Nested Square roots Yue Kwok Choy Nested square roots problems are very interesting. In this article, we investigate some mathematical techniques applied to this topic that most senior secondary school students can understand. 1. √1 + √1 + √1 + ⋯ (a) We put Then x = √1 + √1 + √1 + ⋯ x 2 = 1 + √1 + √1 + ⋯ x 2 − 1 = √1 + √1 + ⋯ = x x2 − x − 1 = 0 We get a quadratic equation : Solving, we have x= 1+√5 2 ≈ 1.6180339887 … Note that the negative root is rejected since x > 0. 1+√5 You may also notice that √1 + √1 + √1 + ⋯ = = φ, the famous Golden Ratio. 2 (b) It seems that our job is done, but in fact we still need to show the convergence of √1 + √1 + √1 + ⋯. We write x1 = √1 = 1 x2 = √1 + √1 = √2 ≈ 1.4142 x3 = √1 + √1 + √1 = √1 + √2 ≈ 1.5538 x4 = √1 + √1 + √2 ≈ 1.5981 …. We note that: xn = √1 + √1 + √1 + ⋯ + √1 (1) The sequence xn (n square roots) ‘may be’ increasing. (2) xn = √1 + xn−1 1 We apply the Monotone Convergence Theorem, which states that every monotonic increasing (or decreasing) sequence bounded above (below) has a limit. (i) To prove xn is increasing, we use Mathematical Induction. Let P(n) : xn < 𝑥n+1 P(1) is true since x1 = 1 < √2 = x2 Assume P(k) is true for some kN , that is xk < 𝑥k+1 For P(k + 1), From (1) 1 + xk < 1 + 𝑥k+1 …(1) xk+1 = √1 + xk < √1 + xk+1 = xk+2 P(k + 1) is true. By the Principle of Mathematical Induction, P(n) is true nN . (ii) To prove xn is bounded, we also use Mathematical Induction. Let P(n) : xn < 2 (We use 2 here instead of 1+√5 2 to simplify our writing.) x1 = 1 < 2 P(1) is true since Assume P(k) is true for some kN , that is For P(k + 1), From (2) 1 + xk < 3 xk < 2 …(2) xk+1 = √1 + xk < √3 < 2 P(k + 1) is true. By the Principle of Mathematical Induction, P(n) is true nN . Finally we use the Monotone Convergence Theorem, xn has a limit, and Quiz lim n→∞ xn = 1+√5 2 . In (1) – (3) below, you may omit the proof of convergence. (1) Show that √a + √a + √a + ⋯ = 1+√1+4a (a > 0) 2 2 +4a (2) Show that √a + b√a + b√a + ⋯ = b+√b (3) Show that √2 − √2 + √2 − √2 + ⋯ = √5−1 2 2 (4) Find the mistake to prove that In (1), take limit But obviously, a 0, lim a→0 (a, b > 0) 1=0: √a + √a + √a + ⋯ = lim 1+√1+4a = 1 a→0 2 √0 + √0 + √0 + ⋯ = 0 1=0. 2 2. √2 + √2 + √2 + ⋯ We change our direction to employ trigonometry to tackle this. π cos 4 = π √2 2 3. π 4 √2(1+sin ) cos 8 = π (for those who don’t know radian, take π radian = 180o ) 2 π 8 = √2+√2 √2(1+sin ) = 2 (Here we use half-angle formula. Please fill in the calculations.) √2+√2+√2 …. cos 16 = 2 Hence, lim √2 + √2 + √2 + ⋯ = 2 n→∞ cos 2 π 2n =2 √1 + 2√1 + 3√1 + 4√1 + ⋯ The Indian mathematics genius Ramanujan once published this problem in the Indian Mathematical journal. He waited for more than 6 months, but no one came forward with a solution. He then discovered that: x+n+a = √ax + (n + a)2 + x√a(x + n) + (n + a)2 + (x + n)√a(x + 2n) + (n + a)2 + (x + 2n)√… … (3) This problem is then a special case where a = 0, n = 1, x = 2. We don’t discuss the story here as it is a bit involved. Readers interested may study: http://en.wikipedia.org/wiki/Denesting_radicals#Infinitely_nested_radicals However, we can still carry out our informal investigation on √1 + 2√1 + 3√1 + 4√1 + ⋯ Firstly study that, x1 = √1 = 1 x2 = √1 + 2√1 = √3 ≈ 1.7321 x3 = √1 + 2√1 + 3√1 ≈ 2.2361 3 x4 = √1 + 2√1 + +3√1 + 4√1 ≈ 2.5598 We therefore guess that : xn = √1 + 2√1 + +3√1 + 4√1 + ⋯ n√1 → 3 We then apply the following identity: n = √1 + (n − 1)(n + 1), for n ≥ 1 repeatedly: 3 = √1 + 2 × 4 = √1 + 2√3 × 5 = √1 + 2√1 + 3√1 + 4 × 6 = ⋯ = √1 + 2√1 + 3√1 + 4√1 + ⋯ + (k − 1)√1 + k(k + 2) ?? = √1 + 2√1 + 3√1 + 4√1 + ⋯ The proof above may already look for most readers, but I still put “??” in the last step, as it seems to be a not so mathematical “deduction”. Studying the last expression above more thoroughly, we may guess a more general formula : 4 = √1 + 3√1 + 4√1 + ⋯ 5 = √1 + 4√1 + 5√1 + ⋯ … n = √1 + (n − 1)√1 + n√1 + (n + 1)√1 + ⋯ 4 If we assume 3 = √1 + 2√1 + 3√1 + 4√1 + ⋯ is proved, we can begin our induction as follows: n = √1 + (n − 1)√1 + n√1 + (n + 1)√1 + ⋯ Let P(n) : Assume P(k) is true for some kN , that is k = √1 + (k − 1)√1 + k√1 + (k + 1)√1 + ⋯ For P(k + 1), From (4), …(4) k 2 = 1 + (k − 1)√1 + k√1 + (k + 1)√1 + ⋯ k 2 − 1 = (k − 1)√1 + k√1 + (k + 1)√1 + ⋯ (k − 1)(k + 1) = (k − 1)√1 + k√1 + (k + 1)√1 + ⋯ k + 1 = √1 + k√1 + (k + 1)√1 + ⋯ P(k + 1) is true. By the Principle of Mathematical Induction, P(n) is true nN . So we have proved n = √1 + (n − 1)√1 + n√1 + (n + 1)√1 + ⋯ which is a step closer to the Ramanujan formula given by (3). Quiz 3 Find √1 + 3√1 + 3√1 + ⋯ 5