STEP 3 2022 Examiners' Report and Mark Scheme
STEP 3 2022 Examiners' Report and Mark Scheme
STEP 3 2022 Examiners' Report and Mark Scheme
Mathematics 3 [9475]
2022
Examiners' Report
Mark Scheme
STEP MATHEMATICS 3
2022
Examiners' Report
STEP 3 Introduction
One question was attempted by well over 90% of the candidates two others by about 90%,
and a fourth by over 80%. Two questions were attempted by about half the candidates and
a further three questions by about a third of the candidates. Even the other three received
attempts from a sixth of the candidates or more, meaning that even the least popular
questions were markedly more popular than their counterparts in previous years.
This was the most popular question with 94% attempting it, and it was also the most
successful with a mean mark of nearly 14/20. Apart from very occasional inaccuracies, part
(i) was always successfully done.
The first summation result in part (ii) was usually successfully done, though there was some
poor summation notation which let some candidates down. The second summation was
completed successfully by virtue of some heavy algebra or, more efficiently, by seeing the
connection to the first result dividing the quartic by t4 and comparing the quartic for the
reciprocals with that in part (i). Some candidates were penalised for not justifying their
result, having clearly worked backwards from part (iii).
Part (iii) was well done except when candidates disregarded their result from part (ii).
A lot of candidates managed to correctly interpret the implication of the curves touching
at two distinct points in terms of the roots for t and the consequent result for the product
of the four roots, but then struggled to reach the required result by algebra or poorly
justified geometric arguments.
Question 2
The fifth most popular question, being attempted by just a little over half the candidates, it
was the fourth most successful with a mean score of 9/20.
Whilst the algebra associated with this question was not difficult, the logic and
communication required was certainly too much for many students.
In part (i) it required some justification that 𝑎𝑎 had to be even. Contradiction or infinite
descent could be used but either way the argument had to be made clear. Claiming “this can
be continued forever” or moduli were always decreasing would eventually get to zero was
not good enough. Successful candidates were able to explain why the integer nature of the
solutions was vital to reach a contradiction.
In part (ii) many candidates were able to see that this was a similar problem to the first one,
and most observed that divisibility by three was now the key idea.
In part (iii), many candidates were able to consider the remainders when divided by 3, but
again many struggled to communicate clearly an argument leading to the final
contradiction.
By part (iv) most candidates were expecting to recreate the original equation again and the
fact that this did not happen meant some came to a dead halt. Other were either oblivious
to the issue or were bluffing their way through as a slightly more subtle argument was now
required.
Question 3
This was very popular, being attempted by over 90%, but not very successfully, with a mean
score of about 5.5/20. In part (i), candidates generally obtained a correct equation for x or
y, but then failed to properly justify the manipulation of the inequality. Whilst the quartic
was frequently correctly obtained in part (ii), there were a number of different incorrect
assumptions or assertions made regarding the two stationary points being repeated roots or
the value of the quartic having different signs at the two stationary points. It was also
common that the case when c is negative was not considered. Whilst it was not uncommon
for candidates to argue incorrectly for part (iii) that the three equations were equivalent to
𝑑𝑑𝑑𝑑 0
the curve C2 in part (ii) having one stationary point, (often using = ), in contrast, a
𝑑𝑑𝑑𝑑 0
pleasing number of candidates who made little progress in (ii) past obtaining the quartic,
approached part (iii) by simply attempting to solve the equations by elimination, earning full
or close to full marks.
Question 4
This was the fourth most popular question being attempted by more than four fifths of the
candidates, with a moderate degree of success scoring a mean of 9/20.
Part (i) suffered from incorrect flows of logic in the inductive and base cases, as well as
failure to mention anything about not dividing by zero.
In part (ii) many ignored the instruction to use the Maclaurin series, and used de L’Hopital’s
Rule to their cost, and some ignored the higher order terms.
Part (iii) was generally well done, though the most common error was not justifying the
evaluation of the product using a geometric series in the exponent.
For part (iv), the best attempted route was to use an imaginary substitution which led to
mostly successful solutions. Some candidates attempted to prove an analogous
trigonometric identity using similar arguments to the previous parts, however losing marks
for not sufficiently fleshing out the details, and some attempted to use Osborn’s Rule, often
with insufficient justification or stating that it was being used. Once the identity was
achieved, the calculation was generally done well if the candidate progressed this far.
Question 5
This was only very slightly less popular than question 3, but it was the third most successful
with a mean of just under half marks.
Part (i) was well done, with a variety of methods used, the most common being by a
substitution of 𝑒𝑒 𝑥𝑥 . In this part, the most frequent errors were showing insufficient working
1+𝑒𝑒𝑎𝑎
to fully justify the given result, not spotting how to simplify , and incorrectly
1+𝑒𝑒−𝑎𝑎
1
integrating ∫ 𝑑𝑑𝑑𝑑 to get ln (1 + 𝑒𝑒 𝑥𝑥 ).
1+𝑒𝑒𝑥𝑥
Part (ii) was generally found to be the hardest. There was a range of responses to the first
requirement from concise use of the Fundamental theorem of Calculus to long, often
imprecise, paragraphs of text. Candidates attempting proof by contradiction tended to be
more successful if they used a sketch to back up their argument. The second result saw
many different methods used. The most common mistakes were not showing enough
working when using a u=-x substitution, not showing that the argument can be reversed,
𝑎𝑎
and using an incorrect argument such as ∫−𝑎𝑎 𝑔𝑔 (𝑥𝑥) = 0 → 𝑔𝑔 (𝑥𝑥) = 0 (to which 𝑔𝑔(𝑥𝑥) = 𝑠𝑠𝑠𝑠𝑠𝑠 𝑥𝑥
is a counterexample). Many candidates did not see the link with the first requirement of the
part. However, the final result of this part was usually done well.
Candidates found part (iii) easier than part (ii), the most common mistake was not realising
that ℎ(𝑥𝑥) = ℎ(−𝑥𝑥) holds, even though this was stated in the question.
Part (iv) was generally done well, with the most common mistakes being neglecting to show
that the functions satisfied the conditions in part 3 or omitting a factor of 2. A few
candidates did not use the results from the previous parts, instead using other methods,
which, as the question stated “hence”, gained no credit for this part.
Question 6
About half the candidates attempted this, but it was one of the least successful with a mean
score of one quarter marks.
Many candidates managed the opening ‘show that’ in part (i) but the limit attempt had
varying levels of success, and a common error was division by a quantity that was not
necessarily nonzero.
In part (ii), diagrams were regularly lacking, often being drawn extremely small with the
most salient details omitted.
In part (a), very few indicated from where the second term in the expression for 𝑥𝑥 arose.
Most attempts appealed to a diagram but did not indicate the pertinent angles.
Many formed the correct equation in (b), but a large number forgot to account for the
periodicity; those that remembered to do so largely did so correctly.
Many who got to (c), erroneously evaluated a 0/0 limit and then argued that the cotangent
was the answer they wanted. However, pleasingly others did spot the zeros and
manipulated the trigonometry effectively.
Question 7
More than a third attempted this, marginally more successfully than questions 3 and 6.
Many attempts were restricted to part (i). The first result was generally achieved, and whilst
the second result was often obtained, quite a few had difficulties doing so because they
overlooked that n was a unit vector and what this implied. Far fewer correctly drew and
labelled the diagram required in part (i) because they failed to appreciate the magnitudes of
the three vectors and that two were perpendicular.
Parts (ii) and (iii), when attempted, saw candidates fall into two camps. A small number
could see what both transformations were and using the considerations suggested in (ii) in
part (iii) as well, could justify their answers. However, a larger number had some idea what
the transformations might be, but often failed to define them precisely, and likewise failed
to justify their conclusions, even given the approach to use in (ii).
Question 8
This was the least popular Pure question, being attempted by marginally fewer than
question 7, but by more than any of the Mechanics or Probability and Statistics. The mean
mark was 6/20.
Generally, part (i) was done well and candidates used binomial expansions accurately,
manipulating their results to find the two required expressions. A few did not gain full
credit through providing insufficient working for the result given in the question.
More than half the candidates progressed no further than attempting part (ii) and, of those
who did attempt it, often stopped part of the way through, although there were some very
well-reasoned attempts. Most candidates attempting part (ii) substituted a = sec(θ) into
their sin expansion but found it difficult to complete the argument to explain why k had to
be even. Of those who got further and successfully managed to show the given results,
often the relevance of those results was not appreciated, and some candidates attempted
to prove irrationality by quoting the irrationality of 𝜋𝜋, despite the fact that the question
stated 𝜃𝜃 was measured in degrees. Very few candidates gained full credit for this part.
Those candidates who gained full credit in part (ii) also did well in (iii).
Question 9
The most popular of the Applied questions, with a third of candidates attempting it; it was
the second-best scoring question on the paper with a mean score of just above half marks.
The question relied mainly on the use of conservation of momentum and Newton’s
experimental law of impact. Most candidates made a very good start with several scoring
full or close to full marks in the first part. The difficulties arose later when dealing with the
three-particle situation in part (ii). Very few candidates were able to take a step back and
see how this problem linked to part (i), resulting in long pages of algebraic manipulation
which were inefficient and rarely correct. A good diagram would have made the link so
much more obvious!
Question 10
Along with question 12, this was the least popular question on the paper with a sixth of
candidates trying it, and scoring one third marks, on average.
Part (i) was done well, demonstrating good use of Hooke’s law, and resolving forces. It was
failing to think about right angled triangle trigonometry that created most problems.
In part (ii), many candidates got the signs of their potential energies wrong. Of those
candidates who got to the correct expression for 𝑝𝑝 most were able to find the maximum
value correctly but very few were able to explain why the physical situation resulted in a
restricted domain for the function. Showing the value of 𝑝𝑝 must be 0.7 to one significant
figure was rarely done well as many candidates used known approximations to the given
surds without justifying the accuracy of these approximations.
Question 11
A quarter of the candidates attempted this, scoring a mean of one third marks. Of these,
about a quarter made little or no progress. However, there were also several very good
attempts achieving most or all of the marks available.
In part (i)(a) the majority of candidates noticed the symmetry of the distribution and were
therefore able to answer this part well, although errors such as omitting the binomial
coefficients and forgetting that X could take the value 0 were made in some cases.
In part (i)(b) most candidates were able to see that the modulus sign in the sum effectively
meant that the calculation of 𝛿𝛿 should be split into two sums. However, in several cases
candidates simply observed that the given result followed from the two sums by symmetry
without sufficient justification to earn the marks.
Almost all candidates who attempted part (i)(c) were able to show the first result by
2𝑛𝑛
applying the definition of � � and then cancelling terms. A small number of candidates
𝑟𝑟
argued the result by viewing the two expressions as representing different ways of counting
the same total number of things. For the next part of (i)(c) the majority of candidates split
the sum and then applied the previous result to the second term. Many candidates,
however, did not pay sufficient attention to the case where 𝑟𝑟 = 0 and ended up with an
incorrect term in the sum. Many candidates jumped straight to the given answer at this
point and therefore did not show sufficient detail to earn the remaining marks for this part.
Many of the candidates who progressed further with this part dealt with the two sums
2𝑛𝑛 2𝑛𝑛 − 1 2𝑛𝑛 − 1
separately, but some used the fact that � � = � � +� � to rearrange into a
𝑟𝑟 𝑟𝑟 − 1 𝑟𝑟
sum of differences, most of which then cancelled out.
Candidates who had completed part (i)(c) well were able to apply the same methods to the
case in part (ii) and this part was generally completed well, although a small number of
candidates failed to notice that the expression for the mean in terms of 𝑛𝑛 had changed.
Question 12
The least popular question on the paper, it was also the least successful with a mean score
of just under one quarter marks. Many attempts did not make much progress beyond the
first part. Candidates with a good understanding of how to calculate the expectation of a
function of a random variable generally made very good progress.
In part (i) many candidates were able to calculate the length of the chord, although many
used the cosine rule on an isosceles triangle to reach 𝑎𝑎√2 − 2 cos 2𝜃𝜃, making that
integration a little harder. A significant number of candidates who attempted this part
omitted to include the probability density function when integrating to calculate the
expected value. A small number of candidates chose to consider the length of the chord as a
random variable and calculated its probability density function, from which they could then
calculate the expected value. While this approach was in general successful it was a
significantly more complicated approach.
In part (ii), many candidates were able to work out the probability density function. Several
candidates struggled to find an expression for the length of the chord and so failed to make
any further progress from this point. Those that did were often able to complete the
calculation of the expected value correctly. In a small number of cases, candidates
attempted to calculate the probability density function for the length of the chord in order
to calculate the expected value. In this case care needs to be taken with the limits of the
integration as the shortest possible length for such chords needs to be calculated. A good
number of candidates were able to rearrange the expected value in part (ii) into the
requested form and many were then able to complete part (iii) successfully, although a
number of attempts again omitted the probability density function and other attempts
multiplied the function by 𝑡𝑡 before integrating.
STEP MATHEMATICS 3
2022
Mark Scheme
1. (i) At intersections
𝑐𝑐 2
(𝑐𝑐𝑐𝑐 − 𝑎𝑎)2 + � − 𝑏𝑏� = 𝑟𝑟 2
𝑡𝑡
M1
Expanding brackets, collecting like terms and multiplying through by 𝑡𝑡 2 (𝑡𝑡 ≠ 0) gives
2𝑎𝑎𝑎𝑎 2𝑏𝑏𝑏𝑏 𝑐𝑐 2
𝑐𝑐 2 − + ( 𝑎𝑎2 + 𝑏𝑏 2 − 𝑟𝑟 2 ) − 3 + 4 = 0
𝑡𝑡 𝑡𝑡 𝑡𝑡
which has roots 𝑡𝑡 𝑖𝑖 and thus M1
Thus
4
1 2 2
� = (𝑏𝑏 − 𝑎𝑎2 + 𝑟𝑟 2)
𝑡𝑡 𝑖𝑖 2 𝑐𝑐 2
𝑖𝑖=1
M1
(𝑡𝑡 1𝑡𝑡2 𝑡𝑡3 + 𝑡𝑡 2 𝑡𝑡3 𝑡𝑡4 + 𝑡𝑡 3𝑡𝑡4 𝑡𝑡1 + 𝑡𝑡 4 𝑡𝑡1𝑡𝑡2 )2 − 2𝑡𝑡 1𝑡𝑡2 𝑡𝑡3𝑡𝑡 4(𝑡𝑡1𝑡𝑡2 + 𝑡𝑡 1𝑡𝑡3 + 𝑡𝑡 1𝑡𝑡 4 + 𝑡𝑡 2 𝑡𝑡3 + 𝑡𝑡 2𝑡𝑡 4 + 𝑡𝑡 3 𝑡𝑡4 )
=
(𝑡𝑡 1𝑡𝑡2 𝑡𝑡3𝑡𝑡 4)2
M1 A1 A1
2 2
= (𝑏𝑏 − 𝑎𝑎2 + 𝑟𝑟 2 )
𝑐𝑐 2
A1
(iii)
4 4
𝑐𝑐 2 2 2
� 𝑂𝑂𝑃𝑃𝑖𝑖 = � � 𝑐𝑐 2 𝑡𝑡 𝑖𝑖 2 +
2
2 � = 𝑐𝑐 2 � 2 (𝑎𝑎2 − 𝑏𝑏 2 + 𝑟𝑟 2) + 2 (𝑏𝑏 2 − 𝑎𝑎2 + 𝑟𝑟 2 )� = 4𝑟𝑟 2
𝑡𝑡 𝑖𝑖 𝑐𝑐 𝑐𝑐
𝑖𝑖=1 𝑖𝑖=1
Alternative E1 B1 as before
2𝑎𝑎 𝑎𝑎
𝑡𝑡 1 + 𝑡𝑡 2 + 𝑡𝑡 3 + 𝑡𝑡 4 = ⇒ 𝑡𝑡 1 + 𝑡𝑡 2 =
𝑐𝑐 𝑐𝑐
2𝑏𝑏 2𝑏𝑏
𝑡𝑡 1 𝑡𝑡2𝑡𝑡 3 + 𝑡𝑡 2 𝑡𝑡3𝑡𝑡 4 + 𝑡𝑡 3 𝑡𝑡4 𝑡𝑡1 + 𝑡𝑡 4 𝑡𝑡1𝑡𝑡2 = ⇒ 𝑡𝑡 1 2𝑡𝑡2 + 𝑡𝑡 2 2𝑡𝑡1 + 𝑡𝑡 12𝑡𝑡2 + 𝑡𝑡 2 2𝑡𝑡1 =
𝑐𝑐 𝑐𝑐
𝑏𝑏
⇒ (𝑡𝑡 1 + 𝑡𝑡 2)𝑡𝑡1𝑡𝑡2 =
𝑐𝑐
M1 A1
𝑏𝑏
So 𝑡𝑡 1𝑡𝑡 2 = 𝑎𝑎 = ±1 and hence 𝑎𝑎 = ±𝑏𝑏 and so the centre of 𝐶𝐶2 is (𝑎𝑎, ±𝑎𝑎 ) as required.
M1 A1
Alternative E1 B1 as before
1 1 1 1 1 1
𝑡𝑡 1 2 + 𝑡𝑡 2 2 + 𝑡𝑡 32 + 𝑡𝑡 4 2 = 2(𝑡𝑡12 + 𝑡𝑡 2 2) = 2 � 2+
�
2 = 𝑡𝑡 2 + 2+ 2+
𝑡𝑡 2 𝑡𝑡 1 1 𝑡𝑡 2 𝑡𝑡 3 𝑡𝑡 42
M1 A1
2 2
and thus (𝑎𝑎2 − 𝑏𝑏 2 + 𝑟𝑟 2 ) = (𝑏𝑏 2 − 𝑎𝑎2 + 𝑟𝑟 2 ) so 𝑎𝑎2 = 𝑏𝑏 2 and 𝑎𝑎 = ±𝑏𝑏
𝑐𝑐2 𝑐𝑐2
M1 A1
2. (i)
If
𝑎𝑎3 + 2𝑏𝑏 3 + 4𝑐𝑐 3 = 0
then
So we may repeat the argument with, say, 2𝑞𝑞 = 𝑏𝑏 and then having done so repeat the whole
argument with 2𝑟𝑟 = 𝑐𝑐 . E1
Thus ∃𝑝𝑝, 𝑞𝑞, 𝑟𝑟 integers with 2𝑟𝑟 = 𝑐𝑐 , |𝑟𝑟| < |𝑐𝑐 | and 𝑝𝑝 3 + 2𝑞𝑞 3 + 4𝑟𝑟 3 = 0 .
So if there were to be a set of such integers a, b, c, there would be a set of such integers p, q, r with
smaller modulus satisfying the same result. This argument may be repeated ad infinitum leading to
the conclusion that there is no least modulus set of integers which is not possible as an infinitely
decreasing sequence of positive integers cannot exist being bounded by 1. (alternatively, assume a,
b, c to be smallest modulus, then we have a contradiction) E1 Hence no such a, b, c exist. *B1 (4)
(ii) If 9𝑎𝑎3 + 10𝑏𝑏 3 + 6𝑐𝑐 3 = 0 , then 10𝑏𝑏 3 = −9𝑎𝑎3 − 6𝑐𝑐 3 = 3(−3𝑎𝑎3 − 2𝑐𝑐 3 )
Thus ∃𝑞𝑞 where 3𝑞𝑞 = 𝑏𝑏, with 𝑞𝑞 an integer and |𝑞𝑞 | < |𝑏𝑏 | M1 A1 and 9𝑎𝑎3 + 270𝑞𝑞 3 + 6𝑐𝑐 3 = 0
which can be divided by 3 to give 3𝑎𝑎3 + 90𝑞𝑞 3 + 2𝑐𝑐 3 = 0.
It would follow that 2𝑐𝑐 3 = −3𝑎𝑎3 − 90𝑞𝑞 3 = 3(−𝑎𝑎3 − 30𝑞𝑞 3 ) and so ∃𝑟𝑟 where 3𝑟𝑟 = 𝑐𝑐, with 𝑟𝑟 an
integer and |𝑟𝑟| < |𝑐𝑐| .
Substituting for 𝑐𝑐 , 3𝑎𝑎3 + 90𝑞𝑞 3 + 54𝑟𝑟 3 = 0 leading to 𝑎𝑎3 + 30𝑞𝑞 3 + 18𝑟𝑟 3 = 0 .
We may repeat the argument with 3𝑝𝑝 = 𝑎𝑎 leading to 27𝑝𝑝3 + 30𝑞𝑞 3 + 18𝑟𝑟3 = 0 which on division
by 3 gives 9𝑝𝑝3 + 10𝑞𝑞 3 + 6𝑟𝑟 3 = 0 , the original equation with a, b, c replaced by p, q, r. A1
So the conclusion can be drawn in the same way as in part (i). (‘by descent’) E1 (4)
(iii) (3𝑛𝑛 ± 1)2 = 9𝑛𝑛 2 ± 6𝑛𝑛 + 1 = 3(3𝑛𝑛 2 ± 2𝑛𝑛) + 1 B1 Every integer may be written as 3𝑛𝑛 − 1 ,
3𝑛𝑛 or 3𝑛𝑛 + 1 . We have shown that the square of an integer which is not a multiple of 3 is one
more than a multiple of 3, and if an integer is a multiple of 3 then it can be written 3𝑛𝑛 and
(3𝑛𝑛)2 = 9𝑛𝑛2 = 3(3𝑛𝑛 2) which is a multiple of 3. Thus the sum of two integers can only be either a
multiple of 3, one more than a multiple of 3, or two more than a multiple of 3 depending on whether
the two integers are multiples of 3, exactly one is a multiple of 3 or neither is a multiple of 3
respectively. Hence the result that the sum of two squares can only be a multiple of three if each of
the integers is a multiple of three. E1
If 𝑎𝑎2 + 𝑏𝑏 2 = 3𝑐𝑐 2 , by the result just deduced,
∃𝑝𝑝 , 𝑞𝑞 where 3𝑝𝑝 = 𝑎𝑎 , 3𝑞𝑞 = 𝑏𝑏 and |𝑝𝑝| < |𝑎𝑎| , |𝑞𝑞 | < |𝑏𝑏 | M1
Substituting for a and b, 9𝑝𝑝 2 + 9𝑞𝑞 2 = 3𝑐𝑐 2 so 𝑐𝑐 2 = 3(3𝑝𝑝2 + 3𝑞𝑞 2) meaning that 𝑐𝑐 2 is a multiple of
3 and hence 𝑐𝑐 is a multiple of 3.
So ∃𝑟𝑟 where 3𝑟𝑟 = 𝑐𝑐, with 𝑟𝑟 an integer and |𝑟𝑟| < |𝑐𝑐| , and substituting for c and dividing by 9 ,
𝑝𝑝 2 + 𝑞𝑞 2 = 3𝑟𝑟 2 which is the original with a, b ,c replaced by p, q, r. As in (i) and (ii), the required
result follows by descent. E1 (4)
(iv) (4𝑛𝑛 ± 1)2 = 16𝑛𝑛 2 ± 8𝑛𝑛 + 1 = 4(4𝑛𝑛2 ± 2𝑛𝑛) + 1 so, the square of an odd integer is one more
than a multiple of four. M1 (2𝑛𝑛)2 = 4𝑛𝑛 2 so the square of an even integer is a multiple of four. M1
Thus, the sum of the squares of three non-zero integers must be 0, 1, 2 or 3 more than a multiple of
four as the integers are all even, two even and one odd, one even and two odd, or all odd
respectively. A1
Thus if 𝑎𝑎2 + 𝑏𝑏 2 + 𝑐𝑐 2 = 4𝑎𝑎𝑎𝑎𝑎𝑎 , 𝑎𝑎, 𝑏𝑏 , and 𝑐𝑐 must all be even. B1
Thus ∃𝑝𝑝, 𝑞𝑞, 𝑟𝑟 integers with 2𝑝𝑝 = 𝑎𝑎 , 2𝑞𝑞 = 𝑏𝑏 , 2𝑟𝑟 = 𝑐𝑐 , and |𝑝𝑝| < |𝑎𝑎| , |𝑞𝑞 | < |𝑏𝑏 | , |𝑟𝑟| < |𝑐𝑐 | . M1
So, if 𝑎𝑎2 + 𝑏𝑏 2 + 𝑐𝑐 2 = 4𝑎𝑎𝑎𝑎𝑎𝑎 , 4𝑝𝑝2 + 4𝑞𝑞 2 + 4𝑟𝑟 2 = 32𝑝𝑝𝑝𝑝𝑝𝑝 , which simplifies to
𝑝𝑝 2 + 𝑞𝑞 2 + 𝑟𝑟 2 = 8𝑝𝑝𝑝𝑝𝑝𝑝. (Alternatively, 𝑎𝑎2 + 𝑏𝑏 2 + 𝑐𝑐 2 = 2𝑛𝑛 𝑎𝑎𝑎𝑎𝑎𝑎 , 𝑎𝑎2 + 𝑏𝑏 2 + 𝑐𝑐 2 = 2𝑛𝑛+1 𝑎𝑎𝑎𝑎𝑎𝑎 )
M1
The argument can be repeated with 𝑝𝑝, 𝑞𝑞 , and 𝑟𝑟 all being even integers with the multiple of the RHS
being a power of two greater than 4. E1 Thus the result follows by descent. E1 (8)
3. (i) 𝑎𝑎𝑥𝑥 2 + 𝑏𝑏𝑏𝑏𝑏𝑏 + 𝑐𝑐𝑦𝑦 2 = 1
𝑎𝑎(4𝑎𝑎𝑎𝑎 − 𝑏𝑏 2 )𝑥𝑥 2 = 𝑏𝑏 2
A1
4𝑎𝑎𝑎𝑎 − 𝑏𝑏 2 > 0
giving
𝑏𝑏 2 < 4𝑎𝑎𝑎𝑎
as required.
(Alternatively, as 2𝑎𝑎𝑎𝑎 = −𝑏𝑏𝑏𝑏 , (−𝑏𝑏𝑏𝑏)2 + 2𝑏𝑏 (−𝑏𝑏𝑏𝑏)𝑦𝑦 + 4𝑎𝑎𝑎𝑎𝑦𝑦 2 = 4𝑎𝑎, (4𝑎𝑎𝑎𝑎 − 𝑏𝑏 2 )𝑦𝑦2 = 4𝑎𝑎 for
M1A1)
E1 (4)
A1
which is a turning point.
So for 4𝑏𝑏 3 𝑐𝑐𝑐𝑐 4 − 8𝑏𝑏 3 𝑥𝑥 3 − 𝑎𝑎𝑐𝑐 3 = 0 to have two solutions, if 𝑐𝑐 > 0 , the turning point needs to be a
81𝑏𝑏3 27𝑏𝑏3
minimum below the x axis and so − − 𝑎𝑎𝑐𝑐 3 < 0 E1 and if 𝑐𝑐 < 0 , the turning point needs to
4𝑐𝑐3 𝑐𝑐3
81𝑏𝑏3 27𝑏𝑏3
be a maximum above the x axis and so − 𝑐𝑐3 − 𝑎𝑎𝑐𝑐 3 > 0 . E1 Thus, in either case
4𝑐𝑐3
multiplication by 4𝑐𝑐 3 yields
(iii) These are three simultaneous equations in two unknowns so we may solve for two of them and
substitute into the third. The first equation rules out 𝑥𝑥 = 0 as the third equation would imply
𝑦𝑦 = 0 , given that 𝑎𝑎𝑎𝑎𝑎𝑎 ≠ 0 and thus 𝑎𝑎𝑦𝑦 3 + 𝑏𝑏𝑥𝑥 2 𝑦𝑦 + 𝑐𝑐𝑐𝑐 ≠ 1 as required.
If we consider 2𝑏𝑏𝑏𝑏𝑏𝑏 + 𝑐𝑐 = 0 and 3𝑎𝑎𝑦𝑦 2 + 𝑏𝑏𝑥𝑥 2 = 0 , the second if these implies that as 𝑏𝑏 > 0 ,
then 𝑎𝑎 < 0. E1
Multiplying the second of these by 4𝑏𝑏𝑦𝑦 2 , 12𝑎𝑎𝑎𝑎𝑦𝑦 4 + 4𝑏𝑏 2 𝑥𝑥 2𝑦𝑦2 = 0 and substituting from the first
of these two equations,
12𝑎𝑎𝑎𝑎𝑦𝑦 4 + 𝑐𝑐 2 = 0
M1
Thus
4 −𝑐𝑐 2
𝑦𝑦 = ± �
12𝑎𝑎𝑎𝑎
A1
and so
𝑐𝑐 4 12𝑎𝑎𝑎𝑎 4 −3𝑎𝑎𝑐𝑐 2
𝑥𝑥 = ∓ � 2 = ∓�
2𝑏𝑏 −𝑐𝑐 4𝑏𝑏3
A1
Substituting these in 𝑎𝑎𝑦𝑦 3 + 𝑏𝑏𝑥𝑥 2 𝑦𝑦 + 𝑐𝑐𝑐𝑐 = 1 , having first multiplied it by 𝑦𝑦 ,
gives
−𝑐𝑐 2 𝑐𝑐 2 𝑐𝑐 2 4 −𝑐𝑐 2
+ − = ±�
12𝑏𝑏 4𝑏𝑏 2𝑏𝑏 12𝑎𝑎𝑎𝑎
which simplifies to
𝑐𝑐 2 4 −𝑐𝑐 2
− = ±�
3𝑏𝑏 12𝑎𝑎𝑎𝑎
M1
4𝑎𝑎𝑐𝑐 6 + 27𝑏𝑏 3 = 0
as required. *A1 (6)
(Alternative: The first two equations were combined to give 4𝑏𝑏 3𝑐𝑐𝑐𝑐 4 − 8𝑏𝑏 3 𝑥𝑥 3 − 𝑎𝑎𝑐𝑐 3 = 0 in part (ii).
M1
The second and third can be combined to give 4𝑏𝑏 3𝑥𝑥 4 + 3𝑎𝑎𝑐𝑐 2 = 0 M1
𝑐𝑐 3 𝑎𝑎
Thus 𝑥𝑥 = − 𝑏𝑏 � 2 A1
1
and 𝑦𝑦 = 3 A1
√4𝑎𝑎
(ii)
𝑦𝑦 𝑦𝑦 1
= = →1
sinh𝑦𝑦 𝑦𝑦 3 𝑦𝑦 5 𝑦𝑦 2𝑦𝑦 4
𝑦𝑦 +
3! + 5! + ⋯
1+
3! 5! + ⋯
+
as 𝑦𝑦 → 0 . E1
letting 𝑛𝑛 → ∞ ,
and using the result shown from the use of the Maclaurin series that
𝑥𝑥
2𝑛𝑛
𝑥𝑥 → 1
sinh 𝑛𝑛
2
we have
sinh 𝑥𝑥 𝑥𝑥 𝑥𝑥 𝑥𝑥
= cosh cosh ⋯ cosh 𝑛𝑛 ⋯
𝑥𝑥 2 4 2
as required. E1 (2)
1
√2+√2
1 �√2+ 1
2− 2 3 𝑥𝑥 3 𝑥𝑥 �√2 √2+1
(iii) Letting 𝑥𝑥 = ln 2 , sinh𝑥𝑥 = 2
= 4
, cosh 2 = 2
= 2√2
, cosh 4
= 2
=
2�√2
M1
etc.
Thus
3
4 3 √2 + 1 �√2 + 1
= × × ⋯
ln2 2√2 2�√2
2� �√2
M1 A1
and
1 4 √2 + 1 �√2 + 1 4 1 + √2 1 + �√2
= × × ⋯= × × ×⋯
ln 2 2√2 2�√2 2 2
2 � �√2 ��
2√2�√2 √2⋯
A1
The denominator of the first fraction is
1 1 1 1
2 × 22 × 24 × ⋯ = 21+2+4+⋯ = 22 = 4
E1
So
1 1 + √2 1 + �√2
= × ×⋯
ln 2 2 2
as required.
𝑖𝑖𝑖𝑖
(iv) Substituting 𝑥𝑥 = 2
in M1
sinh 𝑥𝑥 𝑥𝑥 𝑥𝑥 𝑥𝑥
= cosh cosh ⋯ cosh 𝑛𝑛 ⋯
𝑥𝑥 2 4 2
and using sinh𝑖𝑖𝑖𝑖 = 𝑖𝑖 sin 𝑥𝑥 , cosh 𝑖𝑖𝑖𝑖 = cos 𝑥𝑥 , M1
𝑖𝑖𝑖𝑖
sinh 2 𝑖𝑖𝑖𝑖 𝑖𝑖𝑖𝑖 𝑖𝑖𝑖𝑖 𝑖𝑖 𝜋𝜋 𝜋𝜋 𝜋𝜋
= cosh cosh ⋯ cosh = = cos cos ⋯ cos ⋯
𝑖𝑖𝑖𝑖 4 8 2𝑛𝑛+1 𝑖𝑖𝑖𝑖 4 8 2𝑛𝑛+1
2 2
M1 A1 A1
√2
1+ 2 �2+√2
, cos = 2 cos 2 − 1 and thus cos = �
𝜋𝜋 √2 𝜋𝜋 𝜋𝜋 𝜋𝜋
cos = =
4 2 4 8 8 2 2
�2+√2 � 2+�2+√2
1+
=�
𝜋𝜋
and similarly, cos 2
= etc. M1 A1 M1
16 2 2
So
2 √2 �2 + √2 � 2 + �2 + √2
= × × ⋯
𝜋𝜋 2 2 2
as required. *A1 (9)
(Alternatively, by induction
𝑥𝑥 𝑥𝑥 𝑥𝑥 𝑥𝑥
sin𝑥𝑥 = 2𝑛𝑛 cos cos ⋯ cos 𝑛𝑛 sin 𝑛𝑛
2 4 2 2
M1 A1 E1 (as for (i)
𝑦𝑦
As → 1 as 𝑦𝑦 → 0 ,
sin 𝑦𝑦
sin 𝑥𝑥 𝑥𝑥 𝑥𝑥 𝑥𝑥
= cos cos ⋯ cos 𝑛𝑛 ⋯
𝑥𝑥 2 4 2
M1A1
𝜋𝜋
and then, substituting 𝑥𝑥 = 2 M1 result follows as before A1M1A1. )
5. (i)
𝑎𝑎 𝑎𝑎
1 𝑒𝑒 −𝑥𝑥 −𝑥𝑥 + 1)]𝑎𝑎 = ln �
𝑒𝑒 𝑎𝑎 + 1
� 𝑑𝑑𝑑𝑑 = � 𝑑𝑑𝑑𝑑 = [ − ln( 𝑒𝑒 −𝑎𝑎 � = ln 𝑒𝑒 𝑎𝑎 = 𝑎𝑎
1 + 𝑒𝑒 𝑥𝑥 𝑒𝑒 −𝑥𝑥 + 1 𝑒𝑒 −𝑎𝑎 + 1
−𝑎𝑎 −𝑎𝑎
M1 A1 *A1(3)
Alternative 1.
𝑎𝑎 1 𝑎𝑎 1+𝑒𝑒𝑥𝑥 𝑒𝑒𝑥𝑥 𝑒𝑒𝑎𝑎 +1
∫−𝑎𝑎 𝑑𝑑𝑑𝑑 = ∫−𝑎𝑎 − 𝑑𝑑𝑑𝑑 = [𝑥𝑥 − ln(1 +)]𝑎𝑎−𝑎𝑎 = 2𝑎𝑎 − ln� � = 2𝑎𝑎 − ln 𝑒𝑒 𝑎𝑎 = 𝑎𝑎
1+𝑒𝑒𝑥𝑥 1+𝑒𝑒𝑥𝑥 1+𝑒𝑒𝑥𝑥 𝑒𝑒 −𝑎𝑎 +1
M1 A1 *A1(3)
Alternative 2.
Substitute 𝑢𝑢 = 𝑒𝑒 𝑥𝑥 ,
𝑎𝑎 1 𝑒𝑒𝑎𝑎 1 1 𝑒𝑒𝑎𝑎 1 1 𝑎𝑎 𝑒𝑒𝑎𝑎 +1
∫−𝑎𝑎 1+𝑒𝑒𝑥𝑥 𝑑𝑑𝑑𝑑 = ∫𝑒𝑒 −𝑎𝑎 1+𝑢𝑢 𝑑𝑑𝑑𝑑 = ∫𝑒𝑒 −𝑎𝑎 𝑢𝑢 − 1+𝑢𝑢 𝑑𝑑𝑑𝑑 = [ln 𝑢𝑢 − ln(1 + 𝑢𝑢)]𝑒𝑒𝑒𝑒 −𝑎𝑎 = 2𝑎𝑎 − ln �𝑒𝑒− 𝑎𝑎+1� =
𝑢𝑢
M1 A1
2𝑎𝑎 − ln 𝑒𝑒 𝑎𝑎 = 𝑎𝑎
*A1(3)
Alternative 3.
Substitute 𝑢𝑢 = 1 + 𝑒𝑒 𝑥𝑥 ,
𝑎𝑎 1 1+𝑒𝑒𝑎𝑎 1 1 𝑒𝑒𝑎𝑎 1 1 𝑎𝑎 𝑒𝑒𝑎𝑎 +1
∫−𝑎𝑎 1+𝑒𝑒𝑥𝑥 𝑑𝑑𝑑𝑑 = ∫1+𝑒𝑒 −𝑎𝑎 𝑢𝑢 𝑑𝑑𝑑𝑑 = ∫𝑒𝑒 −𝑎𝑎 𝑢𝑢 − 1+𝑢𝑢 𝑑𝑑𝑑𝑑 = [ln 𝑢𝑢 − ln(1 + 𝑢𝑢)]𝑒𝑒𝑒𝑒 −𝑎𝑎 = 2𝑎𝑎 − ln �𝑒𝑒− 𝑎𝑎+1� =
𝑢𝑢
M1 A1
2𝑎𝑎 − ln 𝑒𝑒 𝑎𝑎 = 𝑎𝑎
*A1(3)
Alternative 4.
𝑎𝑎 𝑎𝑎 0 𝑎𝑎 0
1 1 1 1 1
� 𝑥𝑥 𝑑𝑑𝑑𝑑 = � 𝑥𝑥 𝑑𝑑𝑑𝑑 + � 𝑥𝑥 𝑑𝑑𝑑𝑑 = � 𝑥𝑥 𝑑𝑑𝑑𝑑 + � . −𝑑𝑑𝑑𝑑
1 + 𝑒𝑒 1 + 𝑒𝑒 1 + 𝑒𝑒 1 + 𝑒𝑒 1 + 𝑒𝑒 −𝑥𝑥
−𝑎𝑎 0 −𝑎𝑎 0 𝑎𝑎
𝑎𝑎 𝑎𝑎 𝑎𝑎
𝑎𝑎
1 1 1 1 1 + 𝑒𝑒 −𝑥𝑥 + 1 + 𝑒𝑒 −𝑥𝑥
=� 𝑑𝑑𝑑𝑑 + � 𝑑𝑑𝑑𝑑 = � + 𝑑𝑑𝑑𝑑 = � −𝑥𝑥 ) 𝑑𝑑𝑑𝑑
1 + 𝑒𝑒 𝑥𝑥 1 + 𝑒𝑒 −𝑥𝑥 1 + 𝑒𝑒 𝑥𝑥 1 + 𝑒𝑒 −𝑥𝑥 0 (1 + 𝑒𝑒 )(1+ 𝑒𝑒
𝑥𝑥
0 0 0
M1
𝑎𝑎
2 + 𝑒𝑒 −𝑥𝑥 + 𝑒𝑒 −𝑥𝑥
=� 𝑑𝑑𝑑𝑑 = [𝑥𝑥]𝑎𝑎0 = 𝑎𝑎
0 2 + 𝑒𝑒 −𝑥𝑥 + 𝑒𝑒 −𝑥𝑥
A1 *A1 (3)
(ii)
Suppose
� 𝑔𝑔 (𝑥𝑥) 𝑑𝑑𝑑𝑑 = 𝐺𝐺(𝑥𝑥) + 𝑐𝑐
Then if
𝑎𝑎
M1
0 𝑎𝑎
1 1
⇔ � . − 𝑑𝑑𝑑𝑑 + � 𝑑𝑑𝑑𝑑 = 𝑎𝑎
1 + 𝑓𝑓(−𝑥𝑥) 1 + 𝑓𝑓(𝑥𝑥)
𝑎𝑎 0
M1 A1
𝑎𝑎
1 1
⇔ � + − 1 𝑑𝑑𝑑𝑑 = 0
1 + 𝑓𝑓(−𝑥𝑥) 1 + 𝑓𝑓(𝑥𝑥)
0
M1 A1
1 1
⇔ + − 1 = 0 ∀𝑥𝑥
1 + 𝑓𝑓(−𝑥𝑥) 1 + 𝑓𝑓(𝑥𝑥)
E1 E1
(iii)
𝑎𝑎 0 𝑎𝑎 0 𝑎𝑎
ℎ(𝑥𝑥) ℎ(𝑥𝑥) ℎ(𝑥𝑥) ℎ(−𝑥𝑥) ℎ(𝑥𝑥)
� 𝑑𝑑𝑑𝑑 = � 𝑑𝑑𝑑𝑑 + � 𝑑𝑑𝑑𝑑 = � . −𝑑𝑑𝑑𝑑 + � 𝑑𝑑𝑑𝑑
1 + 𝑓𝑓(𝑥𝑥) 1 + 𝑓𝑓(𝑥𝑥) 1 + 𝑓𝑓(𝑥𝑥) 1 + 𝑓𝑓(−𝑥𝑥) 1 + 𝑓𝑓(𝑥𝑥)
−𝑎𝑎 −𝑎𝑎 0 𝑎𝑎 0
M1
𝑎𝑎
𝑎𝑎
ℎ(𝑥𝑥) ℎ(𝑥𝑥)
= � + 𝑑𝑑𝑑𝑑 = � ℎ(𝑥𝑥) 𝑑𝑑𝑑𝑑
1 + 𝑓𝑓(−𝑥𝑥) 1 + 𝑓𝑓(𝑥𝑥) 0
0
(iv)
𝜋𝜋 𝜋𝜋 𝜋𝜋
2 2 2
𝑒𝑒 −𝑥𝑥 cos 𝑥𝑥 𝑒𝑒 −𝑥𝑥 cos 𝑥𝑥 cos 𝑥𝑥
� 𝑑𝑑𝑑𝑑 = � 𝑥𝑥 + 𝑒𝑒 −𝑥𝑥 𝑑𝑑𝑑𝑑 = 2 � 𝑑𝑑𝑑𝑑
cosh 𝑥𝑥 𝑒𝑒 1 + 𝑒𝑒 2𝑥𝑥
𝜋𝜋 𝜋𝜋 𝜋𝜋
−2 −2 2 −2
M1 A1
cos 𝑥𝑥 satisfies the conditions for ℎ(𝑥𝑥) in part (iii) and 𝑒𝑒 2𝑥𝑥 satisfies the conditions for 𝑓𝑓(𝑥𝑥) in part
(ii). E1
Therefore,
𝜋𝜋 𝜋𝜋
2 2
𝑒𝑒 −𝑥𝑥 cos 𝑥𝑥 𝜋𝜋
� 𝑑𝑑𝑑𝑑 = 2 � cos 𝑥𝑥 𝑑𝑑𝑑𝑑 = 2 [sin𝑥𝑥 ]20 = 2
𝜋𝜋
cosh 𝑥𝑥
−2 0
M1 A1 (5)
6. (i)
𝛼𝛼 2
cos (𝜃𝜃 + 𝛼𝛼) − cos 𝜃𝜃 = cos 𝜃𝜃 cos 𝛼𝛼 − sin 𝜃𝜃 sin 𝛼𝛼 − cos 𝜃𝜃 ≈ cos 𝜃𝜃 �1 − � − sin𝜃𝜃 𝛼𝛼 − cos 𝜃𝜃
2
𝛼𝛼 2
= −𝛼𝛼 sin 𝜃𝜃 −
2
cos 𝜃𝜃 as required. M1 *A1
If sin 𝜃𝜃 ≠ 0
𝛼𝛼 2 𝛼𝛼
( )
sin 𝜃𝜃 + 𝛼𝛼 − sin 𝜃𝜃 𝛼𝛼 cos 𝜃𝜃 − sin 𝜃𝜃 cos 𝜃𝜃 − sin 𝜃𝜃
lim 2 2
= lim = lim 𝛼𝛼 = − cot 𝜃𝜃
𝛼𝛼→0 cos(𝜃𝜃 + 𝛼𝛼) − cos 𝜃𝜃 𝛼𝛼→0 𝛼𝛼 2 𝛼𝛼→0
−𝛼𝛼 sin𝜃𝜃 − cos 𝜃𝜃 − sin𝜃𝜃 − cos 𝜃𝜃
2 2
M1 A1 A1
sin(𝜃𝜃+𝛼𝛼 )−sin 𝜃𝜃 cos(𝜃𝜃+𝛼𝛼 )
(Alternative by l’Hopital , lim ( ) = lim − sin(𝜃𝜃+𝛼𝛼) = lim −cot(𝜃𝜃 + 𝛼𝛼) = − cot 𝜃𝜃
cos 𝜃𝜃+𝛼𝛼 −cos 𝜃𝜃
𝛼𝛼→0 𝛼𝛼→0 𝛼𝛼→0
M1 A1 A1 )
If sin 𝜃𝜃 = 0
sin(𝜃𝜃 + 𝛼𝛼) − sin𝜃𝜃 cos 𝜃𝜃 −2
lim = lim 𝛼𝛼 = lim
𝛼𝛼→0 cos (𝜃𝜃 + 𝛼𝛼) − cos 𝜃𝜃 𝛼𝛼→0 − 𝛼𝛼→0 𝛼𝛼
2 cos 𝜃𝜃
M1
→ −∞ as 𝛼𝛼 → +0 and → ∞ as 𝛼𝛼 → −0 A1 (7)
(ii) (a) If 𝑄𝑄0 is the initial point of contact of 𝐶𝐶1 and 𝐶𝐶2 , and if 𝑋𝑋 is the point on 𝐶𝐶2 which was
initially at 𝑄𝑄0 , then if 𝑄𝑄𝑄𝑄𝑄𝑄0 = 𝜃𝜃 , arc 𝑄𝑄𝑄𝑄0 on 𝐶𝐶1 is of length (𝑛𝑛 − 1)𝑎𝑎𝑎𝑎 E1 and this will equal the
arc length 𝑄𝑄𝑄𝑄 on 𝐶𝐶2 . So if 𝑇𝑇 is the centre of 𝐶𝐶2 , 𝑄𝑄𝑄𝑄𝑄𝑄 = (𝑛𝑛 − 1)𝜃𝜃 , and 𝑇𝑇𝑇𝑇 makes an
angle 𝜃𝜃 + (𝑛𝑛 − 1)𝜃𝜃 = 𝑛𝑛𝑛𝑛 with the 𝑥𝑥 axis. E1
Thus the 𝑥𝑥-coordinate of 𝑃𝑃 is 𝑥𝑥(𝜃𝜃) = 𝑛𝑛𝑛𝑛 cos 𝜃𝜃 + 𝑎𝑎 cos (𝑛𝑛𝑛𝑛) = 𝑎𝑎(𝑛𝑛 cos 𝜃𝜃 + cos 𝑛𝑛𝑛𝑛) as required.
Similarly, 𝑦𝑦(𝜃𝜃) = 𝑎𝑎(𝑛𝑛 sin𝜃𝜃 + sin𝑛𝑛𝑛𝑛 ) . M1 *A1 (4)
(b) 𝑂𝑂𝑂𝑂 = (𝑛𝑛 − 1)𝑎𝑎 if and only if (𝑛𝑛 cos 𝜃𝜃 + cos 𝑛𝑛𝑛𝑛)2 + (𝑛𝑛 sin 𝜃𝜃 + sin 𝑛𝑛𝑛𝑛)2 = (𝑛𝑛 − 1)2
That is if 𝑛𝑛 2 + 2𝑛𝑛 cos (𝑛𝑛 − 1)𝜃𝜃 + 1 = 𝑛𝑛 2 − 2𝑛𝑛 + 1 which is cos (𝑛𝑛 − 1)𝜃𝜃 = −1
M1
(Alternatively, 𝑂𝑂𝑂𝑂 = (𝑛𝑛 − 1)𝑎𝑎 only if 𝑛𝑛 cos 𝜃𝜃 + cos 𝑛𝑛𝑛𝑛 = (𝑛𝑛 − 1) cos 𝜃𝜃 i.e. cos 𝑛𝑛𝑛𝑛 = − cos 𝜃𝜃 ,
and 𝑛𝑛 sin𝜃𝜃 + sin 𝑛𝑛𝑛𝑛 = (𝑛𝑛 − 1) sin 𝜃𝜃 i.e. sin𝑛𝑛𝑛𝑛 = −sin 𝜃𝜃 M1
Thus cos(𝑛𝑛 − 1)𝜃𝜃 = − cos 𝜃𝜃 cos 𝜃𝜃 + −sin𝜃𝜃 sin𝜃𝜃 = −1 so (𝑛𝑛 − 1)𝜃𝜃 is an odd multiple of 𝜋𝜋 M1
Result as before A1)
(c)
𝑦𝑦(𝜃𝜃0 + 𝛼𝛼) − 𝑦𝑦(𝜃𝜃0 ) 𝑎𝑎 (𝑛𝑛 sin(𝜃𝜃0 + 𝛼𝛼) + sin𝑛𝑛(𝜃𝜃0 + 𝛼𝛼)) − 𝑎𝑎(𝑛𝑛 sin𝜃𝜃0 + sin 𝑛𝑛𝜃𝜃0 )
lim = lim
𝛼𝛼→0 𝑥𝑥 (𝜃𝜃0 + 𝛼𝛼) − 𝑥𝑥(𝜃𝜃0 ) 𝛼𝛼→0 𝑎𝑎(𝑛𝑛 cos(𝜃𝜃0 + 𝛼𝛼) + cos 𝑛𝑛(𝜃𝜃0 + 𝛼𝛼)) − 𝑎𝑎(𝑛𝑛 cos 𝜃𝜃0 + cos 𝑛𝑛𝜃𝜃0 )
M1
𝛼𝛼 2 𝑛𝑛 2𝛼𝛼 2
𝑛𝑛 �𝛼𝛼 cos 𝜃𝜃0 − sin 𝜃𝜃0 � + �𝑛𝑛𝑛𝑛 cos 𝑛𝑛𝜃𝜃0 − sin 𝑛𝑛𝜃𝜃0 �
2 2
= lim
𝛼𝛼→0 𝛼𝛼 2 𝑛𝑛 2𝛼𝛼 2
𝑛𝑛 �−𝛼𝛼 sin 𝜃𝜃0 − cos 𝜃𝜃0 � + �−𝑛𝑛𝑛𝑛 sin 𝑛𝑛𝜃𝜃0 − cos 𝑛𝑛𝜃𝜃0 �
2 2
M1 A1
𝛼𝛼
cos 𝜃𝜃0 + cos 𝑛𝑛𝜃𝜃0 − (sin 𝜃𝜃0 + 𝑛𝑛 sin 𝑛𝑛𝜃𝜃0 )
= lim 2
𝛼𝛼→0 − (sin𝜃𝜃 + sin 𝑛𝑛𝜃𝜃 ) − 𝛼𝛼 (cos 𝜃𝜃 + 𝑛𝑛 cos 𝑛𝑛𝜃𝜃 )
0 0 2 0 0
𝜃𝜃 𝜃𝜃
and similarly, sin 𝜃𝜃0 + sin 𝑛𝑛𝜃𝜃0 = 2 sin(𝑛𝑛 + 1) 20 cos(𝑛𝑛 − 1) 20 = 0
Further,
cos 𝜃𝜃0 + 𝑛𝑛 cos 𝑛𝑛𝜃𝜃0 = cos 𝜃𝜃0 + 𝑛𝑛(cos(𝑛𝑛 − 1)𝜃𝜃0 cos 𝜃𝜃0 − sin(𝑛𝑛 − 1) 𝜃𝜃0 sin𝜃𝜃0 )
= (1 − 𝑛𝑛) cos 𝜃𝜃0
So
𝑦𝑦(𝜃𝜃0 + 𝛼𝛼) − 𝑦𝑦(𝜃𝜃0 ) (1 − 𝑛𝑛) sin𝜃𝜃0
lim = = tan𝜃𝜃0
𝛼𝛼→0 𝑥𝑥 (𝜃𝜃0 + 𝛼𝛼) − 𝑥𝑥 (𝜃𝜃0 ) (1 − 𝑛𝑛) cos 𝜃𝜃0
M1 A1
The LHS is the gradient of the tangent to the curve at 𝑃𝑃 and the RHS is the gradient of 𝑂𝑂𝑂𝑂 , as
required. E1 (6)
7. (i)
𝑎𝑎 𝑥𝑥 𝑏𝑏𝑏𝑏 − 𝑐𝑐𝑐𝑐
𝑓𝑓(𝒓𝒓 ) = 𝒏𝒏 × 𝒓𝒓 = �𝑏𝑏 � × �𝑦𝑦� = � 𝑐𝑐𝑐𝑐 − 𝑎𝑎𝑎𝑎�
𝑐𝑐 𝑧𝑧 𝑎𝑎𝑎𝑎 − 𝑏𝑏𝑏𝑏
𝑎𝑎 𝑏𝑏𝑏𝑏 − 𝑐𝑐𝑐𝑐
The x-component of 𝑓𝑓 𝑓𝑓 𝒓𝒓 is the x-component of
� ( )� �𝑏𝑏 � × � 𝑐𝑐𝑐𝑐 − 𝑎𝑎𝑎𝑎�
𝑐𝑐 𝑎𝑎𝑎𝑎 − 𝑏𝑏𝑏𝑏
which is 𝑏𝑏 (𝑎𝑎𝑎𝑎 − 𝑏𝑏𝑏𝑏) − 𝑐𝑐(𝑐𝑐𝑐𝑐 − 𝑎𝑎𝑎𝑎) = −𝑥𝑥 (𝑏𝑏 2 + 𝑐𝑐 2 ) + 𝑎𝑎𝑎𝑎𝑎𝑎 + 𝑎𝑎𝑎𝑎𝑎𝑎 as required. M1 *A1
−𝑥𝑥 (𝑏𝑏 2 + 𝑐𝑐 2 ) + 𝑎𝑎𝑎𝑎𝑎𝑎 + 𝑎𝑎𝑎𝑎𝑎𝑎 = −𝑥𝑥 (𝑎𝑎2 + 𝑏𝑏 2 + 𝑐𝑐 2 ) + 𝑎𝑎2 𝑥𝑥 + 𝑎𝑎𝑎𝑎𝑎𝑎 + 𝑎𝑎𝑎𝑎𝑎𝑎 = −𝑥𝑥 + 𝑎𝑎(𝑎𝑎𝑎𝑎 + 𝑏𝑏𝑏𝑏 + 𝑐𝑐𝑐𝑐)
as 𝒏𝒏 is a unit vector. E1
Similarly, the y and z -components of 𝑓𝑓�𝑓𝑓(𝒓𝒓 )� −𝑦𝑦 + 𝑏𝑏 (𝑎𝑎𝑎𝑎 + 𝑏𝑏𝑏𝑏 + 𝑐𝑐𝑐𝑐) and −𝑧𝑧 + 𝑐𝑐 (𝑎𝑎𝑎𝑎 + 𝑏𝑏𝑏𝑏 + 𝑐𝑐𝑐𝑐)
G1 G1 G1 (8)
(ii)
= 𝒓𝒓 cos 𝜃𝜃 + sin𝜃𝜃 𝒏𝒏 × 𝒓𝒓
A1
If 𝒓𝒓 is perpendicular to 𝒏𝒏 , then 𝒓𝒓 , 𝒏𝒏 , and 𝒏𝒏 × 𝒓𝒓 form a mutually perpendicular vector triad.
g maps 𝒓𝒓 to 𝒓𝒓 cos 𝜃𝜃 + sin 𝜃𝜃 𝒏𝒏 × 𝒓𝒓 which represents an anticlockwise rotation by 𝜃𝜃 about an axis
in the direction 𝒏𝒏 as B1 both vectors are of equal magnitude E1 and are at angle of 𝜃𝜃 to each other
E1 and are both perpendicular to 𝒏𝒏 . E1 (7)
(iii)
ℎ(𝒔𝒔) = −𝒔𝒔 − 2𝑓𝑓(𝒔𝒔) = −𝒔𝒔 − 2�(𝒏𝒏. 𝒔𝒔)𝒏𝒏 − 𝒔𝒔� = 𝒔𝒔 − 2(𝒏𝒏.𝒔𝒔)𝒏𝒏
So, h represents a reflection M1 in the plane through the origin perpendicular to 𝒏𝒏 A1
Justification. If 𝒓𝒓 is as in (ii).
B1
So any vector in the plane through the origin perpendicular to 𝒏𝒏 is invariant under h, E1 and any
vector in the direction of 𝒏𝒏 is reversed. E1 (5)
8. (i)
By de Moivre,
sin(𝑘𝑘𝑘𝑘) = � � cos 𝑘𝑘−1 𝜃𝜃 sin 𝜃𝜃 − �𝑘𝑘� cos 𝑘𝑘−3 𝜃𝜃 sin3 𝜃𝜃 + �𝑘𝑘� cos 𝑘𝑘−5 𝜃𝜃 sin5 𝜃𝜃 − ⋯
𝑘𝑘
1 3 5
𝑘𝑘 𝑘𝑘
= sin 𝜃𝜃 cos 𝑘𝑘−1 𝜃𝜃 �𝑘𝑘 − � �tan2 𝜃𝜃 + � �tan4 𝜃𝜃 − ⋯�
3 5
M1
𝑘𝑘 𝑘𝑘
= sin 𝜃𝜃 cos 𝑘𝑘−1 𝜃𝜃 �𝑘𝑘 − � � (sec 2 𝜃𝜃 − 1) + � � (sec 2 𝜃𝜃 − 1)2 − ⋯ �
3 5
as required. *A1
Similarly, equating real parts,
𝑘𝑘 𝑘𝑘
cos(𝑘𝑘𝑘𝑘) = cos 𝑘𝑘 𝜃𝜃 − � � cos 𝑘𝑘−2 𝜃𝜃 sin2 𝜃𝜃 + � � cos 𝑘𝑘−4 𝜃𝜃 sin4 𝜃𝜃 − ⋯
2 4
𝑘𝑘 𝑘𝑘
= cos 𝑘𝑘 𝜃𝜃 �1 − � � (sec 2 𝜃𝜃 − 1) + � � (sec 2 𝜃𝜃 − 1)2 − ⋯ �
2 4
B1 (6)
(ii)
sin(𝑘𝑘𝑘𝑘) = 0 ⇒ sin𝜃𝜃 cos 𝑘𝑘−1 𝜃𝜃 �𝑘𝑘 − �𝑘𝑘� (sec 2 𝜃𝜃 − 1) + �𝑘𝑘� (sec 2 𝜃𝜃 − 1)2 − ⋯ � = 0
3 5
Thus, if 𝑘𝑘 were odd,
𝑘𝑘
Let = 𝑛𝑛 .
2
𝑛𝑛 𝑛𝑛
By the second result of (i), cos (𝑛𝑛𝑛𝑛) = cos 𝑛𝑛 𝜃𝜃 �1 − � � (sec 2 𝜃𝜃 − 1) + � � (sec 2 𝜃𝜃 − 1)2 − ⋯ �
2 4
1 𝑛𝑛 𝑛𝑛
= �1 − �2� (𝑎𝑎2 − 1) + �4� (𝑎𝑎2 − 1)2 − ⋯ �
𝑎𝑎𝑛𝑛
M1
As before, the bracketed expression is odd, being the sum of one odd number (the first which is 1)
and the remainder even, and thus not zero, so cos(𝑛𝑛𝑛𝑛) ≠ 0 which is a contradiction. A1
Thus, there is no least integer 𝑘𝑘 for which sin(𝑘𝑘𝑘𝑘) = 0 , dM1 and hence that 𝑘𝑘𝑘𝑘 = 180𝑝𝑝 , i.e. that
180𝑝𝑝
𝜃𝜃 = 𝑘𝑘
. Hence 𝜃𝜃 is irrational. E1 (4)
(iii) Suppose there is a positive odd integer 𝑘𝑘 such that sin(𝑘𝑘𝑘𝑘) = 0 and sin(𝑚𝑚𝑚𝑚) ≠ 0 for all
integers 𝑚𝑚 with 0 < 𝑚𝑚 < 𝑘𝑘.
𝑘𝑘 𝑘𝑘
Then sin(𝑘𝑘𝑘𝑘) = sin 𝜑𝜑 cos 𝑘𝑘−1 𝜑𝜑 �𝑘𝑘 − � � tan2 𝜑𝜑 + � � tan4 𝜑𝜑 − ⋯�
3 5
As before in (ii), the bracketed expression is odd and thus not zero, sin𝜑𝜑 ≠ 0 and as
1
cot 𝜑𝜑 = 𝑏𝑏 ≠ 0 , cos 𝜑𝜑 ≠ 0 . Hence a contradiction. E1
M1
i.e.
𝑣𝑣1 + 𝑘𝑘𝑣𝑣2 = 𝑢𝑢 (1)
Newton’s experimental law of impact gives
𝑣𝑣 2 − 𝑣𝑣1 = 𝑒𝑒𝑒𝑒 (2)
M1
( )
(1) − 𝑘𝑘(2) gives 𝑣𝑣1(1 + 𝑘𝑘) = 𝑢𝑢(1 − 𝑘𝑘𝑘𝑘) and hence 𝑣𝑣1 = 𝑢𝑢 1−𝑘𝑘𝑘𝑘 as required. *A1
(1+𝑘𝑘)
𝑢𝑢 (1+𝑒𝑒)
(1) + (2) gives 𝑣𝑣2 (𝑘𝑘 + 1) = 𝑢𝑢(1 + 𝑒𝑒) and hence 𝑣𝑣2 = (1+𝑘𝑘)
as required. *A1 (4)
1
𝐷𝐷 1 𝐷𝐷
Time for B to reach wall is 𝛽𝛽𝛽𝛽
and the time to then return to point 2
𝐷𝐷 from wall is 2
𝑒𝑒𝑒𝑒𝑒𝑒
1
1 𝐷𝐷
Time for A to reach point 𝐷𝐷 from wall is 2
2 𝛼𝛼𝛼𝛼
Thus
1 1
2 𝐷𝐷 𝐷𝐷 𝐷𝐷
= +2
𝛼𝛼𝛼𝛼 𝛽𝛽𝛽𝛽 𝑒𝑒𝑒𝑒𝑒𝑒
M1 A1
which simplifies to
1 1 1 1 1
= + = �1 + �
2𝛼𝛼 𝛽𝛽 2𝑒𝑒𝑒𝑒 𝛽𝛽 2𝑒𝑒
Hence
𝑒𝑒
𝛼𝛼 = 𝛽𝛽 � �
1 + 2𝑒𝑒
A1
Thus
𝑒𝑒
(1 − 𝑘𝑘𝑘𝑘) = (1 + 𝑒𝑒) � �
1 + 2𝑒𝑒
𝑒𝑒 1 + 2𝑒𝑒 − 𝑒𝑒 − 𝑒𝑒 2
𝑘𝑘𝑘𝑘 = 1 − (1 + 𝑒𝑒) � �=
1 + 2𝑒𝑒 1 + 2𝑒𝑒
M1
and so
1 + 𝑒𝑒 − 𝑒𝑒 2
𝑘𝑘 =
𝑒𝑒 (1 + 2𝑒𝑒)
as required. *A1 (5)
(ii) The first collision (between A and B) is as in part (i).
The second collision (between B and C) is as in part (i) as the ratio of masses is the same but 𝑢𝑢 is
replaced by 𝛽𝛽𝛽𝛽 .
Thus, after two collisions, A has speed 𝛼𝛼𝛼𝛼 , B has speed 𝛼𝛼𝛼𝛼𝛼𝛼 , and C has speed 𝛽𝛽 2 𝑢𝑢 . M1 A1
The condition that B and C collide half the distance from the wall is as in (i) (𝐷𝐷 = 3𝑑𝑑)
So
1 + 𝑒𝑒 − 𝑒𝑒 2
𝑘𝑘 =
𝑒𝑒 (1 + 2𝑒𝑒)
E1
Equating the times of A and B to reach the point of simultaneous collision, we have
5 3
2 𝑑𝑑 𝑑𝑑 𝑑𝑑
= +2
𝛼𝛼𝛼𝛼 𝛽𝛽𝛽𝛽 𝛼𝛼𝛼𝛼𝛼𝛼
M1 A1
Therefore
5 2 3
= +
𝛼𝛼 𝛽𝛽 𝛼𝛼𝛼𝛼
5𝛽𝛽 = 2𝛼𝛼 + 3
A1
So, substituting for 𝛼𝛼 and 𝛽𝛽 ,
5(1 + 𝑒𝑒) 2(1 − 𝑘𝑘𝑘𝑘)
= +3
(1 + 𝑘𝑘) (1 + 𝑘𝑘)
Thus,
Factorising we have,
(2𝑒𝑒 − 1)(4𝑒𝑒2 + 7𝑒𝑒 + 3) = 0
further
(2𝑒𝑒 − 1)(𝑒𝑒 + 1)(4𝑒𝑒 + 3) = 0
M1
1
𝑒𝑒 > 0 so 𝑒𝑒 = 2 as required. *A1 (11)
10. (i)
𝐵𝐵𝐵𝐵 = 2𝑎𝑎 cos 𝜃𝜃
and hence
sin 𝜃𝜃
𝑠𝑠1 =
(2cos 𝜃𝜃 − 1)
as required. *A1
By symmetry,
cos 𝜃𝜃
𝑠𝑠2 =
(2sin 𝜃𝜃 − 1)
B1 (5)
1 1
[Both divisions are valid as both extensions are positive and so cos 𝜃𝜃 > 2 and sin 𝜃𝜃 > 2 ] @
(ii)
The GPE of the particle is −𝑊𝑊 × 𝐵𝐵𝐵𝐵 sin𝜃𝜃 = −2𝑊𝑊𝑊𝑊 sin𝜃𝜃 cos 𝜃𝜃
M1 A1
So
1
𝑝𝑝 = (sin𝜃𝜃 + cos 𝜃𝜃 )
2
A1 (5)
1 √2
The expression is a maximum when 𝜃𝜃 = 45𝑜𝑜 when (sin 𝜃𝜃 + cos 𝜃𝜃 ) =
2 2
*B1 which is attainable
1 1
and a minimum when 𝜃𝜃 = 30𝑜𝑜 or 60𝑜𝑜 (from @) M1 E1 when (sin𝜃𝜃 + cos 𝜃𝜃 ) = �1 + √3 � M1
2 4
*A1 (7) which cannot be attained.
√2 1
So ≥ 𝑝𝑝 > 4 �1 + √3 �
2
4 3 2 √3 1
64 < 75 ⇒ < ⇒ < ⇒ 0.65 < �1 + √3�
25 16 5 4 4
M1
9 1 2 3 √2
> = ⇒ 0.75 = >
16 2 4 4 2
M1
√2 1
Thus, 0.75 > ≥ 𝑝𝑝 > �1 + √3 � > 0.65 which shows that 𝑝𝑝 = 0.7 correct to one significant
2 4
figure. *A1 (3)
11. (i) (a) As the coin is fair, the distribution is binomial and symmetric,
so 𝑃𝑃(𝑋𝑋 = 𝑟𝑟) = 𝑃𝑃(𝑋𝑋 = 𝑁𝑁 − 𝑟𝑟) = 𝑃𝑃(𝑋𝑋 = 2𝑛𝑛 − 𝑟𝑟)
Therefore,
𝑛𝑛−1 𝑛𝑛−1 2𝑛𝑛
E1
1 = 𝑃𝑃(𝑋𝑋 ≤ 𝑛𝑛 − 1) + 𝑃𝑃(𝑋𝑋 = 𝑛𝑛) + 𝑃𝑃(𝑋𝑋 ≥ 𝑛𝑛 + 1) = 2𝑃𝑃(𝑋𝑋 ≤ 𝑛𝑛 − 1) + 𝑃𝑃(𝑋𝑋 = 𝑛𝑛)
Hence,
1
𝑃𝑃(𝑋𝑋 ≤ 𝑛𝑛 − 1) = �1 − 𝑃𝑃(𝑋𝑋 = 𝑛𝑛)�
2
E1 (2)
(b)
1
𝜇𝜇 = 𝑁𝑁𝑁𝑁 = 2𝑛𝑛 × 2 = 𝑛𝑛 (or by symmetry) B1
𝑛𝑛−1 2𝑛𝑛
2𝑛𝑛 1 2𝑛𝑛 2𝑛𝑛 1 2𝑛𝑛
𝛿𝛿 = 𝐸𝐸(|𝑋𝑋 − 𝜇𝜇|) = � (𝑛𝑛 − 𝑟𝑟) � � � � + � (𝑟𝑟 − 𝑛𝑛) � � � �
𝑟𝑟 2 𝑟𝑟 2
𝑟𝑟=0 𝑟𝑟=𝑛𝑛+1
M1
𝑛𝑛−1 2𝑛𝑛
1 2𝑛𝑛 2𝑛𝑛 � �1�
2𝑛𝑛
= � (𝑛𝑛 − 𝑟𝑟) �2𝑛𝑛� � � + � (𝑟𝑟 − 𝑛𝑛) �
𝑟𝑟 2 2𝑛𝑛 − 𝑟𝑟 2
𝑟𝑟=0 𝑟𝑟=𝑛𝑛+1
𝑛𝑛−1 𝑛𝑛−1
1 2𝑛𝑛 1 2𝑛𝑛
= � (𝑛𝑛 − 𝑟𝑟) �2𝑛𝑛� � � + � (𝑛𝑛 − 𝑠𝑠) �2𝑛𝑛� � �
𝑟𝑟 2 𝑠𝑠 2
𝑟𝑟=0 𝑠𝑠 =0
M1
𝑛𝑛−1 𝑛𝑛−1
2𝑛𝑛 1 2𝑛𝑛 2𝑛𝑛 1
=2 � ( 𝑛𝑛 − 𝑟𝑟 � � � = � (𝑛𝑛 − 𝑟𝑟) � � 2𝑛𝑛−1
) �
𝑟𝑟 2 𝑟𝑟 2
𝑟𝑟=0 𝑟𝑟=0
M1 *A1(2)
𝑛𝑛−1 𝑛𝑛−1 𝑛𝑛−1
1 1 2𝑛𝑛 1
𝛿𝛿 = � (𝑛𝑛 − 𝑟𝑟) �2𝑛𝑛� 2𝑛𝑛−1 = � 𝑛𝑛 � � 2𝑛𝑛−1 − � 𝑟𝑟 � � 2𝑛𝑛−1
2𝑛𝑛
𝑟𝑟 2 𝑟𝑟 2 𝑟𝑟 2
𝑟𝑟=0 𝑟𝑟=0 𝑟𝑟=0
M1
𝑛𝑛−1 𝑛𝑛−1
1 2𝑛𝑛 2𝑛𝑛 1
= 2𝑛𝑛−1 �𝑛𝑛 � � � − � 𝑟𝑟 � � 2𝑛𝑛−1 �
2 𝑟𝑟 𝑟𝑟 2
𝑟𝑟=0 𝑟𝑟=1
𝑛𝑛−1
1 1 2𝑛𝑛 2𝑛𝑛 − 1 1
= 2𝑛𝑛−1 �𝑛𝑛 �22𝑛𝑛 − � �� − � 2𝑛𝑛 � � 2𝑛𝑛−1 �
2 2 𝑛𝑛 𝑟𝑟 − 1 2
𝑟𝑟=1
M1 M1
𝑛𝑛−2
𝑛𝑛 1 2𝑛𝑛 2𝑛𝑛 − 1
= �22𝑛𝑛−1 − � � −2� � ��
22𝑛𝑛−1 2 𝑛𝑛 𝑟𝑟
𝑟𝑟=0
𝑛𝑛−2 2𝑛𝑛−1
𝑛𝑛 1
− �2𝑛𝑛� − � �2𝑛𝑛 − 1� − � �
2𝑛𝑛 − 1�
= �22𝑛𝑛−1 �
22𝑛𝑛−1 2 𝑛𝑛 𝑟𝑟 𝑟𝑟
𝑟𝑟 =0 𝑟𝑟=𝑛𝑛+1
M1
𝑛𝑛 1
= �22𝑛𝑛−1 − �2𝑛𝑛� − �22𝑛𝑛−1 − �2𝑛𝑛 − 1� − �2𝑛𝑛 − 1���
22𝑛𝑛−1 2 𝑛𝑛 𝑛𝑛 − 1 𝑛𝑛
𝑛𝑛 1 2𝑛𝑛
= � − � � + 2�2𝑛𝑛 − 1��
22𝑛𝑛−1 2 𝑛𝑛 𝑛𝑛
M1
But
(2𝑛𝑛 − 1)! 2𝑛𝑛 (2𝑛𝑛 − 1)! 1 (2𝑛𝑛)! 1 2𝑛𝑛
�2𝑛𝑛 − 1� = = = = � �
𝑛𝑛 𝑛𝑛! (𝑛𝑛 − 1)! 2𝑛𝑛 𝑛𝑛! (𝑛𝑛 − 1)! 2 𝑛𝑛! 𝑛𝑛! 2 𝑛𝑛
M1
Thus
𝑛𝑛 1 2𝑛𝑛 𝑛𝑛
𝛿𝛿 = � � = 2𝑛𝑛 �2𝑛𝑛�
22𝑛𝑛−1 2 𝑛𝑛 2 𝑛𝑛
as required. *A1 (7)
(Alternative
𝑛𝑛 𝑛𝑛 𝑛𝑛
1 1 2𝑛𝑛 − 1� 1
𝛿𝛿 = �(𝑛𝑛 − 𝑟𝑟) �2𝑛𝑛� 2𝑛𝑛−1 = 𝑛𝑛 � � � 2𝑛𝑛−1 − 2𝑛𝑛 � �
2𝑛𝑛
𝑟𝑟 2 𝑟𝑟 2 𝑟𝑟 − 1 22𝑛𝑛−1
𝑟𝑟=0 𝑟𝑟=0 𝑟𝑟=1
M1 A1
𝑛𝑛 𝑛𝑛−1
2𝑛𝑛 1 2𝑛𝑛 − 1� 1
= 2𝑛𝑛 � � � 2𝑛𝑛 − 2𝑛𝑛 � �
𝑟𝑟 2 𝑠𝑠 22𝑛𝑛−1
𝑟𝑟=0 𝑠𝑠 =0
M1
= 2𝑛𝑛𝑛𝑛(𝑋𝑋 ≤ 𝑛𝑛) − 2𝑛𝑛𝑛𝑛(𝑌𝑌 ≤ 𝑛𝑛 − 1)
1
(where Y is a binomial variable �2𝑛𝑛 − 1, � ) M1
2
1 2𝑛𝑛 1
= 𝑛𝑛�1 + 𝑃𝑃(𝑋𝑋 = 𝑛𝑛)� − 2𝑛𝑛 × = 𝑛𝑛 � � 2𝑛𝑛
2 𝑛𝑛 2
M1 M1 *A1 )
1 2𝑛𝑛+1
(ii) 𝜇𝜇 = 𝑁𝑁𝑁𝑁 = (2𝑛𝑛 + 1) × = (or by symmetry)
2 2
2𝑛𝑛+1
2𝑛𝑛 + 1 2𝑛𝑛 + 1 1 2𝑛𝑛+1
𝛿𝛿 = 𝐸𝐸(|𝑋𝑋 − 𝜇𝜇|) = � � 𝑟𝑟 − �� �� �
2 𝑟𝑟 2
𝑟𝑟=0
𝑛𝑛
1 2𝑛𝑛 + 1 2𝑛𝑛 + 1
= 2𝑛𝑛 � � − 𝑟𝑟� � �
2 2 𝑟𝑟
𝑟𝑟=0
M1 A1
𝑛𝑛 𝑛𝑛
1 2𝑛𝑛 + 1
= 2𝑛𝑛 � � �2𝑛𝑛 + 1� − � 𝑟𝑟 �2𝑛𝑛 + 1��
2 2 𝑟𝑟 𝑟𝑟
𝑟𝑟=0 𝑟𝑟=0
𝑛𝑛
1 2𝑛𝑛 + 1
= 2𝑛𝑛 � × 22𝑛𝑛 − � 𝑟𝑟 �2𝑛𝑛 + 1��
2 2 𝑟𝑟
𝑟𝑟=1
𝑛𝑛
1 2𝑛𝑛 + 1
= 2𝑛𝑛 � × 22𝑛𝑛 − �(2𝑛𝑛 + 1) � 2𝑛𝑛 ��
2 2 𝑟𝑟 − 1
𝑟𝑟=1
M1
(2𝑛𝑛 + 1) 2𝑛𝑛
= � �
22𝑛𝑛+1 𝑛𝑛
A1 (5)
which can alternatively be written as
(2𝑛𝑛 + 1) (2𝑛𝑛)! (2𝑛𝑛 + 1)! (2𝑛𝑛 + 1)! (𝑛𝑛 + 1) (𝑛𝑛 + 1) 2𝑛𝑛 + 1
= = = 2𝑛𝑛+1 = 2𝑛𝑛+1 � �
22𝑛𝑛+1 𝑛𝑛! 𝑛𝑛! 22𝑛𝑛+1 𝑛𝑛!𝑛𝑛! 2 (𝑛𝑛 + 1)!𝑛𝑛! 2 𝑛𝑛
(Alternative
𝑛𝑛
2𝑛𝑛 + 1 2𝑛𝑛 + 1 1
𝛿𝛿 = � � − 𝑟𝑟� � � 2𝑛𝑛
2 𝑟𝑟 2
𝑟𝑟=0
M1 A1
𝑛𝑛 𝑛𝑛
2𝑛𝑛 + 1 1 1
= (2𝑛𝑛 + 1) � � � 2𝑛𝑛+1 − (2𝑛𝑛 + 1) � � 2𝑛𝑛 � 2𝑛𝑛
𝑟𝑟 2 𝑟𝑟 − 1 2
𝑟𝑟=0 𝑟𝑟=1
M1
𝑛𝑛−1
1 2𝑛𝑛 1
= (2𝑛𝑛 + 1) − (2𝑛𝑛 + 1)� � � 2𝑛𝑛
2 𝑠𝑠 2
𝑠𝑠 =0
1 1 1 2𝑛𝑛 1
= (2𝑛𝑛 + 1) − (2𝑛𝑛 + 1) � − � � 2𝑛𝑛 �
2 2 2 𝑛𝑛 2
M1
(2𝑛𝑛 + 1) 2𝑛𝑛
= � �
22𝑛𝑛+1 𝑛𝑛
A1 )
12. (i)
1
If 𝐴𝐴𝐴𝐴𝐴𝐴 = 𝜃𝜃 , then the probability distribution function for 𝜃𝜃 𝑓𝑓(𝜃𝜃) = 2𝜋𝜋 on [0,2𝜋𝜋]
𝜃𝜃
𝐴𝐴𝐴𝐴 = 2𝑎𝑎 sin
2
M1 A1
2𝜋𝜋
𝜃𝜃 1
𝐸𝐸(𝐴𝐴𝐴𝐴) = � 2𝑎𝑎sin 𝑑𝑑𝑑𝑑
0 2 2𝜋𝜋
M1 A1
2𝑎𝑎 𝜃𝜃 2𝜋𝜋
= �−cos �
𝜋𝜋 2 0
4𝑎𝑎
=
𝜋𝜋
A1 (5)
1
(Alternatives replace 𝜃𝜃 with 2𝜑𝜑 , 𝑓𝑓(𝜑𝜑) = 𝜋𝜋 on [0,𝜋𝜋] ,
2 𝜋𝜋
or minor segment 𝐴𝐴𝐴𝐴𝐴𝐴 = 2𝜑𝜑 , 𝑓𝑓(𝜑𝜑) = 𝜋𝜋 on �0, 2 � )
(ii)
𝜋𝜋𝑥𝑥 2 𝑥𝑥 2
𝑃𝑃(𝑅𝑅 ≤ 𝑥𝑥) = =
𝜋𝜋𝑎𝑎2 𝑎𝑎2
M1
Therefore
2𝑥𝑥
𝑓𝑓𝑅𝑅 (𝑥𝑥) =
𝑎𝑎2
for 0 ≤ 𝑥𝑥 ≤ 𝑎𝑎 A1 (2)
If the ends of the chord are 𝑋𝑋 and 𝑌𝑌 , then 𝑂𝑂𝑂𝑂𝑂𝑂 is an isosceles triangle so
4 3 𝑎𝑎
2 − 𝑥𝑥 2 sin2 𝑡𝑡 )2 �
= �− ( 𝑎𝑎
3𝑎𝑎2 sin2 𝑡𝑡 0
4
=− (𝑎𝑎3 cos 3 𝑡𝑡 − 𝑎𝑎3 )
3𝑎𝑎2 sin2 𝑡𝑡
4𝑎𝑎(1 − cos 3 𝑡𝑡 )
=
3 sin2 𝑡𝑡
as required. dM1 *A1 (4)
(1 − cos 3 𝑡𝑡 ) (1 − cos 3 𝑡𝑡 ) 1 + cos 𝑡𝑡 + cos 2 𝑡𝑡 1 cos 𝑡𝑡 (1 + cos 𝑡𝑡 )
= = = +
sin2 𝑡𝑡 (1 − cos 2 𝑡𝑡 ) 1 + cos 𝑡𝑡 1 + cos 𝑡𝑡 1 + cos 𝑡𝑡
M1
1 1 2
𝑡𝑡
= 𝑡𝑡 + cos 𝑡𝑡 = cos 𝑡𝑡 + sec
2 cos 2 2 2 2
M1 *A1 (3)
(Alternative
𝑡𝑡
(1 − cos 3 𝑡𝑡 ) 1 − cos 𝑡𝑡 + cos 𝑡𝑡 (1 − cos 2 𝑡𝑡 ) 2 sin2 2 1 𝑡𝑡
2 + cos 𝑡𝑡
= = 𝑡𝑡 𝑡𝑡 + cos 𝑡𝑡 = sec
sin2 𝑡𝑡 sin2 𝑡𝑡 4 sin2 2 cos 2 2 2 2
M1 M1 *A1 )
giving
4𝑎𝑎 1 𝑡𝑡
𝐿𝐿 (𝑡𝑡 ) = �cos 𝑡𝑡 + sec 2 �
3 2 2
(iii)
𝜋𝜋
2 4𝑎𝑎 1 𝑡𝑡 2
𝐸𝐸�𝐿𝐿 (𝑇𝑇)� = � �cos 𝑡𝑡 + sec 2 � 𝑑𝑑𝑑𝑑
0 3 2 2 𝜋𝜋
M1 A1
𝜋𝜋
8𝑎𝑎 𝑡𝑡 2
= �sin 𝑡𝑡 + tan � ��
3𝜋𝜋 2 0
16𝑎𝑎
=
3𝜋𝜋
M1 A1 (4)
This document was initially designed for print and as such does not reach
accessibility standard WCAG 2.1 in a number of ways including missing text
alternatives and missing document structure.