Standard Limits: x^n and x^{1/n}

Powers of a number below 1 in absolute value, and roots of a positive number, take the stage. The scene checks where each of them gathers.

Keep multiplying a number below 1 and it sinks toward 0. Take roots of a positive number again and again and it gathers at 1. At −1 the sign just flips back and forth. Beyond 1 in size, powers grow without end. Bernoulli's inequality handles every case.

For example, eat half of the remaining 🍰 cake each day. What is left goes one half, one quarter, one eighth — nearly vanishing.

∣x∣<1⇒xn→0|x|<1 \Rightarrow x^n \to 0
x>1⇒x1/n→1x>1 \Rightarrow x^{1/n} \to 1
n1/n→1n^{1/n} \to 1

Examples & Exercises

Prove completely by the ε\varepsilon-NN argument that xn→0x^n \to 0 when ∣x∣<1|x| < 1.

Hint: put r=1/∣x∣>1r = 1/|x| > 1, write r=1+hr = 1 + h, and use Bernoulli's inequality (1+h)n≥1+nh(1+h)^n \ge 1 + nh.

r=1∣x∣>1,r=1+h,∣x∣n≤11+nh→0r = \frac{1}{\lvert x\rvert} > 1, \quad r = 1 + h, \quad \lvert x\rvert^n \le \frac{1}{1 + nh} \to 0
x=0.5x = 0.5, n=20n = 20: the dots sink toward 00 — 0.520≈9.54×10−70.5^{20} \approx 9.54\times 10^{-7}, 0.560≈8.67×10−190.5^{60} \approx 8.67\times 10^{-19}.
Show proof
  1. The case x=0x = 0 is trivial, so assume 0<∣x∣<10 < |x| < 1. Put r=1/∣x∣r = 1/|x|; then r>1r > 1, so r=1+hr = 1 + h with h>0h > 0.

  2. Bernoulli's inequality (1+h)n≥1+nh(1+h)^n \ge 1 + nh (for h>−1h > -1): equality at n=1n = 1, and if (1+h)n≥1+nh(1+h)^n \ge 1 + nh then (1+h)n+1=(1+h)n(1+h)≥(1+nh)(1+h)=1+(n+1)h+nh2≥1+(n+1)h(1+h)^{n+1} = (1+h)^n(1+h) \ge (1+nh)(1+h) = 1 + (n+1)h + nh^2 \ge 1 + (n+1)h, so induction gives it. Hence ∣x∣n=1/rn=1/(1+h)n≤1/(1+nh)|x|^n = 1/r^n = 1/(1+h)^n \le 1/(1+nh).

  3. Given ε>0\varepsilon > 0, the Archimedean property gives a natural N>(1/ε−1)/hN > (1/\varepsilon - 1)/h. For n>Nn > N, nh>1/ε−1nh > 1/\varepsilon - 1, so 1+nh>1/ε1 + nh > 1/\varepsilon, i.e. 1/(1+nh)<ε1/(1+nh) < \varepsilon.

    Hence ∣xn−0∣=∣x∣n≤1/(1+nh)<ε|x^n - 0| = |x|^n \le 1/(1+nh) < \varepsilon — xn→0x^n \to 0.

Prove completely that x1/n→1x^{1/n} \to 1 when x>1x > 1.

Hint: put tn=x1/n−1≥0t_n = x^{1/n} - 1 \ge 0 and apply Bernoulli's inequality to x=(1+tn)nx = (1+t_n)^n.

tn=x1/n−1≥0,x=(1+tn)n≥1+ntn,0≤tn≤x−1n→0t_n = x^{1/n} - 1 \ge 0, \qquad x = (1+t_n)^n \ge 1 + nt_n, \qquad 0 \le t_n \le \frac{x-1}{n} \to 0
x=2x = 2, n=20n = 20: 21/n2^{1/n} descends to 11 from above — 21/20≈1.03532^{1/20} \approx 1.0353, 21/60≈1.01162^{1/60} \approx 1.0116.
Show proof
  1. Put tn=x1/n−1t_n = x^{1/n} - 1. Since x>1x > 1, x1/n>1x^{1/n} > 1 and tn≥0t_n \ge 0.

    By definition x=(1+tn)nx = (1 + t_n)^n.

  2. By Bernoulli's inequality (problem 1), (1+tn)n≥1+ntn(1+t_n)^n \ge 1 + nt_n, so x≥1+ntnx \ge 1 + nt_n, i.e. 0≤tn≤(x−1)/n0 \le t_n \le (x-1)/n.

  3. Ground for the bound (x−1)/n→0(x-1)/n \to 0: 1n→0\frac{1}{n} \to 0 by the Archimedean property (given ε′>0\varepsilon' > 0, pick N>1/ε′N > 1/\varepsilon'; then n>Nn > N gives 1n<ε′\frac{1}{n} < \varepsilon'). Multiplying by the fixed constant (x−1)(x-1) keeps the limit 00, so (x−1)/n→0(x-1)/n \to 0, and the squeeze theorem gives tn→0t_n \to 0, i.e. x1/n=1+tn→1x^{1/n} = 1 + t_n \to 1. (In ε\varepsilon-NN form: for N>(x−1)/εN > (x-1)/\varepsilon, n>N⇒∣x1/n−1∣=tn≤(x−1)/n<εn > N \Rightarrow |x^{1/n} - 1| = t_n \le (x-1)/n < \varepsilon.)

Prove completely that n1/n→1n^{1/n} \to 1.

Hint: put tn=n1/n−1t_n = n^{1/n} - 1 and keep only the (n2)tn2\binom{n}{2}t_n^2 term of the binomial expansion.

tn=n1/n−1,n=(1+tn)n≥(n2)tn2,0≤tn≤2n−1→0t_n = n^{1/n} - 1, \qquad n = (1+t_n)^n \ge \binom{n}{2}t_n^2, \qquad 0 \le t_n \le \sqrt{\frac{2}{n-1}} \to 0
x=2x = 2, n=60n = 60: shows 21/k→12^{1/k} \to 1. The proof below applies the same binomial trick with x=nx = n to get tn≤2/(n−1)t_n \le \sqrt{2/(n-1)} (at n=60n = 60 the bound is 0.18410.1841, the actual t60≈0.0706t_{60} \approx 0.0706).
Show proof
  1. Put tn=n1/n−1t_n = n^{1/n} - 1. Since n≥1n \ge 1, tn≥0t_n \ge 0, and by definition n=(1+tn)nn = (1+t_n)^n.

    The case n=1n = 1 is trivial (11/1=11^{1/1} = 1, i.e. t1=0t_1 = 0); the bound 2/(n−1)\sqrt{2/(n-1)} below divides by n−1n - 1, so assume n≥2n \ge 2 from now on.

  2. For n≥2n \ge 2, the binomial theorem gives (1+tn)n=∑k=0n(nk)tnk≥(n2)tn2(1+t_n)^n = \sum_{k=0}^{n}\binom{n}{k}t_n^k \ge \binom{n}{2}t_n^2 (the dropped terms are all nonnegative). So n≥n(n−1)2tn2n \ge \frac{n(n-1)}{2}t_n^2.

  3. Dividing by n>0n > 0 gives 1≥n−12tn21 \ge \frac{n-1}{2}t_n^2, i.e. (for n≥2n \ge 2) 0≤tn≤2/(n−1)0 \le t_n \le \sqrt{2/(n-1)}. Ground for the bound tending to 00: given ε>0\varepsilon > 0, the Archimedean property gives N>1+2/ε2N > 1 + 2/\varepsilon^2; then n>Nn > N gives 2n−1<ε2\frac{2}{n-1} < \varepsilon^2, i.e. 2/(n−1)<ε\sqrt{2/(n-1)} < \varepsilon.

    Hence 2/(n−1)→0\sqrt{2/(n-1)} \to 0, and the squeeze theorem gives tn→0t_n \to 0, i.e. n1/n=1+tn→1n^{1/n} = 1 + t_n \to 1. Numerical check: at n=60n = 60 the bound is 2/59≈0.1841\sqrt{2/59} \approx 0.1841 and the actual 601/60−1≈0.070660^{1/60} - 1 \approx 0.0706 sits inside it.

3. Sequences