LCFS - Screen Reader Cannot Access Penalty Mix Donut Chart Data

**Describe the Bug**
Within the Penalty tab of the Organisation Dashboard, the screen reader does not read or provide access to the Penalty Mix donut chart. Users relying on assistive technologies cannot access the chart data, legend information, percentages, or category breakdowns presented in the visualisation.

**Expected Behaviour**
Screen reader users should be able to access all information presented in the Penalty Mix chart through accessible text alternatives, summaries, data tables, or appropriately labelled chart elements. The screen reader should communicate the chart title, categories, values, percentages, and overall purpose of the chart.

**Actual Behaviour**
The donut chart is not announced or interpreted by the screen reader, preventing users from accessing the information displayed visually within the chart.

**Implications**
Users who rely on screen readers cannot independently access penalty distribution information available in the Penalty Mix chart. This creates an accessibility barrier and may result in users missing important dashboard insights that are only available through visual representation.

**Steps To Reproduce**
Steps to reproduce the behaviour:

User/Role: Organisation User using a Screen Reader

1. Go to the Organisation Dashboard.
2. Click on the "Penalty log" tab.
3. Navigate to the "Penalty Mix" chart.
4. Use a supported screen reader (e.g., NVDA, JAWS, or VoiceOver).
5. Move focus to the donut chart.
6. Listen to the screen reader output.
7. Observe that the chart content, category breakdowns, values, and percentages are not announced or available to assistive technologies.

Start at the Organisation Dashboard's Penalty tab and reproduce the issue by navigating to the Penalty Mix donut chart with a supported screen reader such as NVDA, JAWS, or VoiceOver. Done means the chart title, purpose, categories, values, percentages, legend information, and breakdowns are available to assistive technologies.

Written by the indexing model from the issue text.