Let $ABCD$ be a square. Let $Dx$ be a ray from vertex $D$ that intersects side $BC$ internally at point $E$. Draw $BH$ perpendicular to ray $Dx$, where $BH$ intersects $Dx$ at point $F$ and intersects the extension of side $DC$ at point $H$. Draw line segment $HE$ which intersects side $AB$ at point $K$.
Prove the following:
Quadrilateral $DKBH$ is an isosceles trapezoid.
If $\angle CDx = 15°$, then triangle $DKE$ is equilateral.
If $\angle CDx = 22.5°$, then triangle $DHK$ is isosceles.
Context and Motivation:
This is a problem of my own construction. I arrived at this configuration and solution through the following reasoning process:
I started with a square $ABCD$ and drew ray $Dx$. Then, by constructing the perpendicular from $B$ to ray $Dx$, I noticed that certain angles are transferred within the figure. After drawing segment $HE$, I observed that its extension intersects side $AB$ at point $K$, which led me to conjecture that other segments in the figure might also be equal. Specifically, I conjectured that segments $AK$ and $CH$ are equal, and through triangle comparisons and angle equalities, I arrived at the desired result.
Subsequently, I discovered that for special values of angle $\angle CDx$, special triangles are formed.
Additional observations: Upon re-examining the figure, I notice that:
- Point $E$ is the orthocenter of triangle $BDH$ (since segments $DF$ and $BC$ are altitudes)
- Therefore, the diagonals $KH$ and $BD$ of the isosceles trapezoid $DKBH$ are perpendicular
- Quadrilateral $EFHC$ is cyclic
- Segment $FC$ is the angle bisector of $\angle EFH$
I am interested in:
- Verification of my proofs
- Alternative approaches or more elegant solutions
- Further properties or generalizations of this configuration
- Whether this construction appears in the geometric literature
Thank you for your constructive comments.
My attempt:
Proof of (1):
Triangles $\triangle DCE$ and $\triangle BCH$ are congruent because:
- $\angle DCE = \angle BCH = 90°$ (since $BC \perp DH$) ... (1)
- $DC = BC$ (sides of the square) ... (2)
- $\angle EDC = \angle CBH$ ... (3) (both are acute angles with perpendicular sides, or equivalently, they are complementary to $\angle BHD$)
Therefore $DE = BH$ ... (4) and $EC = CH$ ... (5)
From (5), triangle $\triangle ECH$ is right-angled and isosceles, so $\angle CHE = \angle CEH = 45°$.
Since $\angle CEH = \angle KEB = 45°$ (vertically opposite angles), in the right triangle $\triangle KBE$ we have $\angle EKB = 45°$, which means $\triangle KBE$ is right-angled and isosceles, thus $BK = BE$ ... (6)
Since $AB = BC$ ... (7), subtracting (6) from (7) memberwise gives $AK = EC$ ... (8)
Triangles $\triangle KAD \cong \triangle ECD$ because they are right-angled, $AD = DC$ (sides of the square), and by (8), $AK = EC$.
Therefore $KD = DE$ ... (9)
From (4) and (9): $DK = BH$ ... (10)
Since $KB \parallel DH$ (both perpendicular to $AB$) and angles $\angle KDH$ and $\angle BHD$ are acute, their sum is less than $180°$, so $DK$ and $BH$ are not parallel ... (11)
From (10), (11), and $KB \parallel DH$, quadrilateral $KBHD$ is an isosceles trapezoid. $\square$
Proof of (2):
Given: $\angle CDx = 15°$
From the congruent triangles $\triangle KAD \cong \triangle ECD$ (proven in part 1): $$\angle ADK = \angle CDE = 15°$$
Therefore: $$\angle KDE = 90° - 15° - 15° = 60°$$
From (9), triangle $\triangle KDE$ is isosceles with vertex angle $60°$, hence $\triangle KDE$ is equilateral. $\square$
Proof of (3):
Given: $\angle CDx = 22.5°$
From the congruent triangles, $\angle ADK = \angle CDE = 22.5°$
Therefore: $$\angle KDE = 90° - 22.5° - 22.5° = 45°$$
Thus: $$\angle KDH = \angle KDE + \angle EDH = 45° + 22.5° = 67.5°$$
In triangle $\triangle KDH$, since $\angle DHK = 45°$ (from the right isosceles triangle $\triangle CEH$): $$\angle DKH = 180° - 67.5° - 45° = 67.5°$$
Therefore $\angle KDH = \angle DKH = 67.5°$, so triangle $\triangle HKD$ is isosceles. $\square$ Questions for the community:
- Are there simpler or more elegant proofs for any of these results?
- Can these results be generalized to other angles or configurations?
- Is this construction already known in the geometric literature?
Any feedback or alternative approaches would be greatly appreciated!
