to provide adequate signal-to-noise ratio

Because it has an emitter with two beta-decay con­stants, the time response of a rhodium self-powered detector to step changes in neutron flux involves two

Figure 3 18 shows I(t)/I0 vs time on a semilogarithmic plot for a step decrease m thermal-neutron flux on a rhodium detector For the first 3 min after the step change, the detector signal is dominated by the 42-sec half-life (61-sec time constant), beyond 6 mm after the step change, the

rhodium detector signal follows the 4 4-min decay of 10 4m Rh slgna[ reaches 90% of the step change in less than 3 min The у-intercepts of the two half-life curves in Fig 3.18 correspond to the coefficients of each exponential in Eq 3 13 Figure 3 19 shows the time responses of a rhodium detector to step decreases (Eq 3 13) and to step increases (Eq 3 14) in flux, plotted on a linear scale

(f) Connecting Cables. The connecting cables required to bring the self-powered detector signals out of the core should produce as little background signal as possible As shown in Figs 3 11 and 3 12, the cables commonly used are 40 or 62 mils in outside diameter with magnesia insulation and Inconel lead wire and sheath The neutron cross section for Inconel is negligible, so neutron-induced signals are negligible The major sources of noise signals are

Compton electrons and photoelectrons generated when gamma rays are absorbed or scattered by the lead wire and sheath The noise signals are prompt in responding to changes m incident gamma flux

Compton electrons and photoelectrons originating in the wire and absorbed in the sheath produce positive background signals Electrons originating in the sheath and absorbed in the lead wire produce negative background signals Equal absorption of electrons in the wire and sheath would result in a zero background signal, ideal for maximum neutron signal The larger mass and surface area of the sheath tend to make a negative background signal for most connecting cables Appropriate selection of emitter and sheath materials and dimensions should make it possible to build cables with essentially zero background signal